Data Driven Based Research On Load Control System Performance Assessment Of Thermal Power Units

Load control system is one of the most important thermal power control system, and a typical multi-variable control system. In order to provide an automated online program that can automatically monitor the performance of the control system and performance evaluation, keep abreast of the operational status of the unit, improve the automation level of power plant, This paper utilizes multivariate data-driven techniques to study the performance evaluation method of thermal power unit load control system.First of all, this paper proposes a traditional time domain index based performance evaluation method of thermal power unit load control system. This method utilizes the dynamic performance indicators(overshoot, peak time, maximum dynamic deviation, squared error points, etc.) that are used to describe the process control system, defining the dynamic quality evaluation for thermal power load control systems, according to the operating data in variable condition of unit output variables to calculate the value of performance indicators and divide the performance evaluation grade.Secondly, according to the on-line operating data propose covariance indicators based control system performance evaluation method. The method uses the covariance matrix of the system output data as an evaluation function, selecting the optimal period output data as a baseline data, to be evaluated output data as the monitoring data, by calculating the evaluation function ratio of the reference data with the monitoring data, to obtain an index value related to the control system performance. At the same time through eigenvalue analysis transforme the covariance index into the product of the generalized eigenvalues of matrix output. Using statistical inference theory to make the index value calculated by the portion of the sample data more in line with the general characteristics of the control system.Finally, taking MATLAB/GUI as the development platform, designing thermal power unit load control system performance evaluation simulation interface. Verified by model simulation data and field data in a power plant, evaluation results confirmed the validity and reliability of the two performance indexes.