High-voltage direct-current transmission subsequent commutation failure analysis method considering trigger angle deviation
Through system transient modeling and time-dividing analysis, the trigger angle deviation caused by phase-locked loop transient error is identified as the reason for the subsequent phase commutation failure of the LCC-HVDC system, which solves the lack of multi-electrical volume modeling and quantitative analysis in the prior art. The problem of the system improves the stability and safety of the system after failure.
Patent Information
- Application Number
- CN202410947349.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-16
- Publication Date
- 2024-11-08
AI Technical Summary
The existing LCC-HVDC system lacks effective multi-electrical volume modeling and quantitative analysis of the entire process after the inverter side AC failure, resulting in unclear mechanism of subsequent phase commutation failure, and the existing literature ignores the quasi-steady state formula for the first time The applicability after the phase exchange failure lacks comprehensive analysis.
Using system transient modeling, taking into account the dynamic characteristics of multiple electrical quantities and the role of DC control system, through time-separation modeling and transient commutation process analysis, the trigger angle deviation caused by phase-locked loop transient error is identified as subsequent commutation failure. The main reasons for this, and the impact of different fault moments is analyzed from the controller interaction characteristics.
In-depth analysis and accurate modeling of the subsequent phase exchange failure mechanism of the LCC-HVDC system was realized, the accuracy of the proposed method was verified, and the safe and stable operation ability of the system was improved.
Smart Images

Figure FT_1 
Figure FT_2 
Figure FT_3
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of high-voltage transmission networks. Background Art
[0002] High-voltage direct current transmission (LCC-HVDC) based on grid-commutated converters has been widely used in long-distance, large-capacity power transmission due to its advantages such as large transmission capacity, fast power regulation speed and low operating cost. However, LCC does not have the ability of self-commutation and is highly dependent on the AC power grid to provide commutation voltage for commutation. When an AC fault occurs on the inverter side, commutation failure is prone to occur. If the AC fault that caused the first commutation failure is not cleared in time, the large fluctuation of electrical quantities during system recovery may cause subsequent commutation failures, seriously endangering the safe and stable operation of the system. Therefore, it is of certain theoretical and engineering application value to establish a mathematical model that conforms to the transient commutation process of LCC-HVDC and conduct in-depth research on the mechanism of subsequent commutation failure.
[0003] In summary, although certain research results have been achieved in the study of the subsequent commutation failure mechanism of LCC-HVDC, the existing literature ignores the fact that the quasi-steady-state formula is not applicable to the transient process of the system after the first commutation failure, and lacks multi-electrical quantity modeling and quantitative analysis of the entire transient process of the system after the fault. At the same time, in the analysis of the commutation failure mechanism, most of them only unilaterally consider the adverse effects of the electrical quantity or the DC control system on the system recovery, and lack the study of the subsequent commutation failure mechanism that integrates multiple electrical quantities. Summary of the invention
[0004] The purpose of the present invention is to provide a method for analyzing subsequent commutation failure of high voltage direct current transmission taking into account trigger angle deviation by deeply analyzing the subsequent commutation failure mechanism starting from system transient modeling.
[0005] The steps of the present invention are: S1, converter valve and DC area mathematical model When the system operates normally, the rectifier and inverter valve areas satisfy equation (1): Where N is the number of pole pairs; k r and X r are the transformer ratio and leakage reactance of the rectifier side respectively; k i and X i are the transformer ratio and leakage reactance of the inverter side respectively; α r is the trigger angle on the rectifier side; β i The Echizen trigger angle on the inverter side; Differentiate equation (1) and get u dr and u di The differential equation is: where didr / dt and dα r / dt expression is shown in the mathematical model of DC control system below; Based on the DC system equivalent circuit model, the DC region KCL and KVL differential equations are listed as follows: Where R d and C are the DC line equivalent resistance and capacitance to ground respectively; L d is the sum of the rectifier-side smoothing reactor and the line reactance; u C is the capacitor voltage; S2. DC control system mathematical model The inverter side constant DC current control output trigger angle command value α r It is expressed as: α r =π-[k p1 (I dr_or d drm )+k i1 ∫(I dr_ord -i drm )dt] (4) where k p1 and k i1 They are the control proportion and integral gain of the fixed DC current on the rectifier side and the inverter side respectively; The DC control loop normalization process is described as: where i drm is the measured value of the DC current on the rectifier side; G1 and T1 are the proportional