Faster-than-real-time unit start-up state prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Power plants face challenges in monitoring and predicting abnormal working conditions of units due to the lack of a method for self-adaptive, faster-than-real-time prediction and estimation of start-up states, leading to delayed detection of anomalies and potential equipment failures.
Innovation Solution
A method that calculates the slope of measured values from event recording sequences and historical scatter point distributions to provide real-time predictions and warnings, allowing for proactive monitoring and prevention of anomalies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual monitoring of analog measurement points is used, then the watchman can detect abnormalities when alarm values are reached, but the response time is extremely short and the watchman often has no time to deal with it before tripping occurs
Solution Approach 1:
The system performs preliminary action by calculating the measured value slope and predicting future trends before the alarm threshold is reached. The gradient prediction module computes predicted values based on historical data and current slope, enabling early warning of potential abnormalities before they become critical alarm conditions, thus providing advance response time.
Solution Approach 2:
The patent introduces an intermediary prediction mechanism between manual monitoring and alarm triggering. The gradient prediction module acts as an intermediary that continuously analyzes trends and provides early warnings, bridging the gap between passive alarm-based detection and active predictive monitoring, thereby extending the watchman's response time.
2Measurement precision
If gradient alarm is used to monitor instantaneous jumps, then unstable instantaneous jumps can be detected, but comprehensive time sequence actions and historical statistics cannot be utilized for prediction
Solution Approach 1:
The system merges gradient alarm detection with historical statistics analysis by integrating the gradient prediction module with the historical scatter record table. The prediction mechanism combines real-time gradient calculation with historical distribution statistics, merging instantaneous jump detection capabilities with comprehensive historical data utilization to enable predictive monitoring.
Solution Approach 2:
The system performs preliminary action by pre-calculating and storing historical scatter point distribution records, which are then utilized by the gradient prediction module. This preliminary preparation of historical statistics enables the system to perform predictive analysis by comparing current trends against historical patterns, thus preventing information loss.
3Ease of operation
If the monitoring system relies on watchman to periodically review simulation diagrams, then state monitoring can be performed, but the process is inefficient and abnormalities are only detected when alarm values are reached
Solution Approach 1:
The system implements self-service by automatically performing gradient calculation, trend prediction, and abnormality detection without requiring manual watchman intervention. The gradient prediction module autonomously analyzes measurement data, computes predicted values based on historical statistics, and generates early warnings, enabling the monitoring system to serve itself and significantly improving monitoring efficiency.
Solution Approach 2:
The patent replaces the mechanical manual monitoring system with an automated computational system. The gradient prediction module uses algorithmic calculations based on historical scatter records and current measurement slopes to predict future states, substituting manual visual inspection with automated mathematical modeling, thereby dramatically improving monitoring productivity.
Data Source
AI summary
The present disclosure provides a self-adaption faster-than-real-time method for estimating a working condition start-up state of a unit. Firstly, an event recording sequence, an analog measurement point ID, an analog measurement point first-level alarm value and a history scatter point distribution record are read from a time sequence event record table, an analog measurement point table, an alarm threshold table and a historical scatter record table. A measured value slope of the analog measurement point is calculated according to an event recording relative time. Then, an abnormal state estimated value is calculated based on the history scatter point distribution record, the analog measurement point first-level alarm value and the measured value slope of the analog measurement point. Finally, a pre-warning is sent to remind a watchman to perform a preventive operation when the abnormal state estimated value is above the threshold.