DISTRIBUTED GENERATION CONTROL FOR MICROGRID DURING ISLANDING

Abstract

System and methods are disclosed to handle power imbalance in one or more distributed generation (DG) units: detecting islanding at time t1; selecting Phasor Measurement Unit (PMU) measurements during a pre-defined time window [t1-τ, t1]; checking the time window for an abrupt voltage change; if no sudden change of voltage is detected, determining an average value of sPMUt between t1-τ and t1 and using the average as a best estimate of a system power imbalance; if a sudden change of voltage is detected at time instant t2, determining the average value of sPMUt between t1-τ and t2 and using the average as a best estimate of the system power imbalance; and adjusting a power reference of the DG units based on the best estimated system power imbalance.

Description

BACKGROUND

This application relates to distributed generation control for microgrids.

The increasing penetration of distributed energy sources (DER) brings about the concept of microgrid (MG), in which distributed generation (DG) units and load are integrated in a low voltage (LV) network to enhance the reliability of the electric power systems. Integration of individual DGs not only benefits utilities and customers economically but also guarantees the continuity of the electricity supply under emergency conditions. Various types of DGs exist in a MG, including micro-turbine, photovoltaic (PV) system, fuel cell, battery storage, etc. These DGs are interfaced to the MG either through power electronic converters or through ac rotating machines. Compared to conventional rotating machines, DGs has the advantages of faster control capabilities with higher degree of flexibility. However, control of DGs remains a challenging area to explore, since they may bring as many problems as they may solve.

An MG should work under the grid-tied mode and the islanded mode. Under normal operating conditions, a MG is connected to a medium voltage network, either absorbing power from or injecting power into the main grid. Under the emergency mode, the MG is disconnected from the main grid and works autonomously, in a similar way to physical islands. The smooth transition between these two modes is of vital importance to realize the claimed benefits of MG.

SUMMARY

System and methods are disclosed to handle power imbalance in one or more distributed generation (DG) units: detecting islanding at time t₁; selecting Phasor Measurement Unit (PMU) measurements during a pre-defined time window [t₁τ, t₁]; checking the time window for an abrupt voltage change; if no sudden change of voltage is detected, determining an average value of s_PMU^t between t₁-τ and t₁ and using the average as a best estimate of a system power imbalance; if a sudden change of voltage is detected at time instant t₂, determining the average value of s_PMU^t between t₁-τ and t₂ and using the average as a best estimate of the system power imbalance; and adjusting a power reference of the DG units based on the best estimated system power imbalance.

Advantages of the preferred embodiments may include one or more of the following. The control strategy minimizes transients during the islanding process, and therefore, ensures the angle and voltage stability of the MG during and subsequent to the islanding process. The MG can operate under both grid-tied mode and islanded mode with a smooth transition between these two modes for high MG reliability. The system addresses DG controls for the smooth islanding of MG with a capability to minimize the transients and maintain MG stability. The system knows when to initialize and how to quantify the claimed controls to achieve the desired performance. The innovative DG control scheme, applied together with DG droop controls, minimizes system transients in the islanding process and ensures MG voltage and angle stability.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows a one-line diagram of an exemplary MG system.

FIG. 2 shows an exemplary control diagram of a PV system.

FIG. 3 shows an exemplary process for a Phasor Measurement Unit (PMU) based distributed generation control for a microgrid during an islanding process.

FIG. 4 shows another exemplary process for PMU based distributed generation control for a microgrid during the islanding process.

DETAILED DESCRIPTION

An innovative DG control scheme for MG islanding is discussed below. The control scheme, applied together with DG droop controls, minimizes system transients in the islanding process and ensures MG voltage and angle stability.

A model of an exemplary three-phase, single-feeder micro-grid with an one-line diagram is depicted in Error! Reference source not found. As FIG. 1 shows, the main grid is represented by a synchronous generator with transmission line L1-L2 and transformer T1. The MG is connected to the main grid at point of common coupling (PCC) via circuit breaker CB1. MG islanding can be achieved by opening CB1. Three DGs of different types are modeled: DG1-DG3. DG1 is modeled as a single-mass synchronous generator equipped with excitation and governor control. In practice, it represents either a diesel generator or a gas-turbine-generator unit. DG2 is a dispatchable power source interfaced to the feeder through a voltage-sourced converter. DG2 use decoupled d-q control and its active and reactive power outputs can be controlled independently. In real world, DG2 may also represent an energy storage unit (i.e., battery, capacitor, fly-wheel system) interfaced at the converter dc bus.

