MMC subunit switch fault and voltage sensor fault cooperative monitoring method and system

Abstract

The invention provides an MMC subunit switch fault and voltage sensor fault cooperative monitoring method and system. The method comprises the steps that faults are detected and positioned; reading the capacitor voltage and the bridge arm current of the fault subunit; and the faults are subdivided, and corresponding processing is carried out aiming at different faults. On the basis of a data driving method, a subunit capacitor voltage correlation coefficient is defined by utilizing the difference between the capacitor voltage of a fault subunit and the capacitor voltage of a normal subunit, whether a fault occurs or not can be quickly diagnosed, and if the fault occurs, the fault subunit can be directly positioned; according to the method, a fault distinguishing and judging standard is established, faults are divided into I-type faults and II-type faults through an index Ax according to analysis of fault characteristics, and then different strategies are adopted for further fault division; according to the method, higher harmonic and buffeting influence on the system when other functions are selected can be effectively inhibited.

Description

Technical Field

[0001] The present invention relates to the technical field of voltage sensors, and in particular to a method and system for collaboratively monitoring MMC subunit switch faults and voltage sensor faults. Background Art

[0002] In recent years, modular multilevel converters (MMCs) have been widely researched both domestically and internationally. Compared to traditional multilevel converters, MMCs not only inherit their topological structure but also offer significant technical advantages in terms of transmission losses and fault protection. Therefore, they have become the topology of choice for converter stations in flexible HVDC transmission systems. Due to the large number of IGBTs contained in MMCs, they undoubtedly carry a higher risk of failure. Failure to monitor and locate these failures in a timely manner could lead to even greater cascading failures. Furthermore, to distribute the switching control signals to each subunit, each subunit requires a voltage sensor to measure the subunit capacitor voltage, which also becomes a potential failure point. These two faults share similar characteristics, making distinguishing and accurately locating them a key challenge in ensuring the reliable operation of MMCs.

[0003] In recent years, fault monitoring and maintenance of MMCs have been widely studied both domestically and internationally. However, existing research primarily focuses on switch fault detection. There is a lack of technologies for detecting voltage sensor faults and distinguishing them from switch faults. Existing diagnostic methods for subunit switch open-circuit faults can be categorized into two main approaches: sensor methods and model methods. Model methods directly select appropriate algorithm models, train them based on data, and compare the electrical characteristics of subunits under normal operating conditions with those of the subunit under test to monitor and locate subunit open-circuit faults. This approach offers the advantages of high cost-effectiveness and robustness. However, limited research is currently underway on voltage sensor faults, with existing diagnostic methods primarily based on sliding mode observers for fault monitoring. However, since subunit switch open-circuit faults and voltage sensor faults exhibit similar fault characteristics, and the aforementioned technologies only address a single fault, there is a high risk of false positives for both faults. Therefore, a method for diagnosing both faults simultaneously is particularly necessary.

[0004] Therefore, it is necessary to propose a new technical solution to improve the above technical problems. Summary of the Invention

[0005] In view of the defects in the prior art, the object of the present invention is to provide a method and system for collaboratively monitoring MMC subunit switch faults and voltage sensor faults.

[0006] According to the present invention, a method for collaboratively monitoring MMC subunit switch faults and voltage sensor faults is provided, the method comprising the following steps:

[0007] Step S1: Detect and locate the fault;

[0008] Step S2: reading the capacitor voltage and bridge arm current of the faulty subunit;

[0009] Step S3: Segment the faults and perform corresponding processing for different faults.

[0010] Preferably, in step S1:

[0011] Each bridge arm has N subunits, defining the capacitance-voltage correlation coefficient:

[0012]

[0013] in, is the length of the sliding window, m and n are any two subunits on a bridge arm of the MMC, and is the average value of the capacitor voltage of subunits m and n in the window; by checking whether ρm,n ≈ 1, it is possible to detect whether a fault has occurred and locate the subunit where the fault is located.

[0014] Preferably, in step S2:

[0015] The estimated values ​​of the state variables are obtained based on the measured values ​​of the system's external variables. The state space system is expressed as:

[0016]

[0017] Where, is the observed state vector, u is the input vector, f is the switching function, A is the state gain matrix, B is the input gain matrix, and K is the observer gain vector;

[0018] The capacitor voltage output by the subunit is written as a switching function:

[0019]

[0020] At the same time, the voltage-current relationship of the capacitor is listed as follows:

[0021]

[0022] Since the three phases of MMC are symmetrical, combining formula (6) and formula (7), we get:

[0023]

[0024] in, and Respectively represent the observed current values ​​of the upper and lower bridge arms of phase j, represents the capacitance voltage observation value α of the i-th subunit of phase jpj With α nj is the bridge arm current observer switching function, and the hyperbolic tangent function tanh(x) is selected to express it as follows:

[0025]

[0026] Among them, K1 is the switching gain;

[0027] Combining formula (6) and formula (8), the sub-unit capacitor voltage observer of phase j is constructed as:

[0028]

[0029] Among them, K2 is the switching gain.

