基于支持向量机和遗传算法的电站锅炉水冷壁高温腐蚀建模与燃烧优化

Building Model on High Temperature Corrosion of Water Wall in Power Station Boiler and Combustion Optimization Based on Support Vector Machine and Genetic Algorithms

电站锅炉水冷壁的高温腐蚀是锅炉机组安全的严重威胁,强还原性气氛是导致高温腐蚀的主要原因,针对还原性气氛的特征指标CO浓度进行建模,并结合优化算法实现燃烧优化是控制还原性气氛的有效方法。应用支持向量机算法建立电站锅炉水冷壁周围CO浓度模型,利用热态实炉试验数据对模型进行训练和校验。该模型对不同试验工况下的CO浓度作出了较准确预测,应用该模型和遗传算法对锅炉进行以降低水冷壁周围CO浓度为目标的燃烧优化,结果表明,通过优化使CO浓度有比较明显的降低,与运行工况对比优化结果具有较高可信性,说明支持向量机模型与遗传算法的结合为锅炉水冷壁周围CO浓度的控制提供了有效工具。

High temperature corrosion of water wall is a threat to security production of power station boiler,and is an important problem for the producers.As CO concentration is an important factor for the high temperature corrosion,building a model to predict CO concentration is a good way to optimize combustion.In this paper,A support vector machine（SVM） model predicting CO concentration was developed and verified with experimental data.Good predicting performance of this SVM model was achieved.Low CO concentration were accomplished by using genetic algorithms and SVM model to optimize operating parameters.The results showed that support vector machine and genetic algorithm are good tools for building model to optimize combustion.