DC VOLTAGE DROP CONTROL METHOD WITH DEAD-BAND FOR HVDC GRIDS BASED ON DC VOLTAGE FIDUCIAL NODE

Abstract

The present invention discloses a fiducial node DC voltage based DC voltage droop control method with dead-band for HVDC grids. Two levels of DC voltage control e.g. primary and secondary DC voltage regulation are introduced to realize load sharing and DC voltage control in HVDC grids. In the process of primary DC voltage regulation, the power flow regulation ability of the entire HVDC grids can be significantly improved, and the DC voltage and stability of the HVDC grids will be quickly controlled and guaranteed for the benefit of droop characteristic. Secondary DC voltage regulation is achieved by by introducing the load-DC voltage controller. In the process of secondary DC voltage regulation, the burden of accommodating power imbalance by the DC voltage fiducial node will be alleviated, thus improving the ability to resist disturbances of the entire HVDC grids.

Claims

1. A DC voltage droop control method with dead-band for HVDC grids based on DC voltage fiducial node, comprising: classifying converter nodes in the voltage source converter (VSC) based HVDC grids into power adjustable node and power unadjustable node; designating the converter node featured with maximum capacity as the DC voltage fiducial node; controlling the converter at the DC voltage fiducial node through constant DC voltage control mode; controlling the power unadjustable nodes through constant AC voltage and constant frequency control mode; controlling the power adjustable node except for the DC voltage fiducial node through DC voltage droop with dead-band control mode.

2. The method of claim 1, wherein the step of classifying converter nodes in voltage source converter (VSC) based HVDC grids into power adjustable node and power unadjustable node further comprising: classifying converter nodes accessing the AC grid as the power adjustable nodes; classifying converter nodes directly connected to loads or solely integrated with renewable energies as the power unadjustable node.

3. The method of claim 1, wherein the step of controlling the power adjustable node except for the DC voltage fiducial node through DC voltage droop with dead-band control mode further comprising: determining a local droop power regulation reference P*.sub.droop according to the measured local DC voltage U.sub.dc for any power adjustable node except the voltage fiducial node; determining the local power reference P*.sub.dc according to the power regulation reference ΔP*.sub.dc generated by the load-DC voltage controller at regular intervals; generating the d-axis current reference i*.sub.d using the proportional-integral (PI) controller with input signal P*.sub.droop+P*.sub.dc−P.sub.dc; wherein P.sub.dc is the measured local active power at the power adjustable node; generating a three-phase AC voltage modulation signal using the newly-obtained d-axis current reference i*.sub.d and a ready q-axis current reference i*.sub.g.

4. The method of claim 3, wherein the step of determining the local droop power regulation reference P*.sub.droop further comprising: if U.sub.dc>U.sub.dcmax, then $P_{\mathrm{droop}}^{*}=\frac{U_{\mathrm{dcmax}}-U_{\mathrm{dc}}}{K};$ if U.sub.dc<U.sub.dcmin, then $P_{\mathrm{droop}}^{*}=\frac{U_{\mathrm{dcmin}}-U_{\mathrm{dc}}}{K};$ else if U.sub.dcmin<U.sub.dc<U.sub.dcmax, then P*.sub.droop=0; wherein K is the slope of a pre-defined DC voltage droop curve; the U.sub.dcmax and U.sub.dcmin are the maximum and minimum voltage of the dead-band respectively.

5. The method of claim 4, wherein the dead-band voltage U.sub.dcmax and U.sub.dcmin are determined respectively by the maximum and minimum steady-state local DC voltage taking all operation conditions into consideration.

6. The method of claim 3, wherein the step of determining the local power regulation reference P*.sub.dc further comprising: obtaining the local power regulation reference P*.sub.dc through the following formula:
P*.sub.dc(k+1)=P*.sub.dc(k)+ΔP*.sub.dc(k+1); wherein ΔP*.sub.dc(k+1) is the power regulation reference generated by the load-DC voltage controller at k+1 instant, P*.sub.dc(k+1) is the local power reference at k+1 instant, P*.sub.dc(k) is the local power reference at k instant, k is the natural number and P*.sub.dc(0) is determined by the initial power flow when k=0.

