A method and system for optimizing high-power DC boost control effect
Through neural network analysis of high-power DC boost output and inverter output waveform data, a simulation model is built for simulation, solving the problem of identification of packaged high-power DC boost control system, realizing accurate identification of the state of the control system, and improving the Grid power quality.
Patent Information
- Application Number
- CN202411025396.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-07-30
AI Technical Summary
The problem of identifying the control effect of the packaged high-power DC boost control system is difficult to locate the cause of the fault, resulting in power quality problems such as excessive harmonics at the grid access point.
By reading high-power DC boost output waveform monitoring data and inverter output waveform monitoring data, perform neural network analysis, build simulation models for simulation and data training, and use neural network models to identify whether the high-power DC boost control system is in the Normal state, realize penetration analysis.
Accurately identify the status of the packaged high-power DC boost control system, helping operation and maintenance personnel to quickly locate problems and improve the stability of power quality in the power grid.
Smart Images

Figure CN118554780B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of circuit control, and in particular to a method and system for optimizing high-power direct current boost control effect. Background Art
[0002] High-power DC boost control systems are usually used for large-scale photovoltaic grid-connected power generation. After the low-voltage DC power generated by photovoltaics is collected, it is converted into DC power of a higher voltage level through a high-power DC boost system, and then converted into AC power through an inverter before being connected to the grid. Since photovoltaic power generation equipment, high-power DC boost systems and inverters are all power electronic equipment, they are typical harmonic sources and reactive equipment. When they are connected to the grid, it is very easy to cause power quality problems such as excessive harmonics at the grid access point.
[0003] Research shows that the impact of photovoltaic grid-connected power generation systems on the power quality of the power grid is closely related to the performance of the high-power DC boost control system and the inverter control performance. As long as the performance of either the high-power DC boost control system or the inverter control system is poor, power quality problems may occur.
[0004] However, for the operation and maintenance personnel of the photovoltaic grid-connected power generation system, both the high-power DC boost control system and the inverter control system are encapsulated "black boxes". When the photovoltaic grid-connected power generation system has power quality problems, it is difficult for the operation and maintenance personnel to locate the cause of the fault, whether it is the high-power DC boost control system or the inverter control system.
[0005] In response to this technical problem, the present invention performs neural network analysis based on the high-power DC boost output waveform monitoring data and inverter output waveform monitoring data that can be obtained by operation and maintenance personnel, accurately identifies whether the packaged high-power DC boost control system is in a normal state, and realizes penetrating analysis of the packaged high-power DC boost control system. Summary of the invention
[0006] (1) Technical issues to be resolved
[0007] The object of the present invention is to provide a high-power DC boost control effect optimization method and system to solve the control effect identification problem of a packaged high-power DC boost control system.
[0008] (2) Technical solution
[0009] To achieve the above object, on the one hand, the present invention provides a method for optimizing the control effect of a high-power DC boost, the method comprising the following steps:
[0010] S1, read the basic parameters of the high-power DC boost model and the inverter model, build the high-power DC boost simulation model and the inverter simulation model on the simulation platform; combine the high-power DC boost simulation model and the inverter simulation model to obtain a grid-connected simulation model.
[0011] S2, read the number of control parameters in the high-power DC boost simulation model and the number of control parameters in the inverter simulation model, calculate the number of simulations N, set the first simulation scene to the Nth simulation scene, perform N simulation simulations on the high-power DC boost simulation model and the inverter simulation model, and record the first simulation data group to the Nth simulation data group obtained by the simulation; use the first simulation data group to the Nth simulation data group to train the neural network model to obtain the first neural network model.
[0012] S3, monitor the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data within a preset first monitoring period; perform data preprocessing on the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data to obtain monitoring data samples, input the monitoring data samples into the first neural network model, and obtain a high-power DC boost control effect identification result.
[0013] S4: judging whether a high-power DC boost control parameter needs to be adjusted according to the high-power DC boost control effect identification result.