coefficient and time constant of the first-order inertia link respectively; I dN is the DC current reference value; From equations (4) and (5), we can get α r The differential equation is: Output DC current reference value I d_ord That is the DC current command value on the rectifier side where u dim and i dim are the measured values of DC voltage and DC current on the inverter side respectively; k v 、b v They are the slope and intercept of the three-stage VDCOL curve; Considering VDCOL action dr_ord and α′ r The differential equation is: Where T2 and T3 are udi and i di First-order inertia link time constant; u dN is the DC voltage reference value; Inverter side CC control output Echizen firing angle command value β CC It is expressed as: β CC =k p2 (I dr_ord -i dim -0.1)+k i2 ∫(I dr_ord -i dim -0.1)dt (9) where k p2 and k i2 They are the CC control proportional and integral gains on the inverter side respectively; Differentiating equation (9) yields β CC Differential Equations Similarly, the inverter side CEA control output is higher than the trigger angle command value β CEA It is expressed as: β CEA =k p3 (γ ref +γ CEC -γ * )+k i3 ∫(γ ref +γ CEC -γ * )dt (11) where k p3 and k i3 They are the inverter side CEA control proportional and integral gains respectively; From equations (8) and (11), we can get γ considering CEC action: CEC and β CEA Differential Equations where k i3 is the integral gain, k CEC is the proportionality coefficient; S3. Transient commutation process modeling S31, CP1: That is, the three valve groups are turned on, LCC fails to change phase, valve 4 is turned on, valve 2 fails to shut off and continues to be turned on, at this time valves 2, 3, and 4 are turned on at the same time This process dci and i di The satisfying relationship is as follows: where u ab and u cbThey are the commutation line voltages of valve 4 and valve 2 respectively; will u di Substituting into formula (3), we can get i under the three valve groups conduction state: di Differential Equations: S32, CP2: That is, the four valve groups are turned on. When valve 3 switches to valve 5 in the next phase change process, valves 2, 3, 4, and 5 are turned on at the same time. From formula (3), we can get i di The differential equation is expressed as: S33, CP1: that is, the three valve groups are turned on. After the short circuit of phases c and b occurs continuously, the reverse voltage causes valves 5 and 6 to be turned off. S34, CP3: that is, the two valve groups are turned on. When the commutation from valve 1 to valve 3 fails in the previous commutation process, valve 3 fails to be turned off and continues to be turned on. When the valve current i1 decreases to 0, valves 2 and 3 are turned on. This process di and i di There are the following relationships: Substituting equation (16) into equation (3), we can get i under the two valve groups are on state di Differential Equations: S35、CP 1 / 3 : That is, the 3 / 2 valve groups are turned on, and under the continuous action of VDCOL, i di Rapidly lowered, the phase change process from valve 2 to valve 4 proceeds smoothly; S36, CP4: That is, 0 valve groups are turned on. When valve 3 starts to switch to valve 5, due to I dr_ord Reduced to I dmin =0.55, under the CC control of the rectifier side i di Overshoot approaches 0, causing the commutation valve to fail to form a passage, and the commutation from valve 3 to valve 5 fails. di / dt=0; S4, subsequent commutation failure mechanism: S41, commutation voltage phase change The voltage drop longitudinal component ΔU and transverse component δU are expressed as: Among them, P ac , Q ac and Q C Respectively expressed as: where u di0 is the no-load DC voltage of LCC on the inverter side; From formula (19), we can get: s2 with U G2 The phase angle difference δ is expressed as: Simplify formula (21) to: In formula (1), di Substituting the expression into formula (22), we get P ac About Beta i The partial derivative of is: S42, trigger angle deviation Phase-locked loop error transfer function θ e (s) is: Where ξ is the damping ratio, ω n is the natural frequency, From formula (22), we get θ Us2 (s) is expressed as: Substituting equation (25) into equation (24) and performing inverse Laplace transform, we can obtain the time domain expression of the phase-locked loop transient error θ e (t): From the above formula, when the AC bus voltage on the inverter side shifts forward in phase, the transient error of the phase-locked loop will cause the phase of the phase-locked loop output to PLL Lag behind θ Us2 ; The maximum value of the actual amount of commutation voltage time area S pro_realm and the commutation voltage time area requirement S req It is expressed as: where kU s2 is the effective value of the AC bus voltage on the inverter side after the fault; γ min Subsequent commutation failure critical turn-off angle; the maximum commutation current i allowed by the system dimax for: The trigger angle deviation caused by the transient error of the phase-locked loop is expressed as: Δα=θ e (t) (29) In the time-phase modeling, Δα is added to equation (30): γ=β CEA -Δα-μ (30) Where μ is the inverter side commutation overlap angle; S5. Fault occurs during commutation: When the fault occurs during commutation, S promax It consists of two parts: S1 and S2. where i dN is the DC current reference value; i di (α1+μ2) is the DC current at the end of commutation; Find S promax For the fault time t f The partial derivative is: Where k is the fault severity; S6, the fault occurs during non-commutation period S promax It is composed only of S2 and is expressed as: where i d (α1) is the DC current at the start of commutation.