7. The method of claim 3, wherein the step of implementing the power regulation reference generated by the load-DC voltage controller at regular intervals further comprising: determining the power deviation reference ΔP*.sub.B by ΔP*.sub.B=P*.sub.dcB−P.sub.dcB; wherein P*.sub.dcB and P.sub.dcB are the power reference and the actual active power measured at the DC voltage fiducial node respectively; determining the power deviation reference ΔP*.sub.U by ΔP*.sub.U=K.sub.u(U*.sub.dcB−U.sub.dcB); wherein U*.sub.dcB is the pre-defined DC voltage reference, U.sub.dcB is the actual DC voltage measured at the DC voltage fiducial node, K.sub.u is the proportional gain; determining the total power deviation reference ΔP*.sub.grid by ΔP*.sub.grid=ΔP*.sub.U−ΔP*.sub.B; distributing the total power deviation reference ΔP*.sub.grid as power regulation reference to every power adjustable node accordingly at regular interval; wherein the distribution criterion further satisfies ΔP*.sub.grid=ΔP*.sub.dc1+ΔP*.sub.dc2+ . . . +ΔP*.sub.dcn+ΔP*.sub.dcB; where ΔP*.sub.dc1˜ΔP*.sub.dcn is the power regulation reference of power adjustable node 1 to node n except DC voltage fiducial node, ΔP*.sub.dcB is the power regulation reference of the DC voltage fiducial node, n is the total number of the power adjustable nodes except DC voltage fiducial node.

8. The method of claim 7, wherein the step of determining the local power reference P*.sub.dcB further comprising: obtaining the P*.sub.dcB through a formula of:
P*.sub.dcB(k+1)=P*.sub.dcB(k)+ΔP*.sub.dcB(k+1); wherein ΔP*.sub.dcB(k+1) is the power regulation reference of the DC voltage fiducial node at k+1 instant; P*.sub.dcB(k+1) is the local power reference of the DC voltage fiducial node at k+1 instant; P*.sub.dcB(k) is the local power reference of the DC voltage fiducial node at k instant; k is the natural number and P*.sub.dcB(0) is determined by the initial power flow when k=0.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0040] FIG. 1 shows a block diagram of the voltage droop controller with dead-band;

[0041] FIG. 2 shows a block diagram of the load-DC voltage controller;

[0042] FIG. 3 shows a characteristic curve of the droop controller with dead-band;

[0043] FIG. 4 shows a structure diagram of the four-terminal VSC based HVDC test system;

[0044] FIG. 5(a) shows a graph of active power of the four converters when the power reference of converter 1 changes;

[0045] FIG. 5(b) shows a graph of DC voltage of the four converters when the power reference of converter 1 changes;

[0046] FIG. 5(c) shows a graph of active power reference of converter 2, 3, 4 when the power reference of converter 1 changes;

[0047] FIG. 6(a) shows a graph of active power of the four converters when converter 4 is out of service;

[0048] FIG. 6(b) shows a graph of DC voltage of the four converters when converter 4 is out of service;

[0049] FIG. 6(c) shows a graph of active power reference of converter 2, 3, 4 when when converter 4 is out of service.

DETAILED DESCRIPTION OF THE INVENTION

[0050] First of all, all the converter nodes in the voltage source converter (VSC) based HVDC grids are classified into two groups, namely, power adjustable node and power unadjustable node. The detailed standards for the converter node classification are: those converter nodes connected to main AC grids are power adjustable nodes, whereas those converter nodes directly connected to loads or solely integrated with renewable energies are power unadjustable nodes. Among the power adjustable nodes, the node whose converter featured with maximum capacity is designated as the DC voltage fiducial node. The DC voltage deviation of the HVDC grids is thus defined as the deviation from this fiducial voltage.

[0051] DC voltage in HVDC grids has similar characteristics to the frequency in the AC grids, both of which are the measure of power balance in the network. In AC grids, load sharing and frequency control are realized by primary and secondary frequency regulation. Similarily, two levels of DC voltage control can be introduced to HVDC grids. Here we introduce primary and secondary DC voltage regulation to realize load sharing and DC voltage control in HVDC grids.

