Short-term scheduling of gas-fired CHP plant with thermal storage using optimization algorithm and forecasting models. (1st March 2021)

• Record Type: Journal Article
• Authors: Żymełka, Piotr; Szega, Marcin
• Is Part Of: Energy conversion and management. Volume 231(2021)
• Publication Date: 2021-03-01
• Subjects: Combined heat and power -- Scheduling -- Optimization -- Forecasting -- Simulation modelling; Direct energy conversion -- Periodicals; Energy storage -- Periodicals; Energy transfer -- Periodicals; Énergie -- Conversion directe -- Périodiques; Direct energy conversion; Periodicals; 621.3105
• DOI: 10.1016/j.enconman.2021.113860
• Languages: English
• ISSNs: 0196-8904

Abstract

Highlights: The concept of a computer-based tool for production optimization in the CHP plant of day-ahead planning is presented. The mathematical model of the gas-fired CHP plant was developed. The MLP (Multilayer Perceptron) forecasting model for electricity price was used. The artificial neural network was developed to predict the heat load in the DHN. The evolutionary method was used for solving the optimization problem. Abstract: Accurate production planning is an important aspect of the combined heat and power plants operating on the electricity market. The complexity of production planning and scheduling depends mainly on the scale of the power system. Nowadays, the role of modern computer systems seems to be crucial and significantly affects optimal production planning in CHP plants. The production scheduling process must take into account the relationship between the production of heat and electricity in cogeneration units. An important aspect of optimal scheduling of power systems is precisely forecasting heat demand in district heating networks and electricity prices on the market. In this paper, an optimization-based model for short-term scheduling of gas-fired CHP plant with heat accumulator is presented. The optimization model consists of a detailed simulation model of a cogeneration plant which is combined with an evolutionary algorithm. The optimization objective is to maximize the total gross margin for the day-ahead horizon of the CHP operation. An artificial neural network model is used for predicting heat demand in the district heating network. Different forecast models were tested for the electricity price forecast – extreme learning machines, multi-layer perceptron, auto-ARIMA, and triple exponential smoothing methods. The presented results show that the developed computer-based tool is efficient and effective for short-term scheduling of CHP plant with gas turbines and heat accumulator.