Energy recovery maximisation modelling subject to constrained cooling

Abstract

The heat rejection cycle in a metal production plant involves the transfer of thermal energy from warm (1200°C –1500°C) flue gas to a closed cooling water cycle. This closed cycle is operated to, furthermore, remove the absorbed thermal energy from the circulating cooling water.Monitoring and controlling this heat-absorption-and-rejection cycle is critical. Process stability requires maintaining the temperatures of both the flue gas and cooling water in respective design operational envelopes and is also critical for preventing corrosion in the gas cooler, and overall equipment integrity. To maintain both temperature envelopes, the absorbed waste heat in the cooling water is transferred to the environment via an air-cooled heat exchanger (ACHX) and/or used to generate power through a Rankine cycle (PGEN).

Several studies have focused on evaluating the performance of waste heat recovery Rankine cycles and air-cooled heat exchangers, mostly considering these systems in isolation. These studies typically aim to maximise parameters such as heat recovery efficiency. Relevant studies on combined systems have also demonstrated the use of design conditions and non-fluctuating heat sources. The combined control of power generation and air-cooled heat exchangers; however, specifically the dynamic distribution of flow to increase energy recovery while sustaining the operational envelope with a fluctuating heat source, remains underexplored. Addressing this gap can improve the operational efficiencies of combined heat and power systems that are commonly deployed in industry. Furthermore, filling this gap also provides insights into practical solutions for enhancing power generation performance and reducing operational costs. To achieve this objective, representative models for the PGEN, ACHX, gas cooler, and heat loss components were formulated. Internal control philosophies have also been formulated for controllable components. This ensures that phenomena such as trips during periods of low heat availability and constrained restart procedures are accounted for over the time horizon of the plant-specific waste-heat profile. These models are then used in the novel control philosophies of this thesis to maximise net power generation without breaching the operational temperature envelope.

The first novel philosophy improves on the current philosophy by providing a framework for the dynamic flow adjustment between the PGEN and ACHX with the aim of maximising net power generation. This philosophy maximises the power generation for the available heat in the cooling water. The second novel philosophy improves on the first by solving for increased available heat in the cooling water through adjustment of the cooling water flow rate. This is followed by dynamically adjusting the flow distribution between the PGEN and ACHX. This philosophy provides the framework that ensures maximum power generation for the available heat in the flue gas. The third novel formulation improves on the second by adjusting the approach to allow for athermal energy storage system to dampen the heat source fluctuations and reduce the number of low heat availability trips the PGEN experiences. The approach from this philosophy maximises power generation over the time horizon by managing the charging and discharging of the thermal energy storage with the cooling water exiting the gas cooler.

The frameworks provided by the novel control philosophies is solved for on a historical dataset and increased the energy transferred from flue gas to cooling water from 92.24% to 99.57%. The average power generation over the time horizon also increased from 1.134 MW to 1.485 MW when applying the first philosophy. The second and third philosophy resulted in average power generation of 1.615 and 1.631 MW. The energy recovery ratio, defined as the ratio of electrical energy generated to the available energy in the flue gas, increased from 8.66% for the Work’s philosophy to 11.34%, 12.38%, and 12.96% for the first, second, and third novel philosophies, respectively. These results were obtained through simulation-based evaluations of the proposed approaches.