一种基于卷积神经网络的热工数据分析方法
By using a convolutional neural network model to perform in-depth analysis of thermal power plant data, the problem of DCS systems being unable to deeply mine thermal data in thermal power plants has been solved, enabling automatic fault analysis and early warning, and reducing the risk of accidents.
CN116342318BActive Publication Date: 2026-07-17STATE GRID CHANGYUAN HANCHUAN FIRST POWER CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- STATE GRID CHANGYUAN HANCHUAN FIRST POWER CO LTD
- Filing Date
- 2023-01-05
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing DCS systems in thermal power plants can only provide simple alarm functions for thermal data, and cannot perform in-depth analysis. Furthermore, software malfunctions can easily lead to failure to provide timely warnings, resulting in accidents.
Method used
A thermal data analysis method based on convolutional neural networks is adopted. By establishing a convolutional neural network model, massive thermal data is analyzed and predicted to achieve deeper fault detection and early warning.
Benefits of technology
It enables in-depth mining of thermal data, automatically analyzes the causes of failures, reduces the need for human resources, and provides alternative alarms when the DCS system fails, thereby reducing the probability of accidents in thermal power plants.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN116342318B_ABST
Abstract
本发明涉及数据处理领域，具体涉及一种基于卷积神经网络的热工数据分析方法。该方法包括，将各热工数据的所有不同原因故障类别标注标签值后分为训练集和测试集；搭建卷积神经网络模型，将训练集输入至所述卷积神经网络模型进行正向传播以及反向传播来调整卷积神经网络模型的参数，得到训练好的卷积神经网络模型；将测试集输入至训练好的卷积神经网络模型中进行预测，并进行向前预测以及softmax运算，将概率值中的最大概率值作为最终预测值，且对测试集的准确率进行图形化。本发明通过建立卷积神经网络模型来对海量热工数据进行分析，从而不仅能达到在DCS系统发生故障时进行补充预警措施，而且能对热工数据进行更深一步的挖掘。
Need to check novelty before this filing date? Find Prior Art