首页|基于PCA-ISSA-GRU的燃煤电厂供电煤耗计算研究

基于PCA-ISSA-GRU的燃煤电厂供电煤耗计算研究

扫码查看
随着国内电力体制与市场交易机制的变革,燃煤电厂之间的竞争压力越来越激烈,供电煤耗作为衡量电厂经济效益的重要指标,其精准计算就显得愈发重要.提出一种基于PCA-ISSA-GRU方法的供电煤耗计算模型,首先采用滑动窗口法对数据进行稳态筛选,采用主成分分析法(PCA)对处理的数据进行特征筛选,选择最相关的输入参数.其次对机组的外部环境进行分析,采用K-means方法最终确定8种不同的工况.最后为了使模型计算更加精确,采用改进的麻雀算法(ISSA)对门控循环单元(GRU)的超参数进行寻优.以上海某600MW机组的历史数据进行验证,并对不同组合模型之间的预测精度进行对比.结果表明,本文的模型供电煤耗计算与实际相吻合,平均误差为1.32g/(kW·h),相对误差在±1%,模型计算精度高,泛化能力强,适用于燃煤电厂供电煤耗的计算.同时,综合评价指标对比显示,本文构建的预测模型比其它的预测模型精度更高,效果更好.
Research on Coal Consumption Calculation of Coal-fired Power Plant Power Supply Based on PCA-ISSA-GRU
With the transformation of the domestic power system and market trading mechanism,the competition pressure between coal-fired power plants is increasing.As an important indicator to measure the economic benefits of power plants,the accurate calculation of power supply coal consumption is becoming increasingly important.This article proposes a power supply coal consumption calculation model based on the PCA-ISSA-GRU method.Firstly,the sliding window method is used to perform steady-state screening on the data,and the principal component analysis(PCA)method is used to perform feature screening on the processed data,selecting the most relevant input parameters.Secondly,analyze the external environment of the unit and use the K-means method to ultimately determine 8 different operating conditions.Finally,in order to make the model calculation more accurate,the improved Sparrow Algorithm(ISSA)is used to optimize the hyperparameters of the Gated Recurrent Unit(GRU).Verify the historical data of a 600MW unit in Shanghai and compare the prediction accuracy between different combination models.The results show that the calculation of coal consumption for power supply using the model in this article is consistent with the actual situation,with an average error of 1.32 g/(kW·h)and a relative error of±1%.The model has high calculation accuracy and strong generalization ability,making it suitable for calculating coal consumption for coal-fired power plants.At the same time,the comparison of comprehensive evaluation indicators shows that the prediction model constructed in this article has higher accuracy and better performance than other prediction models.

coal consumption for power supplythermal power unitsbig data processingimproving the Sparrow Algorithmgate control loop unit

赵钊、茅大钧、陈思勤

展开 >

上海电力大学自动化工程学院,上海 200090

华能国际电力股份有限公司上海石洞口第二电厂,上海 200942

供电煤耗 火电机组 大数据处理 改进麻雀算法 门控循环单元

中国华能集团有限公司2022年度科技项目

HNKJ22-HF22

2024

汽轮机技术
哈尔滨市汽轮机厂有限责任公司

汽轮机技术

CSTPCD北大核心
影响因子:0.368
ISSN:1001-5884
年,卷(期):2024.66(5)