Research On Fault Diagnosis Of Wind Turbines Based On Artificial Intelligence Algorithm

The development of science and technology is inseparable from energy.As the speed of technological development continues to increase,the use of energy is also increasing,which makes people have to consider the problem of energy depletion.At the same time,the massive use of fossil energy has also polluted the environment.Now these two issues have become the focus of attention of all countries in the world.Clean and environmentally friendly renewable energy has become the focus of attention.In order to coordinate the relationship between environmental protection and social development,and to meet the energy supply needed for social development,renewable and clean environmental protection is undoubtedly a necessary condition for the alternative energy source.Wind power is one of them..All walks of life have invested a lot of resources and energy in wind power generation,and have achieved remarkable results.However,there are still many problems.Due to the harsh working environment of the wind turbine,long-term operation in a large temperature difference and multi-condition environment,and the structure of the wind turbine is complicated,various failures will inevitably occur.This makes it a lot of trouble and difficulty in repairing the faulty wind turbine,and therefore requires more investment in the repair of the faulty wind turbine to ensure the safe and stable operation of the wind power generation facility,and ultimately leads to a sharp increase in cost.Finding a better way to monitor and diagnose the operating state of a wind turbine will undoubtedly improve its safe and stable operation and alleviate the pressure caused by excessive maintenance costs.In order to be able to identify the faults more quickly and accurately by monitoring the data and applying intelligent algorithms to minimize the labor cost,this paper chooses the neural network as the basic algorithm and collects the wind turbines provided by China Datang Corporation.The vibration signal is a data set.After the data processing,the neural network is trained and simulated.In order to make up for some shortcomings of neural network in processing data,this paper first adds wavelet function to optimize neural network,and optimizes wavelet neural network by genetic algorithm and particle swarm optimization respectively,and takes various parameters in wavelet neural network.The genetic algorithm and the particle swarm algorithm are used to calculate the initial parameters of the wavelet neural network.In the following process,the information contained in each index is integrated to solve the fault diagnosis problem,and compared with other traditional intelligent algorithms applied to wind turbine fault diagnosis.The genetic algorithm optimization wavelet neural network algorithm mentioned in this paper is The accuracy of wind power fault diagnosis is the highest,the accuracy rate is 95.20%,and the accuracy of particle swarm optimization wavelet neural network algorithm is 95.18%,which is similar to the performance of genetic algorithm optimization wavelet neural network algorithm,but particle swarm optimization wavelet The operation speed of the neural network algorithm is significantly faster than the genetic algorithm optimization wavelet neural network algorithm.