The Research Of Multi-spectral Radiation Temperature Measurement Of Turbine Blades Based On Improved NSGA-?

Accurate measurement of the temperature is of vital importance in monitoring.Multi-spectral radiation temperature measurement is a useful method.The equations are established through the information from every channel,when we calculate the temperature the emissivity will be also solved.So the result has less dependence on the emissivity.At present,we ordinarily use the genetic algorithm to calculate the temperature by multi-spectral radiation and have received a good effect.This paper introduces a variety of algorithms to calculate the temperature.In this paper,we introduce the adaptive double population genetic algorithm,different evolution,particle swarm optimization and the NSGA-II to compare with the traditional genetic algorithm.Although the NSGA-II has better stability and accuracy,there are still some problems.So we proposed a new scheme:(1)As the probability of crossover and mutation in NSGA-II is fixed,we introduce an improved one with adaptive adjustment.The probability will change with the evolution of the population.Then the accuracy of temperature will be improved and the stability will also be enhanced.(2)The crossover and mutation strategy of NSGA-II is not closely combined with the information of the population.There is certain blindness in the search for the best solution.In this paper,we compromise the NSGA-II and the different evolution algorithm to guide the direction of the population.In addition,we modified the crosser factor to guide the direction of evolution to a better field.Besides,the accuracy of solution will also be raised.(3)As the elite strategy of NSGA-II can only save the individual of single type,we propose a new one based on the traditional one.The improved strategy holds the individual more selectively and enriches the population.It makes us solve the target temperature more accurately.