[0014] Further, the method of reading the number of control parameters in the high-power DC boost simulation model and the number of control parameters in the inverter simulation model, calculating the number of simulations N, setting the first simulation scenario to the Nth simulation scenario, performing N simulation simulations on the high-power DC boost simulation model and the inverter simulation model, and recording the first simulation data group to the Nth simulation data group obtained by the simulation includes:
[0015] Read the number of control parameters in the high-power DC boost simulation model n 1, n 1>1; Read the number of control parameters in the inverter simulation model n 2, n 2>1; according to the preset number of segments D, the simulation number N is obtained by the simulation number calculation formula; the simulation number calculation formula is:
[0016] .
[0017] According to the number of segments 、The number of control parameters in the high-power DC boost simulation model, the number of control parameters in the inverter simulation model, and the pre-set control parameter limits of the high-power DC boost simulation model and the control parameter limits of the inverter simulation model are used to construct the first simulation scenario to the Nth simulation scenario, and obtain the control parameter value matrix of the inverter simulation model.
[0018] According to the first simulation scenario to the Nth simulation scenario, N simulation simulations are performed on the grid-connected simulation model to obtain the first waveform data of high-power DC boost output to the Nth waveform data of high-power DC boost output and the first waveform data of inverter output to the Nth waveform data of inverter output.
[0019] According to the inverter output first waveform data to the inverter output Nth waveform data and the inverter simulation model control parameter value matrix, the first attribution value to the Nth attribution value are obtained through the control attribution algorithm; the first attribution value to the Nth attribution value are respectively combined with the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data and the inverter output first waveform data to the inverter output Nth waveform data to obtain the first simulation data group to the Nth simulation data group.
[0020] Further, the method of obtaining the first attribution value to the Nth attribution value by controlling the attribution algorithm according to the inverter outputting the first waveform data to the inverter outputting the Nth waveform data and the inverter simulation model control parameter value matrix includes:
[0021] Fourier decomposition is performed on the inverter output first waveform data to the inverter output Nth waveform data respectively, and the fundamental wave amplitude and total harmonic distortion rate of the inverter output first waveform data to the inverter output Nth waveform data are obtained.
[0022] According to the fundamental amplitude and total harmonic distortion rate of the first waveform data output by the inverter to the Nth waveform data output by the inverter, a first abnormality detection algorithm is used to obtain the first attribution value to the Nth attribution value; the first abnormality detection algorithm is: if and ,but , otherwise, parameter attribution identification is performed; among them, is the preset lower limit of fundamental amplitude, is the preset upper limit of fundamental amplitude, is the preset upper limit of total harmonic distortion rate, Output the first i The fundamental amplitude of the waveform data, Output the first i Total harmonic distortion of waveform data, For the i Attribution value, i=1, 2...N; the parameter attribution identification steps are: if j Take [1, n 2], all integer values in the range satisfy ,but ,otherwise ;in, is the control parameter value matrix of the inverter simulation model. j Line i The elements of the column, j =1,2…… n 2, The inverter is preset j Control parameter standard lower limit, The inverter is preset j Control parameter standard upper limit.
[0023] Furthermore, the method of using the first to Nth simulation data groups to train a neural network model to obtain a first neural network model includes:
[0024] The average value of the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data is calculated in sequence to obtain the first DC waveform average value. To the average value of the Nth DC waveform ; Sequentially obtain the variance of the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data to obtain the variance of the first DC waveform To the Nth DC waveform variance .
[0025] Fourier decomposition is performed on the first waveform data output by the inverter to the Nth waveform data output by the inverter in sequence, thereby obtaining the total harmonic distortion rate of the first waveform output by the inverter. Total harmonic distortion rate of the inverter output Nth waveform ; Calculate the effective value of the inverter output first waveform data to the inverter output N waveform data in sequence to obtain the effective value of the inverter output first waveform Until the inverter outputs the effective value of the Nth waveform .
[0026] Establish the to , to , to , to and the mapping relationship between the first attribution value to the Nth attribution value, taking the first attribution value to the Nth attribution value as labels, to , to , to , to As a sample, the neural network model is input and trained to obtain the first neural network model.
[0027] Furthermore, the method of preprocessing the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data to obtain monitoring data samples, inputting the monitoring data samples into the first neural network model, and obtaining the high-power DC boost control effect identification result includes:
[0028] The high-power DC boost output waveform monitoring data is averaged to obtain the DC waveform monitoring average value. ; Obtain the variance of the high-power DC boost output waveform monitoring data to obtain the DC waveform monitoring variance .
