基于SVR和GA的锅炉运行氧量基准值的优化确定

借助现场运行数据,根据锅炉运行氧量的特性,建立了基于支持向量回归的锅炉运行氧量预测模型,结果表明:SVR模型具有较高的回归精度和较好的泛化能力,能够有效地对不同工况下的锅炉氧量进行预测.在此基础上进行二次建模,获得了运行氧量、供电煤耗率与各运行参数之间的关系模型,并结合全局寻优的遗传算法,以机组的供电煤耗率为优化目标对输入参数进行寻优,确定了优化后的锅炉运行氧量基准值.计算结果表明该模型具有较高的准确性,通过全局寻优得到的氧量值具有可操作性,很好地解决了锅炉变工况运行参数基准值的确定问题.

Optimization of boiler operation oxygen content based on support vector regression and genetic algorithms

Using the data of boiler operation,a support vector regression(SVR) model of the boiler oxygen content property was developed based on gas oxygen characteristic.The results show that the model based on SVR has more accurate and forcible generalization ability.The model can predict the oxygen content accurately under different conditions.After that another SVR model describing the relationship between oxygen content,unit's power supply coal consumption rate and operation parameters was also built.Combined with the optimization algorithms(GA),the fiducial oxygen content can be determined according to the optimal object of the unit's power supply coal consumption rate.Results confirm that this method has high calculating precision and it is suitable for engineering purposes due to its maneuverability.It also offers a method for determining the operating oxygen content under actual variant conditions.