Stress Identification Of Space Structure Based On Multi-agent

In the traditional structural health monitoring, sensors are distributed and data processing is centralized, which causes a security problem and an inefficient operation. The concept of multi-agent system comes out of artificial intelligence. It is especially suitable for dealing with distributed problem which needs dividing function. Based on this, this paper introduces the concept of multi-agent to structural health monitoring, especially introduces to the stress identification in detail, to propose the stress identification of space structure based on multi-agent.This paper establishes a stress recognition method based on multi-agent. The function structure of multi-agent system is analyzed, and the junior agent and senior agent are divided in function module. Then the component of Simulink in MATLAB is used to simulate the system. The results of simulation are applied in dome-like structure. Stress identification results are obtained through simulating the process of multi-agent stress identification, which proves the feasibility. Then the method of multi-agent and the traditional centralized function are compared in function simulation, security probability and operational efficiency to demonstrate the superiority of multi-agent method.To improve the identification precision, the method of stress identification of multi-agent is optimized. The calculated sensors are introduced, and the method of stress identification of senior agent is optimized. Through the comparison of identification precision by two methods, the recognition precision by optimized method is ascending. Noise resistance analysis considers two cases, where sensors are all influenced by the noise and the individual sensor is influenced by the noise. In both cases, the identification precisions by the method introduced with the calculated sensors is better. And through the measurements of the Shenzhen Bay Stadium Center, the paper demonstrates the identification accuracy is improved by introducing the calculated sensors. In both cases of the noise, the identification precisions by the method of multi-agent before optimization, the method of multi-agent after optimization and the centralized function method are contrasted. The method of multi-agent after optimization gets the best identification precision. So the optimized method is proved a stronger noise resistance and a stronger adaptability to the actual project. The robustness of optimized method is also better.The paper dicusses the application of multi-agent system on Zhuhai opera house based on different target. Including based on the identification accuracy, based on the reliability and based on the economy. Based on reliability and economy, the paper also discusses in different numbers of junior agents, how the multi-agent method is used. Through the discussion of these methods, the paper demonstrates that the use of multi-agent method is different based on different target cases. The identification precision, the accuracy of reliability and the consumption of different cases is discussed to facilitate the multi-agent system in the practical engineering application.