| 基于时间集分割的蒸汽流量预测模型 |
| |
| 引用本文： | 王梦柯,何利力.基于时间集分割的蒸汽流量预测模型[J].教育技术导刊,2020,19(5):88-93. |
| |
| 作者姓名： | 王梦柯 何利力 |
| |
| 作者单位： | 浙江理工大学 信息学院，浙江 杭州 310018 |
| |
| 摘 要： | 由于制造业生产数据具有较强时段性，相关工艺蒸汽流量预测方法精确度不高，无法有效节能降耗。针对该问题，提出基于时间集分割的蒸汽流量预测模型。基于工艺生产情况及原始数据的时段性，将日生产工艺流量时间集划分为工单稳定生产时段、工单启动后/结束前时段、非工单时段，采用逐点回归模型预测工单稳定生产时段，利用曲线补全模型预测工单启动后/结束前时段。非工单时段无生产，因此无需进行预测。综合逐点回归—曲线补全预测结果，得出日工艺用蒸汽流量。实例研究表明，相较于传统未分割时间集的单一预测模型，基于时间集分割的逐点回归—曲线补全组合预测方法精确度达 94%以上。基于时间集分割的组合模型不仅预测精度高且较稳定，可为蒸汽生产与实时调度提供决策依据。
|
| 关 键 词： | 时段性 时间集划分 逐点回归 曲线补全 |
| 收稿时间： | 2019-07-09 |
Research on Steam Flow Forecasting Model Based on Time Set Segmentation |
| |
| WANG Meng-ke,HE Li-li.Research on Steam Flow Forecasting Model Based on Time Set Segmentation[J].Introduction of Educational Technology,2020,19(5):88-93. |
| |
| Authors: | WANG Meng-ke HE Li-li |
| |
| Institution: | Information Institute，Zhejiang Sci-Tech University，Hangzhou 310018，China |
| |
| Abstract: | The research on accurately predicting the steam flow of the next day process based on the time set segmentation model provides a scientific basis for the boiler opening strategy，so as to achieve energy saving. Based on the process production situation and the time period of the original data，the daily production process flow time set is divided into the work order stable production period, the work order start/end time period and the non-work order time period. On this basis，the point-by-point regression model is used to predict the stable production time of the work order，and the curve completion model predicts the time period after the work order is started/before the end of the work order. The composite point-by-point regression-curve completion prediction results are used to derive the daily process steam flow. The case study shows that the accuracy of the point-by-point regression-curve complement combination prediction method based on time set segmentation is more than 94% compared with the single prediction model of the traditional undivided time set. The combined model based on time set segmentation not only has high prediction accuracy and is stable，but also provides decision-making basis for steam production and real-time scheduling. |
| |
| Keywords: | time period time set division point-by-point regression curve completion |
|
| 点击此处可从《教育技术导刊》浏览原始摘要信息 |
| 点击此处可从《教育技术导刊》下载免费的PDF全文 |
