Optimisation of mine secondary fan systems 
R Fair 
orcid.org/0000-0002-5286-9045 
Dissertation accepted in fulfilment of the requirements for the 
degree Master of Engineering in Mechanical Engineering at the 
North-West University 
Supervisor:  Prof M Kleingeld 
Graduation:   November 2020 
Student number: 33596484 
 
ABSTRACT………………………………………………………………….. 
The rising cost of electricity in South Africa has driven deep-level gold mines to save on costs. 
The optimisation of mine ventilation systems has shown significant potential for energy savings. 
A literature study showed that no work has been done on the optimisation of deep-level gold mine 
secondary fan systems with the aid of simulation. From this problem statement the objectives of 
this study were defined as: 
? Develop a method for the optimisation of deep-level gold mine secondary fans systems 
with the aid of simulation. 
? Evaluate the feasibility of the developed methodology by implementing it on a case study. 
A methodology was developed to construct and calibrate a simulation model for the optimisation 
of a deep-level gold mine secondary fan system. This methodology can be summarised into four 
steps: 
? Gathering of information,  
? Construction of the model,  
? Calibration and verification of the model, and  
? Use the model to optimise the system. 
This methodology was applied to a case study on a deep-level gold mine in South Africa. A 
secondary fan system was identified and a simulation model was constructed. The simulation was 
calibrated using measured data and was used to evaluate the feasibility of optimisation 
techniques on this system.  
Six optimised proposals were developed and presented to the mine. The most beneficial 
optimised proposal was selected and implemented on the system. A cost saving of  
R 2 million per year was achieved with a total airflow reduction of 87 kg/s.  
The results after implementation were used to validate the simulation model and methodology. It 
was found that the simulation model had an average error of 11.7%. This verified the feasibility 
of using the methodology to optimise secondary fan systems in deep-level gold mines. 
Keywords: Deep-level gold mine, ventilation, secondary fan system, optimisation, simulation. 
  
i 
 
ACKNOWLEDGEMENTS 
I would like to express my sincere gratitude to the following parties: 
? Firstly, I would like to thank my Almighty God for blessing me with the talents, willpower, 
and guidance in the completion of my studies. 
? I would like to thank ETA Operations (Pty) Ltd and its sister companies for the financial 
support to complete this study. Special thanks to prof. Eddie Mathews and  
prof. Marius Kleingeld for the opportunity to complete my post-graduate studies. 
? I would like to thank my family, friends, and co-workers for keeping me positive and 
focused. Your support is much appreciated. 
 
  
ii 
 
TABLE OF CONTENTS 
ABSTRACT………………………………………………………………….. ..................................... I 
ACKNOWLEDGEMENTS ......................................................................................................... II 
TABLE OF CONTENTS ........................................................................................................... III 
LIST OF FIGURES .................................................................................................................... V 
LIST OF TABLES ................................................................................................................... VII 
NOMENCLATURE ................................................................................................................. VIII 
LIST OF ABBREVIATIONS ..................................................................................................... IX 
CHAPTER 1 INTRODUCTION .................................................................................................. 1 
1.1 Mine ventilation systems ................................................................................. 1 
1.2 State of the art ................................................................................................ 10 
1.3 Problem statement and objectives ................................................................ 20 
1.4 Overview of the study .................................................................................... 21 
CHAPTER 2 METHODOLOGY ............................................................................................... 22 
2.1 Introduction .................................................................................................... 22 
2.2 A method for mine ventilation system analysis ........................................... 22 
2.3 Development and verification of simulation ................................................. 24 
2.4 Optimisation of a secondary fan system ...................................................... 30 
2.5 Implementation ............................................................................................... 31 
iii 
 
2.6 Summary ......................................................................................................... 32 
CHAPTER 3 RESULTS .......................................................................................................... 34 
3.1 Introduction .................................................................................................... 34 
3.2 Secondary fan system simulation ................................................................. 35 
3.3 Optimisation of the secondary fan system ................................................... 41 
3.4 Implementation ............................................................................................... 55 
3.5 Interpretation of results ................................................................................. 57 
3.6 Summary ......................................................................................................... 59 
CHAPTER 4 CONCLUSION ................................................................................................... 60 
4.1 Summary ......................................................................................................... 60 
4.2 Recommendations.......................................................................................... 61 
BIBLIOGRAPHY ..................................................................................................................... 63 
ANNEXURES … ..................................................................................................................... 69 
APPENDIX A VENTILATION MEASUREMENT ..................................................................... 70 
APPENDIX B SIMULATION ACCURACY .............................................................................. 78 
APPENDIX C SWW9 IN-STOPE VENTILATION .................................................................... 82 
  
iv 
 
LIST OF FIGURES 
Figure 1: Top gold-producing countries [5]. ............................................................................... 1 
Figure 2: Global gold trends [7], [8]............................................................................................ 2 
Figure 3: Gold production in South Africa [4]. ............................................................................ 2 
Figure 4: Indexed electricity price inflation [2]. ........................................................................... 4 
Figure 5: Example of a basic mine ventilation system [23]. ........................................................ 5 
Figure 6: Fan performance curve. ............................................................................................. 8 
Figure 7: Scalable method for mine ventilation networks [46]. ................................................. 23 
Figure 8: Step A - D of the scalable method. ........................................................................... 24 
Figure 9: Step E - G of the scalable method. ........................................................................... 27 
Figure 10: Final steps of the scalable method. ........................................................................ 32 
Figure 11: Simplified ventilation layout of Mine A. .................................................................... 35 
Figure 12: Mine A 2010 level. .................................................................................................. 36 
Figure 13: Simulation skeleton. ............................................................................................... 40 
Figure 14: Simulation of Proposal 1. ........................................................................................ 43 
Figure 15: Simulation of Proposal 2. ........................................................................................ 45 
Figure 16: Simulation of Proposal 3. ........................................................................................ 47 
Figure 17: Simulation of Proposal 4. ........................................................................................ 48 
Figure 18: Simulation of Proposal 5. ........................................................................................ 50 
Figure 19: Simulation of Proposal 6. ........................................................................................ 51 
Figure 20: Proposal 6 operational changes. ............................................................................ 56 
Figure 21: Implementation differences. .................................................................................... 57 
v 
 
Figure 22: Davis anemometer [57]........................................................................................... 70 
Figure 23: Traversing haulage with Davis anemometer [58]. ................................................... 71 
Figure 24: Laser distance meter. ............................................................................................. 71 
Figure 25: Pitot tube [59]. ........................................................................................................ 72 
Figure 26: Vertical manometer [59].......................................................................................... 73 
Figure 27: Whirling hygrometer. .............................................................................................. 74 
Figure 28: Barometer [60]. ....................................................................................................... 75 
Figure 29: Pitot tube traverse lines [24]. .................................................................................. 76 
Figure 30: Pitot tube measuring points [24]. ............................................................................ 76 
Figure 31: Puff-puff method [24]. ............................................................................................. 77 
Figure 32: SWW9 in-stope ventilation. ..................................................................................... 82 
 
  
vi 
 
LIST OF TABLES 
Table 1: Analysis of critical literature. ...................................................................................... 19 
Table 2: Optimisation techniques used in the literature. .......................................................... 20 
Table 3: Studies that discussed the development of a ventilation simulation model. ............... 20 
Table 4: KPI selection. ............................................................................................................ 38 
Table 5: Simulation boundaries. .............................................................................................. 39 
Table 6: Baseline simulation results. ....................................................................................... 40 
Table 7: Simulation results of Proposal 1. ............................................................................... 44 
Table 8: Simulation results of Proposal 2. ............................................................................... 46 
Table 9: Simulation results of Proposal 3. ............................................................................... 47 
Table 10: Simulation results of Proposal 4. ............................................................................. 49 
Table 11: Simulation results of Proposal 5. ............................................................................. 50 
Table 12: Simulation results of Proposal 6. ............................................................................. 52 
Table 13: Simulation proposal comparison. ............................................................................ 53 
Table 14: Proposal advantages and disadvantages. ............................................................... 54 
Table 15: Proposal implementation results. ............................................................................ 56 
Table 16: Pitot tube measuring points [24]. ............................................................................. 77 
Table 17: Baseline simulation accuracy. ................................................................................. 78 
Table 18: Proposal 6 simulation results. ................................................................................. 79 
Table 19: Simulation results as implemented. ......................................................................... 80 
  
  
vii 
 
NOMENCLATURE 
 
 
Symbol Description Units 
   
Ak Actual data point - 
Q Airflow quantity m3/s 
P Barometric pressure kPa 
Tdb Drybulb temperature °C 
H Humidity % 
Error% Percentage error % 
Sk Simulated data point - 
k Specific data point - 
K Total number of data points - 
Twb Wetbulb temperature °C 
   
   
   
   
   
 
 
 
  
viii 
 
 
LIST OF ABBREVIATIONS 
DSM Demand side management 
DXF Drawing eXchange format 
KPIs Key performance indicators 
MAE Mean absolute error 
MVSSA Mine Ventilation Society of South Africa 
PTB Process ToolBox 
RAW Return airway 
SCADA Supervisory control and data acquisition 
3D Three dimensional 
VSD Variable speed drives 
VOD Ventilation on demand 
  
  
  
  
 
 
  
ix 
 
CHAPTER 1 INTRODUCTION 
1.1 Mine ventilation systems  
1.1.1 Gold mining in South Africa 
Gold was first discovered in South Africa on the Witwatersrand in 1884, and the first large mining 
company was formed in 1886 [1]. Since then gold mining has played a significant role in the 
development of the South African economy [2], [3]. South Africa was a world leader in gold 
production for many years with a peak production of 1 000 tonnes in 1970. In 2017 138 tonnes of 
gold was produced which led to R 67.6 billion in sales. [4] Today South Africa is ranked seventh 
amongst the top gold-producing countries, as seen in Figure 1 [5]. 
500
450
400
350
300
250
200
150
100
50
0
Country
 
Figure 1: Top gold-producing countries [5]. 
Globally, the gold mining sector is growing steadily with the fluctuating gold price. Figure 2 
illustrates the positive growth of this industry, with the exception of the financial crises in 2008 [6]. 
1 
Tonnes gold produced
 
3500 20000
18000
3000
16000
2500 14000
12000
2000
10000
1500
8000
1000 6000
4000
500
2000
0 0
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Year
Global gold production Gold price
 
Figure 2: Global gold trends [7], [8]. 
The gold industry continues to make a substantial contribution to the South African economy while 
facing significant challenges. These challenges have led to the decline of gold production as well 
as the number of related employees in South Africa, as shown in Figure 3 [2]. 
300 180000
160000
250
140000
200 120000
100000
150
80000
100 60000
40000
50
20000
0 0
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
Year
Gold production Number of Employees
 
Figure 3: Gold production in South Africa [4]. 
2 
SA gold production [metric tonnes] Production [metric tonnes]
Number of employees Gold price [Rand per ounce]
 
The increasing cost of mining decreases gold production. As a result, 75% of South African gold 
mines were unprofitable in 2017 [2]. Some of the challenges faced by South African gold mines 
include: 
? Volatile gold price and rand-dollar exchange [9], 
? rising production costs, and 
? operational challenges. 
South African gold mines have the highest operational costs of any country [2]. In 2018 the global 
average production cost per ounce of gold was US$818 compared to the South African cost of 
US$1,035. The average gold price of that time was US$1,257 per ounce. Gold mines are very 
focused on reducing production costs to retain maximum profit margins without the loss of 
production [6]. Domestic input costs of South African gold mines are rising at rates above inflation. 
Significant contributors to the cost of production include the cost of electricity (around 20%) and 
employee compensation (around 53%) [2].  
Cost of electricity 
Most of South Africa’s gold mines are deep-level operations with some as deep as 4 km 
underground. Eskom is a South African public utility [10]. These gold mines use around 16 % of 
Eskom’s annual power supply, and an increase in cost has a significant impact on the financial 
viability of these mines. The Eskom electricity tariffs have increased with a double-digit 
percentage rate over the past decade, as seen in Figure 4 [2].  
These increases in electricity costs are driving mines to lower production costs and implement 
energy efficiency projects [6]. 
3 
 
350
300
250
200
150
100
50
0
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
Year
 
Figure 4: Indexed electricity price inflation [2]. 
 
Energy initiative projects 
Gold mining is an electrical energy-intensive process. The increases in electricity costs have 
driven mines to find new ways to save electrical energy. The type of equipment most frequently 
subjected to a faulty application are the fans. This equipment usually run for 24 hours a day, and 
therefore have an essential effect on the overall efficiency of a mine [11]. Studies have shown 
that there is significant scope for energy initiative projects on ventilation systems [3], [12], [13], 
[14], [15], [16]. 
 
1.1.2 Deep-level mine ventilation 
Ventilation is the control of air movement. It is necessary for removing heat and dispersing 
contaminants [17], [18]. This is vital for breathing and creating an environment where worker 
efficiency and safety can be at its maximum. A mine must ensure adequate ventilation by creating 
suitable pathways for the air to flow. These pathways should make provision for air to flow down 
the mine and to the working areas as well as suitable pathways for air to exit the mine when it is 
no longer suitable for use. Pathways for air to enter the mine are known as intakes or downcasts. 
Pathways for air to exit the mine are known as exhausts, upcasts or return airways (RAW) [19], 
[20], [21], [22]. Figure 5 illustrates a basic mine ventilation system. 
4 
Electricity price inflation
 
 
Figure 5: Example of a basic mine ventilation system [23]. 
 
Mine ventilation control devices 
The air flowing through a mine follows the path of least resistance. It becomes necessary to 
implement control devices to make sure the air reaches the various working areas. The main 
control element is the surface fans, and in most cases multiple fans are used in parallel.  
Underground air is distributed effectively with the following devices [19]: 
1. Airlocks 
Airlocks are ventilation doors placed in series in travel ways where there is a large pressure 
difference. Airlocks allow workers and trains to travel without air short-circuiting. 
2.  Regulators  
Regulators are restrictions in an airway which limit the quantity of air through that area. 
3. Stoppings 
Stoppings are permanent or temporary seals used to channel airflow.  
  
