In this paper, the main circuit mathematical model of T-type three-level inverter is established, and a multi-objective optimal control method of T-type three-level inverter model prediction is studied, which mainly includes output voltage control and neutral point potential balance control. Firstly, the load voltage and current are sampled, and the output voltage prediction value is calculated with the sample value, so that the error between the predicted voltage and the reference voltage is the most In the small switch state, the midpoint current of DC bus is measured at the same time, and the deviation value of midpoint potential is obtained. According to the charging and discharging conditions of corresponding DC side capacitor in this switch state, the switch state which reduces the deviation value of midpoint potential is optimized to act on the inverter in the next sampling period, so as to realize output voltage control and midpoint potential balance control. Finally, it is verified by Matlab / Simulink simulation and experiment. The simulation and experiment results show that the method can achieve the output voltage control and neutral point potential balance control well, and has the characteristics of simple principle and easy to digital realization. © 2020 IEEE.

This paper explores interaction patterns for multienergy demand management and proposes an integrated demand response (IDR) mechanism for industrial integrated energy system. The incentive payment is provided not only for curtailing interruptible electric load, but also for increasing flexible heating and cooling loads. The coupling characteristics of multi-energy demands are modeled to reflect the mutual influence between the consumption behaviors for different energy sources. Besides, based on the coupling of CCHP (Combined Cooling, Heating and Power) energy outputs, the demand-supply interaction model is also proposed, which could change the electric generation by stimulating heating and cooling demands. The optimization model of IDR is then established with the objective of minimizing total dispatch cost. Simulation results show that the IDR operator, CCHP and consumers could all benefit from this mechanism. Furthermore, the impact of the incentive price on the dispatch cost is also analyzed.
© 2019 IEEE.

This paper proposes a novel strategy to apply multi-agent systems (MAS) and consensus tracking control on synchronous generators (SGs) in a micro-grid (MG) based on sliding mode strategy in presence of an external disturbance. Cooperative control of SGs has undeniable advantages on MGs which guarantee optimal collaboration among various parts. Sliding mode controllers resist against disturbances meanwhile, achieving finite time synchronization for the nonlinear states of SGs. Stability analysis is proved based on Lyapunov direct method. Simulation results illustrate the effectiveness of the presented method to apply consensus protocol for multi-agent generators in MGs. © 2019 IEEE.

In view of the high coupling degree of regional integrated energy system, a bilayer interaction strategy, consisting of energy suppliers, distribution networks, and users, is proposed. Game interaction strategy includes two aspects: scheduling and bidding. The independent system operator (ISO) coordinates all adjustable resources. Depending on the quotation price and multi-energy load prediction, ISO minimises the total energy cost, which realises the complementary of the multi-energy in the cooperative game. Under the assumption of incomplete information and bounded rationality, this study designs bidding functions and pay-as-bid settlement protocols. On this basis, according to history scheduling data and units' characteristics, agents for energy suppliers pursue maximum interests. Also, the non-cooperative bidding process in multi-energy market is simulated by using Q-learning algorithm. Finally, the evolutionary process of the bilayer competitive game model is studied by practical example, and the existence local Nash equilibrium of the strategy is also proven.
© 2019 Institution of Engineering and Technology.All right reserved.

As new energy with high permeability is connected to the distribution network through power electronic equipment, the droop control in the past can no longer meet the needs of coordinated control of wind power, photovoltaic and other distributed equipment in the AC-DC hybrid distribution network. Therefore, this paper proposes a decentralized coordinated control strategy based on dynamic multi-agent to improve the traditional droop control. Firstly, the distributed controller in AC-DC hybrid distribution network is regarded as an agent, and the frequency deviation of each distributed device can be obtained through the communication and interaction of multiple agents. Then, the reference value of DC voltage in the droop controller is improved to support the system frequency. In addition, by adjusting the active power instruction of dynamic multi-agent, the power distribution of distributed controllers is realized. At last, a typical model of the medium and low voltage AC-DC distribution network is built in the PSCAD/EMTDC software to verify the effectiveness of the proposed method, compared with the traditional droop control.

With the growing installation of distributed energy resources (DER), much attention has been drawn by the power prediction of photovoltaic generators (PV) and wind turbines (WT). Due to changeable weather conditions and the uncertainty of equipment operations, the power grid requires an accurate forecast method to obtain the efficiency of renewable integration. This paper proposes a decision tree-based prediction learner for PV and WT power, which forecasts the renewable output with bootstrap aggregation. The proposed learner is based on mixed tree with bootstrap aggregation and an improved attribute selection algorithm, which is parameterized by cross validation and immune to noisy data. The effectiveness of the proposed methodology is demonstrated through real-world data with different prediction windows. © 2018 IEEE

This paper proposes an interactive bidding strategy for smart distribution networks with clustered aggregators to effectively coordinate the energy and profit of each entity. In the proposed one-leader n-follower bidding model, there are two levels, where the upper level stands for the distribution operator (DO) to secure the operation quality and bid with energy aggregators (EAs), and the lower level represents each EA to bid for interactive energies with the DO and other EAs. The proposed interactive bidding mechanism takes advantages of both renewable generation and energy storage system into account, where the priority-based decision making is applied to the lower level to elaborately map the bidding mechanism and lead each EA's market interaction. Case studies on the IEEE 33-bus test feeder and a practical distribution grid in China demonstrate the effectiveness of the proposed methodology.