A method for quickly locating and diagnosing an inverter switch tube open circuit fault and a related device

Abstract

The present invention is a method for rapid location diagnosis of an inverter switch tube open circuit fault and a related device, the method comprising: performing a fast S transformation FST on a specific harmonic of an open circuit fault of a T-type three-level grid-connected inverter to obtain the time domain dynamic characteristics of each harmonic of the current; performing a dq transformation on the three-phase current according to the time domain dynamic characteristics to solve the load current size, and then normalizing the three-phase current so that its harmonic characteristics are not affected by load changes; receiving the normalized three-phase current, training the collected current characteristic data of a known fault period, obtaining a random forest model, and using the random forest model to automatically identify and locate the open circuit fault online in real time. The present invention has the advantages of accurate fault location, reliable diagnosis results, good anti-interference performance, high intelligence, not being restricted by grid load, and adapting to different grid-connected environments.

Description

Technical Field

[0001] The present invention relates to the technical field of power system grid, and in particular to a method for quickly locating and diagnosing an inverter switch tube open circuit fault and a related device. Background Art

[0002] Chinese patent CN106597195A proposes an online diagnosis method for open-circuit faults of power tubes in locomotive inverters, and Chinese patent CN110208721A proposes a fault diagnosis method for cascaded three-level inverters. These methods are only applicable to direct power supply systems, and their common defects are: limited application scope and poor grid adaptability.

[0003] In the prior art, people have mature experience in detecting and protecting short-circuit faults of inverter switches; however, there is still a lack of rapid and accurate positioning and diagnosis measures for open-circuit faults of inverter switches. Especially in the T-type three-level inverter switch system, after an open-circuit fault occurs, it will cause many problems in the running power grid system, such as increased power, reduced efficiency, DC polarization and harmonic resonance. The factors that restrict the diagnosis and identification of open-circuit faults include: a large number of switch tubes, a large impact of load, complex fault characteristics, and a low degree of intelligence. Summary of the invention

[0004] The purpose of the present invention is to provide a method and related device for quickly locating and diagnosing open-circuit faults of inverter switch tubes, which can quickly obtain the time domain dynamic characteristics of each harmonic of the current, and at the same time perform dq transformation and normalization processing on the three-phase current so that its harmonic characteristics are not affected by load changes, and realize real-time automatic identification and positioning of open-circuit faults.

[0005] A method for quickly locating and diagnosing an inverter switch open circuit fault comprises the following steps:

[0006] The fast S-transform FST is performed on the specific harmonics of the open-circuit fault of the T-type three-level grid-connected inverter to obtain the time domain dynamic characteristics of each harmonic of the current;

[0007] Performing dq transformation on the three-phase current according to the time domain dynamic characteristics, solving the load current magnitude, and then normalizing the three-phase current so that its harmonic characteristics are not affected by load changes;

[0008] Receive the normalized three-phase current, train the collected current characteristic data of the known fault period to obtain a random forest model, and use the random forest model to automatically identify and locate the open circuit fault online and in real time.

[0009] Furthermore, the fast S transform FST is performed on the specific harmonics of the open circuit fault of the T-type three-level grid-connected inverter to obtain the time domain dynamic characteristics of each harmonic of the current, including: performing a fast Fourier transform FFT on the collected signal, and then performing an inverse fast Fourier transform iFFT on the frequency sampling points in the first half thereof to obtain the dynamic time domain characteristics of each harmonic.

[0010] Furthermore, the three-phase current is subjected to dq transformation to solve the load current magnitude, and then the three-phase current is subjected to normalization processing, including: performing abc-dq transformation on the sampled three-phase current to obtain the d-axis current I of the inverter current d (t) and q-axis current I q (t), using the empirical formula Obtain the current normalization coefficient I(t), and then use the current normalization coefficient I(t) to normalize the three-phase current: NI m (t) = I m (t) / I(t), where I m (t) is the effective value of the mth harmonic, NI m (t) is the harmonic feature after normalization.

