Frequency Distribution Calculation Of Photovoltaic Power System Based On Markov Chain And Probabilistic Power Flow

Nowadays,with the increasing severity of global warming,concepts related to energy conservation,emission reduction,environmental protection,and sustainable development have become global issues.In order to achieve the goal of reducing carbon emissions,major countries have successively set annual targets for reducing carbon emissions.Renewable energy sources such as photovoltaic and wind power are widely regarded as an inevitable trend to replace thermal power generation,and the proportion of renewable energy in the future power system will continue to rise.However,due to the random fluctuation characteristics of most renewable energy outputs,it is difficult to predict and control,which brings a lot of randomness to the analysis and calculation of power systems.Therefore,research on probabilistic power flow calculation for handling random inputs has become increasingly rich.Although renewable energy generation represented by photovoltaic has strong random fluctuations in a short period of time,its output data exhibits certain cyclical patterns in specific cycles of days,months,and even years.Therefore,by establishing a probability model for random quantities at various time periods within a cycle and combining it with power system frequency adjustment calculations,it is possible to simulate and calculate the variation patterns of system frequency,node voltage,and other parameters within a cycle.This is of great significance for the operation and planning design of power systems containing renewable power generation.In this regard,a calculation method based on discrete time Markov chain and probabilistic power flow calculation is proposed in this thesis,which aims to calculate the change of power system frequency distribution and the average time of frequency overrun in a specific period.The following are the main contents of the thesis:(1)Based on the primary frequency regulation characteristics of the power system,a state space was constructed with the system frequency at each time as the parameter.Among them,the normal frequency is a non recurrent state,while the frequency exceeding the limit and the deterministic power flow do not converge to an absorbing state.By utilizing the frequency modulation characteristics of the power system and the probabilistic power flow algorithm,the state transition matrix between adjacent moments was obtained.At the same time,the impact of the parameter settings of this method on the calculation results was analyzed,as well as whether the method has scalability.It was also explored whether parallel computing resources can be utilized when calculating the state transition matrix.(2)Merge the state transition matrices at all adjacent moments into a total state transition matrix.According to the transfer characteristics of Markov chain,the system frequency distribution after each initial frequency distribution at any time is calculated.The absorption probability and average absorption time of each absorption state can also be calculated according to other characteristics of the Markov chain.An application scenario was proposed to search for the optimal frequency stability of the system under primary frequency regulation for a certain capacity photovoltaic power station to be connected to the system.(3)The method proposed in this thesis was applied to a New England 39 node power system with photovoltaic power stations.A probability model for load and photovoltaic output was established based on a specific period of days.Then,a comparative analysis was conducted between the method proposed in this thesis and the comparative method,verifying the accuracy of the calculation results of the method proposed in this thesis.We compared the computational performance of our method under different probability power flow algorithms and verified that using parallel acceleration can significantly improve computational speed.