Control and Stability of Microgrid in Steady and Transient States
image credit: © Leowolfert | Dreamstime.com
- Oct 29, 2020 2:05 am GMTOct 27, 2020 5:32 pm GMT
- 703 views
This item is part of the Special Issue - 2020-10 - Distributed Energy Resources, click here for more
As per the definition of Microgrid, “ it is a cluster DERs and loads connected to Utility Grid or Distributed network which operates as an independent entity within its electrical boundaries which can work either in Grid connected or islanded mode. The distributed generators, battery storage systems have to be integrated to Microgrid through parallel interfaced inverters with all its controllers. As the inverters are not having inertia because of power electronics they have to be referred to synchronous frame to make the DGs stable. In this paper it is proposed communication- less controllers which are robust. In this method, the harmonic analyser is used to reduce the harmonics level in the system especially when it is switched over from grid to islanding and islanding to grid mode.
The Microgrid and the grid are connected through Point of Common Coupling (PCC). This is the junction between Microgrid and main grid through which key parameters are monitored for a smooth change over operation modes without any unstability. This is the centre point for any connection or disconnection process and the Microgrid monitors the relevant parameters on both sides of the PCC. The basic purpose of the Synchronization controller is this. This block controls the voltage amplitudes, phase and the frequencies on both sides of the PCC via static switch using techniques such as the implementation of second order generalized integrators and phase-locked loops (PLL).
Islanding is of two types, they are intentional or un-intentional. The intentional is meant for routine check purpose either on Grid or on Microgrid. Un-intentional is meant for faults either on Grid or on Microgrid. If there is any perturbance on the utility grid, the fault current should not circulate in the Microgrid, hence it is to be islanded and vice-versa. The islanding Detection Method used in this paper is of robust type SRF-PLL, which can include virtual impedance compensation and can take care of NDZ also.
Energy equilibrium should be maintained during transition from one mode another. The batteries should be in fully charged condition as the Microgrids generally have small reserve. To insure stability, first the non- essential loads are to be cut off, as a first step to maintain frequency and voltage in the Microgrid.
In grid connected mode, if the DGs can not feed power to all the loads , the grid will do that job. If the generated power is excess than the load power and the battery charging power, the additional power is exported to the grid on an agreed tariff to Utility.
2. Microgrid Model
Fig.1 Single Line diagram of a Microgrid with a typical DG and Storage
The Microgrid network model which is shown in Fig.1, consists of a power control centre, in which grid incomer is connected through a static transfer switch for smooth transition between the transitions and all outgoing breakers are connected to DG interfaced inverters (PV and Battery Energy storage System)and loads. All these lines are connected via distribution lines or cables depending on requirement. The loads are segregated as essential and non-essential loads which are prioritized based on the customer’s specifications.
The Microgrid has to operate either with utility grid synchronized or has to independently work in islanded mode sharing the load power and maintaining voltage and frequency at PCC. The islanding is detected much before instability of Microgrid with the islanding detection method and brings back stability through smooth changeover from grid mode to island mode. Also it should reduce introduction of transients, through filters.
The battery energy storage system is utilized to control voltage and frequency and DGs will try to supply active and reactive powers to the load proportional to their ratings. The reverse droop control method will cover this phenomena.
The loads are also prioritized as essential and non-essential loads. In the event of very serious perturbations the non-essential loads can be avoided by means of automatic load shedding panel. The Microgrid stability can be achieved through the control of voltage and frequency.
3. Inverter controller model
The heart of any Microgrid is the controller. The controller must be capable enough to be stable in synchronous mode, to be stable in transition from grid to island mode, stable in island mode and must be ready to synchronize to grid again for exporting and importing for economical benefits . Nowadays , the Microgrids may be owned by consumers and they can become prosumers by supplying power to another Microgrid with agreed tariff arrangements. There are many islanding techniques which are listed in references. A few of them are discussed below.
They are first classified as,
1. Remote islanding technique (depends on communication based) 2. Local islanding technique.
Again the local islanding techniques are devided into two types,.
- Active islanding,
- Passive islanding
The active islanding methods are used to overcome the limitations of the passive methods in which NDZ is taken care. Generally, active methods intentionally introduce perturbances at PCC, to find out the variations of voltage, frequency and impedance values. With this, it will be confirmed that the Microgrid is islanded.
*Sandia active frequency swift with positive feedback.
* Impedance measurement or harmonic injection.
* Detection of impedance at a specific frequency or monitoring of harmonic distortion
* Sandia Voltage Shift.