[0011] Furthermore, the collected current characteristic data of known fault periods are trained to obtain a random forest model, and the random forest model is used to automatically identify and locate open circuit faults online in real time, including:

[0012] N samples of known fault types are used for training and testing. P harmonic features are selected for each sample. N samples are sampled with replacement from the N training samples for N times to form a training set. The samples that are not sampled are used as predictions to evaluate their errors.

[0013] A classification and regression tree using a random forest decision tree randomly selects q features for each node. Each node on the decision tree is determined based on these features. The best splitting method is calculated based on the q features. Each tree will grow completely without the need for pruning, and finally the fault diagnosis and location results are obtained.

[0014] An inverter switch tube open circuit fault rapid positioning and diagnosis device, comprising: a sampling and signal conditioning circuit module, a DSP digital signal processor module and an open circuit fault identification and positioning module;

[0015] The sampling and signal conditioning circuit module is used to perform a fast S transformation FST on the specific harmonics of the open circuit fault of the T-type three-level grid-connected inverter to obtain the time domain dynamic characteristics of each harmonic of the current;

[0016] The DSP digital signal processor module is used to perform dq transformation on the three-phase current according to the time domain dynamic characteristics, solve the load current size, and then normalize the three-phase current so that its harmonic characteristics are not affected by load changes;

[0017] The open circuit fault identification and positioning module is used to train the collected current characteristic data of known fault periods to obtain a random forest model, and use the random forest model to automatically identify and locate open circuit faults online in real time.

[0018] Furthermore, the sampling and signal conditioning circuit module is specifically used to: perform fast Fourier transform FFT on the collected signal, and then perform inverse fast Fourier transform iFFT on the first half of the frequency sampling points therein to obtain the dynamic time domain characteristics of each harmonic.

[0019] Furthermore, the DSP digital signal processor module is specifically used to: perform abc-dq transformation on the sampled three-phase current to obtain the d-axis current I of the inverter current d (t) and q-axis current I q (t), using the empirical formula Obtain the current normalization coefficient I(t), and then use the current normalization coefficient I(t) to normalize the three-phase current: NI m (t) = I m (t) / I(t), where I m (t) is the effective value of the mth harmonic, NI m (t) is the harmonic characteristic after normalization

[0020] Furthermore, the open circuit fault identification and positioning module is specifically used for:

[0021] N samples of known fault types are used for training and testing. P harmonic features are selected for each sample. N samples are sampled with replacement from the N training samples for N times to form a training set. The samples that are not sampled are used as predictions to evaluate their errors.

[0022] A classification and regression tree using a random forest decision tree randomly selects q features for each node. Each node on the decision tree is determined based on these features. The best splitting method is calculated based on the q features. Each tree will grow completely without the need for pruning, and finally the fault diagnosis and location results are obtained.

[0023] A system for quickly locating and diagnosing an inverter switch tube open circuit fault, comprising: a computer-readable storage medium and a processor; the computer-readable storage medium is used to store executable instructions; the processor is used to read the executable instructions stored in the computer-readable storage medium and execute the inverter switch tube open circuit fault quickly locating and diagnosing method.

[0024] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the method for quickly locating and diagnosing an inverter switch tube open circuit fault.

[0025] The beneficial technical effects of the present invention are: by performing a fast S-transform FST on the specific harmonics of the open-circuit fault, the time-domain dynamic characteristics of each harmonic of the current can be quickly obtained; at the same time, the three-phase current is subjected to dq transformation and normalization processing, so that its harmonic characteristics are not affected by load changes; in addition, the collected current characteristic data is trained and a random forest model is established, so that the open-circuit fault can be automatically identified and located in real time. The present invention also has the advantages of accurate fault location, reliable diagnosis results, good anti-interference performance, high intelligence, not restricted by power grid load, and adaptability to different grid-connected environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a flow chart for rapid location and diagnosis of open circuit faults according to the present invention;

[0027] Figure 2 A circuit diagram of a real-time anti-interference T-type three-level grid-connected inverter fault-tolerant structure in an embodiment of the present invention;

[0028] Figure 3 The open circuit and three-phase current waveform and spectrum diagram of the real-time anti-interference T-type three-level grid-connected inverter Sa1 in the embodiment of the present invention;

[0029] Figure 4 Figure A is the process of Sa1 tube fault diagnosis and location under load change conditions;

[0030] Figure 5 Figure B is the process of Sa1 tube fault diagnosis and location under load change conditions;

[0031] Figure 6 Figure C is the process of Sa2 tube fault diagnosis and location under load change conditions;

[0032] Figure 7 Figure D is the process of Sa2 tube fault diagnosis and location under load change conditions. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] The present invention is further described in detail below through specific implementation examples in conjunction with the accompanying drawings.