5 
 
4. Brattices 
Brattices are fire-proof sheets that can be fixed to the roof, sides, and floor to provide temporary 
stopping to guide the air. 
5. Over-cast/ under-cast 
Over-casts/ under-casts are air bridges that allow two airstreams to pass each other without 
mixing [19].  
6. Ducts 
Ducts are a temporary air pipe system that is used to direct air to specific locations with or without 
the aid of fans. 
7. Primary fans 
The primary fans are fans that have a significant impact on the airflow in a mine, such as the 
surface fans. 
8. Secondary fans 
Secondary fans, or booster fans, are installed in series to the primary fans and are used to 
overcome flow resistance. 
9. Development fans 
Development fans are used to ventilate areas where there is no through-flow of air.  
10. District fans 
District fans are used to direct air into a specific area [23]. 
Ventilation surveys  
The objective of a ventilation survey is to measure the data points required for simulation 
calibration: 
? Airflow [m3/s] 
? Temperature [°C] 
? Humidity [%] 
6 
 
? Pressure [kPa] 
According to the practices of the Mine Ventilation Society of South Africa (MVSSA), various 
instruments and instructions are required to measure each data point, as discussed in Appendix 
A [24]. 
1.1.3 Ventilation fan performance curves 
A fan performance curve plots the relationship between the essential characteristics of a fan. 
Figure 6 shows an example of a fan curve for a fan with a rating of 1 MW.  
The static pressure curve is the most commonly used fan curve and compares the relationship 
between the static pressure and the volume flow rate of a fan at constant impeller speed. The fan 
power curve is used in conjunction with the static pressure curve to determine the power used by 
the fan for each operating point [25].  
The guide vane angle on some fans can be adjusted to control the fan and reduce the flow, static 
pressure, and power used by the fan. Guide vanes are stationary blades installed on the inlet or 
outlet of a fan. If the vanes are installed on the inlet side of the fan, they produce a pre-swirl of 
the air [26]. This reduces the pressure and quantity of the fan by reducing the work done by the 
fan [27]. A guide van angle of 0% represents a fan operating at full load.  
The red marker on Figure 6 indicates an example operating point where the fan is providing  
210 m3/s at a static pressure of 3.75 kPa with the guide vane angle set to 0% [23]. The four static 
pressure curves correspond to the four fan power curves of different guide vane angles. The blue 
dot represents the fan power according to the flow and guide vane positions. The fan power of 
940 kW is read off the secondary axis. 
The system resistance curve responds according to the red arrows in Figure 6. Higher system 
resistance would move the operating point left on the static pressure curve which would result in 
increased pressure and decreased volume flow. Lower system resistance would move the 
operating point right on the static pressure curve which would result in decreased pressure and 
an increased volume flow [23]. 
7 
 
 
Figure 6: Fan performance curve. 
 
1.1.4 Ventilation simulation software 
Mine ventilation simulation software is used to predict the airflow distribution of a mine ventilation 
network. These simulation software packages have many benefits, which include: 
? Mine ventilation planning [28]. 
? Improved decision making – these packages allow the user to test complex scenarios 
without having to spend additional resources on implementation. 
? Exploring possibilities – these packages can be used to determine the limits of a ventilation 
system and help evaluate new procedures. 
? Problem diagnostics – existent or potential. [29] 
Studies have shown that thermo-hydraulic ventilation simulation is a viable operational change 
verification technique [30], [31], [32], [33]. Thermo-hydraulic solvers take into account the 
temperature variation and pressure drop of a pipeline or airway [34]. Discussed below is a 
summary of some of the available ventilation simulation software that incorporate a thermo-
hydraulic solver which is capable of solving incompressible flow. 
8 
 
Ventsim [35] 
Ventsim is a widely used three dimensional (3D) mine ventilation software used for the design 
and control of ventilation systems. Models can be built on the mine Drawing eXchange Format 
(DXF) layouts. DXFs allows the software to create a simulation network with accurate dimensions. 
Vuma-3D [36] 
Vuma-3D is a widely used 3D mine ventilation simulation software that is used in the design and 
optimisation of ventilation networks. This software incorporates empirical data measured in mines 
to assist and speed up the calibration process. Models can be built on the mine DXF layouts.  
Process Toolbox (PTB) 
PTB is a 3D mine ventilation software used for the optimisation of mine ventilation networks. 
Models can be built on the mine DXF layouts.  
Selection of software 
A simulation software package that incorporates a thermo-hydraulic solver and a mass flow 
balance is adequate for ventilation optimisation. The level of data required is used to select the 
software package [37]. Software that is capable of analysing compressible flow can be used on 
most ventilation networks. Software that is capable of analysing incompressible flow is to be used 
for deep-level mines [29].  
Simulation accuracy 
This section discusses how the accuracy of a simulation is measured. The accuracy of the 
baseline simulation can be used to verify that the simulation will produce accurate results. 
According to Friedenstein, Cilliers and van Rensburg [38], one of the most effective methods to 
measure simulation model accuracy is the mean absolute error (MAE). The equation for 
determining the MAE percentage error can be seen in Equation 1 [39]. 
 ?1 ?? ? ?
(1) 
?
?????% = ? | | (100%) ? ??
?=1
 
9 
 
Here: 
Error%   Percentage error [%] 
K   Total number of data points [-] 
k   Specific data point [-] 
Ak   Actual data point [-] 
Sk   Simulated data point [-] 
 
1.2 State of the art 
1.2.1 Preamble 
This section will provide an overview of previous research conducted on the optimisation of 
ventilation networks using simulation. The research in the following subsection had different 
objectives and was applied to different mines. All the research selected for comparison had the 
following objective in common which ensures that they are relevant to this study: 
? Optimisation of a mine ventilation system to save on costs or to improve underground 
conditions. 
The research in the final section included the effectiveness of secondary fans in mining ventilation. 
The following points were discussed for each study: 
? Title. 
? A brief overview of the study. 
? A brief description of how the ventilation system was optimised. 
? How each study is relevant and how it can be applied to secondary fan system 
optimisation. 
1.2.2 Optimisation of mine ventilation  
The studies in this section applied different techniques to mine ventilation optimisation problems. 
These studies show that mine ventilation optimisation has significant potential for energy savings 
and improved underground conditions. 
10 
 
1. Wu and Topuz (1998) [40] 
Title : Analysis of mine ventilation systems using operations research methods 
Overview : The study focused on the use of a programming model to optimise the 
locations of ventilation control devices. Models were proposed with the 
criterion to minimise the overall cost. 
Optimisation  : With the primary purpose of reducing costs, mathematical models were 
developed to optimise the system. Sizes and locations of ventilation control 
devices were optimised. 
Relevance : This study optimised the ventilation system by determining the optimal 
location for ventilation control devices such as fans and regulators. It 
explains how airflow can be natural, controlled, and semi-controlled to 
achieve the optimal solution. This study explains the use of aerodynamic 
resistance in a ventilation system which can be used to define the airflow 
of a secondary fan system and understand how ventilation control devices 
impact the system. 
2. Kukard (2006) [12] 
Title : Research on reducing costs of underground ventilation networks in South 
African mines 
Overview : The study focused on reducing energy costs on mine ventilation fans. 
Improved fan use proved to be financially beneficial. This study showed 
that significant energy savings are possible with the optimisation of mine 
ventilation network. 
Optimisation  : Fans were replaced with more efficient fans, and load shifting energy 
initiatives were implemented. 
11 
 
Relevance : This study showed that replacing auxiliary fans with an improved design 
fan would lead to energy saving. This could be applied to a secondary fan 
system by replacing auxiliary or booster fans with improved design fans.  
3. Chatterjee, Zhang and Xia (2015) [13] 
Title : Optimization of mine ventilation fan speeds according to ventilation on 
demand and time of use tariff 
Overview : The study implemented ventilation on demand (VOD) to reduce energy 
costs. There is a significant opportunity for energy in mine ventilation 
networks.  
Optimisation  : A mathematical model was built to determine optimal fan speeds for 
minimum energy costs.  
Relevance : VOD can be applied to a secondary fan system with the use of variable 
speed drives (VSD) on a secondary fan system. 
4. Pritchard (2008) [41]  
Title : Methods to improve efficiency of mine ventilation systems 
Overview : The study focused on finding methods to improve air usage underground 
to yeild environmental and financial benefits. It showed that there are many 
ways to improve the air usage in a mine. 
Optimisation  : Various methods of optimisation are discussed.  
Relevance : The optimisation techniques discussed in this study can be applied to the 
optimisation of a secondary fan system. These include: 
12 
 
? Day-to-day examinations 
? Improved utilisation of RAWs 
? Use of booster fans 
? Main fan blade replacement 
? VOD 
5. Nie et al. (2018) [42] 
Title : Heat treatment and ventilation optimization in a deep mine 
Overview : The study focused on reducing underground working temperatures of a 
gold mine. This was achieved by installing an air-cooling system as well 
as by optimising the ventilation network. Ventilation system optimisation 
could improve the working environment. 
Optimisation  : Ventilation resistance was reduced by clearing critical airways of waste 
rock and water. Ventilation control equipment was improved. 
Relevance : The amount of waste surrounding a secondary fan system can be reduced 
to increase the flow area and improve airflow. Fan sizing optimisation and 
improved usage of ventilation control devices can be applied to a 
secondary fan system to ventilation.   
1.2.3 Optimisation of mine ventilation with the aid of simulation 
The studies in this section applied different techniques to mine ventilation optimisation specifically 
with the aid of simulation modelling. These studies showed that ventilation optimisation is vital for 
energy savings and improved underground conditions. A ventilation simulation model can be a 
useful tool to verify changes.  
 
 
 
13 
 
6. Webber - Youngman (2005) [43]  
Title : An integrated approach towards the optimization of ventilation, air cooling 
and pumping requirements for hot mines 
Overview : In this study, a predictive simulation was used to establish an optimisation 
schedule for a mine ventilation system. It was shown that the use of 
simulation models is vital. 
Optimisation  : The ventilation system was optimised by implementing VOD. VSD’s were 
used to vary the amount of air supplied to the underground network.  
Relevance : The process for the development of the simulation model can be applied 
to the simulation model for a secondary fan system. The optimisation of 
the air quantities used by fans with the VSD’s can be applied to a 
secondary fan system. 
7. Smit (2017)  
Title : Reducing electrical costs for a mine ventilation system with the aid of 
simulation software 
Overview : This study focused on determining whether simulation software could be 
used to aid the optimisation of a mine ventilation system to reduce costs. 
The effects of ventilation configuration changes were accurately predicted 
with the simulation. 
Optimisation  : Fan operating speeds were changed and the affect it had on the 
underground conditions were compared to the predicted simulation 
conditions. 
14 
 
Relevance : The ventilation optimisation techniques, mentioned below, are discussed 
in this study: 
? Ventilation fan impeller improvements 
? VOD by installing VSD’s on the fans 
? Demand side management (DSM) strategies 
These techniques can be applied to a secondary fan system. The method 
to analyse operational changes with simulation can be used on a 
secondary fan system simulation.  
8. Swart (2003) [14]  
Title : Optimising the operation of underground mine refrigeration plants and 
ventilation fans for minimum costs 
Overview : The study focused on developing a simulation model that was used to 
determine the optimum working point of a mine ventilation system. It 
showed that significant savings are possible with the optimisation of air-
cooling power supplied to ventilation systems.   
Optimisation  : Fan speeds were optimised for minimum energy usage while staying in the 
boundaries of minimum and maximum temperatures and airflows.  
Relevance : The optimisation and simulation procedure to determine the optimal 
working point of a ventilation system can be applied to a secondary fan 
system.  
9. Develo, Pillalamarry and Garab (2016) [44] 
Title : Improving the ventilation system at Rosh Pinah zinc mine 
15 
 
Overview : In this study, a simulation was used to improve the underground ventilation 
conditions of a zinc mine. It showed that ventilation optimisation with the 
aid of simulation could significantly improve the underground conditions 
and increase cost savings. 
Optimisation  : Desired airflows were achieved by changing fan sizes. Air flow was 
redirected to achieve optimal usage of air. 
Relevance : The optimisation techniques and the process of evaluating ventilation 
improvement proposals with simulation can be applied to a secondary fan 
system.  
10. Wei, Fangping and Huiqing (2011) [33] 
Title : The use of 3D simulation system in mine ventilation management 
Overview : In this study, a simulation model was built to determine the effect of 
operational changes to a complex ventilation network. The simulation 
system provided reliable data for decision making. 
Optimisation  : Underground airway resistance was reduced to reduce fan energy usage. 
Relevance : The process for the development and use of a 3D mine ventilation 
simulation can be applied to a secondary fan system.  
11. Zhang and Suo (2016) [45]  
Title : Study of Coal Mine Ventilation System Optimization based on Ventsim 
16 
 
Overview : Different optimisation schemes were simulated to determine the best 
solution. Simulation proved to be practical for optimisation mine ventilation 
networks. 
Optimisation  : Tunnel renovation to reduce airflow resistance in the mine and fan 
installation to increase airflow. 
Relevance : The process for the development and use of a 3D mine ventilation 
simulation can be applied to a secondary fan system.  
 
12. Nel, Vosloo and Mathews (2018) [37], [46]  
Title : Evaluating complex mine ventilation operational changes through 
simulation 
Overview : In this study, a method to evaluate operational changes with the use of 
simulation was developed. This method allows for improved decision 
making when optimising a mine ventilation network.  
Optimisation  : Optimisation of the ventilation network included changes in the airflow 
direction, relocation of fans, and airway enlargement. 
Relevance : This study develops a detailed methodology for the development of a 
ventilation simulation to evaluate the impact of operational changes on a 
ventilation system. This methodology can be applied to develop a 
ventilation simulation of a secondary fan system which requires 
optimisation. 
 