[0029] Perform Fourier decomposition on the inverter output waveform monitoring data to obtain the inverter waveform monitoring total harmonic distortion rate ; Calculate the effective value of the inverter output waveform monitoring data to obtain the inverter waveform monitoring total harmonic distortion rate .
[0030] The , , , The monitoring data samples are combined to obtain monitoring data samples, and the monitoring data samples are input into the first neural network model to obtain monitoring attribution values; if the monitoring attribution value is 0, "the control system is normal" is output, otherwise "the control system is abnormal" is output.
[0031] Based on the same inventive concept, on the other hand, the present invention also provides a high-power DC boost control effect optimization system, characterized in that the system comprises:
[0032] The simulation model building module is used to read the basic parameters of the high-power DC boost model and the inverter model, build the high-power DC boost simulation model and the inverter simulation model on the simulation platform; combine the high-power DC boost simulation model and the inverter simulation model to obtain a grid-connected simulation model.
[0033] The first neural network model construction module is connected to the simulation model construction module, and is used to read the number of control parameters in the high-power DC boost simulation model and the number of control parameters in the inverter simulation model, calculate the number of simulations N, set the first simulation scene to the Nth simulation scene, perform N simulation simulations on the high-power DC boost simulation model and the inverter simulation model, and record the first simulation data group to the Nth simulation data group obtained by the simulation; use the first simulation data group to the Nth simulation data group to train the neural network model to obtain the first neural network model.
[0034] The high-power DC boost control effect identification module is connected to the first neural network model construction module, and is used to monitor the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data within a preset first monitoring period; the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data are preprocessed to obtain monitoring data samples, and the monitoring data samples are input into the first neural network model to obtain a high-power DC boost control effect identification result.
[0035] The high-power DC boost control parameter adjustment module is connected to the high-power DC boost control effect identification module and is used to determine whether the high-power DC boost control parameter needs to be adjusted according to the high-power DC boost control effect identification result.
[0036] Furthermore, the first neural network model building module includes:
[0037] A data reading module is used to read the control parameter quantity in the high-power DC boost simulation model. n 1, n 1>1; Read the number of control parameters in the inverter simulation model n 2, n 2>1; according to the preset number of segments D, the simulation number N is obtained by the simulation number calculation formula; the simulation number calculation formula is:
[0038] .
[0039] The simulation scene construction module is connected to the data reading module and is used to 、 The number of control parameters in the high-power DC boost simulation model, the number of control parameters in the inverter simulation model, and the pre-set control parameter limits of the high-power DC boost simulation model and the control parameter limits of the inverter simulation model are used to construct the first simulation scenario to the Nth simulation scenario, and obtain the control parameter value matrix of the inverter simulation model.
[0040] The simulation module is connected to the simulation scenario construction module, and is used to perform N simulations on the grid-connected simulation model according to the first simulation scenario to the Nth simulation scenario, to obtain the first waveform data of the high-power DC boost output to the Nth waveform data of the high-power DC boost output, and the first waveform data of the inverter output to the Nth waveform data of the inverter output.
[0041] The simulation array calculation module is connected to the simulation simulation module, and is used to obtain the first attribution value to the Nth attribution value through the control attribution algorithm according to the inverter output first waveform data to the inverter output Nth waveform data and the inverter simulation model control parameter value matrix; the first attribution value to the Nth attribution value are respectively combined with the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data and the inverter output first waveform data to the inverter output Nth waveform data to obtain the first simulation data group to the Nth simulation data group.
[0042] Furthermore, the simulation array calculation module includes:
[0043] The harmonic calculation module is used to perform Fourier decomposition on the inverter output first waveform data to the inverter output Nth waveform data, and obtain the fundamental amplitude and total harmonic distortion rate of the inverter output first waveform data to the inverter output Nth waveform data.
