Niannian Cai

A decentralized control architecture for a microgrid with power electronic interfaces

This paper presents a decentralized control architecture for the autonomous operation of a microgrid with power electronic interfaces. The control architecture is based on a multi-agent system (MAS) in which all agents are hierarchically equal and there is no central agent. Neighbor to neighbor three-step communication algorithms are used by the MAS to determine real and reactive power mismatches and resolve the dispatch of real and reactive power among the agents within the microgrid. The research described here includes the design of control modes for the power electronic interfaces in the microgrid. The proposed method is demonstrated on a system comprising four microsources, and is implemented using MATLAB/SIMULINK.

A hierarchical multi-agent control scheme for a black start-capable microgrid

This paper presents a hierarchical control scheme using a multi-agent system for black start operation of a microgrid with power electronic interfaces. Five types of agents are proposed in this architecture, namely Grid Agent, Central Agent, Generation Agent, Load Agent and Breaker Agent. With this architecture, the multi-agent system is able to coordinate distributed generators (DG) and loads to maintain steady state operation of the microgrid either in grid-connected mode or islanded mode; it can also perform a black start if a seamless transition to the islanded mode fails or if a black start becomes necessary for any other reason. A system with four DG units is demonstrated using MATLAB/SIMULINK to demonstrate the operation of the proposed system.

Economic dispatch in microgrids using multi-agent system

In traditional bulk power system, economic dispatch is usually conducted in a central controller, which has access to global system information. However, characterized as flexible and “plug and play”, microgrids are often controlled in a distributed manner. This paper presents a decentralized architecture of multi-agent system for economic dispatch of distributed generators (DG) in microgrids. In this architecture, all DGs are equipped with identical agents, which can only perceive local information or obtain necessary information from their neighbors. By competing with each other, agents attempt to maximize their own profit, thereby tending to an optimal global solution. This paper first develops necessary conditions for minimal cost operation of microgrids. Then it proposes an innovative multi-agent communication algorithm to implement the minimal principle in multi-agent environment based on the consensus theorem. In the end, a five-agent system and a fifty-agent system are investigated to validate the proposed multi-agent algorithm.

Intelligent methods for smart microgrids

This paper summarizes ongoing research in the application of intelligent methods to the design, modeling, simulation and control of microgrids including optimal design of microgrids, and centralized and decentralized control.

A fast transient stability screening and ranking tool

This paper introduces a fast transient stability screening tool to classify a designated set of contingencies into stable and unstable subsets using direct methods. The proposed method is based on the conservativeness of the Lyapunov function. The classification processes of the contingencies are performed along the solution trajectory towards the Controlling Unstable Equilibrium Point (controlling UEP). The proposed screening tool is intended to reliably capture the unstable contingencies and efficiently reduce the number of contingencies that need further analyses. If a numerical problem is encountered during the computation, the proposed scheme applies homotopy-based approaches to find the desired solution. If the numerical problem can not be solved using homotopy-based methods, the contingency is sent to a time-domain simulator for further analysis. The method was applied on the Western System Coordinating Council (WSCC) test system and results are presented.

Use of homotopy-based approaches in finding Controlling Unstable Equilibrium Points in transient stability analysis

This paper introduces the use of homotopy-based approaches in computing the Controlling Unstable Equilibrium Points (controlling UEPs) in transient stability analysis using direct methods. It is well known that the regions of convergence of the controlling UEPs are very sensitive to the initial guesses, and traditional iterative methods fail to find the correct controlling UEPs if the initial guesses lie outside their regions of convergence. On the other hand, homotopy-based approaches are very reliable in finding solutions because they are globally convergent. However, homotopy-based approaches are intrinsically slow if the initial point is far from the desired solution because these methods map the trajectory of the solution from an easy and known solution to the desired solution. This paper proposes an algorithm that uses a homotopy-based approach with the exit point as an initial point to reliably find the correct controlling UEP. To reduce computational effort, the proposed method uses an approximate exit point rather than computing an accurate exit point as is common practice in finding controlling UEPs. Further, this method eliminates the necessity of computing the Minimum Gradient Point (MGP), which makes the homotopy-based approaches comparable with the other iterative methods in terms of the speed of computation. An explicit characterization of the region of convergence of a controlling UEP and its boundary starting from an exit point for a typical power system is derived. The method is applied on the WECC and the NE 39 test systems to demonstrate its effectiveness in finding the controlling UEPs.

Distributed analytics for steady state operation of autonomous microgrids

It has often been argued that microgrids that predominantly contain customer-owned generation (co-gen) should have autonomous control, i.e., their operation, even in grid-connected mode, should not be controlled by the utility providing the point of common coupling (PCC). Some of the control strategies proposed in recent literature rely on distributed control platforms, such as multi-agent systems (MAS). The analytics, i.e., computational and decision-making processes, that occur on these distributed computation platforms are fundamentally different from those that are performed in traditional control centers. This is due partly to the difference in the computation platforms and partly to the fact that microgrids are dynamic in configuration. This paper describes current research in the application of graph-theoretic and distributed computation concepts toward the development of analytics, such as load balancing and optimal operation, that can be implemented on a distributed platform, for the steady state operation of microgrids.

Optimal branch numbering for UPFC benefit study

With the development of smart grid, power electronic control devices, such as Flexible AC Transmission Systems (FACTS), are introduced into power systems. They can help system operators control the real and reactive power flow and regulate local voltage. Placing Unified Power Flow Controller (UPFC), one of FACTS devices, at certain congested locations of power system, can help alleviate system congestion, and therefore reduce production cost. The number (name) of branches can significantly affect the speed of the algorithm to find the optimal location to place UPFC. Previously, this problem was not paid sufficient attention to. In this paper, sensitivity analysis of the production cost is studied. Based on the sensitivity, the algorithm of optimal branch numbering is proposed. Simulation results on IEEE 300 system presents that by renumbering the branches, the efficiency of genetic algorithm to find the optimal location of UPFC can be greatly improved.

Optimal economic power dispatch in the presence of intermittent renewable energy sources

This paper describes a method for solving the economic power dispatch problem in the presence of renewable energy sources. The method proposed in this paper uses linear programming because linear programming based formulations tend to be flexible, reliable, and faster than their nonlinear counterparts. A linearized network model in the form of DC power flow model is utilized in this paper. Thermal limits of transmission lines and real power constraints have both been considered in the proposed model. In addition, piecewise linear models for generating units cost curves are developed during the realization of the presented work. A micro-power optimization model based on Homer software is used to account for the uncertainty in the output power of the intermittent renewable energy sources. The proposed linear programming based method is demonstrated on the standard IEEE 30 bus system. Test results are reported, discussed, and thoroughly analyzed.

Multi-agent system based voltage regulation in a low-voltage distribution network

Due to the high feed of regenerative power in many low-voltage distribution networks, it is possible that in some nodes inadmissible voltages appear. This paper presents a way to comply with the prescribed voltage band. It is realized by a multiagent system, which executes a three-stage algorithm to adjust the voltage at the nodes of the grid. A possible configuration is presented with a voltage regulated distribution transformer. It regulates the secondary voltage in several steps. The multi-agent system measures the voltages in different points in a network. The three-step algorithm controls the voltage, reactive and real power to comply with the prescribed voltage band. The paper shows a flexible communication structure, which is used by the agents to exchange data. An example network is presented. This displays the effectiveness of the communication cycle. Furthermore, the influence of the three-step algorithm is shown in a test system.