Data-driven Uncertainty Analysis In Power Systems With High Proportion Of Renewable Energy

With the large-scale renewable energy power generation system with power electronic equipment as the grid-connected interface connected to the interconnected power grid,the operation and stability of the power system show significant uncertainty.Renewable energy sources such as photovoltaics and wind power are subject to the constraints of the natural environment,and their power generation is intermittent and random.The power electronic equipment of the interface also affects the dynamic stability characteristics of the power system.With the increase of the penetration rate of renewable energy,the safe and stable operation of the system is severely challenged by uncertainty.Deterministic traditional analysis methods are difficult to apply to power systems with significant uncertainties.As a background,the uncertainty analysis method for studying power systems with high-ratio renewable energy has profound theoretical and practical significance.Traditional uncertainty analysis relies on the probability distribution that can accurately describe the uncertainties of the system,and obtaining accurate probability distribution is considered to be a difficult challenge in practical engineering.In engineering applications,it is often impossible to obtain all the probability information of uncertain factors.Therefore,the probability distribution used in traditional uncertainty analysis itself has uncertainty,that is,high-order uncertainty(contains both the uncertainty itself and the uncertainty of the probability distribution model describing its uncertainty).In this paper,based on the impact of high-order uncertainty of renewable energy on the operation and stability of power systems with high-ratio renewable energy,the data-driven uncertainty analysis theory is used to conduct related research from the perspectives of static security assessment and small signal stability.In terms of static security assessment,the main work is divided into probabilistic power flow based on intrusive data-driven polynomial chaos expansion method and distributionally robust probabilistic statistic security assessment method.1.Probabilistic power flow based on intrusive data-driven polynomial chaotic expansion method is proposed.Aiming at addressing the high-order uncertainty of renewable energy power generation and the nonlinearity of the power flow equations,the high-order moment from statistics information of renewable energy power generation can be used to construct a set of orthogonal basis in arbitrary probability space,and then the state variables such as voltage in the power flow equations are approximated by sums of polynomial basis whose parameters are calculated by the stochastic Galerkin method.2.Distributionally robust probabilistic static power assessment of power system is proposed.Power system operation control relies on load forecasting providing the necessary information,and needs to consider the impact of high-order uncertainty of load forecasting error.The probabilistic static safety assessment problem is modeled as a moment-constrained probability measure optimization problem,that is,the generalized moment optimization problem,which can be solved by the semi-definite programming relaxations.The maximum probability of that the security constraints of the power system are violated under all possible probability distributions that satisfy the moment constraint is obtained by the proposed method.In terms of dynamic stability of small disturbances,the main work is divided into stochastic small signal stability analysis of grid-connected photovoltaic systems based on non-intrusive data-driven polynomial chaos expansion method and probabilistic stability margin assessment of power electronic multi-feed system based on generalized short-circuit ratio.1.Stochastic small signal stability analysis of grid-connected photovoltaic systems based on non-intrusive data-driven polynomial chaos expansion method is proposed.Firstly,considering the broadband oscillation characteristics,the impedance modeling method is used to construct the small disturbance stability analysis model of the photovoltaic power generation system.Secondly,the polynomial basis functions are constructed by using several order moments of the original statistics data,and the system dominant damping ratio are approximated by the sums of polynomial basis using the probability allocation method.2.The generalized short-circuit ratio is a static index which reflects the influence of the interconnection of multi power electronic devices and the strength of AC power grid on the system oscillation,which can be used to quantify the small signal stability margin of the system.The renewable energy power generation in multi infeed system is uncertain,so this paper puts forward a calculation method of probability of multi power electronic based devices infeed system small signal stability margin based on generalized short-circuit ratio.This method only needs the expectation and variance of the uncertain variables,by using the generalized short circuit ratio and the generalized moment theory,transforms the small disturbance stability margin calculation into matrix probability D stability problem,and then into semidefinite programming.An example is given to illustrate the effectiveness of the proposed method.The above researches propose a data-driven uncertainty analysis theory,which enriches the tools and methods for the security and stability assessment of power systems.