[0030] Preferably, in step S3:

[0031] Define the indicator Ax:

[0032]

[0033] When Ax belongs to [-ξ,ξ], ξ is set to 2% of the average voltage value during normal operation of the MMC, that is, Ax is determined to be 0. In this case, the fault is judged to be one of the following three faults: open circuit fault on the subunit, voltage sensor disconnection fault, and voltage sensor stuck fault. This situation is recorded as a Class I fault.

[0034] When Ax is far beyond this range, it indicates that the fault is one of the lower switch open circuit fault and voltage sensor gain fault, and this fault is recorded as a Type II fault.

[0035] Preferably,

[0036] For type I fault, the fault is an upper switch open circuit fault, modify the trigger signal Sij in the sliding mode observer, and set The current i and capacitance vc of the faulty subunit are observed. When a Type I fault occurs in subunit 1, the observer equation is:

[0037]

[0038]

[0039] like and If the fault lasts for 80ms and δ1 and δ2 are the defined current and voltage thresholds, the fault is an upper open circuit fault of the subunit; otherwise, the fault is determined to be a disconnection fault or a stuck fault of the voltage sensor; the indicator Bx is defined as:

[0040]

[0041] When Bx belongs to [-ξ,ξ], the fault is judged to be a voltage sensor disconnection fault, otherwise it is a voltage sensor stuck fault;

[0042] For type II faults, the fault is a lower switch open circuit fault, and the trigger signal Sij in the sliding mode observer is modified to observe the current i and capacitance vc of the faulty subunit; if and If it lasts for 80ms, the false fault is a lower open circuit fault of the subunit, otherwise it is a gain fault of the voltage sensor.

[0043] The present invention also provides a system for collaboratively monitoring MMC subunit switch faults and voltage sensor faults, the system comprising the following modules:

[0044] Module M1: Detect and locate faults;

[0045] Module M2: reads the capacitor voltage and bridge arm current of the faulty subunit;

[0046] Module M3: Subdivide the faults and handle them accordingly.

[0047] Preferably, in the module M1:

[0048] Each bridge arm has N subunits, defining the capacitance-voltage correlation coefficient:

[0049]

[0050] in, is the length of the sliding window, m and n are any two subunits on a bridge arm of the MMC, and is the average value of the capacitor voltage of subunits m and n in the window; by checking whether ρm,n ≈ 1, it is possible to detect whether a fault has occurred and locate the subunit where the fault is located.

[0051] Preferably, in the module M2:

[0052] The estimated values ​​of the state variables are obtained based on the measured values ​​of the system's external variables. The state space system is expressed as:

[0053]

[0054] Where, is the observed state vector, u is the input vector, f is the switching function, A is the state gain matrix, B is the input gain matrix, and K is the observer gain vector;

[0055] The capacitor voltage output by the subunit is written as a switching function:

[0056]

[0057] At the same time, the voltage-current relationship of the capacitor is listed as follows:

[0058]

[0059] Since the three phases of MMC are symmetrical, combining formula (6) and formula (7), we get:

[0060]

[0061] in, and Respectively represent the observed current values ​​of the upper and lower bridge arms of phase j, represents the capacitance voltage observation value α of the i-th subunit of phase j pj With α nj is the bridge arm current observer switching function, and the hyperbolic tangent function tanh(x) is selected to express it as follows:

[0062]

[0063] Among them, K1 is the switching gain;

[0064] Combining formula (6) and formula (8), the sub-unit capacitor voltage observer of phase j is constructed as:

[0065]

[0066] Among them, K2 is the switching gain.

[0067] Preferably, in the module M3:

[0068] Define the indicator Ax:

[0069]

[0070] When Ax belongs to [-ξ,ξ], ξ is set to 2% of the average voltage value during normal operation of the MMC, that is, Ax is determined to be 0. In this case, the fault is judged to be one of the following three faults: open circuit fault on the subunit, voltage sensor disconnection fault, and voltage sensor stuck fault. This situation is recorded as a Class I fault.

[0071] When Ax is far beyond this range, it indicates that the fault is one of the lower switch open circuit fault and voltage sensor gain fault, and this fault is recorded as a Type II fault.

[0072] Preferably,

[0073] For type I fault, the fault is an upper switch open circuit fault, modify the trigger signal Sij in the sliding mode observer, and set The current i and capacitance vc of the faulty subunit are observed. When a Type I fault occurs in subunit 1, the observer equation is:

[0074]

[0075]

[0076] like and If the fault lasts for 80ms and δ1 and δ2 are the defined current and voltage thresholds, the fault is an upper open circuit fault of the subunit; otherwise, the fault is determined to be a disconnection fault or a stuck fault of the voltage sensor; the indicator Bx is defined as:

[0077]

[0078] When Bx belongs to [-ξ,ξ], the fault is judged to be a voltage sensor disconnection fault, otherwise it is a voltage sensor stuck fault;

[0079] For type II faults, the fault is a lower switch open circuit fault, and the trigger signal Sij in the sliding mode observer is modified to observe the current i and capacitance vc of the faulty subunit; if and If it lasts for 80ms, the false fault is a lower open circuit fault of the subunit, otherwise it is a gain fault of the voltage sensor.

[0080] Compared with the prior art, the present invention has the following beneficial effects:

[0081] 1. Based on a data-driven approach, the present invention uses the difference in capacitance and voltage between faulty and normal subunits to define a subunit capacitance-voltage correlation coefficient. This allows for rapid diagnosis of faults and, if a fault occurs, direct location of the faulty subunit.