[0006] The present invention is based on the CIGRE HVDC standard system, and verifies the correctness of the proposed transient modeling method and the subsequent commutation failure mechanism analysis in PSCAD / EMTDC. The present invention has clear principle analysis and high modeling accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0007] Figure 1 It is a LCC-HVDC topology; Figure 2 is the post-fault system response characteristic; where (a) is the inverter side shut-off angle, (b) is the inverter side DC voltage, and (c) is the inverter side DC current; Figure 3 It is the equivalent circuit model of DC transmission line; Figure 4 It is the block diagram of LCC-HVDC control strategy; Figure 5 It is the equivalent circuit of the three-phase bridge converter on the inverter side; Figure 6 is the inverter side valve current waveform after the fault; Figure 7 It is the equivalent circuit of commutation process 1; Figure 8 It is the equivalent circuit of commutation process 2; Fig. 9 It is the equivalent circuit of commutation process 3; Fig.10 It is the equivalent model of the receiving system; Fig.11 is an ideal AC voltage source U G2 Structural equivalent model connected in series with equivalent impedance Z2; Fig.12 is the partial derivative of the active power generated by the inverter side with respect to the Echizen firing angle; Fig.13 is the effect of the DC control system on the phase angle difference; where (a) is the Echizen trigger angle command value on the inverter side, (b) is the total active power generated by the inverter side, and (c) is the phase angle difference; Fig.14 It is the SRF-PLL control structure; Fig.15 It is the trigger angle deviation caused by the transient error of the phase-locked loop; Fig.16 The fault occurs during commutation; Fig.17 It is the partial derivative of the maximum value of the commutation voltage-time area provided with respect to the fault moment; Fig.18 is the controller interaction characteristic during system recovery; (a) is the switch from CEA to CC control, (b) is the switch from CC to CEA control; Fig.19 The fault occurs during non-commutation period. DETAILED DESCRIPTION
[0008] The present invention discloses a method for analyzing the subsequent commutation failure mechanism of a high-voltage direct current transmission system (LCC-HVDC) based on a grid-commutated converter. After an AC fault occurs on the inverter side of the LCC-HVDC, the electrical quantity changes dramatically and the controller interaction is complex, resulting in an unclear mechanism of subsequent commutation failure. To this end, the present invention conducts an in-depth analysis of the subsequent commutation failure mechanism based on system transient modeling. First, based on the mathematical model of LCC transient commutation after the fault, a time-segment modeling method that takes into account the dynamic characteristics of multiple electrical quantities and the effects of the DC control system is proposed; secondly, through the analysis of the commutation voltage phase change, it can be obtained that the trigger angle deviation caused by the transient error of the phase-locked loop is the main cause of the subsequent commutation failure; then, starting from the controller interaction characteristics, the influence of different fault moments on the subsequent commutation failure is further analyzed; finally, the correctness of the proposed modeling method and the analysis of the subsequent commutation failure mechanism is verified by PSCAD / EMTDC simulation comparison.
[0009] The transient modeling method of the present invention taking into account the dynamic characteristics of multiple electrical quantities: ① Time-phase modeling method LCC-HVDC topology Figure 1 As shown. Among them: u dr and i dr are the DC voltage and DC current on the rectifier side respectively; u di and idi are the DC voltage and DC current on the inverter side respectively; U s1 and U s2 are the AC bus voltages on the rectifier side and the inverter side respectively. Figure 1 It can be obtained that the inverter side of the system is composed of LCC converters of YNd1 and YNy0 type connection in series. In the subsequent analysis, this paper takes the D bridge converter as an example for research, and the conclusions obtained are also applicable to the Y bridge converter.
[0010] In order to accurately describe the changing trends of various electrical quantities on the rectifier and inverter sides after a LCC-HVDC fault occurs, this section divides the time domain process after the fault occurs into three stages based on the transient response characteristics of the inverter side shutdown angle, DC voltage, and DC current. Figure 2 shown.
[0011] Depend on Figure 2 It can be obtained that the transient response process after the system fault can be divided into the first commutation failure period, the system recovery period and the subsequent commutation failure period, which correspond to stage 1 (t1-t2), stage 2 (t2-t4) and stage 3 (t4-t5) respectively. Since the turn-off angle γ in stage 1 and stage 3 drops to 0°, the system does not meet the commutation conditions. The error of using the quasi-steady-state formula to simplify the description of the trend of the change of each electrical quantity is large, so it is necessary to use transient commutation process modeling for analysis. In addition, when γ becomes a non-zero value at time t2, the system resumes normal commutation, so stage 2 can be analyzed using the traditional quasi-steady-state formula modeling method.
[0012] ② Mathematical model of converter valve and DC region When the system operates normally, the rectifier and inverter valve areas satisfy equation (1): In formula (1): N is the number of pole pairs; k r and X r are the transformer ratio and leakage reactance of the rectifier side respectively; k i and X i are the transformer ratio and leakage reactance of the inverter side respectively; α r is the trigger angle on the rectifier side; β i It is the Echizen trigger angle on the inverter side.
[0013] Differentiating equation (1) yields u dr and u di The differential equation is shown in formula (2): In formula (2): dr / dt and dα r / dt expression can be found in the DC control system mathematical model below. Since the rectifier side can adopt a quasi-steady-state model, u drSatisfies formula (1). And u di It fluctuates with the change of LCC transient commutation process after fault, and its change trend cannot be simplified by quasi-steady-state expression. di / dt See transient commutation process modeling below.
[0014] The LCC-HVDC DC transmission line consists of smoothing reactors and line DC line devices, which can be simplified as follows: Figure 3 The T-type equivalent circuit is shown.
[0015] Based on the DC system equivalent circuit model, the DC region KCL and KVL differential equations are listed as shown in formula (3): In formula (3): R d and C are the DC line equivalent resistance and capacitance to ground respectively; L d is the sum of the smoothing reactor on the rectifier side (inverter side) and the line reactance; u C is the capacitor voltage.
[0016] ③Mathematical model of DC control system The LCC-HVDC control system adopts a hierarchical control structure, in which the control layer is as follows: Figure 4 shown. Figure 4 Medium: I dr_order is the DC current command value on the rectifier side; β CC and β CEA They are the Echizen firing angle command values generated by the constant current control (CC) and constant extinction angle control (CEA) on the inverter side; γ min and γ CEC They are the minimum turn-off angle within 20ms and the current error controller (CEC) output; the subscript m represents the measured value of each electrical quantity after the first-order inertia link.
