Methodology development for identifying improvement opportunities using discrete simulation modelling

Abstract

The Central Workshops' main responsibility, within an integrated steel manufacturer, is to manufacture and recondition plant-related equipment. In recent years, the Central Workshops have been reorganised and down-scaled on various occasions. This was done without adequate change management, leading to operational deficiencies with long-term and immediate consequences. These consequences include increased manufacturing costs, not adhering to deadlines and not being profitable, which could lead to the closure of the Central Workshops. Improvement opportunities must be identified to ensure the Central Workshops' profitability. Due to the unpredictability of job shop operations, and the fluctuations in production demand, this task is challenging. The unpredictability stems from the frequent arrival of varied brake-down orders from integral plants within the steel manufacturer. These challenges complicate the identification of improvement opportunities. A strategic approach to addressing this complexity entails initially focusing on one specific job. A methodology is developed that aims at identifying improvement opportunities. By addressing bottlenecks and inefficiencies with the use of the developed methodology, operations can be enhanced, enabling effective operations and increasing customer satisfaction, which will in turn have a positive financial impact. The developed methodology is built on a sequence of steps, commencing with an understanding of the system under consideration. The required data is identified, cleansed, and subjected to analysis through the use of relevant statistical analysis tools. Distribution functions are identified that reflect the real-world operations of the system under consideration. The distribution fits are evaluated by utilising Exploratory Data Analysis (EDA) techniques to establish a hypothesis and evaluate possible distribution fits through the use of informative graphs. The relevant hypothesis testing methods are used to formalise the interpretation of the 'ideal' distribution fits. Thereafter, discrete simulation modelling software is used to develop models that reflect the operation under consideration. Firstly, the current state is modelled, whereafter an iterative process is followed to identify bottlenecks within the system and consequently identify improvement opportunities. These bottlenecks are identified through the application of the Theory of Constraints (TOC) principles and a thorough assessment of workstation utilisation. The utilisation result not only provides insights into the system's operational efficiency but also serves as a fundamental performance measure. To further evaluate the iterative model development, additional key performance measures, namely throughput and production costs, were employed. The iterative process continued until satisfactory results were achieved. With effective and adequate results obtained, the proposed methodology is identified as correct and adequate. Although the methodology is developed to analyse a single job within the job shop environment, there is a promising avenue for future research to extend its applicability.