* Variation of Active or Reactive Power.
The above methods are advantageous of excellent reducing or even eliminating the Non-detection zone (NDZ), but in order to achieve their purpose they may reduce the quality of the grid voltage or even lead to instability. Although numerous techniques exist and their implementation varies, most of them are based on inverter based DG topologies. It is desirable for microgrid to be stable and robust as it is for practical engineering application.
2.Passive Islanding :
In these methods, normally NDZ facility is not included. But in this paper a passive islanding detection is proposed with a robust inveretr controller topology which can take care of even NDZ facility with the introduction of virtual impedance in the line.
* Over-voltage/ under-voltage protection
* Over-frequency/ under- frequency protection.
* Voltage phase jump
* Voltage harmonic monitoring.
* Current harmonic monitoring
The efficiency of any Microgrid depends on the topologies of inverters. The control technologies for Microgrid are dependent on the output of the inverter. There are two types of active power (P) – reactive power(Q) control and the other one is voltage-frequency control . To get co-ordination between DGs and energy storage systems, there are several control approaches including normal droop control, reverse droop control, inverter based control, primary energy source control, autonomous control and MAS control . In these type of controllers other than MAS controllers, the parameters are taken from local measurements with no communication from DGs. Hence these are robust and are more efficient.
In this paper the modified SRF-PLL method of islanding detection is passive type, which has taken care of virtual impedance and NDZ, which can be used in both transition modes, that is from grid to islanding and islanding to grid.
The Microgrid consists of DGs which are not having inertia like synchronous machines and hence they can not be connected directly to the grid. It is obvious that they should be connected through inverters. They are basically of DC or AC type of sources like batteries, PV, fuel cells etc or micro-turbine, wind turbines etc. repectively. They have to be interfaced with buck / boost converters and inverters to connect to PCC of Microgrid as Micrigrid is synchronized to main grid at PCC via static switch. To connect to Microgrid or even to operate in island mode to share load and also for smooth transition from grid to island and island to grid, the controllers play a very crucial role.
Fig.2 Matlab / Simulink model of inverter
The Simulink model is shown Figure 2. In this model, the DG consists three blocks: DG, buck / boost converter and an inverter. The DG is connected to the Microgrid via an inductor. The inverter monitors and controls the magnitude and phase of the output voltage V. The active and reactive power from the DG to the PCC is decided by Voltage, system voltage E, and the inductive reactance X.
The equations below shows the relationships between them,
Active power P is given by
Reactive power is given by
In which δ is phase angle,
The stability of controller is crucial for the Microgrid to operate within set parameters and in milliseconds. To reduce the operating time, the controller takes local measurements. The inverter along with its controller is unique to any type of eventualities, to make the microgrid reliable and efficient.
There are basically, two types of methods, they are P-Q and V-f controls.
- PQ Control theory
The normal situation is, Microgrid is in synchronism with main grid. In this mode the DGs are in P-Q control sharing the load as per equations (1), (2) & (3) and the excees power is exported to Grid. For incremental change in real power P is set by the phase angle and reactive power by voltage V.
Fig.3 Droop method of P-Q control
As shown in Figure.3, when there is any change in frequency and voltage, the DG output active and reactive power follows the droop characterstic.
- Droop Control theory
This method consists of two parts,
- Frequency droop Control which is active power to frequency, P-f
- Voltage droop Control which is reactive power to voltage, Q-v
- The relationship is given by the following equations,
is the DG output voltage frequency at no-load.
is the voltage at no-load.
is the slope of P-f curve
is the slope of Q-V curve..
- Q-v Droop Contrl:
The line impedance between DGs is small. Hence, there is flow of circulating currents, which causes voltage error st points.It is evident that the voltage droop control helps in voltage regulation. If the current output of DGs are capacitive, the voltage set point is lower but when the ouput is inductive, the set point is higher. But the limitation is within droop parameters. The characterstic curves are shown in Fig.4 below.
Fig.4 Voltage droop control characterstics Fig.5 Frequency droop control characterstics
- P/f droop control
Lot of circulating currents flow due to unbalanced reactive power sharing in isanding and the frequency droop control will eliminate this phenomena.. In the Fig.5, two DGs are considered with droop characterstics anlyse this criterion. When Microgrid is synchronized with grid, the frequency at PCC is unique, f0 as decided by main grid. During any type of fault the frequency changes to f1 and the load is shared by each DG as per droop characterstics.