[0017] The LCC-HVDC rectifier side is equipped with CC and a minimum trigger angle control (min-αcontroller), where CC controls the output trigger angle command value α r It can be expressed as: α r =π-[k p1 (I dr_ord -i drm )+k i1 ∫(I dr_ord -i drm )dt] (4) In formula (4): k p1and k i1 They are the CC control proportional and integral gains on the rectifier side respectively.
[0018] Taking the CC control on the rectifier side as an example, the DC control loop normalization process can be described as: In formula (5): i drm is the measured value of the DC current on the rectifier side; G1 and T1 are the proportional coefficient and time constant of the first-order inertia link respectively; I dN is the DC current reference value, and is taken as 2kA with reference to the CIGRE HVDC standard model. The inverter side normalization process is consistent with the principle of formula (5), and will not be repeated below.
[0019] Based on the above analysis, we can get α from equations (4) and (5) r The differential equation is:
[0020] It should be noted that when the voltage dependent current order limiter (VDCOL) on the inverter side starts to operate, the output DC current reference value I d_ord is the DC current command value on the rectifier side, and the expression is shown in formula (7). At this time, I dr_order The change over time is not zero. In formula (7): u dim and i dim are the measured values of DC voltage and DC current on the inverter side respectively; k v =0.9, b v =0.19 are the slope and intercept of the three-segment VDCOL curve respectively.
[0021] Therefore, considering the I of VDCOL action dr_ord and α′ r The differential equation is shown in formula (8): In formula (8): T2 and T3 are u di and i di First-order inertia link time constant; u dN It is the DC voltage reference value, which is 500kV.
[0022] The LCC-HVDC inverter side is equipped with CC and CEA control. CC , β CEA The larger one is invested, and a smooth transition from CC control to CEA control is achieved through CEC control.
[0023] Inverter side CC control output Echizen firing angle command value β CC It can be expressed as: β CC =k p2 (I dr_ord -i dim -0.1)+k i2 ∫(I dr_ord -i dim -0.1)dt (9) In formula (9): k p2 and k i2 They are the CC control proportional and integral gains on the inverter side respectively.
[0024] Differentiating equation (9) yields β CC The differential equation is shown in formula (10):
[0025] Similarly, the inverter side CEA control output is higher than the trigger angle command value β CEA It can be expressed as: β CEA =k p3 (γ ref +γ CEC -γ * )+k i3 ∫(γ ref +γ CEC -γ * )dt (11) In formula (11): k p3 and k i3 They are the inverter side CEA control proportional and integral gains respectively.
[0026] when i di dr_ord When CEC is in action, the γ considering CEC action can be obtained from equations (8) and (11): CEC and β CEA The differential equation is shown in formula (12): k i3 is the integral gain, k CEC is the proportionality coefficient. Figure 1 It can be obtained that at the moment t2-t3 in stage 2, γ can be regarded as a constant value γ c unchanged, that is, dγ / dt=0; then, under the continuous action of CEA control, γ increases rapidly, and at t3-t4, γ>>γ ref
[16] , γ is approximately expressed as γ Continue to decrease to the baseline value, that is, dγ / dt=-kγ .
[0027] ④ Transient commutation process modeling During the transient process of the first commutation failure, the inverter-side LCC does not meet the normal commutation conditions. The equation relationship between the converter valve and the DC control loop varies with the commutation process. The quasi-steady-state expression is used to describe u di The error of the change trend is large. Therefore, this section classifies the LCC valve group according to the conduction conditions at different stages and establishes a mathematical model that conforms to the transient commutation process.
[0028] The equivalent circuit of the three-phase bridge converter on the inverter side is as follows: Figure 5 As shown, where L r are the equivalent commutation inductances on the inverter side. Take the case where the AC fault occurs when valve 2 and valve 3 are in the on state. Figure 6 As shown in the figure, there are four commutation processes from the time the fault occurs to the time the system resumes stable operation, namely: 0, 2, 3, and 4 valve groups are turned on. The following is divided into commutation processes (CP) based on the above valve group conditions. 1-4 ) is as follows: (1) CP1 (3 valve groups are on): After the fault occurs, when valve 2 switches to valve 4, the switching voltage drops sharply and does not meet the switching conditions, so the LCC fails to switch. After valve 4 is on, valve 2 fails to shut off and continues to be on. At this time, valves 2, 3, and 4 are on at the same time. The equivalent circuit is as follows: Figure 7 shown.
[0029] Figure 7 Medium: u a / b / c are the a / b / c three-phase voltages converted from the grid side to the valve side; i2 and i4 are the valve currents flowing through V2 and V4 respectively. Figure 7 This process u dci and i di The satisfying relationship is as follows: In formula (13): u ab and u cb They are the commutation line voltages of valve 4 and valve 2 respectively.
[0030] Further arrange the above formula and replace u di Substituting into formula (3), we can get i under the three valve groups conduction state: di Differential Equations: Therefore, when the three valve groups are connected as shown in CP1, you can refer to formula (14) to accurately describe i di Under different valve conduction conditions, only the corresponding commutation voltage needs to be replaced.
