基于正弦余弦算法的汽轮机热耗率预测

Prediction of Steam Turbine Heat Rate Based on Sine Cosine Algorithm

为了准确建立汽轮机热耗率预测模型,提出了一种基于量子位Bloch坐标编码自适应的改进后正弦余弦算法(ASCA)和快速学习网(FLN)综合建模的方法.首先将ASCA算法与经典的人工蜂群优化算法(ABC)、教与学优化算法(TLBO)和正弦余弦算法(SCA)进行比较,最后利用某600MW超临界汽轮机组现场运行数据建立汽轮机热耗率预测模型,并将ASCA算法优化的FLN模型(即ASCA-FLN模型)的预测结果与FLN、ABC-FLN、TLBO-FLN和SCA-FLN模型的预测结果进行比较.结果表明:ASCA-FLN模型具有更高的预测精度和更强的泛化能力,更能准确地预测汽轮机的热耗率.

To accurately predict the heat rate of steam turbines, an integrated modeling method was pro- posed by combining an improved adaptive sine consine algorithm （ASCA） based on Bloch coordinates of qubits with the fast learning network （FLN）. The specific way is to compare the ASCA with classical arti- ficial bee colony （ABC） algorithm, teaching-learning-based optimization （TLBO） algorithm and sine con- sine algorithm （SCA）, and then to set up a prediction model for the heat rate using the operation data col- lected from a 600 MW supercritical steam turbine unit, and finally to contrast the prediction results among the ASCA-FLN, FLN, ABC-FLN, TLBO-FLN and SCA-FLN model. Results show that the ASCA-FLN prediction model has higher prediction accuracy and stronger generalization ability, which therefore can be used to predict the heat rate of steam turbines more accurately.