基于SVR误差补偿技术的神经网络城市污水厂水质预测

城市污水处理是一个复杂的生化处理过程,现在的神经网络技术无法做到对此过程精准建模。为了解决该问题,提出了一种基于SVR误差补偿技术的神经网络水质预测算法。该算法先利用BP神经网络对水质处理过程进行映射,再利用SVR误差补偿模型获得BP网络的预测补偿,进行预测数据校正。为了验证补偿模型的性能,还组织了马尔科夫补偿模型的对比试验。试验结果表明,SVR误差补偿模型可有效提高模型预测的精度,且模型性能优于马尔科夫补偿模型。

Water Quality Prediction of Urban Wastewater Treatment Plant Based on Neural Network SVR Error Compensation Technology

Urban wastewater treatment is a complex biochemical treatment process,and the current neural network technology can not accurately model this process.In order to solve this issue,a neural network wastewater quality prediction algorithm based on SVR error compensation technology was proposed.Firstly,the BP neural network was used to map the wastewater quality treatment process,and then the SVR error compensation model was used to obtain the predictive compensation of BP network to correct the predicted data.In order to verify the performance of the compensation model,a comparative experiment of Markov compensation model was also organized.Experimental results showed that the SVR error compensation model could effectively improve the prediction accuracy of the model,and the performance of the model was better than that of the Markov compensation model.