[0035] according to Figure 2 , the current sampling rate is set to 10kHz, the sampling signal duration is 0.4s, an open circuit fault is set at 0.1s, and the load current is set to 1.5 times the original load at 0.3s. The inverter adopts PQ control. Because a, b, and c are symmetrical, the a phase fault is taken as an example here.

[0036] according to Figure 2 , we know that the faults of Sa1 and Sa4 are symmetrical, and the fault characteristics of Sa2 and Sa3 are symmetrical. Symmetrical fault currents have no other differences except the polarity of the DC component. Therefore, in the example, Sa1 open circuit fault and Sa2 open circuit fault are taken as examples. The fault diagnosis and identification of Sa3 and Sa4 are the same as Sa1 and Sa2, and will not be repeated here.

[0037] The embodiment of the present invention provides a method for quickly locating and diagnosing an inverter switch tube open circuit fault, comprising the following steps:

[0038] Step 1, perform fast S transformation FST on the specific harmonics of the open circuit fault of the T-type three-level grid-connected inverter to obtain the time domain dynamic characteristics of each harmonic of the current; and transmit the collected signals to the DSP digital signal processor module.

[0039] The step 1 specifically includes: performing a fast Fourier transform FFT on the collected signal, and then performing an inverse fast Fourier transform iFFT on the frequency sampling points of the first half of the signal, thereby obtaining the dynamic time domain characteristics of each harmonic; since the frequency sampling points of the current signal are mostly 0 after the fast Fourier transform FFT, which is a typical sparse sequence, the fast S transform FST only samples the frequencies exceeding a certain threshold, and performs an inverse fast Fourier transform iFFT on the frequencies, thereby obtaining the time domain harmonic characteristics.

[0040] According to step 1, when an open circuit fault occurs in the T-type grid-connected inverter, the DC component (Ia0), the fundamental component (Ia1), and the second harmonic component (Ia2) account for more than 95% of the total spectrum energy. Therefore, when the fast S transform FST is used for feature extraction, the calculation amount of the fast S transform FST is 1 fast Fourier transform FFT and 3 inverse Fourier transform iFFT. Because the calculation amount of the fast Fourier transform FFT is equal to that of the inverse Fourier transform iFFT, the feature extraction of the fast S transform FST is only the calculation amount of 4 fast Fourier transform FFT, which fully meets the real-time requirements of fault diagnosis and positioning.

[0041] Step 2, performing dq transformation on the three-phase current according to the time domain dynamic characteristics, solving the load current size, and then normalizing the three-phase current so that its harmonic characteristics are not affected by load changes.

[0042] The dq transformation and normalization processing includes: since the inverter current size and harmonics will change due to the load, it is necessary to normalize the current characteristics; the inverter current d-axis current command Id and q-axis current command Iq are respectively derived from the dq-abc transformation. Similarly, the sampled three-phase current needs to be transformed by abc-dq, and the empirical formula is used. Get the current normalization coefficient, set I m (t) is the effective value of the mth harmonic, then the normalized empirical formula for harmonic characteristics that is not affected by load changes is: NI m (t) = I m (t) / I(t).

[0043] According to step 2, Figure 3 , Figure 4 As shown in the figure, at 0.1s, Sa1 has an open circuit fault. At this time, Ia0, Ia1 and Ia2 all change significantly, showing the fault characteristics. At 0.3s, the load changes, and Ia0, Ia1 and Ia2 all change with the load, causing interference to the feature recognition.

