基于数据融合IGA-RGRNN低阶煤制甲烷产量预测模型

为了提高智能系统的准确性与快速性,针对多传感器网络,提出了一种以融合技术为数据基础与改进遗传算法-广义旋转回归神经网络IGA-RGRNN(improved genetic algorithm and rotated generalized regression neural network)算法相结合的预测模型。利用RGRNN强大的非线性随机变量的处理能力,把预测理论引入改进遗传算法循环中,将该模型应用于低阶煤制甲烷产量预测过程,并对预测模型效果进行实验验证。实验结果表明,基于数据融合IGA-RGRNN低阶煤制甲烷产量预测模型的相对误差最大值为2.99%,相对误差最小值为0.25%,相对误差平均值为1.76%,相较其他预测模型具有泛化能力更强和预测精度更高的优势,为低阶煤制甲烷产量预测提供一种新的途径。

Prediction Model of Methane Yield from Low-rank Coal Based on Data Fusion and IGA-RGRNN Algorithm

For multi-sensor network, a prediction model based on data fusion and united the improved genetic algorithm and rotated general regression neural network (IGA-RGRNN) was proposed, both the accuracy and the rapidity of intelligent systems was raised. With using RGRNN''s strong processing capability to handle nonlinear random variables and introducing prediction theory to the improved genetic algorithm, the proposed model is applied to the prediction process of methane production from low rank coal, and its effect is verified by experiments. Experimental results show that the maximum relative error of the prediction model based on data fusion IGA-GRNN was 2.99%,the minimum relative error was 0.25%, and the average relative error was 1.76%.Compared with other predictive models, the proposed model has advantages of stronger generalization ability and higher prediction accuracy, and it will lead to a new approach for the prediction of methane production from low rank coal.