  
17 
 
13. Kocsis, Hall and Hardcastle (2014) [47] 
Title : The integration of mine simulation and ventilation simulation to develop a 
‘Life-Cycle’ mine ventilation system 
Overview : This study focuses on the integration of ventilation planning and 
optimisation into the overall life cycle of a mine. Mine ventilation simulation 
can be used to provide efficient primary and secondary ventilation 
systems. 
Optimisation  : VOD methods were applied to reduce the power consumed by the fans.  
Relevance : This study incorporates optimisation techniques into the design of a 
ventilation system. The process to incorporate VOD into a ventilation 
system can be applied to a secondary fan system. 
1.2.4 Secondary fan systems 
The study in this section investigates the impact of the use of secondary fan systems. 
14. de Villiers et al. (2019) [23]  
Title : Evaluating the impact of auxiliary fan practices on localised subsurface 
ventilation 
Overview : This study focuses on the impact and effectiveness of secondary 
ventilation systems. It was found that the shortcomings of ventilation 
practises resulted in significant energy inefficiencies. 
Relevance : The methodology and results from this study can be used to evaluate a 
secondary fan system for its impact. The impact and efficiency of a 
secondary fan system can be used to determine whether optimisation of 
the fan system is required.   
18 
 
1.2.5 Summary of critical literature 
The studies in this section show that the optimisation of mine ventilation networks has significant 
opportunity for energy savings or improved underground conditions. The use of simulation models 
can be vital to determine the most optimum solutions. The critical literature is summarised in  
Table 1. From this table it can be seen that very few of the studies focussed or determined the 
feasibility of optimising secondary fan systems. This lack of literature shows a need for more 
research on the application of optimisation techniques to a gold mine secondary fan system with 
the aid of simulation. 
Table 1: Analysis of critical literature. 
Secondary Deep level 
No. Authors Optimisation Simulation fan systems gold mine 
1 Wu and Topuz ?? ?? ?? ??
2 Kukard ?? ?? ?? ??
Chatterjee, Zhang 
3 and Xia ?? ?? ?? ??
4 Pritchard ?? ?? ?? ??
5 Nie et al. ?? ?? ?? ??
Webber -
6 Youngman ?? ?? ?? ??
7 Smit ?? ?? ?? ??
8 Swart ?? ?? ?? ??
Develo, 
9 Pillalamarry, Garab ?? ?? ?? ??
Wei, Fangping and 
10 Huiqing ?? ?? ?? ??
11 Zhang and Suo ?? ?? ?? ??
Nel, Vosloo and 
12 Mathews ?? ?? ?? ??
Kocsis, Hall and 
13 Hardcastle ?? ?? ?? ??
14 de Villiers et al. ?? ?? ?? ??
 
The optimisation techniques used by each study are summarised in Table 2 below. These 
techniques can be applied to a secondary fan system. 
  
19 
 
Table 2: Optimisation techniques used in the literature. 
Optimisation technique Studies 
Ventilation on demand and load shifting 2, 3, 4, 6, 7, 8, and 13 
Fan optimisation 4, 7, 9, and 12 
Eliminating complexity 4 and 5 
Improving air utilisation  1, 2, 5, 9, 10, 11, and 12 
Day-to-day examinations  4 
 
The studies that describe the development of a ventilation simulation model are shown in  
Table 3. These methodologies can be used to develop a simulation model for a secondary fan 
system that will provide accurate results. 
Table 3: Studies that discussed the development of a ventilation simulation model. 
Study Software used 
Study 6  Integrated predictive simulation (IPS) 
Study 8 Mathematical model  
Study 10 Ventsim 
Study 11 Ventsim 
Study 12 VUMA3D 
 
These studies were conducted as simulation software improved over the years, but the common 
key development points from these studies can be summarised in the following steps: 
? Gathering of information 
? Construction of model 
? Calibration and verification of model 
? Use the model to optimise system 
1.3 Problem statement and objectives 
The rising cost of electricity in South Africa has driven deep-level gold mines to save on costs to 
prevent them from becoming unprofitable. There is scope for saving electrical power on large 
energy consumers. Studies have shown that the optimisation of ventilation systems is feasible for 
reducing costs. The problem statement is therefore: more research on the optimisation of deep-
level gold mine secondary fan systems with the aid of simulation is required. 
20 
 
The objectives of this study are: 
? Develop a method for the optimisation of deep-level gold mine secondary fan systems. 
? Evaluate the feasibility of the developed methodology by implementing it on a case study 
with the aid of simulation. 
Successful optimisation of a deep-level gold mine secondary fan system will entail: 
? Improved or consistent environmental conditions with a reduction in the cost of ventilating 
that area, or 
? Improved environmental conditions with a consistent cost of ventilating that area. 
1.4 Overview of the study 
Chapter 1 gave a background of the current gold mining situation in South Africa as well as 
background on mine ventilation and ventilation optimisation. A review of the critical literature was 
completed to formulate an apparent problem and objective for this study. 
Chapter 2 will introduce a method to evaluate a mine ventilation system. It provides a detailed 
explanation of the construction of a mine ventilation simulation as well as a guide on the 
optimisation process. 
Chapter 3 will describe a case study where a secondary ventilation system of a mine was 
optimised. The simulation was constructed, and the optimised solution provided. The results of 
the proposed plan and validation thereof is also discussed. 
Chapter 4 will conclude this study and discuss the limitations and further recommendations for 
the methods developed.  
 
 
 
  
21 
 
CHAPTER 2 METHODOLOGY 
2.1 Introduction 
This section will introduce an accurate solution methodology for the optimisation of a deep-level 
gold mine secondary fan system.  
The necessary process of the secondary fan system optimisation that will be followed throughout 
this study is shown below: 
1. Benchmark of the ventilation system. 
2. Development of simulation to evaluate optimisation proposals. 
3. Optimisation of system and implementation. 
4. Validation of results. 
Section 2.2 of this chapter describes a 2008 study that developed a scalable methodology of 
evaluating operational changes in mining ventilation with the use of simulation. This methodology 
is then adapted to a secondary fan system in the subsequent sections of this chapter. 
Section 2.3 describes a methodology for the development and verification of a secondary fan 
system ventilation simulation, according to the study discussed in Section 2.2. 
Section 2.4 describes a methodology to optimise a secondary fan system. This includes a 
methodology to analyse the optimisation proposal with the simulation developed in Section 2.3.  
Finally, the optimised proposal can be implemented and evaluated to validate the simulation 
results. 
2.2 A method for mine ventilation system analysis 
2.2.1 Preamble 
This section will describe a scalable methodology developed in a 2008 study and how it will be 
applied to a secondary fan system optimisation. 
2.2.2 Evaluating complex mine ventilation operational changes through simulation 
In a study conducted by Nel, Vosloo and Mathews [46], a scalable method was developed for the 
evaluation of operational changes of deep-level mine ventilation. This method conforms to the 
development steps followed in other studies. This methodology was selected as it describes the 
22 
 
development of a ventilation simulation in greater detail than other studies and was verified with 
case studies. Figure 7 shows a framework for evaluating operational changes in ventilation 
networks from this study.  
The development of a ventilation simulation was summarised in the following steps: 
? Gathering of information,  
? Construction of the model,  
? Calibration and verification of the model, and  
? Use the model to optimise the system. 
Devvellopmentt  and  vverriiffiiccattiion
A. Identify 
Available 
operational 
data
changes B1. High level B2. Intermediate B3. Low level (Per 
(Surface and 
(Per level) stope)
underground KPIs)
B. Benchmark 
service delivery 
and system
Limited 
data B4. Conduct manual 
B5. Ventilation 
measurements for 
Incorrect network datarequired data
data
C. Verify data 
(manual 
measurements)
D. Establish KPI 
for operational 
changes
E. Evaluate 
F. Determine G. Construct and 
simulation H. Analyse model 
simulation calibrate v l ti
packages for results Evaluation
boundaries simulation model
capabilities
Conssttrruccttiion  and  ccalliibrrattiion
Implementation of 
I. Evaluation (Final 
operational 
KPI validation)
changes
IImpllementtattiion
 
Figure 7: Scalable method for mine ventilation networks [46]. 
23 
 
2.2.3 Application of the scalable method to a secondary fan system 
The methodology shown in Figure 7 develops a simulation model and evaluates the impact that 
operational changes will have on a mine ventilation system. It does not include steps to optimise 
a ventilation system. This methodology is adapted by adding the optimisation of a secondary fan 
system to step H. The optimisation techniques are evaluated with the simulation before the 
implementation.  
2.3 Development and verification of simulation 
2.3.1 Preamble 
This section describes a methodology for the development and verification of a simulation model 
for a secondary fan system. Steps A to G from Figure 7 will be scaled to the development of a 
secondary fan system simulation.  
2.3.2 Benchmark ventilation system 
It is vital to understand the system and to collect the data required to develop an accurate 
ventilation simulation. Steps A to D on Figure 7 focus on the collection of information to accurately 
define the system and optimisation project. Figure 8 shows these steps that form part of 
benchmarking the ventilation system. 
A. Identify 
Available 
operational 
data
changes B1. High level B2. Intermediate B3. Low level (Per 
(Surface and 
(Per level) stope)
underground KPIs)
B. Benchmark 
service delivery 
and system
Limited 
data B4. Conduct manual 
B5. Ventilation 
measurements for 
Incorrect network datarequired data
data
C. Verify data 
(manual 
measurements)
D. Establish KPI 
for operational 
changes
 
Figure 8: Step A - D of the scalable method. 
24 
 
The first step of the study is to identify operational changes [46]. This study evaluated operational 
changes made to a ventilation system to optimise the system. In some cases, it might not be clear 
what operational changes can be made to optimise the system. In these cases, the inefficiencies 
of the system have to be identified which will indicate the operational changes required to optimise 
the system.  
Comparing the current ventilation system to its ideal running conditions will help identify 
inefficiencies. An efficient system provides adequate ventilation for all the working areas without 
leakages and unnecessary air being down-casted in the mine, all without using excessive 
electrical power.  
This study focuses on the optimisation of secondary fan systems, and it is, therefore, required to 
know which areas of the mine are being ventilated by each secondary fan system. These systems 
can be compared to the ideal running conditions of such a system and suggested operational 
changes can be identified. 
Suggestions to improve ventilation efficiency by: 
? Using less or smaller fans when there is excess air being used. 
? Using higher efficiency fans to provide the same airflow. 
? Seal off leaking ventilation seals and doors. 
? Seal off closed working areas. 
? Rerouting ventilation on routes with less resistance. 
? Install VSDs for ventilation on demand [48]. 
The effects that these operational changes have on the ventilation system will be made clear by 
the simulation results. The detail optimisation of the system is completed once the system is 
calibrated and verified. The complexity of the system will indicate the level of accuracy required 
on the data for the simulation. 
Ventilation network data is required for the calibration of the simulation model and should be 
obtained from the mine. The parameters that are vital for the calibration of a simulation model 
are: 
? Air mass flow [kg/s], 
? Barometric pressure [kPa], 
? Temperature [°C], 
? Humidity [%], and 
25 
 
? Ventilation control devices specifications. 
The accuracy of the available mine data is essential as it determines the accuracy of the 
simulation. Data from available mine data (B1 to B3 in Figure 7) or manual measurements (B4 on 
Figure 7) may be used. There are different levels of available data from high-level to low-level 
which will determine the accuracy of the simulation. The data can be obtained from the SCADA 
system or by manual measurements. 
SCADA data 
A Supervisory Control and Data Acquisition (SCADA) system uses a network of data 
communication for high-level process supervision.  These systems are used to monitor and 
control large and complex networks such as mining ventilation. Data from this system is stored 
and can be accessed to see trends in operating conditions [49]. 
B1. High-level data 
This data is used when conducting a high-level study that does not require a high level of 
accuracy. High-level data includes only the Key Performance Indicators (KPIs) of a mine 
ventilation system. KPIs used differ from mine to mine but usually include the airflow into the mine, 
average stope temperature, average stope airflow, and average face velocity.  
B2. Intermediate-level data 
Intermediate-level data includes the airflow, temperatures and barometric pressures of each mine 
level intake and return. This data can be used to complete mass balances around the intake and 
returns for each level. 
B3. Low-level data 
Low-level is the most detailed data set available for a mine. The airflow, temperatures, and 
barometric pressures are known for all points in the mine. This will yield the most accurate 
simulation models. 
B4. Conduct manual measurements for required data 
Where SCADA data is not available, manual measurements are conducted according to the 
procedures in Appendix A. Manual measurements are also conducted when a higher accuracy 
simulation is required than what the SCADA data is capable of achieving.  
26 
 
The data must be verified to ensure that it is accurate and usable. Available data can be verified 
with manual measurements (Point C in Figure 7). Benchmarking of the service delivery and the 
data verification is an iterative process where the process can be repeated until a satisfactory 
accuracy is achieved.  
The percentage error of each measurement can describe the accuracy of the data. Mass balances 
can be performed on the mass airflow measurements to verify the accuracy of the measurements 
[50].  
The final step of the ventilation system benchmarking is to select the KPIs, step D in Figure 7. 
The KPIs are used to evaluate the simulation objectives. The operational changes which lead to 
a more efficient system are often used as the main KPIs. 
In the case of optimisation of a ventilation system, KPIs may include but are not limited to: 
? Energy consumed by fans, 
? Air quantity usage, 
? Air quantity through working areas, and 
? Temperatures of working areas and travel ways. 
2.3.3 Simulation construction and calibration 
Once a satisfactory benchmark of the ventilation system is developed, the construction and 
calibration of the simulation can commence. Steps E to G in Figure 7 focus on this construction 
and calibration process. Figure 9 shows the steps from the study discussed which forms part of 
the construction and calibration of the secondary fan system simulation. 
E. Evaluate 
F. Determine G. Construct and 
simulation 
simulation calibrate 
packages for 
boundaries simulation model
capabilities
 
Figure 9: Step E - G of the scalable method. 
 
