The Application Of Artificial Neural Network In Gas Examination System

At present, there are always the defects that the poor stability and the low measuring precision because of the cross-sensitivity exist for sensor. Take the pressure sensor and the coal-mine gas for example, the cross-sensitivity is existed for environmental temperature and humidity, the output of the sensor is depended on not only by the target parameters, such as the pressure difference and the methane concentration, but also by the non-target parameters, like the factors as environmental temperature, humidity and electromagnetic interference. If the target parameters are constant but the non-target parameters such as the environmental temperature is changing, as a result, the variation is accompanied with the output. To increase the stability of the target parameters under the non-target parameters with changing, the multi-sensor data fusion technology is combined with the intelligent technology, it is to say, applying the technology which the target parameters and the non-target parameters are monitored by multi-sensors, the data for the output in which the effect of the variation for the environmental factors are eliminated and the function of the sensor is also improved by the process of the computer confusion.Because the nerve network has good non-linearity approaching ability, self-study and auto-adapted performance, it is widely used in each domain now. In recent years, the nerve network is introduced into the sensor. The performance of the sensor is improved in the certain degree through establishing the intelligent sensor system. In view of the fact that there are many kinds of the nerve network structures and the training methods, this paper takes enhancing the sensor stability, the reliability and the usability as a goal, takes the software superiority as the main principle in improving the hardware flaw, mainly studies the application of the BP nerve network and the RBF nerve network in the intelligent sensor system. The research indicated that it can make effective improvement in the sensor's output performance in using the sensor compensation system which this paper designed and the union of the nerve network algorithm.Otherwise, virtual instrument using software to realize the traditional instrument function such as data collection, memory, analysis, display, it supports by the least amount of hardware modules. It does not emphasize reality form of the physical apparatus, and breaks the bound of the defining instrument box by manufacturers, and used the soft panel in the CRT display to replace the original instrument panel. So the measurement parameters and process are controlled by keyboard and mouse. Virtual Instrument offers users a space of fully playing their talents and imagination. Virtual Instrument technology is used in gas detection systems, without additional hardware it can conveniently, fast and accurately detect the gas concentration in mine based on the advantages of the software, In this paper a gas sensors compensation system based on Lab VIEW is designed, it combines algorithm of neural net- work, it can monitor the gas concentration situation of the mine conveniently, fast and intuitively.