基于多维指标的煤电机组控制回路性能评价研究OA

The paper proposes a multi-dimensional indicator-based method for evaluating control loop quality of coal-fired power units.First,an Incremental/Differential/Static(I/D/S)setpoint differential encoding strategy is introduced to automatically identify effective control periods and filter out interference data during setpoint step changes.Second,a six-dimensional evaluation index system is constructed,comprising effective automatic control rate,accuracy rate,rapidity rate,stability rate,oscillation rate,and comprehensive score.Based on ARMA model fitting of system dynamic characteristics,indicator parameters are extracted using the unit impulse response function.Data preprocessing is implemented through Savitzky-Golay(SG)filtering combined with the Pauta criterion,enabling offline evaluation based on DCS historical data.Validation using 20,000 DCS historical data points from a 630MW unit of a certain power plant demonstrates that the proposed method can accurately identify effective control periods,effec-tively distinguish different control states,and verifies the method's effectiveness and engineering practicality.