[0044] The attribution value calculation module is connected to the harmonic calculation module and is used to obtain the first attribution value to the Nth attribution value according to the fundamental amplitude and total harmonic distortion rate of the inverter outputting the first waveform data to the inverter outputting the Nth waveform data by using the first abnormality detection algorithm; the first abnormality detection algorithm is: if and ,but , otherwise, parameter attribution identification is performed; among them, is the preset lower limit of fundamental amplitude, is the preset upper limit of fundamental amplitude, is the preset upper limit of total harmonic distortion rate, Output the first i The fundamental amplitude of the waveform data, Output the first i Total harmonic distortion of waveform data, For the i Attribution value, i =1, 2...N; the parameter attribution identification steps are: if j Take [1, n 2], all integer values in the range satisfy ,but ,otherwise ;in, is the control parameter value matrix of the inverter simulation model. j Line i The elements of the column, j =1,2…… n 2, The inverter is preset j Control parameter standard lower limit, The inverter is preset j Control parameter standard upper limit.
[0045] Furthermore, the first neural network model building module also includes:
[0046] A high-power DC boost output processing module is used to sequentially calculate the average value of the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data to obtain the first DC waveform average value To the average value of the Nth DC waveform ; Sequentially obtain the variance of the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data to obtain the variance of the first DC waveform To the Nth DC waveform variance .
[0047] The inverter output processing module is connected to the high-power DC boost output processing module, and is used to sequentially perform Fourier decomposition on the inverter output first waveform data to the inverter output Nth waveform data, thereby obtaining the total harmonic distortion rate of the inverter output first waveform. Total harmonic distortion rate of the inverter output Nth waveform ; Calculate the effective value of the inverter output first waveform data to the inverter output N waveform data in sequence to obtain the effective value of the inverter output first waveform Until the inverter outputs the effective value of the Nth waveform .
[0048] The neural network training module is connected to the inverter output processing module to establish the to , to , to , to and the mapping relationship between the first attribution value to the Nth attribution value, taking the first attribution value to the Nth attribution value as labels, to , to , to , to As a sample, the neural network model is input and trained to obtain the first neural network model.
[0049] Furthermore, the high-power DC boost control effect identification module includes:
[0050] A high-power DC boost output monitoring processing module is used to calculate the average value of the high-power DC boost output waveform monitoring data to obtain the DC waveform monitoring average value ; Obtain the variance of the high-power DC boost output waveform monitoring data to obtain the DC waveform monitoring variance .
[0051] The inverter output monitoring data processing module is connected to the high-power DC boost output monitoring processing module and is used to perform Fourier decomposition on the inverter output waveform monitoring data to obtain the inverter waveform monitoring total harmonic distortion rate. ; Calculate the effective value of the inverter output waveform monitoring data to obtain the inverter waveform monitoring total harmonic distortion rate .
[0052] The identification result output module is connected to the inverter output monitoring data processing module and is used to , , , The monitoring data samples are combined to obtain monitoring data samples, and the monitoring data samples are input into the first neural network model to obtain monitoring attribution values; if the monitoring attribution value is 0, "the control system is normal" is output, otherwise "the control system is abnormal" is output.
[0053] (3) Beneficial effects
[0054] Compared with the prior art, the present invention has the following beneficial effects:
[0055] By reading the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data, neural network analysis is performed to accurately identify whether the packaged high-power DC boost control system is in a normal state, thereby achieving penetrating analysis of the packaged high-power DC boost control system. BRIEF DESCRIPTION OF THE DRAWINGS
[0056] Figure 1 A flowchart of a high-power DC boost control effect optimization method according to Embodiment 1 of the present invention;
[0057] Figure 2 This is a schematic diagram of the module composition of a high-power DC boost control effect optimization system according to Example 2 of the present invention. DETAILED DESCRIPTION
[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.
[0059] Before giving examples, it is necessary to explain the application scenarios of the present invention. The present invention is applied to the control effect identification of a packaged high-power DC boost control system in a large-scale photovoltaic grid-connected power generation system.
[0060] Example 1
[0061] like Figure 1 As shown, this embodiment provides a method for optimizing the control effect of a high-power DC boost, and the method comprises the following steps:
[0062] S1, read the basic parameters of the high-power DC boost model and the inverter model, build the high-power DC boost simulation model and the inverter simulation model on the simulation platform; combine the high-power DC boost simulation model and the inverter simulation model to obtain a grid-connected simulation model.
