Initial Explosion Signal Detection Of Combustible Gas/air Mixtures In Pipelines

In order to minimize the losses caused by the explosion of combustible gas in the pipeline,it is necessary to use a variety of sensors at the initial stage of the explosion to quickly and accurately determine the occurrence of an explosion.In order to obtain the signal in the explosion environment in many aspects,the main work included the following aspects:(1)Set up experimental system.In this thesis,the length of seven pipe was 21 meters,with a 104 mm diameter.Twelve sensors were arranged on the device.According to the characteristics and working principle of UV sensor,infrared sensor and ion probes,the three kinds of sensors were respectively protected by secondary encapsulation to ensure that the sensors were not damaged in the explosion experiment,so the experimental environment could be detected accurately.A set of collection control system designed by the laboratory was used in the experiment.Explosion signal detection of combustible gas/air mixtures experiment carried out in the above-mentioned system.(2)Design sensor operating circuit.Design a T-type network current to voltage conversion circuit for infrared sensor,which could effectively reduce the bias error of the feedback network and facilitate the resistance configuration.A high-voltage DC UV sensor working circuit was designed.And an ion probe working circuit was designed,which could convert the ion current into a voltage signal.(3)Analyze the experimental results.According to the experimental results which carried out in the above-mentioned system,the flame propagation velocity in the experimental pipeline increased first,then decreased and then increased.Under this condition,the three sensor signals were analyzed independently.The results showed that the rise time of the infrared sensor waveform was fast,and the first derivative can react quickly and accurately to the time of explosion characteristic signal generation.The UV sensor waveform was in the form of a pulse wave whose initial pulse wave responded quickly to the explosion.The ion probe waveform could reflect the time of the explosion,but the waveform was more complex.(4)Construct multi-signal fusion algorithm based on BP neural network.Based on the data obtained from the two sensors in the experiment,the multi-signal fusion algorithm basedon BP neural network was designed.The training data of the BP neural network were 2000 for each sensor in the experiment.And 200 data in the experimental were used as BP neural network simulation input.After training the neural network.The accuracy of the identification of the characteristic signal can reach 100%.