论文部分内容阅读
在含油污泥进行资源化处理过程中,针对处理目标受多个因素影响的实际,为了解决工艺之间的耦合问题,采用正交实验的方法来解决,并把主要参数作为优化对象,把含油污泥的脱水率作为评价目标,通过采用GA-BP算法对含油污泥耦合工艺正交实验参数进行了线性与非线性分析。在采用遗传算法优化神经网络的权值和阈值的基础上,用优化后的权值和阈值对测试样本和训练样本进行了预测。预测结果表明,预测误差都有明显减小,分别由0.34211减少到0.031549和0.15476减少到0.040682,可见耦合参数趋向于非线性优化。
In the process of resource-based treatment of oily sludge, according to the fact that the treatment target is affected by many factors, in order to solve the problem of the coupling between the processes, orthogonal experiment is used to solve the problem. The main parameters are taken as the optimization object, The dewatering rate of sludge was taken as the evaluation target. The orthogonal experimental parameters of oil sludge coupling process were analyzed by using GA-BP algorithm linearly and non-linearly. Based on the genetic algorithm to optimize the weights and thresholds of neural networks, the test samples and training samples are predicted with the optimized weights and thresholds. The prediction results show that the prediction errors are obviously reduced, decreasing from 0.34211 to 0.031549 and 0.15476 to 0.040682, respectively. It can be seen that the coupling parameters tend to be nonlinear optimization.