The Research On Stochastic Resonance And Optimization Theory And Its Application In Low Concentration Gas Dection

With the deepening research on low concentrations gas detection, more and more detectingmethods were proposed in recent years, such as ultrasonic technology, optical interferenceprinciple, infrared absorption spectroscopy, sagnac effect of light path techniques and so on. TheStochastic Resonance (SR) was a new method under the development for nearly30years indetecting weak signal submerged in strong noise. This paper designed a low concentration gasdetection system based on stochastic resonance. It overcame the deficiency of the traditionalmethods for their low sensitivity, time-consuming, costly and difficult to pop up in lowconcentration gas detection. Stochastic resonance was a non-linear system, which transferredpartial noise energy into signal’s to enhance the ability to detect the weak signal. It couldeffectively improve output signal-to-noise ratio (SNR) and detection characteristics of the system.Aiming at the difficulties in optimizing the system parameters of stochastic resonance inpractical engineering application, this paper proposed an adaptive parameters tuning algorithm,and furthermore integrated with the optimization theory, the intelligent optimization algorithms,i.e. genetic algorithms was proposed.First, the paper introduced the basic theory and background of stochastic resonance andbi-stable stochastic resonance systems, and summarized the intelligent optimization algorithms.Second, it studied the parameters characteristics of the stochastic resonance systems. It analyzedhow system parameters affected the stochastic resonance phenomenon from the physicalconcepts and experimental simulations. It could clearly determine the reorientation (increase ordecrease) of system parameters and laid the theoretical foundation for the system to reachstochastic resonance state quickly.For practical signal in engineering, the traditional method of adjusting the noise intensitycould not effectively detect the weak signal and the sensitivity was low. So it is necessary tostudy the method of tunning system parameters to achieve stochastic resonance. This paperutilized the adaptive parameter tunning algorithm to search the optimal system parameters andthe weighted signal-to-noise ratio (WSNR) as evaluating index. It effectively overcame the difficulties in optimizing the system parameters and improved the detecting sensitivity of thestochastic resonance system. The fast and accurate detection of the weak signals submerged instrong noise and enhance the useful signal had been the focus of stochastic resonance system.Combined with the optimization theory the paper proposed genetic algorithm based on globaloptimization that tuned system parameters adaptively and determined the optimal systemparameters and noise intensity simultaneously. Through the experiment in low concentrationsammonia detection, the output SNR, detecting speed and accuracy of the system had been greatlyimproved. Through the regression analysis we could found the maximum WSNR and the gasconcentration had good line relationship.This paper provided a new method for low concentration gas detection. It also proposed thefeasible and effective algorithms on the system parameters adaptive selection and increasing thedetecting speed and accuracy of stochastic resonance system. The designed low concentrationgas detection system based on stochastic resonance and the proposed intelligent optimizationalgorithms have good application prospects.