Efficient planning of energy production and maintenance of large-scale combined heat and power plants

In this study, an efficient optimization framework is presented for the simultaneous planning of energy production and maintenance in combined heat and power plants, and applied in the largest coal-fired cogeneration plant of Kazakhstan. In brief, the proposed optimization model considers: (i) unit commitment constraints for boilers and turbines; (ii) minimum and maximum runtimes as well as minimum idle times for boilers and turbines; (iii) bounds on the operating levels for boilers and turbines within desired operating regions; (iv) extreme operating regions for turbines; (v) energy balances for turbines; (vi) total electricity and heat balances for satisfying the corresponding demands for electricity and heat (for each heat network); and (vii) maintenance tasks for units that must occur within given flexible time-windows. The minimization of the annual total cost of the cogeneration plant constitutes the optimization goal here, and consists of startup and shutdown costs, fixed operating and fuel costs, maintenance costs, and penalties for deviation from heat and electricity demands, and penalties for turbines for operating outside the desired operating regions. An extensive data analysis of historical data has been performed to extract the necessary input data. In comparison to the implemented industrial solution that follows a predefined maintenance policy, the solutions derived by the proposed approach achieve reductions in annual total cost more than 21% and completely avoid turbines operation outside their desired operating regions. Our solutions report substantial reductions in startup/shutdown, fuel and fixed operating costs (about 85%, 15%, and 13%, respectively). The comparative case study clearly demonstrates that the proposed approach is an effective means for generating optimal energy production and maintenance plans, enhancing significantly the resource and energy efficiency of the plant. Importantly, the proposed optimization framework could be readily applied to other cogeneration plants that have a similar plant structure.