[0082] 2. The present invention establishes a fault differentiation and judgment standard. Based on the analysis of fault characteristics, the faults are first divided into Class I faults and Class II faults using the indicator Ax, and then different strategies are adopted to further classify the faults.

[0083] 3. The present invention designs a new sliding mode observer to select the hyperbolic tangent function tanh(x) with smooth and continuous characteristics, which can effectively suppress the high-order harmonics and chattering effects on the system when other functions are selected. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings:

[0085] Figure 1 This is a diagram of the fault diagnosis process of the present invention;

[0086] Figure 2 This is the basic topological structure diagram of the MMC of the present invention;

[0087] Figure 3 This is a topological diagram of the half-bridge subunit of the present invention;

[0088] Figure 4 This is a forward current diagram of the half-bridge subunit of the present invention in a locked state;

[0089] Figure 5 This is a reverse current diagram of the half-bridge subunit of the present invention in a locked state;

[0090] Figure 6 This is a forward current diagram of the half-bridge subunit of the present invention in the on-state;

[0091] Figure 7 This is a reverse current diagram of the half-bridge subunit of the present invention in the on state;

[0092] Figure 8 This is a forward current diagram of the half-bridge subunit of the present invention in a cut-off state;

[0093] Figure 9 This is a reverse current diagram of the half-bridge subunit of the present invention in a cut-off state;

[0094] Figure 10 This is a characteristic diagram of the switch open circuit fault of the present invention;

[0095] Figure 11 This is a fault characteristic diagram of the voltage sensor SF of the present invention;

[0096] Figure 12 This is a diagram showing the observed and actual values ​​of the bridge arm current and capacitor voltage of the upper switch sliding mode observer of the present invention;

[0097] Figure 13 This is a characteristic diagram of a switch open circuit fault under the present invention;

[0098] Figure 14 This is the fault characteristic diagram of the voltage sensor GF of the present invention;

[0099] Figure 15 This is a diagram showing the observed and actual values ​​of the bridge arm current and capacitor voltage of the switch sliding mode observer under the present invention. DETAILED DESCRIPTION

[0100] The present invention will be described in detail below with reference to specific embodiments. The following examples will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those skilled in the art, several changes and improvements can be made without departing from the scope of the present invention. These all fall within the scope of protection of the present invention.

[0101] Example 1:

[0102] According to the present invention, a method for collaboratively monitoring MMC subunit switch faults and voltage sensor faults is provided, the method comprising the following steps:

[0103] Step S1: Detect and locate the fault; each bridge arm has N subunits, and the capacitance-voltage correlation coefficient is defined as:

[0104]

[0105] in, is the length of the sliding window, m and n are any two subunits on a bridge arm of the MMC, and is the average value of the capacitor voltage of subunits m and n in the window; by checking whether ρm,n ≈ 1, it is possible to detect whether a fault has occurred and locate the subunit where the fault is located.

[0106] Step S2: Read the capacitor voltage and bridge arm current of the faulty subunit; obtain the estimated value of the state variable based on the measured value of the system's external variables. The state space system is expressed as:

[0107]

[0108] Where, is the observed state vector, u is the input vector, f is the switching function, A is the state gain matrix, B is the input gain matrix, and K is the observer gain vector;

[0109] The capacitor voltage output by the subunit is written as a switching function:

[0110]

[0111] At the same time, the voltage-current relationship of the capacitor is listed as follows:

[0112]

[0113] Since the three phases of MMC are symmetrical, combining formula (6) and formula (7), we get:

[0114]

[0115] in, and Respectively represent the observed current values ​​of the upper and lower bridge arms of phase j, represents the capacitance voltage observation value α of the i-th subunit of phase j pj With α nj is the bridge arm current observer switching function, and the hyperbolic tangent function tanh(x) is selected to express it as follows:

[0116]

[0117] Among them, K1 is the switching gain;

[0118] Combining formula (6) and formula (8), the sub-unit capacitor voltage observer of phase j is constructed as:

[0119]

[0120] Among them, K2 is the switching gain.

[0121] Step S3: Segment the faults and handle them accordingly; define the indicator Ax:

[0122]

[0123] When Ax belongs to [-ξ,ξ], ξ is set to 2% of the average voltage value during normal operation of the MMC, that is, Ax is determined to be 0. In this case, the fault is judged to be one of the following three faults: open circuit fault on the subunit, voltage sensor disconnection fault, and voltage sensor stuck fault. This situation is recorded as a Class I fault.

[0124] When Ax is far beyond this range, it indicates that the fault is one of the lower switch open circuit fault and voltage sensor gain fault, and this fault is recorded as a Type II fault.