[0052] For primary DC voltage regulation, the converter at the DC voltage fiducial node is controlled through constant DC voltage control mode;The converters at the power unadjustable node are controlled through constant AC voltage and constant frequency control mode;All the other converters at the power adjustable node except DC voltage fiducial node are controlled through DC voltage droop control method with dead-band control mode.

[0053] Primary DC voltage regulation is the inherent response of the DC voltage droop control when HVDC grids subjected to disturbances. Generally, the primary DC voltage regulation functions within a duration of 500 milliseconds after the disturbance disappears. After this 500 milliseconds interval, the secondary (load-DC voltage control) DC voltage regulation takes effect. In this embodiment, the power reference of the power adjustable station is supposed to be refreshed every 500 milliseconds by the secondary DC voltage regulation system, similar to the updation of power reference of automatic generation control (AGC) power plant every certrain seconds realized by secondary frequency regulation in AC grids.

[0054] FIG. 1 shows the block diagram of the voltage droop controller with dead-band. In the process of the primary DC voltage regulation, the updated local power reference P*.sub.droop+P*.sub.dc is calculated according to the measured local active power P.sub.dc and DC voltage U.sub.dc, where P*.sub.dc(k+1)=P*.sub.dc(k)+ΔP*.sub.dc(k+1), ΔP*.sub.dc(k+1)=0;

[0055] FIG. 3 shows the characteristic curve of the droop controller with dead-band. In FIG. 3, K is the slope of the pre-defined DC voltage droop line; and U.sub.dcmax and U.sub.dcmin are the maximum and minimum voltage of the dead-band respectively. Specifically, the dead-band voltage U.sub.dcmax and U.sub.dcmin are determined respectively by the maximum and minimum steady-state local DC voltage taking all operation conditions into consideration. Similar to the constant frequency dead-band and speed droop coefficient parameters of AGC in AC grids, U.sub.dcmax, U.sub.dcmin and K remains constant in the process of primary and secondary DC voltage regulation. P*.sub.dc, P*.sub.dc1, P*.sub.dc2 and P*.sub.dc3 are the local power references updated by secondary DC voltage regulation system every 500 milliseconds, where P*.sub.dc(k+1)=P*.sub.dc(k)+ΔP*.sub.dc(k+1), ΔP*.sub.dc(k+1)≠0;

[0056] The constant DC voltage controlled fiducial node must be set up first in the normal operation of the HVDC grids. Since the DC voltage fiducial node serves as a slack bus in the HVDC grids, the power through this node varies with loads. In order to alleviate the burden of accommodating power imbalance by the DC voltage fiducial node, as well as to improve the ability of resisting disturbances in the entire HVDC grids, the present disclosure introduces secondary DC voltage regulation, which is also referred to load-DC voltage control.

[0057] FIG. 2 shows a block diagram of the load-DC voltage controller. The total power deviation reference ΔP*.sub.grid consists two parts: the first one is the power deviation reference ΔP*.sub.B determined by ΔP*.sub.B=P*.sub.dcB−P.sub.dcB, where P*.sub.dcB and P.sub.dcB are the power reference and the actual active power measured at the DC voltage fiducial node respectively; Also, P*.sub.dcB(k+1)=P*.sub.dcB(k)+ΔP*.sub.dcB(k+1). The second part is the power deviation reference ΔP*.sub.U determined by ΔP*.sub.U=K.sub.u(U*.sub.dcB−U.sub.dcB), where U*.sub.dcB is the pre-defined fiducial DC voltage reference, U.sub.dcB is the actual DC voltage measured at the DC voltage fiducial node, K.sub.u is the proportional gain. The total power deviation reference ΔP*.sub.grid is determined by ΔP*.sub.grid=ΔP*.sub.U−ΔP*.sub.B. During the process of secondary DC voltage regulation, ΔP*.sub.grid is distributed as power regulation reference to every power adjustable station accordingly at regular interval.

[0058] The implementation of the fiducial node DC voltage based DC voltage droop control method with dead-band will be illustrated further by the four-terminal VSC based HVDC test system (4) shown in FIG. 4. Note that the sign convention applies equally to active and reactive power: positive power means that converter draws power from AC grids.