[0044] Step 3, receiving the normalized three-phase current, training the collected current characteristic data of the known fault period to obtain a random forest model, and using the random forest model to automatically identify and locate the open circuit fault online in real time.

[0045] The training of current characteristic data to obtain a random forest model includes: using N samples of known fault types for training and testing, selecting p harmonic features for each sample, sampling N times from the N training samples in a sampling method with replacement to form a training set, and using the unsampled samples as predictions to evaluate their errors; using a classification and regression tree of a random forest decision tree, randomly selecting q features for each node, each node on the decision tree is determined based on these features, and calculating the best splitting method based on these p features, so that each tree will grow completely without the need for pruning, and finally obtaining fault diagnosis and positioning results.

[0046] According to step 3, Figure 5 , Figure 6As shown in the figure, after normalization, Ia0, Ia1 and Ia2 change slightly by about one cycle and then return to the previous fault characteristic state at 0.32s. Random forest identifies the normalized features, and the fault index in phase a is Sa1 or Sa4. The polarity of Ia0 is "-", which is judged as an open circuit fault in the upper bridge arm Sa1. The fault is correctly identified and located. Similarly, Figure 5 As shown in the figure, at 0.1s, Sa2 has an open circuit fault. At this time, Ia0, Ia1 and Ia2 all change significantly, showing the fault characteristics. At 0.3s, the load changes, and Ia0, Ia1 and Ia2 all change with the load, causing interference to the feature recognition.

[0047] According to step 3, Figure 7 As shown in the figure, after normalization, Ia0, Ia1 and Ia2 change slightly by about one cycle and then return to the previous fault feature state at 0.32s. Random forest identifies the normalized features, and the fault index in phase a is Sa2 or Sa3. The polarity of Ia0 is "-", which is judged as an open circuit fault of the front tube Sa2. The fault is correctly identified and located.

[0048] The embodiment of the present invention also provides a device for quickly locating and diagnosing an open-circuit fault of an inverter switch tube, comprising: a sampling and signal conditioning circuit module, a DSP digital signal processor module and an open-circuit fault identification and locating module.

[0049] The sampling and signal conditioning circuit module is used to perform a fast S transformation FST on the specific harmonics of the open circuit fault of the T-type three-level grid-connected inverter to obtain the time domain dynamic characteristics of each harmonic of the current.

[0050] The DSP digital signal processor module is used to perform dq transformation on the three-phase current according to the time domain dynamic characteristics, solve the load current size, and then normalize the three-phase current so that its harmonic characteristics are not affected by load changes.

[0051] The open circuit fault identification and positioning module is used to train the collected current characteristic data of known fault periods to obtain a random forest model, and use the random forest model to automatically identify and locate open circuit faults online in real time.

[0052] An embodiment of the present invention also provides a system for quickly locating and diagnosing an inverter switch tube open circuit fault, comprising: a computer-readable storage medium and a processor; the computer-readable storage medium is used to store executable instructions; the processor is used to read the executable instructions stored in the computer-readable storage medium and execute the inverter switch tube open circuit fault quickly locating and diagnosing method.

[0053] A computer-readable storage medium having a computer program stored thereon, which implements the inverter switch tube open circuit fault rapid location diagnosis method when executed by a processor. Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0054] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0055] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0056] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0057] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for quickly locating and diagnosing an inverter switch open circuit fault, characterized in that: The steps include: The fast S-transform FST is performed on the specific harmonics of the open-circuit fault of the T-type three-level grid-connected inverter to obtain the time domain dynamic characteristics of each harmonic of the current; Performing dq transformation on the three-phase current according to the time domain dynamic characteristics, solving the load current magnitude, and then normalizing the three-phase current so that its harmonic characteristics are not affected by load changes; Receiving the normalized three-phase current, training the collected current characteristic data of the known fault period to obtain a random forest model, and using the random forest model to automatically identify and locate the open circuit fault online and in real time; The three-phase current is subjected to dq transformation to obtain the load current, and then the three-phase current is subjected to normalization processing, including: performing abc-dq transformation on the sampled three-phase current to obtain the d-axis current I of the inverter current d (t) and q-axis current I q (t), using the empirical formula Obtain the current normalization coefficient I(t), and then use the current normalization coefficient I(t) to normalize the three-phase current: NI m (t) = I m (t) / I(t), where I m (t) is the effective value of the mth harmonic, NI m (t) is the harmonic feature after normalization.