The software used to simulate the operational changes should be selected to suit the level of the 
solution required. For high accuracy results, a 3D software package that solves mass balances 
and incorporates a thermal-hydraulic solver should be selected. Suitable simulation software 
27 
 
packages are discussed in Section 1.1.4. Other factors which could be compared when selecting 
software are listed below: 
? Cost of simulation software. 
? Training required. 
? Availability of software on the mine. 
The next step is to select the simulation boundaries (Point F in Figure 7). These boundaries are 
the individual constraints of each parameter being simulated. The simulation boundaries are to 
be determined for each KPI. This includes the service delivery, operational, and technical 
boundaries of the project and should be selected per the optimisation objectives. 
The boundaries can include but are not limited to the legal ventilation limits, planned production 
targets, velocities, and shaft diameters [51]. The legal mine ventilation limits determined by the 
MVSSA are as follows [52]:  
Maximum working wet-bulb temperature [°C] 32.5 
Maximum working dry-bulb temperature [°C] 37.0 
Minimum working face velocity [m/s] 0.4 
With all the data gathered, the software selected, and the boundaries defined, the model can be 
constructed (Point G on Figure 7). The first step is to construct a simulation skeleton which is the 
first iteration of the calibration process. A simulation skeleton is an un-calibrated model of the 
ventilation system that includes average sizing values for the components. This can be 
constructed from digital underground mine layouts, and shaft and average tunnel dimensions can 
also be incorporated. It is essential to have the length of the airflow paths constructed correctly.  
The ventilation control devices can be added to the simulation. This includes the doors, seals, 
regulators, ducting, and fans. An average fan operating point can be programmed into the model 
as part of the first iteration of the calibration. This skeleton or first iteration should, according to 
the average sizing values, have airflow in the correct directions as a result of the pressure 
differences. Average sizing and operating conditions of components can be obtained from 
manufacturers.  
28 
 
For the following iterations of the model, the actual verified data can be incorporated into the 
simulation model by adjusting component parameters. The sections below will discuss how to 
calibrate all the vital parameters. 
Air mass flow [kg/s] and barometric pressure [kPa] 
These two parameters are calibrated simultaneously as they are directly correlated to each other. 
For the correct flow through shafts, tunnels, ducting, and stopes, the airflow resistance is adjusted. 
Lower resistance will result in higher airflow, and a higher resistance will result in lower airflow. 
Lower resistance will result in a lower pressure difference, and a higher resistance will result in a 
higher pressure difference. The operating point of the fans can also be adjusted to make sure that 
they produce the correct airflow and pressure difference. 
Temperature [°C] 
The increase in dry-bulb temperature from the air flowing through the mine is calibrated by 
adjusting the amount of heat added to the air by each component. For shafts, tunnels, and stopes, 
the virgin rock temperature and heat transfer coefficients can be adjusted to obtain the correct 
temperature gain per section. For ducting, the heat transfer coefficients can be adjusted. 
Humidity [%] 
The humidity can be calibrated by adjusting the amount of moisture that is added to the air by 
each component. As humidity is the ratio between the dry-bulb and wet-bulb temperature, 
calibrating the wet-bulb temperature would have the same effect. 
Calibration controllers 
Some mine ventilation simulation software, like PTB, have digital calibration controllers to aid this 
process. This component dynamically controls other components to replicate the operation of an 
actual ventilation network [38]. Each controller works like a control system which adjusts the 
parameters to obtain a set point. Calibration controllers could save a lot of time in the calibration 
process as the user does not have to iterate the simulation manually.  
2.3.4 Verification of simulation 
To verify the simulation model, the simulation data can be compared to the actual network data 
to determine the accuracy [53]. Each data point measured should be compared to the simulation 
at that location. The MAE method is used to determine the collective accuracy of the simulation, 
29 
 
according to Equation 1. The simulation should correspond to within 5% of the actual network 
data to ensure accurate results [38]. 
2.4 Optimisation of a secondary fan system 
2.4.1 Preamble 
With a calibrated simulation model, the operational change suggestions discussed in  
Section 2.3.2 can be implemented in the model and simulated to determine their impacts. In this 
section, these operational changes are discussed in more detail as well as how they are applied 
to optimise a ventilation system. Evaluating the results from the different optimisation techniques 
will also be discussed.  
2.4.2 Optimisation techniques 
As part of step H in Figure 7, the optimisation techniques are evaluated with the calibrated 
simulation. The methods to optimise mine ventilation systems shown in Table 2 are discussed 
below: 
Ventilation on demand and load shifting 
On some mines, the ventilation demand may change throughout the day or with seasonal 
changes. The control of fans and cooling utilities can be used to prevent excessive air flows and 
cooling. Limiting the supply to what the mining area needs at specific points in time will lead to 
power savings.  
Fan optimisation 
Replacing fans with more efficient fans will reduce the power used to ventilate the same area.  
The use of multiple fans underground could give rise to the possibility that fans are working 
against each other. Eliminating fans or improved usage of fans can eliminate this problem.  
Eliminating complexity 
Over time most ventilation structures will leak. Minimising the number of ventilation structures will 
reduce the likelihood of leaks forming in the future. Planning less complicated air pathways will 
not only reduce air leaks but will save on ventilation equipment costs. 
  
30 
 
Improving air utilisation 
Having adequate flow area is vital to support the required airflow. The air pressure and fan powers 
increase exponentially with increased air velocity so the most economical solution will be at lower 
air velocities. Utilising the airways in the correct manner can lead to a more efficient system.  
Day-to-day examinations 
Mine ventilation operations are checked on a day-to-day basis to ensure that the ventilation 
control devices are in proper working condition. 
2.4.3 Evaluation with simulation 
The result from the optimised proposals can be analysed by comparing them to the KPIs selected 
in Section 2.3.2 for each data category. 
Each operational change made must result in conditions which satisfy the service delivery and 
operational efficiency KPIs. This analysis is also used to refine the simulation and will help detect 
any user input faults.  
The optimised proposals, which satisfy the KPIs, are then compared to each other. The best 
proposal can then be selected. The selection will depend on the objective of the optimisation. 
Other factors which might influence the selection despite the KPIs are listed below: 
? Ease of implementation 
? Cost of implementation 
? The lifespan of working areas 
? Safety concerns 
2.5 Implementation  
2.5.1 Preamble 
This section describes the implementation of the optimised proposal on the mine ventilation 
system. An evaluation of the implemented proposal will be completed, and the simulation results 
validated. This section falls under the final two steps of the framework shown in Figure 7. 
Figure 10 shows the steps from the study discussed in Section 2.2.2, which forms part of the 
implementation of the optimised proposals. 
31 
 
Implementation of 
I. Evaluation (Final 
operational 
KPI validation)
changes
 
Figure 10: Final steps of the scalable method. 
2.5.2 Implementation of the optimised proposal 
The optimised proposal selected from evaluations with the calibrated simulation is implemented 
on the mine ventilation system. A detailed plan of the proposal is to be given to the ventilation 
construction teams to ensure that the proposal is implemented correctly. The plan should include 
all the operational changes made to the ventilation system. 
2.5.3 Evaluation of the implemented proposal 
A part of step I in Figure 7, the implemented proposal results are evaluated to validate that the 
KPIs are met. Data has to be gathered on the mine ventilation system after the proposal is 
implemented. Data can be gathered in the same way as with the ventilation system benchmarking 
in Section 2.3.2. The gathered data can then be compared to the simulation proposal to ensure 
that all the KPI and optimisation objectives have been met.  
2.5.4 Validation of secondary fan system simulation  
The accuracy of the secondary fan system simulation can then be determined by comparing the 
simulation results to the actual results. [14] The MAE method is used to determine the percentage 
error.  
2.6 Summary 
This chapter presents a scalable method for the optimisation and evaluation of mine secondary 
fan systems. Implementing this method on mine secondary fan systems will promote improved 
efficiency and service delivery.  
Section 2.2 of this chapter described a 2008 study that developed a scalable methodology of 
evaluating operational changes in mining ventilation with the use of simulation. This methodology 
was then applied to a secondary fan system in the subsequent sections of this chapter. 
Section 2.3 described a methodology for the development and verification of a secondary fan 
system ventilation simulation, according to the study discussed in Section 2.2. 
32 
 
Section 2.4 described a methodology to optimise a secondary fan system. This included a 
methodology to analyse the optimisation proposal with the simulation developed in Section 2.3.  
Finally, the optimised proposal was implemented and evaluated to validate the simulation results. 
 
  
33 
 
CHAPTER 3 RESULTS 
3.1 Introduction 
3.1.1 Preamble 
This chapter will apply the method discussed in Chapter 2 to a mine complex. A simulation is 
constructed and is used to optimise a secondary fan system. The best-optimised proposal is 
implemented on the mine complex, and the simulation is verified. 
3.1.2 Case study 
The ventilation network of a deep-level gold mine in South Africa was selected as a case study. 
The mine will be referred to as Mine A for the purpose of this study. Mine A has active mining 
areas on six levels with depths ranging between 1 650 and 2 010 meters under the surface. Mine 
A has an extensive and complicated ventilation network with 49 active working areas, the furthest 
being over 2.3 km from the intake.  
Mine A has two downcast shafts and two upcast shafts. Five shaft (5#) downcast and upcast are 
the main intake and return for Mine A. Four shaft (4#) downcast and upcast are secondary shafts 
intake and return and is mainly used for pumping mine water. 5# and 4# are connected with 
airways on 2010 level (2010L). A simplified ventilation layout of Mine A is shown in Figure 11. 
This study will focus on the secondary fan system on 2010L highlighted in the Figure 11. 
3.1.3 Application of method 
The scalable method shown in Figure 7 was applied to a secondary fan system in Mine A. A 
simulation was used to evaluate proposals to optimise this system. The proposal which best suits 
the objective was implemented, and the simulation results were verified. 
 
34 
 
5# Downcast 5# Upcast 4# Upcast
1650L
1750L
1810L
1870L
1940L
2010L
730 kW Booster fan
Secondary fan system
 
Figure 11: Simplified ventilation layout of Mine A. 
 
3.2 Secondary fan system simulation 
3.2.1 Preamble 
A simulation of the secondary fan system will be used to evaluate optimisation proposals. In order 
to develop this simulation, the system has to be investigated to obtain a benchmark. From the 
benchmark, the simulations can be constructed and calibrated. The simulation will be developed 
and calibrated according to the steps in Figure 8. 
3.2.2 Benchmark ventilation system 
Mine A has multiple secondary fan systems on all its levels to overcome the resistance of the 
extensive ventilation network. The secondary fan system on 2010L makes use of a 730 kW 
booster fan which exhausts the air to 4#. This fan is an old centrifugal fan with a high maintenance 
cost. An initial audit on the fan showed that it was extracting 170 kg/s. This fan was initially 
35 
 
installed to ventilate a large portion of the mine. Due to the downscaling of mining in this area, the 
amount of air required has reduced. Figure 12 shows a top view of Mine A 2010L. The area 
enclosed by the rectangle is ventilated by the secondary fan system. The black DXF lines 
represent the haulages and crosscuts of 2010L. 
Closed off section
5# Down-cast
5# Up-cast
Secondary fan system
2010L SWW9
2010L SWW12
Booster fan
To secondary shaft (4#)
 
Figure 12: Mine A 2010 level. 
 
The amount of air that was flowing through the two working areas (2010L SWW9 and 2010L 
SWW12) amounted to 81.2 kg/s. These two working areas, and a return from 1940L, are the only 
areas which the booster fan had to ventilate. The additional air that it was extracting was air from 
unused areas or through leaks. This was identified as an inefficient system as it was extracting 
around 70 kg/s of air which was not required.  
This inefficient secondary fan system identified on 2010L of Mine A can be optimised with 
operational changes to reduce airflow. The main problem with this system is that it was extracting 
air from unused areas. Sealing of leaks and unused working areas would improve the air usage 
36 
 
for this secondary fan system. The improved air usage allows for the removal or replacement of 
the booster fan with smaller fans to reduce the unnecessary air being extracted. 
The booster fan had a high maintenance cost of around R120 000 per year and failure of this fan 
would risk the entire section being unventilated. This was an additional incentive to remove or 
replace the booster fan with smaller fans. 
With a basic understanding of the system, more detailed data can be gathered, as in step B in 
Figure 7. Mine A does not have their ventilation system data on the SCADA as they are not 
adequately equipped to do so. The ventilation department conducts ventilation audits monthly, 
but these are mostly spot-checked with kestrel vanes.  
The simulation results around the booster fan and affected working areas need to be accurate to 
reduce the risk of decreased production. 2010L and the area affected by the fan requires low-
level data for accurate results. As this was not available, manual measurements needed to be 
taken. The rest of the mine can be calibrated using the ventilation department data. As a result of 
the downscaling, the booster fan impacts a relatively isolated area. An objective of the 
optimisation is to reduce the airflow through the secondary fan system. The booster fans are in 
series with the main surface fans. The reduction in airflow would increase the main fan pressure, 
as seen in Figure 6. The increase in pressure would lead to improved ventilation of the rest of the 
mine because of an increase in airflow through other areas. 
Manual measurements were conducted on the SWW section of 2010L, 1940L, and 1870L, which 
is where the secondary fan system extracts from. These levels have over 2.5 km of travelling from 
downcast to the last crosscut and over 45 data points were measured to define the system. 
To verify that the data was correct, air mass balances were calculated where possible, as 
discussed in step C of Figure 7. The measured data was also compared to the mine ventilation 
department audits. Points where the air mass balances did not add up, or when the measured 
data and ventilation department data did not concur, were re-measured. 
As in step D in Figure 7, the KPIs are selected to evaluate the simulation objectives. The KPIs in 
Table 4 were selected for the 2010L secondary fan system.   
 