[0125] For type I fault, the fault is an upper switch open circuit fault, modify the trigger signal Sij in the sliding mode observer, and set The current i and capacitance vc of the faulty subunit are observed. When a Type I fault occurs in subunit 1, the observer equation is:

[0126]

[0127]

[0128] like and If the fault lasts for 80ms and δ1 and δ2 are the defined current and voltage thresholds, the fault is an upper open circuit fault of the subunit; otherwise, the fault is determined to be a disconnection fault or a stuck fault of the voltage sensor; the indicator Bx is defined as:

[0129]

[0130] When Bx belongs to [-ξ,ξ], the fault is judged to be a voltage sensor disconnection fault, otherwise it is a voltage sensor stuck fault;

[0131] For type II faults, the fault is a lower switch open circuit fault, and the trigger signal Sij in the sliding mode observer is modified to observe the current i and capacitance vc of the faulty subunit; if and If it lasts for 80ms, the false fault is a lower open circuit fault of the subunit, otherwise it is a gain fault of the voltage sensor.

[0132] The present invention also provides a collaborative monitoring system for MMC subunit switch failures and voltage sensor failures. The collaborative monitoring system for MMC subunit switch failures and voltage sensor failures can be implemented by executing the process steps of the collaborative monitoring method for MMC subunit switch failures and voltage sensor failures. That is, those skilled in the art can understand the collaborative monitoring method for MMC subunit switch failures and voltage sensor failures as a preferred implementation of the collaborative monitoring system for MMC subunit switch failures and voltage sensor failures.

[0133] Example 2:

[0134] The present invention also provides a system for collaboratively monitoring MMC subunit switch faults and voltage sensor faults, the system comprising the following modules:

[0135] Module M1: Detects and locates faults; each bridge arm has N subunits, and defines the capacitance-voltage correlation coefficient:

[0136]

[0137] in, is the length of the sliding window, m and n are any two subunits on a bridge arm of the MMC, and is the average value of the capacitor voltage of subunits m and n in the window; by checking whether ρm,n ≈ 1, it is possible to detect whether a fault has occurred and locate the subunit where the fault is located.

[0138] Module M2: Reads the capacitor voltage and bridge arm current of the faulty subunit; obtains the estimated value of the state variable based on the measured value of the system's external variables. Its state space system is expressed as:

[0139]

[0140] Where, is the observed state vector, u is the input vector, f is the switching function, A is the state gain matrix, B is the input gain matrix, and K is the observer gain vector;

[0141] The capacitor voltage output by the subunit is written as a switching function:

[0142]

[0143] At the same time, the voltage-current relationship of the capacitor is listed as follows:

[0144]

[0145] Since the three phases of MMC are symmetrical, combining formula (6) and formula (7), we get:

[0146]

[0147] in, and Respectively represent the observed current values ​​of the upper and lower bridge arms of phase j, represents the capacitance voltage observation value α of the i-th subunit of phase j pj With α nj is the bridge arm current observer switching function, and the hyperbolic tangent function tanh(x) is selected to express it as follows:

[0148]

[0149] Among them, K1 is the switching gain;

[0150] Combining formula (6) and formula (8), the sub-unit capacitor voltage observer of phase j is constructed as:

[0151]

[0152] Among them, K2 is the switching gain.

[0153] Module M3: Segment the faults and handle them accordingly; define the indicator Ax:

[0154]

[0155] When Ax belongs to [-ξ,ξ], ξ is set to 2% of the average voltage value during normal operation of the MMC, that is, Ax is determined to be 0. In this case, the fault is judged to be one of the following three faults: open circuit fault on the subunit, voltage sensor disconnection fault, and voltage sensor stuck fault. This situation is recorded as a Class I fault.

[0156] When Ax is far beyond this range, it indicates that the fault is one of the lower switch open circuit fault and voltage sensor gain fault, and this fault is recorded as a Type II fault.

[0157] For type I fault, the fault is an upper switch open circuit fault, modify the trigger signal Sij in the sliding mode observer, and set The current i and capacitance vc of the faulty subunit are observed. When a Type I fault occurs in subunit 1, the observer equation is:

[0158]

[0159]

[0160] like and If the fault lasts for 80ms and δ1 and δ2 are the defined current and voltage thresholds, the fault is an upper open circuit fault of the subunit; otherwise, the fault is determined to be a disconnection fault or a stuck fault of the voltage sensor; the indicator Bx is defined as:

[0161]

[0162] When Bx belongs to [-ξ,ξ], the fault is judged to be a voltage sensor disconnection fault, otherwise it is a voltage sensor stuck fault;

[0163] For type II faults, the fault is a lower switch open circuit fault, and the trigger signal Sij in the sliding mode observer is modified to observe the current i and capacitance vc of the faulty subunit; if and If it lasts for 80ms, the false fault is a lower open circuit fault of the subunit, otherwise it is a gain fault of the voltage sensor.

[0164] Example 3:

[0165] The present application relates to fault diagnosis of subunits of a modular multilevel converter, and designs a method for simultaneously monitoring subunit switch faults and voltage sensor faults in combination with a sliding mode observer.

[0166] The present invention proposes a method that can quickly detect the occurrence of a fault and promptly locate the corresponding subunit simply by using the subunit capacitor voltage and bridge arm current. By adding a sliding mode observer, the fault type is further subdivided, and the three types of faults, namely, open circuit faults of the upper and lower switches of the MMC subunit and voltage sensor, can be simultaneously located and distinguished.

