工业电阻炉多参数能耗建模与预测

电阻炉温度变化存在非线性、大延迟的特点,建立精确的能耗数学模型比较困难.为解决理论建模复杂且不具备实时性的问题,提出了一种基于数据驱动的电阻炉多参数能耗预测方法.首先,通过分析电阻炉工作阶段的能耗特性,建立了电阻炉理论能耗预测模型;然后,利用粒子群优化算法对支持向量回归的超参数进行寻优,建立了基于支持向量回归的多参数能耗预测模型;最后,对比了支持向量回归、高斯过程回归、自适应模糊神经推理系统模型在单参数及多参数条件下的能耗预测结果.实验结果表明,基于粒子群优化下的支持向量回归多参数能耗预测方法具有更好的预测效果.

Multi-parameter energy consumption modeling and prediction of an industrial resistance furnace

It is difficult to establish an accurate mathematical model of energy consumption for the temperature variation of a resistance furnace due to its nonlinear and large delay characteristics. In order to solve the problem of complexity and not real-time performance of theoretical modeling, a data driven based multi-parameter energy consumption prediction approach of the resistance furnace is developed in this paper. Firstly, the theoretical energy consumption prediction model of the resistance furnace is established by analyzing the energy consumption characteristics of the resistance furnace in the working stage. Then, the particle swarm optimization algorithm is used to optimize the hyper-parameters of support vector regression, and a multi-parameter energy consumption prediction model based on support vector regression is established. Finally, the energy consumption prediction results of support vector regression, gaussian process regression, and adaptive network-based fuzzy inference system models under single parameter and multi-parameter conditions are compared. The experimental results show that the support vector regression multi-parameter energy consumption prediction method based on particle swarm optimization has better prediction effect.