 
 
37 
 
Table 4: KPI selection. 
KPI Unit  Reason 
2010L SWW9 air mass flow kg/s Adequate airflow is required for ventilating this area 
2010L SWW9 wet bulb temperature °C Temperatures may not exceed legal limits 
2010L SWW9 dry bulb temperature °C Temperatures may not exceed legal limits 
2010L SWW12 air mass flow kg/s Adequate airflow is required for ventilating this area 
2010L SWW12 wet bulb temperature °C Temperatures may not exceed legal limits 
2010L SWW12 dry bulb temperature °C Temperatures may not exceed legal limits 
Power used by the secondary fan system kW Reduce power consumption 
Air mass flow though the secondary fan system kg/s Reduce unnecessary air being used 
1940L RAW air mass flow kg/s Remain unaffected 
 
3.2.3 Simulation construction and calibration 
This section will focus on the construction, calibration and verification of the simulation for the 
Mine A secondary fan system on 2010 level. This section will follow the steps in Figure 9.  Process 
Toolbox (PTB) is 3D mine ventilation simulation software that solves mass balances and 
incorporates a thermal-hydraulic solver. This system is also capable of simulating large and 
complex systems and proves a suitable software to use. 
Simulation boundaries were determined for each KPI and are shown in Table 5. According to  
step F in Figure 7, these boundaries were later used to evaluate the success of the operational 
changes that were made. The wet and dry-bulb temperatures are measured at the end of the 
working area. The maximum allowable temperatures were selected according to the legal limit in 
South Africa. The ventilation department of Mine A specified the flow through the working areas. 
According to their past measurements, the required air mass flow through the working area to 
ensure 0.4 m/s air velocity at the working face is 30 kg/s and 14 kg/s for SWW12 and SWW9, 
respectively.  
The requirement for the other KPIs are to minimise the electrical power used by the secondary 
fan system and to reduce unnecessary airflow. There is no minimum limit as long as the other 
simulation boundaries are met. Air is returned from 1940 level via a RAW which is ventilated by 
the secondary fan system which should remain unaffected.   
 
 
38 
 
Table 5: Simulation boundaries. 
Item KPI Unit  Boundary 
1 SWW9 Working area air mass flow kg/s 14 
2 SWW9 Working area wet bulb temperature °C 32.5 
3 SWW9 Working area dry bulb temperature °C 37 
4 SWW12 Working area air mass flow kg/s 30 
5 SWW12 Working area wet bulb temperature °C 32.5 
6 SWW12 Working area dry bulb temperature °C 37 
7 Power used by the secondary fan system kW Minimise 
8 Air mass flow though the secondary fan system kg/s Minimise 
9 1940L RAW air mass flow kg/s 23 
 
A simulation skeleton of the mine was constructed in PTB with components set to industry-
standard specifications and ventilation control devices included. Over 5000 components were 
used to construct this simulation. Figure 13 shows the PTB components built over the mine 
underground layouts for the area ventilated by the secondary fan system. The two working areas 
on 2010 level are marked 2010L SWW9 and 2010L SWW12. The air returned from 1940L travel 
through a RAW to 2010L marked 1940L RAW. These three areas and the booster fan, marked 
730 kW booster fan, make up the 2010L secondary fan system. The main haulage is an intake 
and the haulage to 4# is a return airway. 
The data was programmed into the baseline simulation model, and an iterative approach was 
followed to calibrate the model. Calibration controllers were used to speed up the calibration 
process. Components were calibrated according to the parameters discussed as part of step G 
on Figure 7. The calibrated baseline simulation results for the KPIs are compared to the actual 
measured data in Table 6.  
The power used by the 730 kW booster fan could not be measured at the time of the manual 
measurements. The power used was determined with a PTB simulation for a fan running at the 
same conditions. A fan efficiency of 65% was assumed for the 730 kW fan when calculating the 
original power consumption. This is slightly lower than an ideal fan efficiency of 70% [23]. A full 
set of results is seen in Appendix B. 
 
39 
 
 
Figure 13: Simulation skeleton. 
 
Table 6: Baseline simulation results. 
Item KPI Unit Actual  Baseline simulation 
1 2010L SWW9 air mass flow kg/s 42.2 40.4 
2 2010L SWW9 wet-bulb temperature °C 32.0 28.8 
3 2010L SWW9 dry-bulb temperature °C 36.5 34.3 
4 2010L SWW12 air mass flow kg/s 40.8 40.9 
5 2010L SWW12 wet-bulb temperature °C 29.0 30.0 
6 2010L SWW12 dry-bulb temperature °C 33.0 33.2 
7 Power used by the secondary fan system kW 691.0 691.0 
8 Air mass flow through the secondary fan system kg/s 180.5 178.1 
9 1940L return air mass flow kg/s 28.1 27.4 
 
40 
 
3.2.4 Verification of simulation 
The baseline simulation was verified by comparing the simulation outputs to the actual measured 
data, as in Table 6. The MAE method was used to calculate the accuracy of the baseline 
simulation, according to Equation 1, and showed that the baseline had an error of less than 5%. 
This is within the recommended accuracy limits according to literature [38], and the simulation is 
deemed accurate. Appendix B shows the full data set used to determine the accuracy of the 
baseline simulation. 
3.3 Optimisation of the secondary fan system 
3.3.1 Preamble 
This section will focus on the optimisation of the secondary fan system and will follow the steps 
outlined in Figure 10. 
3.3.2 Optimisation techniques 
The main objective of the optimisation was to reduce the energy used by the secondary fan 
system and to reduce the amount of air being extracted. This had to be achieved while meeting 
the other simulation boundaries, as seen in Table 5. To reduce energy and reduce the airflow 
through the secondary fan system the 730 kW fan had to be removed or replaced.   
The methods to improve the efficiency of mine ventilation systems and how they can be applied 
to this system are discussed below. These techniques were applied to the simulation in different 
proposals discussed in the next section. 
Ventilation on demand and load shifting 
Mine A is poorly instrumented on 2010L where the secondary fan system is located, and as a 
result there was no equipment measuring the ventilation conditions of this area. This will make 
controlling the fans difficult. The working areas also require a reasonably constant airflow so this 
optimisation technique will not be evaluated. 
Fan optimisation 
As the 730 kW booster fan used an excessive amount of electrical power, this technique was 
applied in all the proposals. The booster fan was removed and replaced in the simulation with 
41 
 
smaller fans to reduce power consumption. The size and location of the new fans were 
determined according to the simulation results.  
Eliminating complexity 
The secondary fan system could be simplified to reduce the likelihood of leaks forming in the 
future. The system was simplified by reducing the number of return airways in some of the 
proposals. 
Improving air utilisation 
There is excessive air flowing through the working areas ventilated by this secondary fan system. 
Improved air usage was simulated in some of the proposals. 
Day-to-day examinations 
This technique is already part of the everyday operations of the mine ventilation department. 
3.3.3 Evaluation with simulation 
The calibrated simulation model was then used to simulate proposed optimised solutions to 
determine the impact of operational changes. The methodology in Figure 7 was adapted and the 
evaluation of optimisation techniques was added to step H. A total of six proposals were simulated 
with different ventilation flows and fan configurations according to the optimisation techniques. 
Other factors which affected the implementation of each proposal, such as the difficulty of 
implementation, are also discussed for each proposal.  
The ambient condition temperature measured for the baseline simulation was a dry bulb 
temperature of 26°C with a humidity of 55%, which was used to calibrate the model. A worst-case 
scenario of 27°C with a humidity of 55%, the average day-time dry-bulb temperature for a summer 
day, was assumed for the proposals. 
Proposal 1 
The first two proposals simulate returning the air used by the work areas in this section to 5#. A 
PTB layout of Proposal 1 is illustrated in Figure 14. The main haulage is a return airway and the 
walkway to 4# has no airflow. 
42 
 
 
Figure 14: Simulation of Proposal 1. 
The return airway to 4# is sealed off, and the 730 kW booster fan is decommissioned. The air in 
Proposal 1 is being extracted by two 75 kW booster fans and four 45 kW booster fans closer to 
the working area. Table 7 shows the KPIs for the simulation compared to the actual data before 
the operational changes were implemented. The KPIs meet the requirements set out in Table 5. 
The secondary fan power demand decreased by 539 kW which would lead to an electricity cost 
saving of R 3.77 million per year according to 2019 electricity costs in South Africa [54]. The air 
flowing through the secondary fan system was also reduced by 112.9 kg/s. 
The following optimisation techniques were used in this proposal: 
? Fan optimisation, 
? Eliminating complexity, and 
? Improved air utilisation. 
  
43 
 
Table 7: Simulation results of Proposal 1. 
KPI Unit Actual  Proposal 1  Trend 
2010L SWW9 air mass flow kg/s 42.2 15.5 ??
2010L SWW9 wet-bulb temperature °C 32.0 32.3 - 
2010L SWW9 dry-bulb temperature °C 36.5 33.6 ??
2010L SWW12 air mass flow kg/s 40.8 42.2 ??
2010L SWW12 wet-bulb temperature °C 29.0 31.1 ??
2010L SWW12 dry-bulb temperature °C 33.0 33.9 ? 
Secondary fan system power demand kW 691.0 152.0 ??
Air mass flow through the secondary fan 
system kg/s 180.5 67.5 ??
1940L return air mass flow kg/s 28.1 23.9 ??
 
New smaller fans with higher efficiencies are installed to replace the 730 kW booster fan. The 
fans are installed closer to the working areas and in such a way that they do not obstruct travelling. 
Rerouting the air back to 5# simplifies the secondary fans system by eliminating the airway 
between 4# and 5#. Seals and fans are installed in such a way that the air is being used more 
effectively. 
A difficulty that will be faced if this proposal is implemented is that the haulage after the 75 kW 
installed at point number 1 on Figure 14 is an old, unused area of the mine and needs to be 
cleared and properly sealed to ensure an adequate RAW is available. The main walkway to the 
working areas has also been converted into a RAW, meaning travel of workers would be hindered. 
Proposal 2 
The second proposal is similar to the first proposal with the only difference being the fan 
placement for the booster fans ventilating 2010L SWW12. Proposal 2 is illustrated in Figure 15. 
The main haulage is a return airway and the walkway to 4# has no airflow. 
44 
 
 
Figure 15: Simulation of Proposal 2. 
It can be seen that the three 45 kW fans are installed closer to the SWW12 working area at  
point 4. This allows the fans to extract only from the working area to ensure proper ventilation. 
Table 8 shows the KPIs for the simulation compared to the actual data before the operational 
changes were implemented. It can be seen that the KPIs meet the requirements set out in  
Table 5. The results are very similar to Proposal 1 with slightly improved conditions at 2010L 
SWW12. The secondary fan power demand decreased by 539 kW which would lead to an 
electricity cost savings of R 3.77 million per year. The air flowing through the secondary fan 
system was also reduced by 112.9 kg/s. 
Optimisation techniques used for Proposal 2 are the same as those from Proposal 1. This 
proposal would face the same difficulty of clearing the RAW after the 75 kW booster fans installed 
at point 1 on Figure 15. The new location of the 45 kW booster fans could also cause 
complications as it might restrict travel to the working area. The main walkway to the working 
areas has also been converted into a RAW meaning travel of workers would be hindered.  
  
45 
 
Table 8: Simulation results of Proposal 2. 
KPI Unit Actual  Proposal 2 Trend 
2010L SWW9 air mass flow kg/s 42.2 14.5 ??
2010L SWW9 wet-bulb temperature °C 32.0 32.3 - 
2010L SWW9 dry-bulb temperature °C 36.5 33.6 ??
2010L SWW12 air mass flow kg/s 40.8 43.1 ??
2010L SWW12 wet-bulb temperature °C 29.0 31.1 ??
2010L SWW12 dry-bulb temperature °C 33.0 33.9 ??
Secondary fan system power demand kW 691.0 152.0 ??
Air mass flow through the secondary fan system kg/s 180.5 67.5 ??
1940L return air mass flow kg/s 28.1 23.9 ??
 
Proposal 3 
The third proposal replaces the 730 kW booster fan with four smaller booster fans to reduce the 
air returned to 4#. Proposal 3 is illustrated in Figure 16. The main haulage is an intake and the 
haulage to 4# is a return airway. 
Table 9 shows the KPIs for the simulation compared to the actual data before the operational 
changes were implemented. The KPIs meet the requirements set out in Table 5. The secondary 
fan power demand decreased by 554 kW, which would lead to an electricity cost savings of  
R 3.88 million per year. The air flowing through the secondary fan system was also reduced by 
87.3 kg/s. 
It can be seen that the 2010L SWW9 temperatures decreased with less air mass flow. This is only 
possible if the sealing program is implemented correctly and the air ventilating the working area 
comes directly from the intake on 1940L. 
 
46 
 
 
Figure 16: Simulation of Proposal 3. 
Table 9: Simulation results of Proposal 3. 
KPI Unit Actual  Proposal 3 Trend 
2010L SWW9 air mass flow kg/s 42.2 26.6 ??
2010L SWW9 wet-bulb temperature °C 32.0 29.9 ??
2010L SWW9 dry-bulb temperature °C 36.5 31.8 ??
2010L SWW12 air mass flow kg/s 40.8 29.5 ??
2010L SWW12 wet-bulb temperature °C 29.0 30.8 ??
2010L SWW12 dry-bulb temperature °C 33.0 33.6 ??
Power used by the secondary fan system kW 691.0 137.0 ??
Air mass flow through the secondary fan system kg/s 180.5 93.2 ??
1940L return air mass flow kg/s 28.1 25.5 ??
 
47 
 
The only optimisation technique used for this proposal was fan optimisation by installing smaller 
and higher efficiency fans. Any leaking seal will reduce the amount of air ventilating the working 
areas. The seals must be installed correctly and maintained.  
Proposal 4 
The fourth proposal replaces the 730 kW booster fan with two 75 kW booster fans aided by 
booster fans closer to the working areas. Proposal 4 is illustrated in Figure 17. The main haulage 
is an intake and the haulage to 4# is a return airway. 
 
 
Figure 17: Simulation of Proposal 4. 
Table 10 shows the KPIs for the simulation compared to the actual data before the operational 
changes were implemented. It can be seen that the KPIs meet the requirements set out in  
Table 5. The secondary fan power demand decreased by 558 kW, which would lead to an 
48 
 
electricity cost saving of R 3.91 million per year. The air flowing through the secondary fan system 
was also reduced by 111.7 kg/s. 
Table 10: Simulation results of Proposal 4. 
KPI Unit Actual  Proposal 4 Trend 
2010L SWW9 air mass flow kg/s 42.2 15.4 ??
2010L SWW9 wet-bulb temperature °C 32.0 32.4 ??
2010L SWW9 dry-bulb temperature °C 36.5 33.7 ??
2010L SWW12 air mass flow kg/s 40.8 44.5 ??
2010L SWW12 wet-bulb temperature °C 29.0 30.4 ??
2010L SWW12 dry-bulb temperature °C 33.0 33.9 ??
Power used by the secondary fan system kW 691.0 133.0 ??
Air mass flow through the secondary fan system kg/s 180.5 68.8 ??
1940L return air mass flow kg/s 28.1 24.2 ??
 
