Probabilistic Prediction Method For Thermal Ratings Based On Micrometeorology Of Overhead Conductors

With the rapid development of Chinese economy,social electricity consumption and power generation of power system are increasing,which makes the current-carrying capability of transmission lines face challenges.Meanwhile,the large-scale integration of renewable energy power generation into the power grid makes the power grid operation mode more complex and diverse.Under some special operating conditions,the problem that the accommodation of renewable energy is limited by the insufficient current-carrying capability of transmission lines gradually becomes severe.Since the thermal rating represents the current carrying capacity of an overhead conductor,its future variation information is very important to assist operators in exploiting and utilizing the current-carrying capability of an overhead conductor.Therefore,based on the historical data of the meteorological elements around an overhead conductor,it is of great significance to effectively predict the thermal rating of an overhead conductor by using data mining technology.At present,with the wide application of dynamic thermal rating technology,some thermal rating prediction methods have been developed.According to the difference in the form of prediction results,these methods can be divided into two major groups:point prediction and probabilistic prediction.The point prediction methods only provide the expectation value of the thermal rating at future time instant.However,the probabilistic prediction methods can provide the probability distribution of the thermal rating at future time instant,that is,the prediction result contains not only the expectation value but also the variance and other information that can fully reflect the uncertainty of the thermal rating prediction.Therefore,the probabilistic prediction method for the thermal rating of an overhead conductor is worthy of in-depth study.For an overhead conductor,meteorological correlations exist among the meteorological elements in its vicinity,and autocorrelation exists among the time series of the thermal rating.However,existing thermal rating probabilistic prediction methods fail to comprehensively consider the meteorological correlations and the autocorrelation of the thermal rating,which makes these prediction methods need to be optimized,and the probabilistic prediction results can be improved.Moreover,in power system optimization decision-making,to ensure the operation security of an overhead conductor,the minimum thermal rating in each time interval of the forecast horizon should be selected as its current-carrying limit.However,existing thermal rating probabilistic prediction methods mainly take the thermal rating at a specific time instant in the future as the prediction target,and do not consider the thermal rating variations inside the future time interval,which makes the prediction results fail to represent the current-carrying capability of an overhead conductor during the whole time interval.Applying these prediction results to power system optimization decision-making will lead to risky decision-making results,thereby bringing security risks.Against this background,this paper first establishes the thermal rating and the intra-interval minimum thermal rating calculation models based on the heat balance equation,and then conducts data analysis according to the historical data of four meteorological elements,the thermal rating and the intra-interval minimum thermal rating.Besides,a thermal rating probabilistic prediction method is proposed in this paper,which can comprehensively consider the meteorological correlations and the autocorrelation of the thermal rating.In the method,based on the nonparametric quantile regression theory and the Copula theory that is frequently used in correlation modeling,a series of single-period probabilistic predictions of thermal ratings are conducted considering meteorological correlations,and then the joint probabilistic prediction of multiperiod thermal ratings is conducted considering the autocorrelation of the thermal rating.Finally,a probabilistic prediction method for the intra-interval minimum thermal rating is proposed in this paper.In the method,after conducting a series of single-period probabilistic predictions of intra-interval minimum thermal ratings based on the nonparametric and nonlinear quantile regression neural network,the Copula theory is used to consider the autocorrelation of the intra-interval minimum thermal rating and the joint probabilistic prediction of multiperiod intra-interval minimum thermal ratings is conducted.In this paper,the proposed methods are verified by the historical data of meteorological elements around actual overhead conductors.Numerical simulations show that the proposed methods effectively improve the probabilistic prediction results of the thermal rating,and are conducive to assist operators to discreetly exploit and utilize the potential current-carrying capability of the existing power grid.