DG3 is a PV system, whose topology and control diagram are shown in Error! Reference source not found. FIG. 2. It includes multiple PV panels, a buck-boost converter, a dc-link capacitor, and a three-phase bipolar PWM inverter. Various controls are implemented in the PV system, including the maximum power point tracking (MPPT), phase lock loop (PLL), inner-loop dc-link voltage regulator and outer-loop output current controller.

A PMU is simulated at the main feeder to measure both the voltage at PCC and the current flowing into MG. The real-time voltage and current waveforms from PT and CT are transformed using Discrete Fourier Transform (DFT) to obtain the corresponding phasors. Loads are modeled in aggregated form as static passive components as functions of voltage and frequency. Real and reactive parts of the load are defined as (1)-(2).

$\begin{array}{cc}P=P_o·{\left(\frac{V}{V_o}\right)}^{\mathrm{NP}}\left(1+K_{\mathrm{PF}}·{\Delta }_f\right)& \left(1\right)\\ Q=Q_o·{\left(\frac{V}{V_o}\right)}^{\mathrm{NQ}}\left(1+K_{\mathrm{QF}}·{\Delta }_f\right)& \left(2\right)\end{array}$

where P_o, Q_o, V_o are the rated real power, reactive power, and voltage, respectively; P, Q, and V are the actual real/reactive power and voltage of the load; NP and NQ are voltage indexes for real and reactive power; K_PF and K_QF are frequency indexes of the real and reactive power. Parameters used in this study are provided in the Error! Reference source not found.

DGs and load in the system are connected to the main distribution feeder through impedance Z_DG1, Z_DG2, Z_PV, and Z_Load, respectively.

As shown in Error! Reference source not found., under grid-tied mode, the power (s_PMU^t) flowing from the main grid into the MG at any time instant t can be calculated using (3):

s _PMU ^t=3v_PCC^ti_PCC^t<(σ_V−σ_I)  (3)

where v_PCC^t<σ_V and i_PCC^t<σ_I are positive sequence voltage and current phasors measured by PMU.

The following power balance equation (4) can be used:

$\begin{array}{cc}{s}_{\mathrm{PMU}}^t+\sum _{i=1}^Ns_{{\mathrm{DG}}_i}^t-\sum _{j=1}^Ms_{{\mathrm{Load}}_j}^t-S_{\mathrm{loss}}=0& \left(4\right)\end{array}$

where s_DGi^t is generation of the ith DG; s_Loadj^t is the jth load; S_loss is MG system loss; N and M are numbers of DGs and loads in MG.

When MG is disconnected from the main grid, power injection from the grid s_PMU^t is cut off. The MG will go through a sophisticated transient process due to the pre-islanding disturbance, fault and generation-load imbalance. During islanding, voltages and system frequency deviate from the nominal values. Upon detection of such changes, each DGs in the MG will adjust its outputs accordingly based on the pre-defined control schemes to balance the load and to stabilize the system. The most commonly used control scheme is the droop control, the basic idea of which is to adjust DG output based on (5)-(6):

P _{DG i} =P _{DG i} ^ref −K _f ·Δf  (5)

Q _{DG i} =Q _{DG i} ^ref −K _V ·ΔV  (6)

where P_DGi^ref and Q_DGi^ref redefined real and reactive power settings; K_f is the frequency index for real power output while K_V is the voltage index for reactive power output; Δf and ΔV are frequency and voltage deviations from the nominal operating condition.

However, droop control has its limitation when used for MG islanding:

• • Droop control is sort of deviating regulation. The resulted voltage and frequency errors may not be acceptable based on existing standards.
  • Droop control is slow. DGs under droop control only reacts after voltage/frequency error is detected, when the best action time for system stability may have already passed.