[0059] All the converter nodes (9, 10, 11, 12) in the test HVDC grids (4) should be classified first. Apparently, The first converter (5) is integrated solely with renewable energies (like wind farms) (1), as a result, the first converter node (9) is a power unadjustable node and the first converter (5) adopts constant AC voltage control and frequency control. The second converter (6), the third converter (7) and the fourth converter (8) are connected to AC grids (3, 13, 14), thus the second converter node (10), the third node (11) and the forth node (12) belong to power adjustable node. In this test system (4), the fourth converter (8) is chosen as the DC voltage fiducial station for its largest power capacity and adopts constant DC voltage control with fiducial voltage set to ±500 kV. The second converter (6) and the third converter (7) adopt DC voltage droop control with dead-band (shown in FIG. 1). In the following discussion, the present disclosure focuses on the procedure of determining the proposed droop controller parameters.

[0060] For all the converters (5, 6, 7, 8) in the HVDC grids (4), maximum active power P.sub.dcmax and minimum active power P.sub.dcmin should be determined firstly. For each droop with dead-band controller, parameters that should be determined include the slope of the droop line K, DC voltage dead-band values U.sub.dcmax and U.sub.dcmin.

[0061] The maximum and minimum active power P.sub.dcmax and P.sub.dcmin are determined by the converter capacity. In the test system, since the second converter (6) is the sending-end station, its maximum and minimum active power are set to positive converter capacity and zero respectively. The third converter (7) operates in rectifier or inverter mode, as a result, its maximum and minimum active power are set to positive converter capacity and negative converter capacity respectively. The maximum and minimum active power of the fourth converter (8) are set to zero and negative converter capacity respectively. All maximum and minimum active powers of the droop-controlled converters are listed in TABLE 1.

TABLE-US-00001 TABLE 1 the first the second the third the fourth Converter converter converter converter converter Maximum power 1500 3000 1500 0 P.sub.dcmax/MW Minimum power 0 0 −1500 −3000 P.sub.dcmin/MW

[0062] Droop coefficient K is defined as the change in DC voltage that results in 100% change in converter power flow. Generally, the value of K varies from 4% to 5% from the practical engineering point of view. In the present test system, K is set to 4% for the second converter (6) and the third converter (7).

[0063] The maximum and minimum voltage of the dead-band U.sub.dcmax and U.sub.dcmin are determined respectively by the maximum and minimum steady-state local DC voltage taking all operation conditions into consideration. Considering the rational operation conditions for the test system, the four extreme operation conditions listed in TABLE 2 are able to cover all the working conditions. Based on the power reference of the extreme operation conditions and the fiducial DC voltage reference, the maximum and minimum voltage of the dead-band U.sub.dcmax and U.sub.dcmin can be achieved easily by the DC power flow calculation. In this example, the dead-band voltage U.sub.dcmax and U.sub.dcmin of droop-controlled the second converter (6) and the third converter (7) are calculated as U.sub.dc2max=506.423 kV, U.sub.dc2min=501.050 kV, U.sub.dc3max=504.606 kV, U.sub.dc3min=501.225 kV.

[0064] At this point, the parameters of all the droop controllers have been determined completely.

TABLE-US-00002 TABLE 2 the first the second the third the fourth the second the third fiducial Operation converter converter converter converter converter converter node mode P.sub.dc1/MW P.sub.dc2/MW P.sub.dc3/MW P.sub.dc4/MW U.sub.dc2/kV U.sub.dc3/kV U.sub.dc4/kV 1 1500 3000 −1500 −3000 506.423 501.225 500 2 0 3000 0 −3000 505.57 502.089 500 3 1500 0 1500 −3000 502.962 504.606 500 4 0 0 1500 −1500 501.05 502.734 500

[0065] The following disclosure demonstrates the effectiveness of the proposed control method in improving the stability of the entire HVDC grids when subjected to both small disturbances and large disturbances.

[0066] TABLE 3 shows the initial operation status of the test system. The following simulation cases are carried out on the PSCAD/EMTDC platform.

