Chemical laboratory risk early warning method based on discrete Hopfield neural network

Abstract

The invention provides a chemical laboratory risk early warning method based on a discrete Hopfield neural network. Actual results of a ventilation and illumination equipment operation condition a1, a temperature and humidity control equipment operation condition a2, a circuit system operation condition a3, a dangerous chemical storage environment condition a4, a dangerous chemical chemical property preservation condition a5, an experimental instrument safety inspection condition a6, an experimental instrument safety operation condition a7, a safety indication sign condition a8, a fireproof and explosion-proof equipment operation condition a9, an experimental environment sanitary condition a10, an emergency equipment condition a11 and an emergency evacuation channel condition a12 are encoded and input into the constructed discrete Hopfield neural network, and the risk early warning of the multi-index assessment system for the safety condition of the chemical laboratory in the colleges and universities is realized through the characteristic of keeping unchanged after multiple iterations. The problems that a large amount of data is needed for simulation, mobility is poor, and a construction method is complex are solved, and visual, rapid and accurate risk early warning of the laboratory safety condition is achieved.

Description

technical field [0001] The invention relates to the field of laboratory risk early warning, in particular to a chemical laboratory risk early warning method based on a discrete Hopfield neural network. Background technique [0002] Laboratory risk early warning is to realize the understanding and evaluation of the current risk status of the laboratory by artificially designing the secondary indicators of the multi-indicator evaluation system of safety status and giving corresponding risk status comments according to the evaluation comparison table. Objective and efficient early warning of laboratory risk status can well maintain the good operation status of the laboratory and reduce the possibility of safety accidents, which is very important for universities. Therefore, the research results of the present invention have broad application prospects. [0003] Laboratory risk early warning is a complex early warning system that includes many indicators such as equipment and e...

AI Technical Summary

Problems solved by technology

But they have the problem of falling into local optimal solutions, overfitting, or incorrectly correlating information leading to fitting errors

Their modeling methods are complex, time and space complexities are high, and a large number of training samples are required for modeling learning to achieve good simulation results. When there are few training samples, the lack of training sample dimensions, or changes in the field of evaluation targets It may cause the neural network to fail to output correct warning results

Images