Since electronically interfaced DGs can adjust their outputs in higher speed and degree of flexibility than the conventional rotating machines, the idea proposed here is to adaptively adjust the output power references of the DGs, immediately after the detection of islanding.

The MG system imbalance can be defined as the total power generation subtracted by total load and system loss. Islanding of MG can be either due to disturbances, such as a fault and its subsequent switching incidents, or due to preplanned switching events. For preplanned islanding, MG power imbalance can be estimated with PMU measurements collected right before the CB opens. For islanding caused by disturbance or fault, it is challenging to estimate the true system power imbalance since the system operating condition may have already deteriorated before the CB opens. Consequently, behaviors of voltage and current measurements at the PCC have to be studied carefully for the best estimation of power imbalance within the MG.

One embodiment uses the following process to estimate the power imbalance for both aforementioned cases.

Assuming islanding is detected at time instant t₁, collect voltage and current phasors at PCC during predefined time window [t₁-τ, t₁].

Process the voltage phasors from time instant t₁ to t₁-τ (backward), and search for abrupt voltage change by evaluating its rate of change: difference the time series and compare the difference with the defined threshold ε, as shown in (7):

Δv_PCC=|v_PCCⁿ−v_PCC^n-1|>ε  (7)

If no abrupt change of voltage is detected, calculate the average value of s_PMU^t within the specified time window [t₁-τ, t₁] using (3). Use this average as the best estimate of the MG power imbalance S_I before islanding.

If abrupt change of voltage is detected at time instant t₂, calculate the average of s_PMU^t in time window [t₁-τ, t₂] using (3). Use this average as the best estimate of the MG system power imbalance S_I.

The controller will send signals to electronically interfaced DGs adjusting their power reference based on the estimated system imbalance S_I, using (8)

$\begin{array}{cc}\Delta S_{{\mathrm{DG}}_i}=\Delta P_{{\mathrm{DG}}_i}+j·\Delta Q_{{\mathrm{DG}}_i}=c_i·S_Is.t.\sum _{i=1}^Kc_i=1& \left(8\right)\end{array}$

where ΔS_DGi, is adjustment to the power reference of the ith DG; K is the total number of electronically interfaced DGs in MG; c_i is the coefficient that determines the percentage of the total system imbalance assigned to the ith DG.

The length of the time window (τ) is critical for the estimation of MG power imbalance. It has been noticed that, if τ is too small, t₂ could be very close to time instant (t₁-τ). For extreme condition, t₂ could even be right before (t₁-τ). Under these conditions, the calculated value of S_PMU will be far away from the power imbalance, and certainly should not be used as the estimated MG system power imbalance. The selection of τ can be based on the islanding detection time. More specifically, the time window length be 2˜3 times the largest possible islanding detection time.

Two case studies have been conducted to verify the proposed control. It is recognized that severity of the transients experienced by the MG during and subsequent to islanding process is highly dependent on several factors, including pre-islanding system operating conditions, DG types in the MG, islanding detection time, type and location of the event triggering islanding. This paper makes no attempt to validate the proposed scheme exhaustively under all circumstances but rather demonstrate its concept and principle. Towards that end, the following assumptions are used for the two exemplary case studies:

Major portion of the load (80˜90%) in MG is supplied by local DGs while the rest is supplied by the utility (main) grid.

Output limits of DG2 are neglected, that is, DG2 can output whatever amount of power the controller requires.

Islanding is detected within 2 to 5 cycles depending on the severity of the event that initiates the process.

FIG. 3 shows an exemplary process for a Phasor Measurement Unit (PMU) based distributed generation control for a microgrid during an islanding process. First, the process records the PMU measurement at the point of common coupling (102). The process then determines islanding (104). If not, the process loops back to 102 and otherwise the process searches the voltage measurement during a time window for an abrupt voltage change (106). The process then changes for an abrupt voltage change (108). If so, the PMU measurement recorded during the disturbance is deleted (110). From 108 or 110, the process estimates a power imbalance using the remaining PMU measurement in the pre-determined time window (112). The process adjusts the power references of DGs based on the estimated system power imbalance (114). The process then performs smoothed islanding process (116).

