Stochastic real-time operation control of a combined heat and power (CHP) system under uncertainty

• Published in: Energy conversion and management Vol. 216; p. 112916
• Main Authors: Olympios, Andreas V., Le Brun, Niccolò, Acha, Salvador, Shah, Nilay, Markides, Christos N.
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 15.07.2020; Elsevier Science Ltd
• Subjects: Algorithms; architecture; Cogeneration; Combined heat and power; Computer simulation; Controllers; Cost control; Decision making; design; Design optimization; Distributed generation; Electricity; electricity costs; Electricity pricing; Energy conservation; energy conversion; Energy costs; Environmental objective; Heat; Intervals; Multiple objective analysis; Objective function; objectives; operating costs; Optimization; Optimization techniques; prediction; Predictions; prices; purchasing; Real time operation; Real-time control; risk; Risk management; Risk reduction; Sizing; Stochastic programming; supermarkets; system optimization; Tradeoffs; Uncertainty; Real-time control; Uncertainty; Distributed generation; Combined heat and power; Optimization; Stochastic programming
• ISSN: 0196-8904; 1879-2227; 1879-2227
• DOI: 10.1016/j.enconman.2020.112916

In this paper we present an effort to design and apply a multi-objective real-time operation controller to a combined heat and power (CHP) system, while considering explicitly the risk-return trade-offs arising from the uncertainty in the price of exported electricity. Although extensive research has been performed on theoretically optimizing the design, sizing and operation of CHP systems, less effort has been devoted to an understanding of the practical challenges and the effects of uncertainty in implementing advanced algorithms in real-world applications. In this work, a two-stage control architecture is proposed which applies an optimization framework to a real CHP operation application involving intelligent communication between two controllers to monitor and control the engine continuously. Since deterministic approaches that involve no measure of uncertainty provide limited insight to decision-makers, the methodology then proceeds to develop a stochastic optimization technique which considers risk within the optimization problem. The uncertainty in the forecasted electricity price is quantified by using the forecasting model’s residuals to generate prediction intervals around each forecasted electricity price. The novelty of the proposed tool lies in the use of these prediction intervals to formulate a bi-objective function that represents a compromise between maximizing the expected savings and minimizing the associated risk, while satisfying specified environmental objectives. This allows decision-makers to operate CHP systems according to the risk they are willing to take. The actual operation costs during a 40–day trial period resulting from the installation of the dynamic controller on an existing CHP engine that provides electricity and heat to a supermarket are presented. Results demonstrate that the forecasted electricity price almost always falls within the developed prediction intervals, achieving savings of 23% on energy costs against only utilizing a boiler and buying electricity from the grid.Whole-day simulations using data from the supermarket are then conducted for representative days of the year to demonstrate that the proposed approach can provide even higher savings – more than 35% on energy costs – while revealing the risk-return trade-off arising from different operation strategies. Conservative choices emerging from the stochastic approach are shown to reduce risk by 7–9% at the expense of much smaller reductions of 1–3% in expected savings, while proper sizing of the CHP engine leads to risk reductions of more than 20%.