[0031] (2) CP2 (4 valve groups are turned on): When valve 3 switches to valve 5 in the next phase change process, valves 2, 3, 4, and 5 are turned on at the same time. The equivalent circuit is as follows: Figure 8 As shown. Since valve 2 and valve 5 are in the same bridge arm, a short circuit occurs on the DC side, u di = 0, and VDCOL starts to act to suppress i di Surge.
[0032] From formula (3), we can get: di The differential equation can be expressed as:
[0033] Therefore, when a DC side short circuit occurs in the LCC, equation (15) can be used to describe i di After the system has a phase c short circuit, when the next phase change interval valve 4 changes phase to valve 6, valves 2, 3, 5, and 6 are turned on, and the phase b of the changeover valve is short-circuited again, i di The differential equation is the same as (15).
[0034] (3) CP1 (three valve groups are turned on): After the short circuit of phases c and b occurs continuously, the reverse voltage causes valves 5 and 6 to be closed. When valve 5 switches to valve 1 in the next phase change process, since valve 5 is closed, phase change does not occur; similarly, when valve 6 switches to valve 2, the system also does not switch phases. Until valve 1 switches to valve 3, the switching valves are not completely switched, and the conduction valves are 1, 2, and 3. This process is i di The form of the differential equation is consistent with equation (14).
[0035] (4) CP3 (two valve groups are turned on): When the commutation from valve 1 to valve 3 fails in the previous commutation process, valve 3 fails to shut off and continues to conduct. When the valve current i1 decreases to 0, valves 2 and 3 are turned on. The equivalent circuit is as follows: Fig. 9 shown.
[0036] Depend on Fig. 9 This process u di and i di There are the following relationships:
[0037] Substituting equation (16) into equation (3), we can get i under the two valve groups conduction state: di Differential Equations:
[0038] Therefore, when the two valve groups are connected as shown in CP3, we can refer to formula (17) to describe i di Changing trends.
[0039] (5)CP 1 / 3(3 / 2 valve groups are on): Under the continuous action of VDCOL, i di The phase change from valve 2 to valve 4 is carried out smoothly. At this stage, the valve group changes from valves 2, 3, and 4 to valves 3 and 4. di The form of the differential equation is consistent with that of (14) 3 valve conduction and (17) 2 valve conduction.
[0040] (6) CP4 (0 valve groups are on): When valve 3 starts to switch to valve 5, due to I dr_ord Reduced to I dmin =0.55, under the CC control of the rectifier side i di Overshoot approaches 0, causing the commutation valve to fail to form a passage, and the commutation from valve 3 to valve 5 fails. di / dt=0.
[0041] Based on the above analysis, the system time-segment modeling method based on the transient commutation process proposed in this section aims to accurately fit the actual simulation results of PSCAD and provide a theoretical basis for mathematical modeling of the transient response process of the LCC-HVDC system after a fault.
[0042] The subsequent commutation failure mechanism of the present invention is as follows: The evolution mechanism of subsequent commutation failure of LCC-HVDC is as follows Fig.10 As shown in the figure, the trigger angle deviation caused by the transient error of the phase-locked loop during system recovery will lead to excessive recovery of DC current, rapid reduction of CEC output, and β CEA The slow increase directly compresses the turn-off angle margin, which is the main reason for the subsequent commutation failure of the system. The following will explain in detail the mechanism of subsequent commutation failure caused by the trigger angle deviation based on the causes and effects of the trigger angle deviation.
[0043] When the system is operating normally, substituting the rated parameters of the CIGRE HVDC standard model into the shutdown angle expression (18) yields: cosγ ref =0.965, cosβ ref =0.789, Therefore, the inverter side has a higher firing angle β i It is the dominant factor affecting the size of γ. i A low level may cause subsequent commutation failure in the system at the end of the recovery phase. In order to deeply analyze the mechanism of subsequent commutation failure in LCC-HVDC, the effect of the changes in various electrical quantities during the system transient recovery on β i The adverse effects
[0044] ① Analysis of commutation voltage phase change based on Figure 1The LCC-HVDC topology structure shown in the figure converts the receiving end AC system into an ideal AC voltage source U G2 The equivalent model of the structure in series with the equivalent impedance Z2 is as follows: Fig.11 shown.
[0045] Fig.11 Middle: P ac and Q ac They are respectively the active power and reactive power sent by the inverter-side AC system to the receiving-end power grid; Q C The reactive power is generated by the reactive power compensation device. Due to the existence of equivalent impedance Z2, U s2 with U G2 There is a voltage drop between s2 As the reference phasor, the voltage drop longitudinal component ΔU and transverse component δU can be expressed as: Among them, P ac , Q ac and Q C Can be expressed as: In formula (20): u di0 It is the no-load DC voltage of LCC on the inverter side.
[0046] From formula (19), we can get: s2 with U G2 The phase angle difference δ can be expressed as:
[0047] Since δ is mainly related to δU, and X>>R in HVDC transmission, the change of δU is mainly determined by P ac Influence, formula (21) can be simplified as:
[0048] From the above formula, we can get P ac The change of will directly affect the phase angle difference δ. On this basis, in order to further study the impact of the DC control system on power transmission during system recovery, u in formula (1) is replaced by di Substituting the expression into equation (22) we can get P ac About Beta i The partial derivative of is:
[0049] like Fig.12 As shown in the figure, substituting the rated parameters of the CIGRE HVDC standard model into the rated parameters, the P ac About Beta i The partial derivatives of are all less than 0, indicating that P acWith β i decreases and increases.