[0167] Aiming at the fault operation and maintenance of offshore wind power converter stations, the present invention innovatively proposes a method for simultaneously diagnosing subunit switch open-circuit faults and voltage sensor faults. Compared with existing technical solutions, the present invention has the following features: establishing a capacitor-voltage correlation coefficient index, which can effectively determine whether a fault has occurred in the circuit and quickly locate the subunit where the fault is located, so as to promptly disconnect the faulty subunit to maintain the normal operation of the MMC; based on the MMC's nearest level approximation modulation strategy, the specific changes in the trigger signals when the upper and lower switch open-circuit faults occur are obtained respectively, facilitating fault simulation; creating a new sliding mode observer, which effectively suppresses the high-order harmonics and chattering effects caused by the original observer on the system; based on the sliding mode observer, different faults with similar fault characteristics can be directly distinguished, with effectiveness and rapidity.

[0168] The technical solution applied for in the present invention is aimed at MMC subunit fault diagnosis, and is a collaborative diagnosis solution for both MMC subunit switch open circuit fault and voltage sensor fault.

[0169] Table 1 below shows the switch trigger signal in the case of an MMC subunit switch open circuit fault. Formulas (1), (2), and (3) are the fault characteristics of voltage sensor disconnect fault (DF), stuck fault (SF), and gain fault (GF), respectively. The specific diagnostic implementation scheme will be described based on this.

[0170]

[0171] Table 1 Subunit switch trigger signal under fault conditions

[0172]

[0173]

[0174]

[0175] The similarities between the characteristics of an open-circuit switch and a voltage sensor fault are as follows: after an open-circuit fault occurs on a subunit switch Tu, if the current is in the forward direction, as in normal conditions, the current flows through Du, charging the capacitor and causing the capacitor voltage to rise. However, if the current is in the reverse direction, D1 is turned on due to the forward voltage applied across D1 and the reverse voltage across Du, causing the capacitor to bypass and the capacitor voltage to remain unchanged. Because offshore wind turbine MMCs typically employ a nearest-level approximation modulation strategy, when the capacitor voltage of the faulty subunit reaches its maximum value among all subunits, switch Tu will remain off when the current is in the forward direction, bypassing the capacitor. However, when the current is in the reverse direction, as previously analyzed, the capacitor will also be bypassed. Therefore, when an open-circuit fault occurs on a subunit switch, the capacitor voltage of the faulty subunit will rise and remain constant after reaching its maximum value.

[0176] Formula (2) also shows that a stuck voltage sensor fault refers to a situation where the voltage sensor output value remains constant at the value at the time of the fault. Consider the case where the capacitor voltage of a subunit reaches its maximum value at the time of the fault. Since its capacitor voltage output value will continue to remain at that value, this situation will result in the same consequences as an open switch on the subunit and exhibit similar fault characteristics. Therefore, distinguishing between an open switch fault on a subunit and a stuck voltage sensor fault is one of the innovations of the present invention.

[0177] After an open-circuit fault occurs in switch Tl in a subunit, if the current is in the forward direction, due to the positive voltage applied across Du and the reverse voltage across Dl, Du is turned on, and current flows through Du, charging the capacitor, causing the capacitor voltage to rise. If the current is in the reverse direction, as in normal conditions, current flows through Tu, discharging the capacitor, causing the capacitor voltage to drop. Under the nearest level approach modulation strategy, when the current is in the reverse direction, the faulty subunit will remain in the on state, discharging the capacitor. Considering an MMC operating in inverter mode, since its positive current interval is longer than its negative current interval, the subunit capacitor voltage will increase more than it decreases within a current cycle, resulting in a periodic and continuous increase in the faulty subunit capacitor voltage.

[0178] Formula (4) also shows that a voltage sensor gain failure refers to a situation where the voltage sensor output value suddenly changes to the actual output value multiplied by a constant. When the gain is greater than or equal to 1.15, the value of the faulty voltage sensor will increase rapidly, but due to the existence of the MMC system's capacitor voltage control strategy, its output value will decrease to normal after several cycles. When the gain is less than or equal to 0.85, the faulty voltage sensor value will also decrease rapidly and increase back to normal after several cycles. This situation also shows similar fault characteristics to the open circuit switch failure of the sub-unit. Therefore, distinguishing the open circuit switch failure of the sub-unit from the voltage sensor gain failure is also one of the research focuses of this invention.

[0179] Fault diagnosis and location method: From the above MMC fault analysis, it can be seen that since subunit switch failures and voltage sensor failures have similar capacitor voltage fault characteristics, the capacitor voltage value cannot be directly used as the detection basis in actual fault diagnosis. However, due to the existence of redundant subunits in the MMC, the first step of this fault detection method is to quickly locate the location of the faulty subunit so that it can be disconnected in time to maintain the normal operation of the MMC. Assuming that each bridge arm has N subunits, the capacitor voltage correlation coefficient is defined as:

[0180]

[0181] in, is the length of the sliding window, m and n are any two subunits on a bridge arm of the MMC, and is the average capacitor voltage of subunits m and n in the window. Based on the fault characteristic analysis above, a subunit switch disconnection fault and the three types of voltage sensor faults will cause the capacitor voltage of the faulty subunit to differ from that of other normal subunits. Therefore, by checking whether ρm,n ≈ 1, we can directly detect whether a fault has occurred and quickly locate the faulty subunit.