2. The method for rapid location and diagnosis of an inverter switch open circuit fault according to claim 1, characterized in that: The method of performing a fast S transform FST on the specific harmonics of the open-circuit fault of the T-type three-level grid-connected inverter to obtain the time-domain dynamic characteristics of each harmonic of the current includes: performing a fast Fourier transform FFT on the collected signal, and then performing an inverse fast Fourier transform iFFT on the frequency sampling points of the first half thereof to obtain the dynamic time-domain characteristics of each harmonic.

3. The method for rapid location and diagnosis of an inverter switch tube open circuit fault according to claim 1, characterized in that: The method of training the collected current characteristic data of known fault periods to obtain a random forest model, and using the random forest model to automatically identify and locate open circuit faults online in real time, includes: N samples of known fault types are used for training and testing. P harmonic features are selected for each sample. N samples are sampled with replacement from the N training samples for N times to form a training set. The samples that are not sampled are used as predictions to evaluate their errors. A classification and regression tree using a random forest decision tree randomly selects q features for each node. Each node on the decision tree is determined based on these features. The best splitting method is calculated based on the q features. Each tree will grow completely without the need for pruning, and finally the fault diagnosis and location results are obtained.

4. A device for quickly locating and diagnosing an inverter switch tube open circuit fault, characterized in that include: Sampling and signal conditioning circuit module, DSP digital signal processor module and open circuit fault identification and positioning module; The sampling and signal conditioning circuit module is used to perform a fast S transformation FST on the specific harmonics of the open circuit fault of the T-type three-level grid-connected inverter to obtain the time domain dynamic characteristics of each harmonic of the current; The DSP digital signal processor module is used to perform dq transformation on the three-phase current according to the time domain dynamic characteristics, solve the load current size, and then normalize the three-phase current so that its harmonic characteristics are not affected by load changes; The open circuit fault identification and location module is used to train the collected current characteristic data of the known fault period to obtain a random forest model, and use the random forest model to automatically identify and locate the open circuit fault online in real time; The DSP digital signal processor module is specifically used to: perform abc-dq transformation on the sampled three-phase current to obtain the d-axis current I of the inverter current d (t) and q-axis current I q (t), using the empirical formula Obtain the current normalization coefficient I(t), and then use the current normalization coefficient I(t) to normalize the three-phase current: NI m (t) = I m (t) / I(t), where I m (t) is the effective value of the mth harmonic, NI m (t) is the harmonic feature after normalization.

5. The device according to claim 4, characterized in that: The sampling and signal conditioning circuit module is specifically used to: perform fast Fourier transform FFT on the collected signal, and then perform inverse fast Fourier transform iFFT on the first half of the frequency sampling points therein to obtain the dynamic time domain characteristics of each harmonic.

6. The device according to claim 4, characterized in that: The open circuit fault identification and positioning module is specifically used for: N samples of known fault types are used for training and testing. P harmonic features are selected for each sample. N samples are sampled with replacement from the N training samples for N times to form a training set. The samples that are not sampled are used as predictions to evaluate their errors. A classification and regression tree using a random forest decision tree randomly selects q features for each node. Each node on the decision tree is determined based on these features. The best splitting method is calculated based on the q features. Each tree will grow completely without the need for pruning, and finally the fault diagnosis and location results are obtained.

7. A system for quickly locating and diagnosing an inverter switch tube open circuit fault, comprising: A computer readable storage medium and a processor; The computer-readable storage medium is used to store executable instructions; The processor is used to read the executable instructions stored in the computer-readable storage medium to execute the inverter switch tube open circuit fault rapid location and diagnosis method according to any one of claims 1 to 3.

8. A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for quickly locating and diagnosing an inverter switch tube open circuit fault according to any one of claims 1 to 3 is implemented.

Images