[0050] Based on the above analysis, the influence of the DC control system on δ is as follows: Fig.13 As shown, due to U G2 It can be regarded as an ideal voltage source, and its voltage waveform is always a sine wave, so the fluctuation of δ can be approximately expressed as U s2 The phase change. From the above figure, we can see that before time t3, P ac With β CEA The increase of δ leads to the decrease of U s2 Phase shift occurs, the actual value of the inverter side trigger angle α real Less than the trigger angle command value α ord , which is beneficial to phase change; after t3, as β CC Continuously decreasing, P ac Quick recovery, U s2 A phase shift occurs, which will cause α real Greater than α ord , which is not conducive to phase switching. s2 The effect of phase advance on trigger angle deviation and commutation margin is shown in Section 2.2 Trigger angle deviation analysis.
[0051] ② Analysis of trigger angle deviation High-voltage direct current transmission systems usually use SRF-PLL to achieve phase locking. In order to improve the anti-interference ability of the phase-locked loop, the harmonic suppression ability is usually enhanced by reducing the bandwidth.
[24] , but this will reduce the dynamic performance of the phase-locked loop to a certain extent. The SRF-PLL control structure is as follows Fig.14 shown.
[0052] Fig.14 Middle: k p and k i are the proportional gain and integral gain of the phase-locked loop PI control, k p =10, k i =50;θ Us2 (s),θ e (s) and θ PLL (s) are PLL input, deviation and output respectively.
[0053] Depend on Fig.14 It can be obtained that the phase-locked loop error transfer function θ e (s) is: In formula (24): ξ is the damping ratio, ω n is the natural frequency,
[0054] When the system fails to switch for the first time, U s2 Positive sequence voltage phase changes with power ΔP ac The fluctuation is equivalent to the phase deviation signal of the phase-locked loop input. From formula (22), we can get θ Us2 (s) can be expressed as:
[0055] Substituting equation (25) into equation (24) and performing inverse Laplace transform, we can obtain the time domain expression of the phase-locked loop transient error θ e (t):
[0056] From the above formula, when the AC bus voltage on the inverter side shifts forward in phase, the transient error of the phase-locked loop will cause the output phase of the phase-locked loop to PLL Lag behind θ Us2 Since LCC-HVDC adopts equal phase interval trigger control, the generation of trigger pulse depends on the trigger angle command value α ord With θ PLL Compared with Fig.15 As shown, when θ PLL Reach α ord When θ Us2 Increased to α in advance real This will result in α ord Less than α real , which directly compresses the γ margin and is not conducive to commutation.
[0057] In addition, from the perspective of the DC control system, when the trigger angle deviation Δα occurs in the system recovery phase, the actual maximum value S of the commutation voltage time area is pro_realm and the commutation voltage time area requirement S req It can be expressed as: In formula (27): kU s2 is the effective value of the AC bus voltage on the inverter side after the fault; γ min The critical turn-off angle for subsequent commutation failure is 7°.
[0058] From the above formula, we can get the maximum commutation current i allowed by the system. dimax for:
[0059] From the above formula, we can get that Δα>0 will lead to i dimax When it is lower than the DC current command value on the inverter side, the combined effect of CC control on the rectifier side and the inverter side will cause i diOver-recovery. During this period, on the one hand, di Increase will lead to S req To ensure successful commutation, the system can only increase the commutation overlap angle μ to increase S pro , γ has a decreasing trend; on the other hand, as the current deviation on the inverter side is gradually eliminated, the CEC output γ CEC Rapidly decreases to 0, causing β CEA The small increase makes it impossible for CEA control to play a key role in regulating γ to ensure sufficient commutation margin, which may lead to subsequent commutation failure.
[0060] Based on the above analysis, the trigger angle deviation caused by the transient error of the phase-locked loop can be expressed as: Δα=θ e (t) (29)
[0061] From the above formula, it can be obtained that Δα is added to formula (30) in the time period modeling to quantitatively determine whether the system has the risk of subsequent commutation failure in the early stage of stage 3. γ=β CEA -Δα-μ (30) In formula (30), μ is the inverter side commutation overlap angle. Considering that the commutation valve has basically restored normal commutation before the subsequent commutation failure occurs in the system, μ is taken as the rated value of 30° in this paper.
[0062] The impact of the fault moment of the present invention on subsequent commutation failure: Based on the analysis of the subsequent commutation failure mechanism, in order to further study the impact of different fault moments on subsequent commutation failure, this section analyzes the two working conditions where the fault occurs during commutation and during non-commutation.
[0063] ①The fault occurs during commutation First, the fault occurs during the commutation period. When the AC fault on the inverter side occurs during the commutation period, the system commutation process changes as follows: Fig.16 shown.
[0064] Fig.16 Middle: t f is the moment of fault occurrence; subscripts 1 and 2 represent the electrical angle before / after the fault and the commutation voltage time area provided respectively. When the fault occurs during the commutation period, S promax It is composed of two parts, S1 and S2, and can be expressed as: In formula (31): i dN is the DC current reference value; i di (α1+μ2) is the DC current at the end of commutation.