TABLE-US-00003 TABLE 3 Converter control method references of the controller the first constant AC voltage and U.sub.s1* = 220 kV; f.sub.0* = 50 Hz; converter frequency control wind power: P.sub.ac1 = 1000 MW; Q.sub.ac1 = 0 Mvar the second d-axis: DC voltage P.sub.dc2* = 2000 MW; Q.sub.ac2* = converter droop control with 0 Mvar dead-band; q-axia: constant reactive power control the third d-axis: DC voltage P.sub.dc3* = −500 MW; Q.sub.ac3* = converter droop control with 0 Mvar dead-band; q-axia: constant reactive power control the fourth d-axis: constant DC U.sub.dc4* = ±500 kV; Q.sub.ac4* = converter voltage control; 0 Mvar q-axia: constant reactive power control

[0067] (A) Small Disturbances:

[0068] In this case, the parameters of load-DC voltage controller (FIG. 2)are shown in TABLE 4. The active power reference will be updated by the secondary DC voltage regulation system (FIG. 2) every 0.5 second.

TABLE-US-00004 TABLE 4 the first the second the third the fourth Converter converter converter converter converter initial power 1000 2000 −500 determined reference by initial P.sub.dc*/MW power flow status desired power 1400 determined determined determined reference by load-DC by load-DC by load-DC P.sub.dc*/MW voltage voltage voltage controller controller controller proportional gain — — — 10 K.sub.u/MW/kV coefficient of — 0% 50% 50% power distribution

[0069] Suppose the test system has already been in steady state operation at t=0 s, the active power reference P*.sub.dc1 of the first converter (5) is changed from 1000 MW to 1400 MW at t=0.1 s. FIG. 5 shows the responses of the test system under this disturbance. Specifically, FIG. 5(a) shows a graph of active power of the four converters; FIG. 5(b) shows a graph of DC voltage of the four converters; FIG. 5(c) shows a graph of active power reference of the second, the third and the fourth converter.

[0070] As can be seen from FIG. 5, since the disturbance caused by power change in the first converter (5) is not so large, the power that the fourth converter (8) draws from DC grids (4) does not exceed its maximum capacity, which means the constant DC voltage control mode of the fourth converter (8) does not change. As a result, the DC voltage in the test system will not fluctuate severely. It is notable that as long as the DC voltage does not exceed the dead-band range, the droop-controlled with dead-band converters (6, 7) remains constant power transference.

[0071] The response of the entire test system caused by this small disturbance can be depicted as follows: at t=0.1 s, the active power injected to the HVDC grids (4) increases by 400 MW, which leads to a tendency of DC voltage increasing. Once the fiducial constant DC voltage controller at the fourth converter (8) detects the increasing DC voltage, the surplus power will be balanced by the fourth converter (8)at an early stage. Before the secondary DC voltage regulation system functions, the second converter (6) and the third converter (7) remains original constant power transference, the fourth converter (8)remains constant DC voltage control and the power transferred through the fourth converter (8)increases by 400 MW. Since the control period of the secondary DC voltage regulation is 0.5 second, the first update of power reference will occur at t=0.6 s. During the process of secondary DC voltage regulation, the power deviation reference ΔP*.sub.U, which is the second part of the total power deviation reference to the load-DC voltage controller, remains zero since the DC voltage has been well controlled to the fiducial reference. The total power deviation reference ΔP*.sub.grid is thus calculated as ΔP*.sub.grid=−ΔP*.sub.B, where the power deviation reference ΔP*.sub.B is the first part of the total power deviation reference to the load-DC voltage controller and ΔP*.sub.B is determined by ΔP*.sub.B=P*.sub.dcB−P.sub.dcB. Then ΔP*.sub.grid is distributed as power regulation reference to every power adjustable node (converter 2, 3 and 4) at 0.5 second interval, according to the coefficients of power distribution listed in TABLE. 4. This means that after t=0.6 s, the second converter (6), the third converter (7) and the fourth converter (8) operate with a updated power reference. As time goes by with another 0.5 second, the second update of power reference will occur at t=1.1 s with the same procedure depicted at the first period of secondary DC voltage regulation. This procedure will be repeated until the HVDC grids (4) enter to the desired operation status. As seen from FIG. 5, the power transferred by the second converter (6), the third converter (7) and the fourth converter (8) ultimately settle to 2000 MW, −700 MW and −2700 MW respectively, which indicates that the expected operation status has been fulfilled.