The optimisation techniques used in this proposal are listed below: 
? Fan optimisation 
? Improved air utilisation 
As with Proposal 2, the booster fans installed close to the working area at point 3 in Figure 17 
could restrict travel to and from the working area.  
Proposal 5 
The fifth proposal is similar to Proposal 3. Proposal 5 is illustrated in Figure 18. The main haulage 
is an intake and the haulage to 4# is a return airway. The new booster fans are placed at  
points 1 and 2 in Figure 18. This ensures that the air returned from 2010L SWW9 and SWW12 is 
evenly distributed.  
Table 11 shows the KPIs for the simulation compared to the actual data before the operational 
changes were implemented. It can be seen that the KPIs meet the requirements set out in  
Table 5. The secondary fan power demand decreased by 555 kW, which would lead to an 
electricity cost saving of R 3.88 million per year. The air flowing through the secondary fan system 
was also reduced by 88.6 kg/s. 
 
49 
 
 
Figure 18: Simulation of Proposal 5. 
Table 11: Simulation results of Proposal 5. 
KPI Unit Actual  Proposal 5 Trend 
2010L SWW9 air mass flow kg/s 42.2 17.7 ??
2010L SWW9 wet-bulb temperature °C 32.0 30.5 ??
2010L SWW9 dry-bulb temperature °C 36.5 32.0 ??
2010L SWW12 air mass flow kg/s 40.8 36.8 ??
2010L SWW12 wet-bulb temperature °C 29.0 30.4 ??
2010L SWW12 dry-bulb temperature °C 33.0 33.6 ??
Power used by the secondary fan system kW 691.0 136.0 ??
Air mass flow through the secondary fan system kg/s 180.5 91.9 ??
1940L return air mass flow kg/s 28.1 25.3 ??
 
  
50 
 
The optimisation techniques used in this proposal are listed below: 
? Fan optimisation 
? Improved air utilisation 
Any leaking seal will reduce the amount of air ventilating the working areas. The seals must be 
installed correctly and maintained. 
Proposal 6 
The sixth proposal is similar to Proposal 5, but the booster fans are installed closer to the working 
area. Proposal 6 is illustrated in Figure 19. The main haulage is an intake and the haulage to 4# 
is a return airway. The new booster fans installed at point 1 and 3 on Figure 19 are all 75 kW fans 
which are larger than necessary but provide a safety factor.  
 
Figure 19: Simulation of Proposal 6. 
Table 12 shows the KPIs for the simulation compared to the actual data before the operational 
changes were implemented. It can be seen that the KPIs meet the requirements set out in  
51 
 
Table 5. The secondary fan power demand decreased by 472 kW, which would lead to an 
electricity cost saving of R 3.31 million per year. The air flowing through the secondary fan system 
was also reduced by 77.9 kg/s. 
The larger fans extract more air from the working areas than other proposals. As mentioned, this 
provides a safety factor and ensures that enough air will be available even if the seals are 
damaged.  
Table 12: Simulation results of Proposal 6. 
KPI Unit Actual  Proposal 6 Trend 
2010L SWW9 air mass flow kg/s 42.2 31.3 ??
2010L SWW9 wet-bulb temperature °C 32.0 30.4 ??
2010L SWW9 dry-bulb temperature °C 36.5 32.6 ??
2010L SWW12 air mass flow kg/s 40.8 51.5 ??
2010L SWW12 wet-bulb temperature °C 29.0 30.0 ??
2010L SWW12 dry-bulb temperature °C 33.0 33.5 ??
Power used by the secondary fan system kW 691.0 219.0 ??
Air mass flow through the secondary fan system kg/s 180.5 102.6 ??
1940L return air mass flow kg/s 28.1 26.0 ??
 
The optimisation techniques used in this proposal are listed below: 
? Fan optimisation 
? Improved air utilisation 
This proposal does not reduce the electrical cost savings by as much as the other proposals. It 
also does not decrease the air mass flow through the secondary fans system by as much, but it 
does provide a safety factor. The increased fans sizes would also provide the option of expanding 
the working area sizes in future without having to install new booster fans. 
Comparison of the proposals 
The simulation results for each proposal and the simulation boundaries seen in Table 5 were 
compared to each other. The results for proposal 1 – 6 are compared in Table 13.  
 
52 
 
Table 13: Simulation proposal comparison. 
KPI Unit Actual  Proposal 1  Proposal 2 Proposal 3 Proposal 4 Proposal 5 Proposal 6 
2010L SWW9 air mass flow kg/s 42.2 15.5 14.5 26.6 15.4 17.7 31.3 
2010L SWW9 wet-bulb temperature °C 32.0 32.3 32.3 29.9 32.4 30.5 30.4 
2010L SWW9 dry-bulb temperature °C 36.5 33.6 33.6 31.8 33.7 32.0 32.6 
2010L SWW12 air mass flow kg/s 40.8 42.2 43.1 29.5 44.5 36.8 51.5 
2010L SWW12 wet-bulb temperature °C 29.0 31.1 31.1 30.8 30.4 30.4 30.0 
2010L SWW12 dry-bulb temperature °C 33.0 33.9 33.9 33.6 33.9 33.6 33.5 
Secondary fan system power demand kW 691.0 152.0 152.0 137.0 133.0 136.0 219.0 
Air mass flow through secondary fan system kg/s 180.5 67.5 67.5 93.2 68.8 91.9 102.6 
1940L return air mass flow kg/s 28.1 23.9 23.9 25.5 24.2 25.3 26.0 
Electricity cost savings R/yr - 3 777 943 3 775 280 3 882 432 3 910 324 3 889 440 3 309 318 
53 
 
The results show that all of the proposals meet the requirements for the air mass flow boundaries, 
which will sustain the 0.4 m/s face velocity required. All the proposals also meet the return 
temperatures for both working areas. All the proposals have managed to reduce the secondary 
fan system power demand and air mass flow through the system. Proposal 4 showed the most 
significant reduction in power demand which leads to an electricity cost saving of R 3.91 million 
per year. Proposal 1 and 2 showed the most significant reduction in air mass flow through the 
secondary fan system with a reduction of 112.9 kg/s.  
These proposal results are closely matched, and all meet the simulation objectives and 
boundaries. The difficulty of implementing the proposal as well as problems that might arise after 
implementation are to be compared to make a decision. The advantages and disadvantages of 
the proposals are compared below in Table 14. 
Table 14: Proposal advantages and disadvantages. 
Proposal  Advantages Disadvantages 
• 5# 2010L RAW needs to be cleared and sealed 
• The highest reduction in mass air-flow through 
off. 
Proposal 1 the system. 
• Difficult to implement.  
• System simplified by eliminating airway to 4#. 
• Increased walkway temperatures. 
• The highest reduction in mass air-flow through • 5# 2010L RAW needs to be cleared and sealed 
the system. off. 
Proposal 2 • System simplified by eliminating airway to 4#. • Difficult to implement.  
• Booster fan locations provide better air mass • Booster fans restrict travel to SWW12. 
flow through the working area. • Increased walkway temperatures. 
• Simplest proposal to implement. • Is reliant on ventilation seal maintenance. 
Proposal 3 • High reduction in secondary fan system power • Air cannot be controlled as accurately. 
demand. • Low reduction in mass air flow through system. 
• The highest reduction in secondary fan system 
power demand. • Most work to implement. 
Proposal 4 
• Booster fan locations provide better air mass • Booster fans restrict travel to SWW12. 
flow through the working area. 
• Simple proposal to implement. 
• Booster fan locations provide better air mass • Is reliant on ventilation seal maintenance. 
Proposal 5 flow control. • Low reduction in mass air-flow through the 
• High reduction in secondary fan system power system. 
demand. 
• Simple proposal to implement. • Low reduction in secondary fan system power 
• Booster fan locations provide better air mass demand. 
Proposal 6 
flow control. • Low reduction in mass air-flow through the 
• Larger fans provide a safety factor. system. 
 
  
54 
 
Proposal selection 
The proposal chosen by the mine to be implemented was proposal 6. It was selected for the 
following reasons: 
? It required less work to implement than the other proposals. 
? It provided a safety factor to reduce the risk of project failure. 
? The underground conditions were improved. 
This proposal did, however, have the lowest reduction in secondary fan system power and air 
mass flow. The simulation boundaries were met, as seen in Table 5. 
 
3.4 Implementation  
3.4.1 Preamble 
In this section, the optimised proposal is selected and implemented. The results are evaluated, 
and the simulation is validated. This section follows the steps in Figure 10. 
3.4.2 Implementation of the optimised solution 
The implementation plan for Proposal 6 was given to the mining construction teams, and the 
proposal was implemented. Figure 19 shows a basic layout for Proposal 6 and only illustrates the 
operational changes made to the fans in the system. There were also changes suggested on the 
other ventilation control devices such as seals and doors. Figure 20 illustrates the implementation 
plan for Proposal 6. 
3.4.3 Evaluation of the implemented solution 
A ventilation audit was conducted after implementing the proposal, and the results are shown in 
Table 15. It can be seen that no data could be obtained for the 2010L SWW9 working area, which 
was a result of Fall of Ground (FOG) in the area making it unsafe to access the area. The mine 
ventilation department completed an in-stope ventilation audit after the implementation of the 
proposal. This audit showed that the new secondary fan system provided conditions which are 
conducive to safe and productive work. Appendix C shows the in-stope audit layout. In addition 
to the electrical cost savings, the cost of maintenance of the new fans will be less than that of the 
730 kW booster fan. 
55 
 
 
Figure 20: Proposal 6 operational changes. 
 
Table 15: Proposal implementation results. 
KPI Unit Simulation Results 
2010L SWW9 air mass flow kg/s 31.3 - 
2010L SWW9 wet-bulb temperature °C 30.4 - 
2010L SWW9 dry-bulb temperature °C 32.6 - 
2010L SWW12 air mass flow kg/s 51.5 37.5 
2010L SWW12 wet-bulb temperature °C 30.0 31.5 
2010L SWW12 dry-bulb temperature °C 33.5 32.9 
Secondary fan system power demand kW 219.0 422.1 
Air mass flow through the secondary fan system kg/s 102.6 93.5 
1940L return air mass flow kg/s 26.0 24.0 
Electricity cost savings R/yr 3 309 318 1 884 484 
 
56 
 
3.5 Interpretation of results 
3.5.1 Preamble 
A few differences were noted after the implementation of the proposal. This section discussed the 
results, discrepancies, and the consequence thereof. 
3.5.2 Discrepancies 
There are a few discrepancies in the results when compared to the simulation results. During the 
audit, it was found that the proposal was not implemented exactly as instructed. This was a result 
of difficulties that were faced by the mine ventilation construction team. The differences are 
indicated in Figure 21.   
 
Figure 21: Implementation differences. 
 
57 
 
Even though the project was not implemented exactly as set out, the results were still positive. 
The working conditions were all conducive to safe and productive work, and all the optimisation 
objectives for this project, listed below, were met: 
? Reduce secondary fan system power demand, 
? Reduce air mass flow through the secondary fan system, and 
? Sustained environmental conditions in working areas. 
The first point in Figure 21 refers to the 730 kW booster fan that was not completely removed. In 
the implementation plan, the booster fan was to be removed entirely. By not removing the booster 
fan, the open RAW cross-section area was reduced from 24.4 m2 to 4.3 m2.  
As seen in Figure 6, the increased resistance of the system resulted in increased static pressure 
of the fans and a decreased volume flow. This would also result in an increase in power 
consumption by the fans. This coincides with the larger power usage measured in Table 15. It is 
also possible that the system has higher resistance between the secondary fan system and 4#, 
further increasing the pressure. The larger power usage could also be the result of an incorrect 
assumption of efficiency when the fan was measured. 
The other three points represent locations where seals were not implemented correctly. This leads 
to air being extracted from incorrect places which reduces the air-flow through the working areas. 
Point 3 and 4 could not be sealed as a result of dangerous working conditions due to FOG in the 
area. 
3.5.3 Validation of simulation results 
To validate the simulation results, the actual measured data after the proposal implementation 
can be compared to the simulation results. The MAE method was used to determine the accuracy 
and it was found that the simulation predicted the implementation results with an error of 21.3%. 
A full comparison is shown in Appendix B.  
21.3% error may seem like a mediocre result, but the proposal was not implemented exactly as 
instructed. When the simulation model is adjusted to the differences shown in Figure 21, the new 
results have an error of 11.7% with the MAE method. A full comparison is shown in Appendix B.  
The average error of the simulation could be a result of the potential for error when following the 
methodology or by the potential for error of the data gathering techniques: 
58 
 
? The vane anemometer used to measure airspeed has an inaccuracy of between  
1% and 5% [55]. 
? The traverse method used for measuring airflow has an inaccuracy of between  
5% and 23% [56]. 
Considering this significant potential for error, the resulting average error of 11.7% is acceptable. 
This validates using this scalable method on deep-level gold mines’ secondary fan systems where 
this level of accuracy is required. 
3.6 Summary  
This section provides a summary of the results of applying the methodology to a case study. 
In Section 3.2, a secondary fan system simulation was developed and calibrated. The simulation 
was verified and had an error of less than 5%. 
In Section 3.3, the simulation was used to apply optimisation techniques to the model. 
Optimisation proposals were simulated and compared.  
In Section 3.4, an optimised proposal was selected and implemented on the secondary fan 
system of Mine A. A ventilation audit was completed on the system after the project was 
implemented and the results were compared to the simulation results. The results showed that 
the project was successful, and that the simulation model had an accuracy of 11.7%. 
  
