Research On Key Technologies Of Air Conditioning Loads Participating In Power System Dispatch And Control

Power fluctuations are becoming more prominent with increasing renewable energy integration into the grid,which requires power systems to be more flexible at peak shaving and frequency adjustment.Demand response as one of the key technologies can effectively restrain power fluctuations,relieve the supply-demand tension and improve energy efficiency of power systems.Air conditioning loads as a kind of flexible resources at demand side have huge power regulation potential and can be manipulated by flexible control methods,which makes them good demand response resources and need to be intensively studied.Air conditioning loads can be divided into two categories according to the types of compressors,namely fixed frequency and inverter air conditioning loads.The two kinds of air conditioning loads differ at control and modeling methods,which means that they cannot be investigated using the same technologies.Based on the thermodynamic models of fixed frequency and inverter air conditioning loads,this dissertation studies the control strategies and aggregated models of the two kinds of air conditioning loads from the lower level to the higher level.And the aggregated parameters of air conditioning loads are identified at the level of power systems.The main contributions are given as follows.1.Based on state space model,a control strategy of fixed frequency conditioning loads is proposed.To improve the control accuracy of air conditioning loads,an aggregated control model of air conditioning loads,in which the compressor time delay is taken into account,is presented to describe the collective behaviors of air conditioning loads with control signals.A probability control strategy is also proposed to reduce adverse impacts on the durability of air conditioning compressor and communication requirements between controller and air conditioning loads.Simulation results demonstrate that the proposed aggregated model of air conditioning loads with control signals can closely follow the actual behaviors of air conditioning loads and the proposed control strategy can reduce the number of switching times without communication burden.2.Aggregated models of fixed frequency air conditioning loads at the level of power systems are proposed.One of the key challenges for air conditioning loads to be integrated into the system level operation is building a compact aggregated model,in which,the primary behaviors of air conditioning loads are accurately captured.In this paper,fixed frequency air conditioning loads are aggregated as a virtual generator,virtual battery and power pulse models according to their different operating behaviors with various control methods.The corresponding strategies are also provided to restrain the cold load pickup,reduce the communication requirement between controller and air conditioning loads,and guarantee end users' comfort levels during the modeling process of aggregated models.This will bring system operator great convenience to manage air conditioning loads and conventional components when the system level decisions are made.Simulation results demonstrate that the proposed aggregated models can accurately follow the actual behaviors of air conditioning loads at the level of power systems.3.The control strategy and aggregated model of inverter air conditioning loads on the power system level are presented.In order to achieve the compatibility of the inverter air conditioning loads with the current dispatch models,this paper attempts to model an inverter air conditioning system as a minute-time scale virtual battery.The comparisons between the virtual battery and the lithium-ion battery are given.In order to protect the end-users' privacy and relieve the computational burden of the centralized control,a hierarchical control framework is designed and an aggregated model of virtual batteries is proposed to handle the heterogeneity of the inverter air conditionings.A finite-horizon optimization model is used to compare the operating performances of the aggregated virtual batteries and the lithium-ion batteries.Simulation results demonstrate that the aggregated virtual battery model works well with the power dispatch model developed for the lithium-ion batteries.In other words,with the virtual battery modeling of air conditioning loads,lithium-ion batteries can be replaced in current dispatch models providing services for the grid.4.Multi-time scale aggregated parameters of air conditioning loads at the level of power systems are accurately identified.Accurate parameters of aggregated models of air conditioning loads are critical to effectively coordinate air conditioning loads with other resources in the system operation.Thus,it is vital to identify accurate parameters of aggregated models of air conditioning loads,which is often a cumbersome task in practical applications due to inaccurate power metering,end-users' random behaviors for consuming energy in real-time,and privacy issues pertaining to the availability and the collection of individuals' data.To address this problem,high dimensional model representation(HDMR)is introduced to estimate the parameters of aggregated models of air conditioning loads using the probability distribution of air conditioning parameters.A numerical simulation study demonstrates that parameters of aggregated models of air conditioning loads can be accurately estimated by HDMR.