[0065] If the system experiences the first commutation failure under a severe fault, then γ in Equation (31) min = 0°. That is, the positive effect of increasing S by reducing γ has reached its upper limit. At this time, the larger S pro is, the larger the shortfall of the commutation voltage time area, and the worse the changing trend of each electrical quantity when the system enters the recovery stage. Calculating the partial derivative of S req with respect to the fault time t promax gives: f In Equation (32): k is the severity of the fault, 0 < k < 1; at the fault occurrence time α < ωt < π, so sin(ωt f ) > 0. As f shown, it holds for different degrees of AC bus voltage dips Fig.17 , so the earlier the fault occurs, the smaller S is. promax
[0066] Based on the above analysis, when the fault occurs during commutation, the earlier the fault occurs, the smaller S promax is, the earlier the surge of i di occurs, and the larger the DC current i di (α1 + μ2) at the end of commutation. The larger the difference between S promax and Sr eq is, which is equivalent to a more severe fault.
[0067] From the perspective of the changing characteristics of the DC control system, when the fault occurs earlier, the drop amplitude of 1u di is deeper under a severe fault, and VDCOL acts earlier to reduce i di . Since i d is inversely proportional to γ, γ recovers to the safe operating range earlier. As Fig.18 shown in (a), the switching process from CEA to CC control will occur earlier, causing i di to enter the recovery process earlier. In stage 2, due to the over - recovery of i di , γ decreases rapidly, resulting in a smaller β CC while β CEA gradually increases. As Fig.18 shown in (b), the switching process from CC to CEA control will also occur earlier. However, since γ CEC rapidly decreases to 0° at the end of stage 2, the increase of β CEA is slow and at a low level, and the system has a risk of subsequent commutation failure. Therefore, the two - time early interaction of the controllers during the system recovery is one of the changing characteristics of the DC control system before the subsequent commutation failure occurs.
[0068] In addition, from the perspective of the adverse effects of trigger angle deviation on commutation, since the system recovery process is inevitable, trigger angle deviation will occur regardless of whether the system has subsequent commutation failure. When the fault occurs later, it is equivalent to a minor fault, and the system recovery period is i di The increase is small, and the phase-locked loop can track U relatively well. s2 Phase, commutation voltage time area actual supply S pro_real Always greater than S req , the risk of subsequent commutation failure in the system is low.
[0069] Therefore, when a fault occurs during commutation, the earlier the fault occurs and the more serious the fault is, the faster the recovery speed of the DC current in phase 2 is, and the more unfavorable it is to commutation.
[0070] ②The fault occurs during non-commutation period Secondly, the fault occurs during the non-commutation period. When the AC fault on the inverter side occurs during the non-commutation period, the system commutation process changes as follows: Fig.19 shown.
[0071] Depend on Fig.19 Available, S promax It is composed only of S2 and can be expressed as: In formula (33): i d (α1) is the DC current at the start of commutation.
[0072] If the system fails to commutate for the first time under a more serious fault, then γ in equation (33) min =0°. Since the DC control system has not yet been regulated at the beginning of the fault, that is, α remains unchanged for a short time, S promax It does not change with the change of fault time. Therefore, when the fault occurs during the non-commutation period, the earlier the fault occurs, the greater the di The earlier the surge process, the d (α1) and i di The larger (α1+μ2) is, the greater the fault starting time i di Bigger, S req With S promax The larger the difference, the more serious the fault. In the case of a more serious fault, as VDCOL acts earlier, the two controller interactions occur earlier during system recovery, resulting in stage 2i di Fast recovery at the same time β i At a low level, the risk of subsequent commutation failure in the system increases.
[0073] Therefore, when a fault occurs during the non-commutation period, the earlier the fault occurs and the more severe the fault is, the faster the recovery speed of the DC current in phase 2 is, and the greater the probability of subsequent commutation failure in the system.
