Thermal power equipment-based state early warning method and system

Abstract

The invention discloses a thermal power equipment-based state early warning method. The method comprises the steps of obtaining historical data of n detected points, and performing sample extraction on the historical data by taking 1/h as preset precision to obtain m learning samples; performing learning on the learning samples by utilizing a memory matrix to obtain an indication model; accordingto a quantity relationship of the detected points, the learning samples and the preset precision, calculating sample integrity of the memory matrix, and according to the sample integrity, changing a quantity of the learning samples to correct the indication model; and performing state early warning operation by utilizing the corrected indication model. According to the method, the state early warning accuracy of thermal power equipment can be improved; and a state of the thermal power equipment is subjected to objective and good evaluation. The invention furthermore discloses a thermal power equipment-based state early warning system, a computer readable storage medium and a thermal power equipment diagnosis apparatus, which have the abovementioned beneficial effects.

Description

technical field [0001] The invention relates to the field of equipment safety monitoring, in particular to a state warning method and system based on thermal power equipment, a computer-readable storage medium and a diagnostic device for thermal power equipment. Background technique [0002] With the rapid development of the electric power industry and the in-depth reform and innovation of electric power enterprises, the requirements for the safety, economy and reliability of power plants are getting higher and higher. It is an urgent requirement for equipment management to master the operation rules of power generation equipment or systems, predict the abnormal state of the equipment in time before the equipment develops a fault, and take measures to eliminate the fault in the bud. [0003] In the thermal power industry, the unit structure is relatively complex. Due to a series of factors such as demand fluctuations and load adjustments, the operating conditions of the equ...

AI Technical Summary

Problems solved by technology

Among the above methods, the data-driven non-parameter learning method makes full use of massive historical data to establish a fault early warning model, which has good advantages in the process of big data learning and processing, but the disadvantages of this type of system are: (1) Field application process Early warning capabilities are uneven, and the accuracy of early warning results is difficult to evaluate objectively and well

(2) The system adopting this method has only general early warning and does not give deeper fault information. Some experience knowledge of users and even proven fault cases cannot be well learned and applied by the system

These shortcomings lead to poor field application effect and cannot provide targeted early warning and diagnostic conclusions

Images