Research And Effectiveness Evaluation On Bird Expelling Strategies

In order to prevent safety accidents caused by bird damage to transmission lines,to reduce power grid losses and to ensure power consumption in daily life,power grid has put a large number of device for repelling birds on transmission towers to drive away birds.For this kind of bird expelling strategies based on practical data,since power grid has not studied the release strategy of device for repelling birds,the release of most device for repelling birds depends on the working experience of front-line workers,whose subjective judgment will lead to high cost and unsatisfactory effect of device for repelling birds.At the same time,with the increasingly sophisticated technology of device for repelling birds,the South China Power Grid will continue to install emerging technology device for repelling birds on transmission towers in the future.Such device for repelling birds to be launched also lack the research of release strategy,which is easy to lead to the increase of cost.In this context,the practical data lack of validity labels for the released device for repelling birds.In this paper,the improved kmodes algorithm is used to cluster the device for repelling birds based on the practical data,and a small amount of device feedback data with known effectiveness is used as the test collection to verify the clustering effect of the algorithm.For the device for repelling birds to be put into operation,the experimental analysis of such device for repelling birds is carried out to test the effect of bird expelling strategies,and then the effectiveness evaluation model is established through the voting-based algorithm fusion classification algorithm to evaluate the effectiveness of bird expelling strategies based on experimental data.The specific work of this paper is as follows:（1）An improved k-modes algorithm is proposed,which solves the problem of the traditional k-modes algorithm to the selection of the initial clustering center by combining genetic algorithm.At the same time,considering the difference between ordered classification variables and disordered classification variables,the distance measurement formula is improved.The experimental results show that the improved k-modes algorithm can stably find the initial clustering center with better clustering effect,and the clustering effect is also better than the traditional k-modes algorithm;（2）Analyze the effectiveness of bird expelling strategies.The bird expelling strategies based on the practical data is clustered by the improved k-modes algorithm to obtain the effectiveness evaluation,and then the anti-bird effect of the bird expelling strategies under different influencing factors is analyzed.The bird expulsion strategy based on experimental data,first of all,analyzes the bird expulsion effect of the bird expulsion strategy under different influencing factors.Then,in order to evaluate the bird expelling effect of each strategy parameter,the behavior of birds in the experiment was quantified into different bird response types by combining with the fuzzy comprehensive evaluation method.The weight is obtained by counting the bird driving effect data of each strategy parameter.In order to make the weight of strategy parameters more consistent with the actual experimental situation,the weight calculation formula of CRITIC（Criteria Importance Though Intercrieria Correlation）is optimized.（3）The voting fusion classification algorithm are used to establish the effectiveness evaluation model.The experimental results show that the voting fusion classification algorithm can effectively evaluate the effectiveness of bird expulsion strategy,and in each classification index,the effectiveness evaluation model constructed by the voting-based algorithm is better than that constructed by other classification algorithms.