[0182] Design of a new sliding mode observer: A sliding mode observer is a system that can obtain estimated values ​​of state variables based on the measured values ​​of the system's external variables. Its state space system can be expressed as:

[0183]

[0184] Where, is the observed state vector, u is the input vector, f is the switching function, A is the state gain matrix, B is the input gain matrix, and K is the observer gain vector.

[0185] The capacitor voltage output by the subunit can then be written as a switching function:

[0186]

[0187] At the same time, according to the voltage and current relationship of the capacitor, it can be listed:

[0188]

[0189] Since the three phases of MMC are symmetrical, only the unidirectional MMC is used as an example to construct the MMC bridge arm current and capacitor voltage sliding mode observer. Combining formula (6) and formula (7), we can get:

[0190]

[0191] in, and Respectively represent the observed current values ​​of the upper and lower bridge arms of phase j, represents the capacitance voltage observation value α of the i-th subunit of phase j pj With α nj is the switching function of the bridge arm current observer. In this paper, the hyperbolic tangent function tanh(x) is used to express it as follows:

[0192]

[0193] Where K1 is the switching gain. The hyperbolic tangent function tanh(x) is chosen because it is smooth and continuous, and can effectively suppress the high-order harmonics and chattering effects of other functions on the system.

[0194] Combining formula (6) and formula (8), the sub-unit capacitor voltage observer of phase j can be constructed as:

[0195]

[0196] Among them, K2 is the switching gain.

[0197] Fault differentiation criteria: Based on the fault characteristic analysis above, an upper subunit open circuit fault has similar fault characteristics to a voltage sensor disconnection fault or a stuck fault, while a lower subunit open circuit fault has similar characteristics to a voltage sensor gain fault. Define the indicator Ax:

[0198]

[0199] To eliminate the effects of harmonics and noise errors, we assume that when Ax falls within the range [-ξ,ξ] (ξ is set to 2% of the average voltage during normal MMC operation), Ax is considered zero. In this case, the fault is determined to be one of the following: an upper subunit open circuit fault, a voltage sensor disconnection fault, or a voltage sensor stuck fault. This is classified as a Type I fault. When Ax is significantly outside this range, it indicates a lower switch open circuit fault or a voltage sensor gain fault. This is classified as a Type II fault.

[0200] For type I fault, assuming that the fault is an upper switch open circuit fault, modify the trigger signal Sij in the sliding mode observer according to Table 1 (replace Sij with SF), and set Observe the current i and capacitance vc of the faulty subunit. Assume that a type I fault occurs in subunit 1. The observer equation at this time is:

[0201]

[0202] like and (δ1 and δ2 are defined current and voltage judgment thresholds) If the condition persists for 80ms, the assumption is correct and the fault is an upper open circuit fault of the subunit; otherwise, the fault is determined to be a disconnection fault or a stuck fault of the voltage sensor. Define the indicator Bx:

[0203]

[0204] When Bx belongs to [-ξ,ξ], the fault is judged to be a voltage sensor disconnection fault, otherwise it is judged to be a voltage sensor stuck fault.

[0205] For type II faults, it is assumed that the fault is a lower switch open circuit fault. The trigger signal Sij in the sliding mode observer is modified according to Table 1. The current i and capacitance vc of the faulty subunit are also observed. and If it lasts for 80ms, the assumption is correct and the fault is a lower open circuit fault of the subunit. Otherwise, it is a gain fault of the voltage sensor. The whole fault diagnosis process is as follows: Figure 1 shown.

[0206] In order to verify the feasibility of the proposed method for collaborative fault monitoring of MMC subunit switches and voltage sensors, a simulation system was built in Matalaab / Simulink environment. Figure 1 The MMC three-phase converter shown in FIG. 1 has a structure of a subunit SMn as shown in FIG. Figure 2 As shown, its working principle in three modes is as follows Figure 3 、 Figure 4 、 Figure 5 As shown, the circuit parameters are shown in Table 3:

[0207] Table 3 Simulation circuit parameters

[0208]

[0209]

[0210] Normal working condition simulation verification: Figure 1 As shown in FIG, when the MMC is in normal working state, assuming that it works in the inverter state, it has the following characteristics during stable operation:

[0211] The DC voltage is maintained at a constant value and an AC voltage is output. The MMC DC voltage is maintained by three phase units connected in parallel, and all three phases use a connection method in which the upper and lower bridge arms are connected in series. Therefore, it is necessary to ensure that the number of sub-units in the active state in each phase unit is equal and unchanged to avoid distortion of the output AC voltage waveform and the generation of inter-phase circulating current. In general, if the total number of sub-units in a phase unit is 2N, the number of sub-units in the active state is N, and the maximum number of levels that can be output is (N+1). Since the number of sub-units in the active state in each phase unit is always a constant value, the output AC voltage can be adjusted by switching the sub-units of each bridge arm on and off. The voltage of each bridge arm can be expressed as:

[0212] u pa + u na = u pb + u nb = u pc + u nc = U dc (15)

[0213] At the same time, due to the symmetry of the three-phase parameters, the DC bus current Id will ideally be evenly distributed among the three phase units, meaning that the DC current flowing through each phase unit is Id / 3. Taking phase a as an example, since the upper and lower bridge arms have the same reactor Larm, the AC current ia will be evenly divided between the upper and lower bridge arms. In summary, the upper and lower bridge arm currents of phase a are:

[0214]

[0215] Fault diagnosis simulation verification:

[0216] Distinguishing between upper switch open circuit fault and voltage sensor SF fault of subunit: Assume that upper switch open circuit fault and voltage sensor SF fault occur in subunit 1 at 0.5105s, respectively. The fault characteristics are as follows: Figure 7 、 Figure 8 As shown in Figure 3, it can be seen that when these two faults occur, parameters such as the three-phase current and the capacitor voltage of the faulty sub-unit will show similar fault characteristics.

