The State Evaluation Method Of Wind Turbine Based On Big Data Modeling And Analysis

In the renewable energy,wind power generation accounts for a large proportion.With the increase of the running time of wind turbine,the aging and wear of its components will be aggravated gradually,which leads to the deterioration of its running state.It is of great significance to grasp the operation state of the equipment in time,to ensure the safe and stable operation,to prolong the service life,and to arrange the operation and maintenance reasonably.Therefore,this paper makes a systematic study on the evaluation method of wind turbine operating state,mainly including the following contents.(1)The evaluation index of the running state of wind turbine is determined.Wind turbine is a complex system composed of several components,and each component affects each other.Therefore,there must be a relationship between the monitor data of each component,and a certain index can not truly reflect the actual state of the corresponding component.According to the structure and working principle of wind turbine,based on the data mining algorithm,this paper determines the state index of wind turbine from the relationship between the physical quantities within the motor.(2)The prediction model of wind turbine state index is established.Considering that the computational units of shallow learning(artificial neural network,support vector machine,etc.) have limited ability to represent complex functions,a combined prediction model based on Kriging and gated recurrent unit is proposed.The linear component is predicted by the Kriging principle,and then the nonlinear component prediction model is constructed based on the gated recurrent unit.Finally,the final prediction value is obtained by combining the two prediction results.(3)The dynamic weight calculation method of wind turbine state index is put forward,and considering the influence of the fixed value as the limit value of the evaluation index,the dynamic limit and the degree of deterioration are put forward.Each state index of wind turbine plays a different role in state assessment,and the influence degree of each index on operation state is also different in different period.Therefore,the dynamic weight is used to represent the influence of different indexes on the evaluation result.Firstly,the fixed weight of each evaluation index is calculated by analytic hierarchy process,and then the real-time weight of each index is determined by partial mutual information method.Finally,the two indexes are combined as dynamic weights.(4)The operation state evaluation model of wind turbine based on cloud model is established.First,the relative error of the prediction value of the state index is calculated according to the combined prediction model,and then the error is normalized as the dynamic deterioration degree of the corresponding index.Then the operation state of the unit is judged based on the similarity cloud and fuzzy comprehensive evaluation method respectively,the conclusions of the two methods are compared and summarized.The above research results include four models of wind turbine status index screening,dynamic weight calculation,state evaluation index prediction and operation state evaluation.Based on the actual operation data of a wind farm,the results are in line with the reality and meet the needs of the project.It has good application value.