摘要
常规企业能耗智能预测模型设定的节点多为独立式的,仅针对一个有效距离进行能耗的预测,整体性不强,预测结果不精准,导致模型的相对预测误差增加,为此提出对基于马尔科夫链的工业企业能耗智能预测模型构建的分析与研究。根据模型构建的需求及标准,提取能耗特征变量智能,采用多层级关联的形式,将独立的预测节点形成连接,营造一个全覆盖式的预测区域,提升预测的整体性,并确保预测结果的精度,完成关联预测节点的部署。以此为基础,构建马尔科夫链多层级预测结构,内部结构还需要增设能耗预测矩阵,采用马尔科夫链执行目标修正实现智能预测。最终测试结果表明:与两种传统预测模型相对比,马尔科夫链能耗智能预测模型测试组预测范围相对较大,且在实际执行的过程中,预测误差可控,被较好地降低在了1.5以下,预测速度快,具有实际的应用意义。
The nodes set in the conventional enterprise energy consumption intelligent prediction model are mostly independent,and the prediction of energy consumption is only for an effective distance,which is not strong in a holistic way and the prediction results are not accurate,leading to an increase in the relative prediction error of the model.According to the requirements and criteria for model construction,energy consumption characteristic variables are extracted intelligently,and a multi-level association is used to form connections between independent prediction nodes,creating a full-coverage prediction area,enhancing the wholeness of prediction and ensuring the accuracy of prediction results,and completing the deployment of associated prediction nodes.Based on this,a Markov chain multi-level forecasting structure is constructed,and the internal structure also requires the addition of an energy consumption forecasting matrix,using Markov chains to perform target corrections to achieve intelligent forecasting.The final test results show that:compared with the two traditional prediction models,the Markov chain energy consumption intelligent prediction model test group has a relatively large prediction range,and in the actual implementation process,the prediction error is controllable,being better reduced to below 1.5,fast prediction speed,and has practical application significance.
作者
李承国
刘爱勇
宁尚武
赵樽波
王超
Li Chengguo;Liu Aiyong;Ning Shangwu;Zhao Zunbo;Wang Chao(Shandong Houfeng Automobile Radiator Co.,Ltd.,Tai'an Shandong 271000,China)
出处
《现代工业经济和信息化》
2023年第6期8-10,13,共4页
Modern Industrial Economy and Informationization
关键词
马尔科夫链
工业企业
能耗
智能预测
模型构建
模型设计
Markov chain
industrial enterprises
energy consumption
intelligent prediction
model construction
model design
