A Machine Learning-Based Model Predictive Control Method for Pumped Storage Systems

Abstract

Integrated systems required for renewable energy use are under development. These systems impose more stringent control requirements. It is quite challenging to control a pumped storage system (PSS), which is a key component of such power systems. Because of the S-characteristic area of the PSS pump turbine, traditional proportional-integral-derivative (PID) control induces considerable speed oscillation under medium and low water heads. PSSs are difficult to model because of their nonlinear characteristics. Therefore, we propose a machine learning (ML)-based model predictive control (MPC) method. The ML algorithm is based on Koopman theory and experimental data that includes PSS state variables, and is used to establish linear relationships between the variables in high-dimensional space. Subsequently, a simple, accurate mathematical PSS model is obtained. This mathematical model is used via the MPC method to obtain the predicted control quantity value quickly and accurately. The feasibility and effectiveness of this method are simulated and tested under various operating conditions. The results demonstrate that the proposed MPC method is feasible. The MPC method can reduce the speed oscillation amplitude and improve the system response speed more effectively than PID control.

Identification of heat exchanger process through Hammerstein and polynomial models

“…The Output-Error (OE) estimator has the advantage of being more readily calculable than the predictionerror estimator. Using SI techniques such as the Hammerstein Weiner, Auto Regressive with Exogenous Input (ARX), Auto Regressive Moving Average with Exogenous Input (ARMX), Box-Jenkins (BJ) and OE models, a mathematical model was designed for a laboratory-based heating system [11][12][13][14][15][16][17]. The BJ model provides the greatest Final Prediction Error (FPE), correlation analysis, percentage of fitness, and loss function according to the simulated results [18][19][20][21].…”

“…The dependence of the parameter on the input and output was found numerically and roughly using polynomials [17]. Input-output data are taken from the Pseudo Random Binary Sequence (PRBS) experiment to find the heat exchanger system using the ARMAX model [13].…”