Visualization and analysis of electrical parameter design based on digital sensors

This paper provides an in-depth study on the application of digital sensor technology in electrical fire detection. It firstly focuses on the design of electrical parameters to improve the accuracy of fire detection. Based on the research results of recurrent neural networks, the article combines the Long Short-Term Memory (LSTM) network and the Gated Recurrent Unit (GRU) neural network. It innovatively proposes an electrical fire feature recognition method based on the LSTM-GRU network. The study also incorporates fuzzy inference techniques to optimize the fire alarm decision-making process, which achieves a hierarchical output for electrical fire detection recognition and alarm decision-making. The experimental results show that the LSTM-GRU network achieves an accuracy of 98.83% in electrical fire classification and recognition, significantly better than the results of using LSTM network or GRU network alone. Regarding electrical fire distance recognition, the relative error of the method only ranges from 0.32% to 2.10%, and its output of fire alarm decision is entirely correct. The study not only verifies the high accuracy and reasonableness of the electrical fire feature recognition method based on digital sensing technology, but also provides a brand new idea for recognizing electrical fire detection and alarm decision-making.