[0217] First, according to the capacitor voltage correlation coefficient ρmn, the fault subunit can be located as subunit 1, and the fault is subdivided at 0.511s. Then, according to the index Ax∈[-ξ,ξ], it can be concluded that the two types of faults are both type I faults. Therefore, the upper switch sliding mode observer is used for subunit 1, that is, assuming that the fault is an open circuit fault of the upper switch of the subunit, the observed and actual values ​​of the bridge arm current and capacitor voltage in the two cases are as follows: Figure 9 shown.

[0218] It can be found that when the fault is the voltage sensor SF fault, there is a bridge arm current The assumption is wrong, and the fault is successfully diagnosed as the voltage sensor SF fault of subunit 1; when the fault is the upper switch open circuit fault, due to and And the duration exceeds 80ms, so it can be judged that the assumption is correct and the fault is successfully diagnosed as the lower open circuit fault of subunit 1.

[0219] Distinguishing between the lower switch open circuit fault and the voltage sensor GF fault of the subunit: Assume that the lower switch open circuit fault and the voltage sensor GF fault occur in subunit 2 at 0.5105s. The fault characteristics are as follows: Figure 10 、 Figure 11 As shown in Figure 3, it can be seen that when these two faults occur, parameters such as the three-phase current and the capacitor voltage of the faulty sub-unit will show similar fault characteristics.

[0220] Similarly, according to the capacitance-voltage correlation coefficient ρmn, the faulty subunit can be located as subunit 2, and according to It can be concluded that the fault type is a type II fault, so the lower switch sliding mode observer is used for subunit 2, that is, assuming that the fault is an open circuit fault of the lower switch of the subunit, the observed values ​​and actual values ​​of the bridge arm current and capacitor voltage in the two cases are as follows: Figure 12 shown.

[0221] It can be found that when the fault is the voltage sensor GF fault, there is a capacitor to carry you The assumption is wrong, and the fault is successfully diagnosed as the voltage sensor GF fault of subunit 2; when the fault is the lower switch open circuit fault, due to and And the duration exceeds 80ms, so it can be determined that the assumption is correct and the fault is successfully diagnosed as a lower open circuit fault of subunit 2.

[0222] Those skilled in the art may understand this embodiment as a more specific description of Embodiment 1 and Embodiment 2.

[0223] Those skilled in the art will appreciate that, in addition to implementing the system and its various devices, modules, and units provided by the present invention in purely computer-readable program code, it is entirely possible to implement the same functions of the system and its various devices, modules, and units provided by the present invention in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; the devices, modules, and units for implementing various functions can also be considered as both software units implementing the method and structures within the hardware component.

[0224] The above describes specific embodiments of the present invention. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art may make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. The embodiments of this application and the features in the embodiments may be combined with each other in any manner unless there is a conflict.

Claims

1. A method for collaboratively monitoring MMC subunit switch faults and voltage sensor faults, characterized in that: The method comprises the following steps: Step S1: Detect and locate the fault; Step S2: reading the capacitor voltage and bridge arm current of the faulty subunit; Step S3: Segment the faults and perform corresponding processing for different faults.

2. The method for collaboratively monitoring MMC subunit switch failure and voltage sensor failure according to claim 1, characterized in that: In the step S1: Each bridge arm has N subunits, defining the capacitance-voltage correlation coefficient: in, is the length of the sliding window, m and n are any two subunits on a bridge arm of the MMC, and is the average value of the capacitor voltage of subunits m and n in the window; by checking whether ρm,n ≈ 1, it is possible to detect whether a fault has occurred and locate the subunit where the fault is located.

3. The method for collaboratively monitoring MMC subunit switch failure and voltage sensor failure according to claim 1, characterized in that: In the step S2: The estimated values ​​of the state variables are obtained based on the measured values ​​of the system's external variables. The state space system is expressed as: Where, is the observed state vector, u is the input vector, f is the switching function, A is the state gain matrix, B is the input gain matrix, and K is the observer gain vector; The capacitor voltage output by the subunit is written as a switching function: At the same time, the voltage-current relationship of the capacitor is listed as follows: Since the three phases of MMC are symmetrical, combining formula (6) and formula (7), we get: in, and Respectively represent the observed current values ​​of the upper and lower bridge arms of phase j, represents the capacitance voltage observation value α of the i-th subunit of phase j pj With α nj is the bridge arm current observer switching function, and the hyperbolic tangent function tanh(x) is selected to express it as follows: Among them, K1 is the switching gain; Combining formula (6) and formula (8), the sub-unit capacitor voltage observer of phase j is constructed as: Among them, K2 is the switching gain.

