粒子群算法在气力输送管道压降预测中的应用

管道压降是气力输送系统设计的一个重要参数,传统的求解方法比较复杂.本文提出了以气体流速、颗粒浓度、混合比等作为神经网络输入,建立管道压降网络模型的方法.为进一步提高管道压降预测准确度,以预测误差作为适应度值,采用粒子群算法对网络权值和阈值寻优,优化神经网络,并利用样本数据训练出了有效的压降预测网络.通过将预测数据和粉料气力输送实验装置的实测数据相比较,结果表明,该方法预测误差小,准确度高,有较高的实用价值.

Application of Particle Swarm Optimization to Pressure Drop Prediction of Pneumatic Transport Pipe

Pipe Pressure drop is an important parameter of pneumatic transport system design,the traditional solution method is relatively complex.This paper proposes a method which establishes the pipe pressure drop predictive network model by taking gas flow rate,particle concentration and mixture ratio as the inputs of the neural network.In order to further improve the predictive precision of pipe pressure drop,the particle swarm algorithm is used to optimize the network weight and the threshold value by taking prediction error as the fitness value.In addition,an effective pressure drop prediction network is trained by using the sample data.By comparing the prediction data with the measured data of the powder pneumatic transport experimental device,the result demonstrates that the method has high precision and relatively high practical value.