[0072] (B) Large Disturbances:

[0073] In this case, the parameters of load-DC voltage controller (FIG. 2) are shown in TABLE 5. Fiducial DC voltage station shall have enough capacity to accommodate power imbalance. As a result, the second converter (6) is selected as the backup fiducial DC voltage station responsible for fiducial DC voltage control when the master fiducial DC voltage station (8) is out of service. When the master fiducial station (8) breaks down, it takes time for the protection system to inform the backup fiducial DC voltage station (6) to take the role of DC voltage control. In this embodiment, the time delay is set to 50 milliseconds, which means the second converter (6) switches to constant DC voltage control 50 milliseconds after the fourth converter (8) is out of service.

TABLE-US-00005 TABLE 5 the second the fourth converter converter (backup (master fiducial DC fiducial DC the first voltage the third voltage Converter converter station) converter station) initial power 1000 2000  −500 determined by reference initial power P.sub.dc*/MW flow status desired power 1000 determined −1200 — reference by load-DC P.sub.dc*/MW voltage controller fiducial DC — ±500 kV — ±500 kV voltage reference U.sub.dc*/kV proportional —  10 — 10 gain K.sub.u/MW/kV coefficient of — 100% 0% 0% power distribution

[0074] Suppose that the test system has already been in steady state operation at t=0 s, the master fiducial DC voltage controlled the fourth converter (8) is out of service at t=0.1 s. FIG. 6 shows the responses of the test system after the fourth converter (8) is out of service. Specifically, FIG. 6(a) shows a graph of active power of the four converters; FIG. 6(b) shows a graph of DC voltage of the four converters; FIG. 6(c) shows a graph of active power reference of converter 2, 3, 4.

[0075] The response of the entire test system caused by this large disturbance can be depicted as follows: when master fiducial station (8) breaks down at t=0.1 s, the active power drawn from the HVDC grids (4) decreases by about 2500 MW, which leads to a quick increase of DC voltage. The increasing DC voltage exceeds the dead-band range of the droop controller, as a result, the primary DC voltage regulations of the second converter (6) and the third converter (7) take effect. Consequently, the power of the second converter (6) injected to the HVDC grids (4) decreases and the power of the third converter (7) drawn from the HVDC grids (4) increases. The backup fiducial the second converter (6) switches to constant DC voltage control at t=0.15 s, meanwhile, the desired power reference of the third converter (7) is set to −1200 MW. Since the control period of the secondary DC voltage regulation is 0.5 second, the first update of power reference occurs at t=0.65 s. During the early stage of secondary DC voltage regulation, the power deviation reference ΔP*.sub.U, which is the second part of the total power deviation reference to the load-DC voltage controller, is determined by ΔP*.sub.U=K.sub.u(U*.sub.dcB−U.sub.dcB). The total power deviation reference ΔP*.sub.grid is thus calculated as ΔP*.sub.grid=ΔP*.sub.U−ΔP*.sub.B, where the power deviation reference ΔP*.sub.B is the first part of the total power deviation reference to the load-DC voltage controller and ΔP*.sub.B is determined by ΔP*.sub.B=P*.sub.dcB−P.sub.dcB. Then ΔP*.sub.grid is 100 percent assigned to the second converter (6) at 0.5 second interval, according to the coefficients of power distribution listed in TABLE.5. This means that after t=0.65 s, the second converter (6) operates with a updated power reference. As time goes by with another 0.5 secondly, the second update of power reference will occur at t=1.15 s with the same procedure depicted at the first period of secondary DC voltage regulation. This procedure will be repeated until the HVDC grids enter to the desired operation status. As seen from FIG. 6, the power transferred by the second converter (6) and the third converter (7) ultimately settle to 200 MW, and −1200 MW respectively, which indicates that the expected operation status has been fulfilled.