Active Distribution Network Optimal Scheduling With Multiple Time Scales And Load Uncertainty

Active distribution network(ADN)is an effective way to realize the integration of distributed renewable energy in the future.In order to cope with the intermittent of renewable energy output and investment costs on the power generation side,the development of demand-side dispatchable resources has become an effective way to resolve contradictions.Firstly this paper summarizes the current research status of optimal dispatching for active distribution networks,and introduces the types of demand response and uncertainty models.Based on those,it conducts research on optimal dispatching that takes into account multiple time scales and load uncertainties.At the previous stage,considering the role of price-based demand response in guiding the load,a time-of-use model with the price elasticity coefficient as the core is established.The peak-valley difference of load,ADN economic benefits,and so on are taken as the objective function.Decision variables include gas turbine output,hourly electricity price and so on.Considering power balance constraints,upper and lower limit constraints of each output unit and so on,a day-ahead scheduling model for active distribution networks is established.Bat algorithm with inertia weights and population diversity whose "early maturity" phenomenon has been improved is used to get a reasonable electricity price and output plan for the next day.In the day-in rolling optimization phase,using the model predictive control method,based on the previously scheduled price and output plan,rolling optimization is played with the minimum deviation between the output of the day-in and day-ahead phase as objective function,and the interruptible load resource can be called to reduce imbalance risk caused by fluctuations of renewable energy and load.The improved IEEE 33-node distribution network is used to verify the effectiveness and economy of the multi-time scale scheduling framework considering demand response.According to different action mechanisms of price-based demand response and incentive-based demand response,uncertainty models are established for them respectively: firstly,fuzzy variables are used to describe the price-based demand response,and robust optimization theory is used to measure actual interruption amount of the interruptible load.Then,according to the difference in response speed of the two demand-side management methods,they are allocated with other output resources in two-stage decision model.After that,based on the theory of fuzzy chance constrained programming and multi-stage robust optimization theory,the improved bat algorithm and entropy weight method are used to solve the model.The results of simulation examples show that the peak-to-valley difference of the load curve after the implementation of TOU has been greatly reduced and the average electricity price has remained stable,which alleviate the day-ahead scheduling pressure while maintaining users’ satisfaction;the day-in rolling scheme can effectively reduce the gap between planned output and actual needed output,reducing the risk of power shortage,and the effect is more obvious in scenarios with greater volatility.In the scheduling model that considers uncertainty,during the day-in optimization stage,the operating cost increases with the increase of the robustness coefficient and the day-in reliability level,of which the day-in reliability has a greater impact;the uncertainty in day-ahead phase will increase the cost of scheduling,but the cost of scheduling of day-in phase will be significantly reduced,which reduces the total cost of scheduling.