5.Matlab / Simulink controller design
As long as the Microgrid is in grid mode the DGs deliver constant output. But in islanding mode, the DGs have share the balance of the load which is shared by grid. The section 4, model already discussed this phenomena. The output power is obtained by voltage magnitude and phase angle of Microgrid. The P-Q controller takes care of this situation. The output reale power of DG is obtained as per P-Q control diagram Fig.6 below.
Fig. 6 P-Q Control Scheme
The active power P of the DG is the input parameter of the controller to control the output voltage phase angle of the inverter. The reactive power Q of DG is to control the output voltage magnitude as per Q-v droop characterstic. The saturation limit blocks are to limit the droop of phase angle and voltage magnitude. System becomes un-stable if the variations are too much.
The P-Q controller model is shown in Figure.6, which is to provide P and Q settings. The signal generator controls whether DG ha sto operate in P-Q or V-f mode. The switch decide the type of droop whether it is P-Q or V-f control.
Fig.6 Reference P-Q generator
6.Simulation Model and Results
The MATLAB / Simulation model is shown in Fig. 7 below.
The model consists of two DGs with 3 loads. Loads 1 and 3 are essential loads, where as load 2 is non-essential. Microgrid is connected to grid via CB1. The battery energy storage must have the capacity to function autonomously in islanded mode with fully charged condition SOC. The islanding detection method is enabled by the controller to operate to islanding mode.
The MATLAB / Simulation model is shown in Fig. 7 below
Fig.7 Simulation model
The switching parameters are given in the table.1 given below.
The two modes grid to islanding and islanding to grid are simulated as below.
- Grid to islanding change over
The sampling period is divided into three sub-stages as detailed below.
First Stage : 0 - 0.3 secs
The Microgrid is connected to grid with P-Q control and the system frequency and voltage is dictated by grid. The DG output powers are as set values.
Secont Stage : 0.5 - 0.6 secs
At 0.6 secs circuit breaker 1 opened bringing Microgrid to islanding mode and controleer attains P-f droop mode. The DGs share the load according to P-f droop characterstics. But the load can be met by DGs totally and the frequency drops to 48Hz bringing the Microgrid to unstable state.
Third Stage : 0.6 - 1.5 secs
At 0.6 secs, the non-essential load is cut off and the system attains stability at 1.2 secs. The DGs started to share the load till the sampling period upto 1.5 secs.
Fig.8 Active power output VS time of DG1 and DG2
Fig.9 Frequency vs Time
Fig.10 Output Voltages of DG1 and DG2
Fig.11 Reactive Power vs Time of DG1 and DG2
The details of results of DG1 and DG2 are given below.
Figure 8 shows the active output power vs time.
Fig.9 shows frequency vs time of the system. The simulation results show that the parameters are as per the set values without much deviation. The frequency wave form is the system frequency. After the operation from grid to island, the system frequency is 49.65 and is stable after the load 2 shedding. The active power and reactive power are1700 and 1200 W.
Fig.10 shows the output voltages of DGs. with respect to time.
Fig.11 shows the reactive power of DGs with respect to time.
In grid mode, the output reactive power is constant but in island mode in island mode it follows P-f mode.
- Resynchronizing from islanding mode to grid mode
During faults on main grid, the Microgrid is isolated and operate in islanding mode in a set time which is millisecs. But the resynchronization is preplanned and will in asmooth transition. During Microgrid transition from grid to island the monitoring variations are large on either side of PCC. Hence without checking the system parameters, it is risk. Hence time is required as per the set time to avoid transients. Most practical way to bring to steady state is to open CB2 and CB3. Then the DGs will feed the local loads. After monitoring the voltage, phase angle and frequency at PCC, close CB2 and CB3. When all these variables are in the specified limits, then reclosing the Microgrid to Grid mode is possible. Then Microgrid turns into P-Q mode.
Microgrid can change over from grid mode to islanding with very little perturbations as per Microgrid definition of DOE, CERTS and IEE 1547 standards. In this paper SRF-PLL methodology of islanding detection is adopted which is proved to be effective, taking care of internal control and primary droop control loops. The results of the simulations shows the effectiveness and validity of proposal. The Microgrid, DGs run in PQ control mode when it is synchronized to the main Grid. This makes the Microgrid, with DGs and loads, behaves like a load to the main Grid. In the event of islanded mode, the DGs share the load according to P/f droop control theory. The appropriate design of controller and operating rules, makes the Microgrid transition states stable and smooth. But to improve the stability, the non-essential loads are to be avoided.
Future work for me and to any researcher is on the lines, “ to achieve better waveforms with less transients and less time of switch over from one mode to another mode”.