4. The method for collaboratively monitoring MMC subunit switch failure and voltage sensor failure according to claim 1, characterized in that: In the step S3: Define the indicator Ax: When Ax belongs to [-ξ,ξ], ξ is set to 2% of the average voltage value during normal operation of the MMC, that is, Ax is determined to be 0. In this case, the fault is judged to be one of the following three faults: open circuit fault on the subunit, voltage sensor disconnection fault, and voltage sensor stuck fault. This situation is recorded as a Class I fault. When Ax is far beyond this range, it indicates that the fault is one of the lower switch open circuit fault and voltage sensor gain fault, and this fault is recorded as a Type II fault.

5. The method for collaboratively monitoring MMC subunit switch failure and voltage sensor failure according to claim 4, characterized in that: For type I fault, the fault is an upper switch open circuit fault, modify the trigger signal Sij in the sliding mode observer, and set The current i and capacitance vc of the faulty subunit are observed. When a Type I fault occurs in subunit 1, the observer equation is: like and If the fault lasts for 80ms and δ1 and δ2 are the defined current and voltage thresholds, the fault is an upper open circuit fault of the subunit; otherwise, the fault is determined to be a disconnection fault or a stuck fault of the voltage sensor; the indicator Bx is defined as: When Bx belongs to [-ξ,ξ], the fault is judged to be a voltage sensor disconnection fault, otherwise it is a voltage sensor stuck fault; For type II faults, the fault is a lower switch open circuit fault, and the trigger signal Sij in the sliding mode observer is modified to observe the current i and capacitance vc of the faulty subunit; if and If it lasts for 80ms, the false fault is a lower open circuit fault of the subunit, otherwise it is a gain fault of the voltage sensor.

6. A collaborative monitoring system for MMC subunit switch faults and voltage sensor faults, characterized in that: The system includes the following modules: Module M1: Detect and locate faults; Module M2: reads the capacitor voltage and bridge arm current of the faulty subunit; Module M3: Subdivide the faults and handle them accordingly.

7. The MMC subunit switch fault and voltage sensor fault coordinated monitoring system according to claim 6, characterized in that: In the module M1: Each bridge arm has N subunits, defining the capacitance-voltage correlation coefficient: in, is the length of the sliding window, m and n are any two subunits on a bridge arm of the MMC, and is the average value of the capacitor voltage of subunits m and n in the window; by checking whether ρm,n ≈ 1, it is possible to detect whether a fault has occurred and locate the subunit where the fault is located.

8. The MMC subunit switch fault and voltage sensor fault coordinated monitoring system according to claim 6, characterized in that: In the module M2: The estimated values ​​of the state variables are obtained based on the measured values ​​of the system's external variables. The state space system is expressed as: Where, is the observed state vector, u is the input vector, f is the switching function, A is the state gain matrix, B is the input gain matrix, and K is the observer gain vector; The capacitor voltage output by the subunit is written as a switching function: At the same time, the voltage-current relationship of the capacitor is listed as follows: Since the three phases of MMC are symmetrical, combining formula (6) and formula (7), we get: in, and Respectively represent the observed current values ​​of the upper and lower bridge arms of phase j, represents the capacitance voltage observation value α of the i-th subunit of phase j pj With α nj is the bridge arm current observer switching function, and the hyperbolic tangent function tanh(x) is selected to express it as follows: Among them, K1 is the switching gain; Combining formula (6) and formula (8), the sub-unit capacitor voltage observer of phase j is constructed as: Among them, K2 is the switching gain.

9. The MMC subunit switch fault and voltage sensor fault coordinated monitoring system according to claim 6, characterized in that: In the module M3: Define the indicator Ax: When Ax belongs to [-ξ,ξ], ξ is set to 2% of the average voltage value during normal operation of the MMC, that is, Ax is determined to be 0. In this case, the fault is judged to be one of the following three faults: open circuit fault on the subunit, voltage sensor disconnection fault, and voltage sensor stuck fault. This situation is recorded as a Class I fault. When Ax is far beyond this range, it indicates that the fault is one of the lower switch open circuit fault and voltage sensor gain fault, and this fault is recorded as a Type II fault.

10. The MMC subunit switch fault and voltage sensor fault coordinated monitoring system according to claim 9, characterized in that: For type I fault, the fault is an upper switch open circuit fault, modify the trigger signal Sij in the sliding mode observer, and set The current i and capacitance vc of the faulty subunit are observed. When a Type I fault occurs in subunit 1, the observer equation is: like If the fault lasts for 80ms and δ1 and δ2 are the defined current and voltage thresholds, the fault is an upper open circuit fault of the subunit; otherwise, the fault is determined to be a disconnection fault or a stuck fault of the voltage sensor; the indicator Bx is defined as: When Bx belongs to [-ξ,ξ], the fault is judged to be a voltage sensor disconnection fault, otherwise it is a voltage sensor stuck fault; For type II faults, the fault is a lower switch open circuit fault, and the trigger signal Sij in the sliding mode observer is modified to observe the current i and capacitance vc of the faulty subunit; if and If it lasts for 80ms, the false fault is a lower open circuit fault of the subunit, otherwise it is a gain fault of the voltage sensor.

Images