FIG. 4 shows another exemplary process for PMU based distributed generation control for the microgrid during the islanding process C.1 for PMU based DG control during islanding process which uses the PMU for the microgrid islanding control. The process C.1 runs the following modules:

C1.1 PMU at PCC

• • To implement the proposed control scheme, a PMU needs to be installed at the point of common coupling to measure both the voltages at PCC and the currents flowing from the main grid into the MG.

C1.2 Power injection from main grid to MG

• • Under grid-tied operating conditions, the complex power s_PMU^t flowing from the main grid into the MG at any time instant can be calculated based on the PMU measurements using (3):

s _PMU ^t=3v_PCC^ti_PCC^t<(σ_V−σ_I)

where v_PCC^t<σ_v and i_PCC^t<σ_I are the positive sequence voltage and current phasors measured by the PMU.

C1.2.1. Power balance of MG

• • Under grid-tied operating conditions, the complex power s_PMU^t from the main grid in addition to the total DG generation is always equal to the total load and system loss:

$s_{\mathrm{PMU}}^t+\sum _{i=1}^Ns_{{\mathrm{DG}}_i}^t-\sum _{j=1}^Ms_{{\mathrm{Load}}_j}^t-S_{\mathrm{loss}}=0$

• • where s_DGi^t is the generation of the i th DG; s_Loadj^t is the j th load; S_loss is the loss within the MG; N and Mare the total number of DGs and loads in the MG.

C1.3 Estimation of power imbalance during islanding

• • When islanding occurs, the power within the MG no longer exists since the grid power injection s_PMU^t disappears abruptly. The MG system will go through a sophisticated transient process, during which the DG outputs and load will be adjusted based on their settings, power ratings, and system droop characteristics. Estimating and compensating the power imbalance can make the islanding process smooth and increase the stability of the system.

C1.3.1 Time window based backward search

• • In order to estimate the power imbalance, the bad PMU data need to be removed. For bad data detection, time window based backward search is proposed. Assuming islanding is detected at time t₁, consider only the PMU measurements during a pre-defined time window [t₁-τ, t₁]. Check the voltage measurements from time instant t₁ to t₁-τ (backward), and try to search for abnormal voltage change.

C1.3.2 Voltage based bad data detection

• • The abrupt change of voltage at PCC is used as a measure to detect the bad data in the PMU measurements. If sudden change of voltage is detected at time instant t₂, the PMU data between t₁ and t₂ are identified as bad data and only the rest of the PMU data are used to estimate the MG system power imbalance.

C1.3.3 Average PMU measurements

• • After deleting the bad data, rest of the data in the searched time window is used to calculate the average power flow from the main grid to the MG. This average value is used as the best estimate of the MG system imbalance.

C1.3.4 DG power reference adjustment

• • After computation of the estimated system imbalance S_PMU, the power reference of electronically DGs can be adjusted based on:

$\Delta S_{{\mathrm{DG}}_i}=\Delta P_{{\mathrm{DG}}_i}+j·\Delta Q_{{\mathrm{DG}}_i}=c_i·{\stackrel{\_}{S}}_{\mathrm{PMU}}$ $s.t.\sum _{i=1}^Kc_i=1$

• • where ΔS_DGi is the suggested adjustment to the power reference of the i th DG; K is the total number of electronically interfaced DG in the system; c_i is a coefficient that determines what percentage of the total system imbalance should be assigned to the i th DG.

C1.4 Simulation based system verification

• • A practical MG system is selected and simulated to verify the proposed control scheme. The microgrid under study consists of three DG units: DG1, DG2 and the PV system. The main grid is simulated by a generator with a transmission line and distribution transformer. By opening circuit breaker at PCC, the microgrid can be disconnected from the main grid.

C1.4.1 Detailed dynamic model of the DG

• • To simulate the transients during the islanding process, detail dynamic models of the DGs and load are constructed with various controls. The control parameters are tuned in such a way that the system has best dynamic characteristics.