59 
 
CHAPTER 4 CONCLUSION 
4.1 Summary 
The literature on mine ventilation optimisation with the use of simulation was analysed and it was 
found that there is limited literature on the optimisation of secondary fan systems. There is no 
literature on the optimisation of deep-level gold mine secondary fan systems, as seen in Table 1.  
The objectives of this study were to: 
? Develop a method for the optimisation of deep-level gold mine secondary fan systems 
with the aid of simulation. 
? Evaluate the feasibility of the developed methodology by implementing it on a case study. 
In Chapter 2, a methodology was introduced. The methodology was developed by adapting a 
study by Nel, Vosloo and Mathews [46] and applying it to a deep-level gold mine secondary fan 
system. This methodology was summarised into four stages: 
? Gathering of information,  
? Construction of the model,  
? Calibration and verification of the model, and  
? Use the model to optimise the system. 
This optimised solution was then implemented on a secondary fan system. Results of the 
implementation are used to verify the methodology. 
In Chapter 3, this methodology was applied to a case study on a deep-level gold mine. Mine A 
has a total of 6 working levels with the deepest at 2,010 meters underground. This mine had a 
730 kW booster fan which was used in a secondary fan system to ventilate various working areas 
on the deepest level of the mine. This booster fan was extracting more air than required, as the 
mine was downscaling the working areas impacted by the fan. This secondary fan system was 
selected for optimisation. 
A baseline simulation was constructed in PTB, according to the methodology, with an average 
error of less than 5%. This simulation was used to optimise the secondary fan system and six 
proposals were presented to the mine. These proposals applied the different optimisation 
techniques from the literature to the secondary fan system. The KPIs for each simulation were 
60 
 
compared to the simulation boundaries to ensure that the proposal met the desired optimisation 
objectives. These proposals were then compared to each other and presented to the mine. 
Proposal 6 was selected for implementation as it met the optimisation objectives, was not a lot of 
work to implement, and provided a safety factor. This proposal included the following optimisation 
techniques: 
? Fan optimisation  
? Improved air utilisation 
This proposal showed improved environmental conditions with a reduction in the cost of 
ventilating that area. 
After the implementation, a ventilation audit was conducted on the area impacted by the 
secondary fan system. Results showed that the simulation proposal had an average error of 
21.3%. The audit also showed that the selected proposal was not implemented exactly as 
instructed. The discrepancies in the implementation caused discrepancies in the data and the 
poor average error.  
The average error was improved to 11.7% by modifying the simulation proposal to how it was 
actually implemented. This verified that it is feasible to use the developed methodology for deep-
level gold mines’ secondary fan systems that require optimisation solutions with this level of 
accuracy.  
4.2 Recommendations 
When the methodology was applied to a case study, a few limitations were identified. The 
methodology was applied to a specific mine secondary fan system. The specifications of this case 
study are listed below: 
? Deep-level gold mine ventilation system, 
? Only the area impacted by the secondary system was simulated, 
? Manual ventilation measurements were used to benchmark the system, 
? VOD optimisation was not applied to the secondary fan system, 
? PTB was selected to simulate the network. 
The methodology developed could be applied to other mining operations to verify it further. 
Further verification is required for the following aspects of the methodology: 
61 
 
? Other types of mining operations, 
? Impact of the optimisation on the entire mine ventilation network, 
? Use of SCADA data for simulation calibration, 
? Application of VOD on secondary fan systems, 
? Use of other mine ventilation simulation software. 
Further recommendations would be to complete a detailed cost analysis. The detailed cost 
analysis will improve the decision-making process when selecting an optimised proposal. The 
pricing of components can be included to determine the cost efficiency thereof.  
 
  
62 
 
BIBLIOGRAPHY  
[1] “Discovery of Gold in 1884,” 2016. [Online]. Available: 
https://www.sahistory.org.za/article/discovery-gold-1884. [Accessed: 21-Apr-2019]. 
[2] This is Gold, “The state of the gold industry,” 2018. [Online]. Available: 
http://www.thisisgold.co.za/#fact-sheets. 
[3] G. Edgar, D. Plessis, L. Liebenberg, and E. H. Mathews, “The use of variable speed drives 
for cost-effective energy savings in South African mine cooling systems,” Appl. Energy, vol. 
111, no. 2013, pp. 16–27, 2013. 
[4] Chamber of Mines of South Africa, “Mine SA 2017 Facts and Figures Pocketbook,” 2017. 
[5] J. Burton, “Which Country Is The World’s Largest Producer Of Gold?,” 2018. [Online]. 
Available: https://www.worldatlas.com/articles/top-14-gold-producers-in-the-world.html. 
[6] D. Nell, “Optimising production through improving the efficiency of mine compressed air 
networks with limited infrastructure,” M.Eng. Dissertation, North-West University, 2017. 
[7] Statista, “Global mine production of gold from 2005 to 2018 (in metric tons),” 2019. [Online]. 
Available: https://www.statista.com/statistics/238414/global-gold-production-since-2005/. 
[8] Index mundi, “Gold Monthly Price - Rand per Troy Ounce,” 2018. [Online]. Available: 
https://www.indexmundi.com/commodities/?commodity=gold&months=360&currency=zar
. [Accessed: 21-Apr-2019]. 
[9] P. N. Neingo and T. Tholana, “Trends in productivity in the South African gold mining 
industry,” J. South. African Inst. Min. Metall., vol. 6, no. November 2015, pp. 8–11, 2016. 
[10] “ESKOM,” 2019. [Online]. Available: http://www.eskom.co.za/Pages/Landing.aspx. 
[Accessed: 22-Apr-2019]. 
[11] Eskom, “The Energy Efficiency series,” 2010. [Online]. Available: 
http://www.eskom.co.za/sites/idm/Documents/121040ESKD_Mining_Brochure_paths.pdf. 
[12] W. C. Kukard, “Research on reducing costs of underground ventilation networks in South 
African mines,” M.Eng. Dissertation, Northwest University, 2006. 
[13] A. Chatterjee, L. Zhang, and X. Xia, “Optimization of mine ventilation fan speeds according 
63 
 
to ventilation on demand and time of use tariff,” Appl. Energy, vol. 146, pp. 65–73, 2015. 
[14] C. Swart, “Optimising the operation of underground mine refrigeration plants and ventilation 
fans for minimum costs,” PhD Thesis, Potchefstroomse Univeristeit vir Chritelike Hoer 
Onderwys, 2003. 
[15] V. R. Babu, T. Maity, and H. Prasad, “Energy saving techniques for ventilation fans used 
in underground coal mines—A survey,” J. Min. Sci., vol. 51, no. 5, pp. 1001–1008, 2015. 
[16] C. Ö. Karacan, “Development and application of reservoir models and artificial neural 
networks for optimizing ventilation air requirements in development mining of coal seams,” 
Int. J. Coal Geol., vol. 72, no. 3–4, pp. 221–239, 2007. 
[17] G. Wei, “Optimization of Mine Ventilation System Based on Bionics Algorithm,” Procedia 
Eng., vol. 26, pp. 1614–1619, 2011. 
[18] J. Cheng, “Assessment of mine ventilation system reliability using random simulation 
method,” Environ. Eng. Manag. J., vol. 15, no. 4, pp. 841–850, 2016. 
[19] J. C. Tien, Practical Mine Ventialtion Engineering, no. 1. Interec Publishing Company, 
1999. 
[20] E. I. Acuña and I. S. Lowndes, “A Review of Primary Mine Ventilation System Optimization,” 
Interfaces (Providence)., vol. 44, no. 2, pp. 163–175, 2014. 
[21] M. C. ?uvar, D. Cioclea, V. Arad, C. Lupu, and N. I. Vlasin, “Method for improving the 
management of mine ventilation networks after an explosion,” Enviromental Eng. Manag. 
J., vol. 16, no. 6, pp. 1373–1381, 2017. 
[22] M. Biffi, D. Stanton, H. Rose, and D. Pienaar, “Ventilation strategies to meet future needs 
of the South African platinum industry,” J. South. African Inst. Min. Metall., vol. 107, no. 1, 
pp. 59–66, 2007. 
[23] D. J. de Villiers, M. J. Mathews, P. Maré, M. Kleingeld, and D. Arndt, “Evaluating the impact 
of auxiliary fan practices on localised subsurface ventilation,” Int. J. Min. Sci. Technol., pp. 
2–10, 2019. 
[24] P. du Toit, Intermediate Certificate in Mine Enviromental Control - Practical Course in MEC 
workbook, Revision 4. The Mine Ventilation Society of South Africa, 2007. 
64 
 
[25] Twin City Fan, “Fan Engineering,” 2018. [Online]. Available: https://www.tcf.com/wp-
content/uploads/2018/06/Understanding-Fan-Curves-FE-2000.pdf. [Accessed: 15-Sep-
2019]. 
[26] R. Papar, A. Szady, V. Martin, and A. Mckane, “Increasing energy efficiency of mine 
ventilation systems,” Lawrence Berkeley Natl. Lab., pp. 611–617, 1999. 
[27] C. J. P. Venter, “The effect of controlling main ventilation fan inlet guide vanes for a deep 
level platinum mine to reduce electrical costs,” M.Eng. Dissertation, North-West University, 
2017. 
[28] S. Bluhm, R. G. Moreby, F. von Glehn, and C. Pascoe, “Life-of-Mine Ventilation and Copper 
Mine,” in The Australian Mine Ventilation Conference, 2013, no. July, pp. 1–3. 
[29] M.-C. ?uvar, V. Arad, C. Lupu, D. Cioclea, and N.-L. Vlasin, “Use of information technology 
in mining ventilation,” Recent Adv. Comput. Sci. Netw., pp. 104–111, 2013. 
[30] E. De Souza, “Optimization of complex mine ventilation systems with computer network 
modelling,” in IFAC Proceedings Volumes, 2007, p. Pages 323-329. 
[31] P. Maré, “Novel simulations for energy management of mine cooling systems,” PhD Thesis,  
North-West University, 2017. 
[32] A. Nikolaev, N. Alymenko, A. Kamenskikh, S. Fet, and V. Nikolaev, “Results of modelling 
of mine ventilation with air curtain installed into downcast and upcast shafts,” Perm J. Pet. 
Min. Eng., vol. 16, no. 3, pp. 291–300, 2017. 
[33] F. Wei, F. Zhu, and H. Lv, “The use of 3D simulation system in mine ventilation 
management,” Procedia Eng., vol. 26, pp. 1370–1379, 2011. 
[34] E. S. Menon, Transmission Pipeline Calculations and Simulations Manual. Elsevier Inc., 
2015. 
[35] “VentSimTM Design | VentSim.” [Online]. Available: https://ventsim.com/ventsim-design/. 
[Accessed: 12-Sep-2019]. 
[36] “Vuma|VUMA Software.” [Online]. Available: https://www.vuma3d.com/. [Accessed: 24-
Sep-2019]. 
[37] A. J. H. Nel, J. C. Vosloo, and M. J. Mathews, “Supplementary file for?: Evaluating complex 
65 
 
mine ventilation operational changes through simulations,” J. Energy South. Africa, vol. 29, 
no. 3. 
[38] B. M. Friedenstein, C. Cilliers, and J. van Rensburg, “Simulating operational improvements 
on mine compressed air systems,” South African J. Ind. Eng., vol. 29, no. November, pp. 
69–81, 2018. 
[39] J. Watkins, “Trade-off between simulation accuracy and complexity for mine compressed 
air systems,” M.Eng. Dissertation, North-West University, 2018. 
[40] X. S. Wu and E. Topuz, “Analysis of Mine Ventilation Systems Using Operations Research 
Methods,” Pergamon, vol. 5, no. 4, 1998. 
[41] C. Pritchard, “Methods to improve efficiency of mine ventilation systems,” Natl. Inst. Occup. 
Saf. Heal., pp. 1–5, 2008. 
[42] X. Nie, X. Wei, X. Li, and C. Lu, “Heat Treatment and Ventilation Optimization in a Deep 
Mine,” Adv. Civ. Eng., vol. 2018, pp. 1–12, 2018. 
[43] R. Webber-Youngman, “An intergrated approach towards the optimization of ventilation, 
air cooling and pumping requirements for hot mines,” PhD Thesis, North West University, 
2005. 
[44] E. Develo, M. Pillalamarry, and E. Garab, “Improving the ventilation system at Rosh Pinah 
zinc mine,” J. South. African Inst. Min. Metall., vol. 116, no. 4, pp. 301–305, 2016. 
[45] J. G. Zhang and C. Y. Suo, “Study of Coal Mine Ventilation System Optimization based on 
Ventsim,” MATEC Web Conf., vol. 44, 2016. 
[46] A. J. H. Nel, J. C. Vosloo, and M. J. Mathews, “Evaluating complex mine ventilation 
operational changes through simulation,” J. Energy South. Africa, vol. 29, no. 3, pp. 22–
32, 2018. 
[47] M. C. ?uvar, C. Lupu, V. Arad, D. Cioclea, V. M. P?sculescu, and N. Mija, “Computerized 
Simulation of Mine Ventilation Networks for Sustainable Decision Making Process,” 
Environ. Eng. Manag. J., vol. 13, no. 6, pp. 1445–1451, 2014. 
[48] E. A. Abdelaziz, R. Saidur, and S. Mekhilef, “A review on energy saving strategies in 
industrial sector,” Renew. Sustain. Energy Rev., vol. 15, no. 1, pp. 150–168, 2011. 
66 
 
[49] “SCADA Systems,” 2019. [Online]. Available: http://www.scadasystems.net/. [Accessed: 
06-Jul-2019]. 
[50] L. U. O. Wengui, X. I. E. Xianping, X. Houzao, C. U. I. Chao, Y. I. N. Xiaoyu, and S. U. 
Meiwei, “Reliability calculation of mine ventilation network,” Procedia Eng., vol. 84, no. 1, 
pp. 752–757, 2014. 
[51] D. S. Euler, “Application of ventilation management programs for improved mine safety,” 
Int. J. Min. Sci. Technol., vol. 27, no. 4, pp. 647–650, 2017. 
[52] R. C. W. Webber, R. M. Franz, W. M. Marx, and P. C. Schutte, “A review of local and 
international heat stress indices, standards and limits with reference to ultra-deep mining,” 
J. South African Inst. Min. Metall., vol. 103, no. 5, pp. 313–323, 2003. 
[53] P. Chang, G. Xu, F. Zhou, B. Mullins, S. Abishek, and D. Chalmers, “Minimizing DPM 
pollution in an underground mine by optimizing auxiliary ventilation systems using CFD,” 
Tunn. Undergr. Sp. Technol., vol. 87, no. November 2018, pp. 112–121, 2019. 
[54] Eskom, “Tarrifs & Charges 2019/2020,” 2020. [Online]. Available: 
http://www.eskom.co.za/CustomerCare/TariffsAndCharges/Pages/Tariffs_And_Charges.
aspx. [Accessed: 05-Sep-2019]. 
[55] K. Siren, G. Rosen, and P. V. Nielsen, Industrial Ventilation Design Handbook. Elsevier 
Inc., 2001. 
[56] I. D. Unwin and M. Phil, “The Measurement of Air Flow in British Coal Mines?: A Historical 
Review,” 2007. [Online]. Available: 
https://pdfs.semanticscholar.org/8b2f/070984e0e566f04406704bd13405c426c804.pdf. 
[Accessed: 19-Oct-2019]. 
[57] “Davis Anemometer | AMS Haden Instrument & Mining Services.” [Online]. Available: 
https://www.amshaden.co.za/shop/davis-anemometer/. [Accessed: 15-Sep-2019]. 
[58] D. S. Kingery, Introduction to mine ventilation principles and practices. Washington: Bureau 
of Mines, 1960. 
[59] “Testo L-type Pitot Tube 40" - IvyTools.com,” 2019. [Online]. Available: 
https://www.ivytools.com/Testo-L-type-Pitot-Tube-40-p/0635-2345.htm. [Accessed: 15-
Sep-2019]. 
67 
 
[60] “GPB 3300 Barometer | AMS Haden Instrument & Mining Services.” [Online]. 
Available: https://www.amshaden.co.za/shop/gpb-3300-barometer/. [Accessed: 15-Sep-
2019]. 
 
