Research On Energy Flow Calculation Of Integrated Energy System Based On Data Driven

Integrated energy system(IES)coupled with various networks can realize the complementary cooperation of different energy sources,it’s an important direction to achieve sustainable development at present.Multiple energy flow(MEF)calculation is the basis of operation analysis and steady-state study of IES,and is an important prerequisite for studying complementary and interactive characteristics of energy sources.The penetration of renewable energy such as wind energy and photovoltaic energy in the IES is significantly increased,and its output is highly uncertain under the influence of surrounding environmental factors.The energy consumption load of electricity,heat and gas is constantly changing over time,showing a characteristic of volatility.It is difficult to calculate IES deterministic performance flow and uncertain performance flow by using model-driven method,and the model complexity and accuracy are contradictory.Therefore,this paper considers applying the data-driven method to the energy flow calculation process of IES.The main work is as follows:(1)In view of the shortcomings of the current Newton method,such as the high Jacobian matrix dimension leading to slow computation speed and dependence of convergence solution on initial value selection,this paper proposes a MEF calculation method for IES based on the integrated strategy of heterogeneous learner.Firstly,considering the characteristics of the IES,different data-driven methods are used to train the MEF calculation model of the system.Then,according to the accuracy and difference degree of the models trained by different data-driven methods,models with excellent performance and differences are selected as the base learner.The heterogeneous learner integration strategy is used to integrate multiple base learners to form the heterogeneous learner integration model.The simulation results show that the proposed model can achieve better accuracy than the model built by the single data-driven method in different examples,and can perform a large number of repetitive calculations quickly and accurately.(2)In view of the uncertainties existing in the current IES,this paper proposes a method for calculating the interval energy flow of the IES based on mechanism-driven and data-driven method.In this method,affine number is used to represent uncertainty.On the theoretical basis of affine arithmetic(AA),inverse trigonometric function quadratic approximation,principal component analysis and system structural properties are used to form affine optimization models of corresponding electricity,heat and gas systems respectively,and a unified affine optimization model is concluded.Then the space branch and bound method is used to solve the optimization problem,and the upper and lower boundaries of the system are obtained.Simulation results verify the efectiveness and efficiency of the proposed method and ensure the global optimal solution.(3)Participate in the development of a simulation toolbox suitable for IES-IES operation toolbox.The toolbox takes MATLAB/Simulink as the simulation platform,with energy flow calculation,optimal energy flow calculation,holomorphic energy flow calculation,real-time simulation functions,this paper participates in the realization of energy flow calculation and real-time simulation functions.Decomposition method is used as the framework for energy flow calculation,and push-back substitution method is used to calculate the initial value of given energy flow calculation,so as to reduce the number of iterations.RT-LAB simulation platform is used to invoke C files in dynamic model library to realize real-time simulation of dynamic model.The simulation results show that IES operation toolbox can be applied to the energy flow calculation of the system,and can realize the dynamic real-time simulation of the system.