C1.4.1.1 PV system topology and controls

• • The PV system modeled includes multiple PV panels, a dc-dc buck-boost converter, a bipolar PWM inverter, and an L filter. Various controls are implemented in the PV modeling, including the maximum power point controller (MPPT), phase lock loop (PLL), inner-loop dc link voltage controller and outer-loop output current controller. The output current of the inverter is synchronized with the grid voltage so that the PV system is operating at unity power factor.

C1.4.1.2 Independent PQ control of the dispatchable source

• • A dispatchable power source is modeled. It is interfaced to the MG through a voltage-sourced converter (VSC). The active and reactive power outputs of this dispatchable source are controlled independently.

C1.4.1.3 Droop control

• • The real power vs. frequency and reactive power vs. voltage droop controls are used for the control of the DGs. The DG output varies with the system frequency and voltage levels following:

P _{DG i} =P _{DG i} ^ref −K _f ·Δf

Q _{DG i} =Q _{DG i} ^ref −K _V ·ΔV

C1.4.1.4 Load modeling

• • Load is modeled in aggregated form in the system as a static passive component which is a function of the voltage and frequency of the system. The real and reactive power of the load are considered separately using:

$P=P_o·{\left(\frac{V}{V_o}\right)}^{\mathrm{NP}}\left(1+K_{\mathrm{PF}}·{\Delta }_f\right)$ $Q=Q_o·{\left(\frac{V}{V_o}\right)}^{\mathrm{NQ}}\left(1+K_{\mathrm{QF}}·{\Delta }_f\right)$

• • where P_o, Q_o, V_o are the rated real power, reactive power, and voltage of the load; P, Q, and V are the actual real/reactive power and voltage of the load; N_P and N_Q are the voltage index for real and reactive power; K_PF and K_QF are the frequency index of the real and reactive power.

The invention may be implemented in hardware, firmware or software, or a combination of the three. Preferably the invention is implemented in a computer program executed on a programmable computer having a processor, a data storage system, volatile and non-volatile memory and/or storage elements, at least one input device and at least one output device.

By way of example, a block diagram of a computer to support the system is discussed next. The computer preferably includes a processor, random access memory (RAM), a program memory (preferably a writable read-only memory (ROM) such as a flash ROM) and an input/output (I/O) controller coupled by a CPU bus. The computer may optionally include a hard drive controller which is coupled to a hard disk and CPU bus. Hard disk may be used for storing application programs, such as the present invention, and data. Alternatively, application programs may be stored in RAM or ROM. I/O controller is coupled by means of an I/O bus to an I/O interface. I/O interface receives and transmits data in analog or digital form over communication links such as a serial link, local area network, wireless link, and parallel link. Optionally, a display, a keyboard and a pointing device (mouse) may also be connected to I/O bus. Alternatively, separate connections (separate buses) may be used for I/O interface, display, keyboard and pointing device. Programmable processing system may be preprogrammed or it may be programmed (and reprogrammed) by downloading a program from another source (e.g., a floppy disk, CD-ROM, or another computer).

Each computer program is tangibly stored in a machine-readable storage media or device (e.g., program memory or magnetic disk) readable by a general or special purpose programmable computer, for configuring and controlling operation of a computer when the storage media or device is read by the computer to perform the procedures described herein. The inventive system may also be considered to be embodied in a computer-readable storage medium, configured with a computer program, where the storage medium so configured causes a computer to operate in a specific and predefined manner to perform the functions described herein.

The invention has been described herein in considerable detail in order to comply with the patent Statutes and to provide those skilled in the art with the information needed to apply the novel principles and to construct and use such specialized components as are required. However, it is to be understood that the invention can be carried out by specifically different equipment and devices, and that various modifications, both as to the equipment details and operating procedures, can be accomplished without departing from the scope of the invention itself

Claims

1. A process to handle power imbalance in one or more distributed generation (DG) units in a microgrid (MG) to supply a main grid, comprising: detecting islanding at time t1;selecting Phasor Measurement Unit (PMU) measurements during a pre-defined time window [t1-τ, t1];checking the time window for an abrupt voltage change;if no sudden change of voltage is detected, determining an average value of sPMUt between t1-τ and t1 and using the average as a best estimate of a system power imbalance;if a sudden change of voltage is detected at time instant t2, determining the average value of sPMUt between t1-τ and t2 and using the average as a best estimate of the system power imbalance; andadjusting a power reference of the DG units based on the best estimated system power imbalance.