Claims
1. A method for analyzing subsequent commutation failure of high voltage direct current transmission taking into account trigger angle deviation, characterized in that: The steps are: S1, converter valve and DC area mathematical model When the system operates normally, the rectifier and inverter valve areas satisfy equation (1): Where N is the number of pole pairs; k r and X r are the transformer ratio and leakage reactance of the rectifier side respectively; k i and X i are the transformer ratio and leakage reactance of the inverter side respectively; α r is the trigger angle on the rectifier side; β i is the Echizen trigger angle on the inverter side; Differentiate equation (1) and get u dr and u di The differential equation is: where di dr / dt and dα r / dt expression is shown in the mathematical model of DC control system below; Based on the DC system equivalent circuit model, the DC region KCL and KVL differential equations are listed as follows: Where R d and C are the equivalent resistance of the DC line and the capacitance to ground respectively; L d is the sum of the rectifier-side smoothing reactor and the line reactance; u C is the capacitor voltage; S2. DC control system mathematical model The inverter side constant DC current control output trigger angle command value α r It is expressed as: α r =π-[k p1 (I dr_ord -i drm )+k i1 ∫(I dr_ord -i drm )dt] (4) where k p1 and k i1 They are the control proportion and integral gain of the fixed DC current on the rectifier side and the inverter side respectively; The DC control loop normalization process is described as: where i drm is the measured value of the DC current on the rectifier side; G1 and T1 are the proportional coefficient and time constant of the first-order inertia link respectively; I dN is the DC current reference value; From equations (4) and (5), we can get α r The differential equation is: Output DC current reference value I d_ord That is, the DC current command value on the rectifier side where u dim and i dim are the measured values of DC voltage and DC current on the inverter side respectively; k v , b v They are the slope and intercept of the three-stage VDCOL curve; Considering VDCOL action dr_ord and α' r The differential equation is: Where T2 and T3 are u di and i di First-order inertia link time constant; u dN is the DC voltage reference value; Inverter side CC control output Echizen firing angle command value β CC It is expressed as: β CC =k p2 (I dr_ord -i dim -0.1)+k i2 ∫(I dr_ord -i dim -0.1)dt (9) where k p2 and k i2 They are the CC control proportional and integral gains on the inverter side respectively; Differentiating equation (9) yields β CC Differential Equations Similarly, the inverter side CEA control output is higher than the trigger angle command value β CEA It is expressed as: b CEA =k p3 (c ref +g CEC -c * )+k i3 ∫(γ ref +g CEC -c * )dt (11) where k p3 and k i3 They are the inverter side CEA control proportional and integral gains respectively; From equations (8) and (11), we can get γ considering CEC action: CEC and β CEA Differential Equations where k i3 is the integral gain, k CEC is the proportionality coefficient; S3. Transient commutation process modeling S31, CP1: That is, the three valve groups are turned on, LCC fails to change phase, valve 4 is turned on, valve 2 fails to shut off and continues to be turned on, at this time valves 2, 3, and 4 are turned on at the same time This process dci and i di The satisfying relationship is as follows: where u ab and u cb They are the commutation line voltages of valve 4 and valve 2 respectively; will u di Substituting into formula (3), we can get i under the three valve groups conduction state: di Differential Equations: S32, CP2: That is, the four valve groups are turned on. When valve 3 switches to valve 5 in the next phase change process, valves 2, 3, 4, and 5 are turned on at the same time. From formula (3), we can get i di The differential equation is expressed as: S33, CP1: that is, the three valve groups are turned on, and after the c and b phases are short-circuited continuously, the reverse voltage causes valves 5 and 6 to be closed; S34, CP3: that is, the two valve groups are turned on. When the previous phase change process fails, valve 1 fails to change phase to valve 3, and valve 3 fails to close and continues to be turned on. When the valve current i1 decreases to 0, valves 2 and 3 are turned on. This process di and i di There are the following relationships: Substituting equation (16) into equation (3), we can get i under the two valve groups are on state di Differential Equations: S35、CP 1 / 3 : That is, the 3 / 2 valve groups are turned on, and under the continuous action of VDCOL, i di Rapidly lowered, the phase change process from valve 2 to valve 4 proceeds smoothly; S36, CP4: that is, 0 valve groups are turned on, when valve 3 starts to switch to valve 5, due to I dr_ord Reduced to I dmin =0.55, under the CC control of the rectifier side i di Overshoot approaches 0, causing the commutation valve to fail to form a passage, and the commutation from valve 3 to valve 5 fails. di / dt=0; S4, subsequent commutation failure mechanism: S41, commutation voltage phase change The voltage drop longitudinal component ΔU and transverse component δU are expressed as: Among them, P ac , Q ac and Q C Respectively expressed as: where u di0 is the no-load DC voltage of LCC on the inverter side; From formula (19), we get: s2 with U G2 The phase angle difference δ is expressed as: Simplify formula (21) to: In formula (1), di Substituting the expression into formula (22), we get P ac About Beta i The partial derivative of is: S42, trigger angle deviation Phase-locked loop error transfer function θ e (s) is: Where ξ is the damping ratio, ω n is the natural frequency, From formula (22), we get θ Us2 (s) is expressed as: Substituting equation (25) into equation (24) and performing inverse Laplace transform, we can obtain the time domain expression of the phase-locked loop transient error θ e (t): From the above formula, when the AC bus voltage on the inverter side shifts forward in phase, the transient error of the phase-locked loop will cause the phase of the phase-locked loop output to PLL Lag behind θ Us2 ; The maximum value of the actual amount of commutation voltage time area S pro_realm and the commutation voltage time area requirement S req It is expressed as: where kU s2 is the effective value of the AC bus voltage on the inverter side after the fault; γ min Critical turn-off angle for subsequent commutation failure; The maximum commutation current i allowed by the system dimax for: The trigger angle deviation caused by the transient error of the phase-locked loop is expressed as: Dα=θ e (t) (29) In the time-phase modeling, Δα is added to equation (30): c = b CEA -Da-m (30) Where μ is the inverter side commutation overlap angle; S5. Fault occurs during commutation: When the fault occurs during commutation, S promax It consists of two parts: S1 and S2. where i dN is the DC current reference value; i di (α1+μ2) is the DC current at the end of commutation; Find S promax For the fault time t f The partial derivative is: Where k is the fault severity; S6, the fault occurs during non-commutation period S promax It is composed only of S2 and is expressed as: where i d (α1) is the DC current at the start of commutation.
