Self-adaptive test method of intelligent prediction algorithm foranalog measurement value

Abstract

The invention provides a self-adaptive test method of an intelligent prediction algorithm for analog measurement value and provides an effective means for technicians to test performance of an intelligent trend judgment algorithm, and can also provide a fault sample for the intelligent trend judgment algorithm. The method comprises the following specific steps: firstly, reading an event recordingsequence condition, an analog measurement point ID and an analog measurement point alarm value from a time sequence event recording table, an analog measurement point table and an alarm threshold table; then obtaining an operation record of a normal operation state of the unit in a statistical period to form analog measurement point measurement value historical statistics based on the switching value signal; calculating to obtain an analog measurement point analog measurement value with a time scale by integrating historical statistics, the analog measurement point alarm value and the analog measurement point current measurement value; and finally, providing the analog measurement point analog measurement value with the time scale to a unit starting state analog measurement value intelligent prediction algorithm for testing, calculating the sensitivity, and when the sensitivity exceeds a threshold value domain, giving an alarm to remind a technician to adjust the algorithm.

Description

technical field [0001] The invention relates to the technical field of electrical equipment, in particular to an adaptive testing method for an intelligent prediction algorithm of an analog measurement value. Background technique [0002] The monitoring board of the watchman in the power plant needs to monitor more than 500 analog measurement points with tripping and exit functions. These analog measurement points with tripping and exit functions are distributed in different simulation diagrams of the upper computer of the monitoring system. It is an impossible task to rely on the on-duty staff to manually track the changing trend and find abnormalities in advance. At present, intelligent technologies such as machine learning are booming, and intelligent technologies provide important technical means in trend judgment and prediction. [0003] However, it cannot be ignored that these intelligent technologies rely on fault samples, and a large number of fault samples are req...

AI Technical Summary

Problems solved by technology

[0004] In the past, the acquisition of fault samples came from debugging and fault handling after overhaul, and the cost of obtaining these fault samples was extremely high. In addition, the numerical characteristics of these fault samples could not completely cover the trend characteristics under fault conditions, which had certain limitations.

Images