  
68 
 
ANNEXURES … 
  
69 
 
APPENDIX A VENTILATION MEASUREMENT 
Velocity and flow area are measured to determine the airflow in haulages and air pipes. The Davis 
anemometer shown in Figure 22 is used to measure the average air velocity in m/s of a haulage 
or duct. It is traversed through the haulage, as seen in Figure 23, while being timed with a 
stopwatch. The Davis anemometer gives a measurement in meters (m) which is then divided by 
the measurement time in seconds (s) to give an average air velocity reading in m/s. 
The air velocity range of a Davis anemometer is 0.5 – 30 m/s and each vane comes with a 
calibration certificate which is used to adjust the readings according to an adjustment factor. The 
measuring time should not be less than 50 seconds, and the anemometer should always be held 
parallel to the airflow without obstruction. The measurements should be taken in laminar flow 
areas, to ensure accuracy, and far away from bends or obstructions in the haulage which could 
cause turbulence. For air velocities lower than 0.5 m/s the puff-puff method can be used. 
 
Figure 22: Davis anemometer [57]. 
70 
 
 
Figure 23: Traversing haulage with Davis anemometer [58]. 
The flow area measured in m2 is multiplied by the air velocity to determine the airflow quantity 
measured in m3/s. A laser distance meter, shown in Figure 24, or measuring tape is used to 
determine flow area. For accurate results, multiple measurements are done of the height and 
width of the same area and then averaged.  
 
Figure 24: Laser distance meter. 
71 
 
The same principles are applied for the measurement of air ducts. The air velocity needs to be 
adjusted according to Equation 2 and the factors for the traverse method of ducts: 
 ????????? =  ? × ?  (2) 
Where: 
?????????  Adjusted air velocity [m/s] 
?   Measured air velocity [m/s] 
?   Correction factor [-] 
? Intake side factor – 0.92 
? Outlet side factor – 0.82 
A centre spot measurement can be taken and should be adjusted according to the following 
factors: 
? Intake side factor – 0.87 
? Outlet side factor – 0.87 
The pitot tube method can be used for places where it is not possible to access the inlet or outlet 
of a duct system. The pitot tube method requires two instruments, a pitot tube and a vertical 
manometer. Figure 25 shows an L-type pitot tube which is used to measure the dynamic pressure 
inside an air duct. The L-type pitot is used in conjunction with a vertical manometer shown in 
Figure 26. 
 
Figure 25: Pitot tube [59]. 
 
72 
 
 
Figure 26: Vertical manometer [59]. 
Rubber tubing is used to connect the manometer and the pitot tube. Average velocity pressure is 
obtained by taking readings inside the duct at distances prescribed by the MVSSA shown in 
Appendix A. The air velocity can be calculated from the average velocity pressure using  
Equation 3. 
 (3) 
2 × ??
? =  ?  
?
Where: 
?   Air velocity [m/s] 
??   Velocity pressure [Pa] 
?   Air density [kg/m3] 
 
A whirling hygrometer is used to measure the temperature and humidity. This instrument, shown 
in Figure 27, measures dry-bulb temperature and wet-bulb temperature. The top bulb 
thermometer measures the dry-bulb temperature. The bottom bulb thermometer has a wet wick 
73 
 
covering its bulb and measures the wet-bulb temperature. Before the temperatures are recorded, 
the instrument is spun at a rate of 3 revolutions per minute for 30 second. The measurement is 
taken in the middle of the flow area and away from any heat sources.  
 
Figure 27: Whirling hygrometer. 
The barometer shown in Figure 28 measures the barometric pressure at each measurement 
point. The static pressure of an underground booster fan can be measured by taking a barometric 
pressure reading on the inlet and outlet side of the fan and subtracting the two. A vertical 
manometer can also be used to measure the static pressure across a fan. 
74 
 
 
Figure 28: Barometer [60]. 
 
To measure the air-flow in a duct, the velocity profile has to be measured. This can be averaged 
to obtain an average velocity pressure inside the duct which can be converted to a velocity using 
Equation 3. 
 
The velocity profile is measured on two traverses shown in Figure 29. The weights are hung 
around the ducts to determine the location where holes should be drilled in the duct as some 
ducts do not have predetermined holes. 
 
The depths of the measuring points can be determined according to Figure 30 and Table 16. 
 
75 
 
 
Figure 29: Pitot tube traverse lines [24]. 
 
 
Figure 30: Pitot tube measuring points [24]. 
 
 
76 
 
Table 16: Pitot tube measuring points [24]. 
Point General 
number case 
1 0.026 d 
2 0.082 d 
3 0.146 d 
4 0.226 d 
5 0.342 d 
6 0.658 d 
7 0.774 d 
8 0.854 d 
9 0.918 d 
10 0.974 d 
 
The puff-puff method can be used to measure air-flow velocities lower than the vane anemometer 
capabilities. Dust is timed over a measured distance to obtain an airflow velocity. This technique 
is shown in Figure 31. 
 
Figure 31: Puff-puff method [24]. 
 
  
77 
 
APPENDIX B SIMULATION ACCURACY 
The accuracy of the simulation was determined with the MAE method according to Equation 1. 
The measured data points are compared to the calibrated simulation results in Table 17. 
Table 17: Baseline simulation accuracy. 
Point Type Actual Simulation Difference (%) 
Air mass flow 11.2 11.44 2.14 
Main walkway from station  Wetbulb temperature 21 21.4 1.90 
Drybulb 28 28.3 1.07 
Station Pressure 107.76 107.68 0.07 
Air mass flow 19.82 20.51 3.48 
After leaking 75kw fan ducts Wetbulb  28 27.62 1.36 
Drybulb 32 30.67 4.16 
Air mass flow 24.07 20.51 14.79 
After SWW6 Wetbulb  29.5 27.67 6.20 
Drybulb 32 30.89 3.47 
Air mass flow 24.78 20.51 17.23 
Before SWW8 Wetbulb  29 27.73 4.38 
Drybulb 32 31.17 2.59 
Air mass flow 28.13 27.37 2.70 
From 1940 RAW Wetbulb  32 28.82 9.94 
Drybulb 36.5 34.15 6.44 
Air mass flow 24.34 24.12 0.90 
After SWW9  Wetbulb  29 27.6 4.83 
Drybulb 32 31.97 0.09 
Air mass flow 42.22 40.37 4.38 
SWW9 Working area Wetbulb  32 28.83 9.91 
Drybulb 36.5 34.3 6.03 
Air mass flow 18.21 18.12 0.49 
After SWW10  Wetbulb  29 27.63 4.72 
Drybulb 32 32.1 0.31 
Air mass flow 11.05 13.18 19.28 
Before SWW12  Wetbulb  30 27.85 7.17 
Drybulb 32 33.87 5.84 
Air mass flow 17.46 17.66 1.15 
SWW12 Crossover intake Wetbulb  29.5 27.86 5.56 
Drybulb 33 33.92 2.79 
Air mass flow 40.82 40.9 0.20 
SWW12 Working area Wetbulb  29 30.03 3.55 
Drybulb 33 33.21 0.64 
78 
 
Air mass flow 57.05 54.08 5.21 
After SWW12 Wetbulb  30 29.55 1.50 
Drybulb 34 34.41 1.21 
Air mass flow 94.25 93.13 1.19 
4# return left Wetbulb  30.5 29.15 4.43 
Drybulb 34 34.34 1.00 
Air mass flow 86.22 84.95 1.47 
4# return right Wetbulb  31 29.12 6.06 
Drybulb 36.5 34.59 5.23 
Pressure before booster fan Pressure 105.21 105.83 0.59 
Booster fan pressure Pressure 3 2.9 3.33 
MAE       4.24 
The Proposal 6 data points are compared to the actual measured results in Table 18. The points 
which could not be measured are marked with grey. 
Table 18: Proposal 6 simulation results. 
Actual 
Point Type results Simulation Difference (%) 
Air mass flow 12.5 7.35 41.20 
Main walkway from station  Wetbulb temperature 26.4 21.6 18.18 
Drybulb 29.4 28.7 2.38 
Station Pressure 107.63 107.8 0.16 
Air mass flow 12 7.36 38.67 
After leaking 75kw fan ducts Wetbulb  27.9 26.73 4.19 
Drybulb 31 30.36 2.06 
Air mass flow 15.4 7.36 52.21 
After SWW6 Wetbulb  27.9 26.87 3.69 
Drybulb 31 30.95 0.16 
Air mass flow 15.4 7.36 52.21 
Before SWW8 Wetbulb  27.9 27.05 3.05 
Drybulb 31 31.69 2.23 
Air mass flow       
From 1940 RAW Wetbulb        
Drybulb       
Air mass flow 13.3 1.6 87.97 
After SWW9  Wetbulb  28.9 30.84 6.71 
Drybulb 31.1 37.53 20.68 
Air mass flow       
SWW9 Working area Wetbulb        
Drybulb       
After SWW10  Air mass flow 20 4.84 75.80 
79 
 
Wetbulb  30 30.9 3.00 
Drybulb 31.8 37.79 18.84 
Air mass flow 20 3.59 82.05 
Before SWW12  Wetbulb  30 30.21 0.70 
Drybulb 31.8 34.66 8.99 
Air mass flow       
SWW12 Crossover intake Wetbulb        
Drybulb       
Air mass flow 31.2 51.49 65.03 
SWW12 Working area Wetbulb  31.5 29.96 4.89 
Drybulb 32.9 33.48 1.76 
Air mass flow 51.2 55.08 7.58 
After SWW12 Wetbulb  31 30.24 2.45 
Drybulb 32 34.81 8.78 
Air mass flow 93.5 102.6 9.73 
4# return total Wetbulb  33 30.7 6.97 
Drybulb 36.1 36.56 1.27 
Booster fans power Power 422.1 219 48.12 
MAE       21.30 
 
The adjusted simulation results are compared to the actual measured results in Table 19. The 
points which could not be measured are marked with grey. 
Table 19: Simulation results as implemented. 
Point Type Actual results Simulation Difference (%) 
Air mass flow 12.5 8.63 30.96 
Main walkway from station  Wetbulb temperature 26.4 21.36 19.09 
Drybulb 29.4 27.89 5.14 
Station Pressure 107.63 107.82 0.18 
Air mass flow 12 8.64 28.00 
After leaking 75kw fan ducts Wetbulb  27.9 26.13 6.34 
Drybulb 31 30.03 3.13 
Air mass flow 15.4 8.64 43.90 
After SWW6 Wetbulb  27.9 26.22 6.02 
Drybulb 31 30.39 1.97 
Air mass flow 15.4 8.64 43.90 
Before SWW8 Wetbulb  27.9 26.39 5.41 
Drybulb 31 31.06 0.19 
Air mass flow       
From 1940 RAW 
Wetbulb        
80 
 
Drybulb       
Air mass flow 13.3 11.91 10.45 
After SWW9  Wetbulb  28.9 28.56 1.18 
Drybulb 31.1 34.46 10.80 
Air mass flow       
SWW9 Working area Wetbulb        
Drybulb       
Air mass flow 20 18.81 5.95 
After SWW10  Wetbulb  30 28.57 4.77 
Drybulb 31.8 34.54 8.62 
Air mass flow 20 20.52 2.60 
Before SWW12  Wetbulb  30 28.79 4.03 
Drybulb 31.8 35.26 10.88 
Air mass flow       
SWW12 Crossover intake Wetbulb        
Drybulb       
Air mass flow 31.2 37.5 20.19 
SWW12 Working area Wetbulb  31.5 30.78 2.29 
Drybulb 32.9 33.75 2.58 
Air mass flow 51.2 58.01 13.30 
After SWW12 Wetbulb  31 30.28 2.32 
Drybulb 32 35.22 10.06 
Air mass flow 93.5 105.52 12.86 
4# return total Wetbulb  33 30.79 6.70 
Drybulb 36.1 37.09 2.74 
Booster fans power Power 422.1 219 48.12 
MAE       11.71 
 
  
81 
 
APPENDIX C SWW9 IN-STOPE VENTILATION  
Figure 32 illustrates a layout of SWW9 working area. It can be seen that the air velocities 
measured inside the working area are 0.4 m/s or above. This makes them compliant to the 
MVSSA limits [52].  
 
Figure 32: SWW9 in-stope ventilation. 
82