2. The process of claim 1, wherein the PMU is installed at a point of common coupling (PCC) to measure voltages at PCC and the currents flowing from the main grid into the MG.

3. The process of claim 1, wherein a complex power sPMUt flowing from the main grid into the MG at any time instant is determined as: sPMUt=3vPCCtiPCCt<(σV−σI)where vPCCt<σv and iPCCt<σI are positive sequence voltage and current phasors measured by the PMU.

4. The process of claim 1, wherein complex power sPMUt from the main grid in addition to the total DG generation is equal to a total load and system loss:

5. The process of claim 1, comprising estimating power imbalance during islanding.

6. The process of claim 5, wherein when islanding occurs, power within the MG no longer exists and the MG adjusts DG outputs and load based on settings, power ratings, and system droop characteristics to make islanding smooth and increase system stability.

7. The process of claim 1, comprising checking voltage measurements backward from time instant t1 to t1-τ, and searching for abnormal voltage change.

8. The process of claim 1, comprising using an abrupt change of voltage at the PCC to detect bad data in PMU measurements.

9. The process of claim 1, comprising: detecting a sudden change of voltage at time instant t2 and identifying PMU data between t1 and t2 as bad data and using the rest of the PMU data to estimate an MG system power imbalance; anddetermining an average power flow from the main grid to the MG as the best estimate of an MG system imbalance.

10. The process of claim 9, wherein after determining the estimated system imbalance SPMU, adjusting the power reference of the DGs based on:

11. The process of claim 1, comprising performing simulation based system verification, wherein the main grid is simulated by a generator with a transmission line and distribution transformer, and wherein opening a circuit breaker at PCC disconnects the microgrid from the main grid.

12. The process of claim 11, comprising applying a dynamic model of the DG.

13. The process of claim 12, wherein the microgrid includes a photovoltaic (PV) system, comprising modeling the PV system with multiple PV panels, a dc-dc buck-boost converter, a bipolar pulse-width-modulation inverter, and a filter.

14. The process of claim 13, comprising modeling controls in the PV modeling, including maximum power point controller (MPPT), phase lock loop (PLL), inner-loop dc link voltage controller and outer-loop output current controller, wherein an output current of the inverter is synchronized with the grid voltage so that the PV system is operating at unity power factor.

15. The process of claim 1, comprising modeling a dispatchable power source interfaced to the MG through a voltage-sourced converter (VSC) and active and reactive power outputs of the dispatchable source are controlled independently.

16. The process of claim 1, wherein real power, frequency and reactive power, and voltage droop controls are used for the control of the DGs whose output varies with the system frequency and voltage levels: PDGi=PDGiref−Kf·Δf QDGi=QDGiref−KV·ΔV

17. The process of claim 1, comprising performing load modeling, wherein a load is modeled in aggregated form as a static passive component and as a function of voltage and frequency of the system.

18. The process of claim 17, wherein real and reactive power of the load are processed separately using:

19. A process to handle power imbalance in one or more distributed generation (DG) units in a microgrid (MG) to supply a main grid, comprising: detecting islanding at time instant t1 and collecting voltage and current phasors at a point of common coupling (PCC) during a predefined time window [t1-τ, t1];processing voltage phasors backward in time from time instant t1 to t1-τ;searching for an abrupt voltage change by evaluating its rate of change;if no abrupt change of voltage is detected, determining an average value of sPMUt within the time window [t1-τ, t1] and using the average as the best estimate of an MG power imbalance SI before islanding;if the abrupt change of voltage is detected at time instant t2, determining the average of sPMUt in time window [t1-τ, t2] and using the average as the best estimate of the MG power imbalance SI;adjusting the DG's power reference based on the estimated MG power imbalance SI as:

20. The process of claim 19, wherein the rate of change is determined by a difference in a time series and comparing the difference with a defined threshold ε: ΔvPCC=|vPCCn−vPCCn-1|>ε