[0063] Exemplarily, the basic parameters of the high-power DC boost model and the inverter model are read, and the high-power DC boost simulation model and the inverter simulation model are built on the PSCAD / EMTDC simulation platform. The basic parameters of the high-power DC boost model include: the number of high-power DC boost bridge arm submodules is 10, the high-power DC boost submodule capacitance is 1300 , the high-power DC boost bridge arm inductance is 10mH, the high-power DC boost capacitor voltage reference value is 1kV, the high-power DC boost pulse modulation frequency is 2.5kHz, and the high-power DC boost high-voltage side voltage reference value is 10kV. The basic parameters of the inverter model include: the inverter DC side voltage reference value is 10kV, the number of inverter bridge arm submodules is 10, the inverter capacitor voltage reference value is 1kV, the inverter carrier frequency is 2.5kHz, and the inverter AC voltage reference value is 10kV.
[0064] S2, read the number of control parameters in the high-power DC boost simulation model and the number of control parameters in the inverter simulation model, calculate the number of simulations N, set the first simulation scene to the Nth simulation scene, perform N simulation simulations on the high-power DC boost simulation model and the inverter simulation model, and record the first simulation data group to the Nth simulation data group obtained by the simulation; use the first simulation data group to the Nth simulation data group to train the neural network model to obtain the first neural network model.
[0065] S3, monitor the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data within a preset first monitoring period; perform data preprocessing on the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data to obtain monitoring data samples, input the monitoring data samples into the first neural network model, and obtain a high-power DC boost control effect identification result.
[0066] S4: judging whether a high-power DC boost control parameter needs to be adjusted according to the high-power DC boost control effect identification result.
[0067] For example, for a traditional packaged high-power DC boost control system, it is difficult for operation and maintenance personnel to view the high-power DC boost control parameters, let alone determine whether the high-power DC boost control parameters need to be adjusted. By adopting the method described in the present invention, after the operation and maintenance personnel obtain the instruction for the output of the high-power DC boost control effect identification result, if the high-power DC boost control effect identification result output is "control system normal", there is no need to adjust the high-power DC boost control parameters, and the operation and maintenance personnel can keep the system running. If the high-power DC boost control effect identification result output is "control system abnormal", it is necessary to adjust the high-power DC boost control parameters, and the operation and maintenance personnel can promptly notify the high-power DC boost control system R&D manufacturer to optimize the high-power DC boost control parameters.
[0068] Further, the method of reading the number of control parameters in the high-power DC boost simulation model and the number of control parameters in the inverter simulation model, calculating the number of simulations N, setting the first simulation scenario to the Nth simulation scenario, performing N simulation simulations on the high-power DC boost simulation model and the inverter simulation model, and recording the first simulation data group to the Nth simulation data group obtained by the simulation includes:
[0069] Read the number of control parameters in the high-power DC boost simulation model n 1, n 1>1; Read the number of control parameters in the inverter simulation model n 2, n 2>1; according to the preset number of segments D, the simulation number N is obtained by the simulation number calculation formula; the simulation number calculation formula is:
[0070] .
[0071] Exemplarily, the number of control parameters in the high-power DC boost simulation model and the number of control parameters in the inverter simulation model are the number of control parameters that can be adjusted in the packaged high-power DC boost control system and the inverter control system. The number of control parameters in the high-power DC boost simulation model is 4, the number of control parameters in the inverter simulation model is 4, the number of pre-set segments is 5, and the number of simulation times N is calculated to be 390625.
[0072] According to the number of segments 、 The number of control parameters in the high-power DC boost simulation model, the number of control parameters in the inverter simulation model, and the pre-set control parameter limits of the high-power DC boost simulation model and the control parameter limits of the inverter simulation model are used to construct the first simulation scenario to the Nth simulation scenario, and obtain the control parameter value matrix of the inverter simulation model.
[0073] Exemplarily, actual values can be used as the control parameter limits of the high-power DC boost simulation model and the control parameter limits of the inverter simulation model, and normalized values can also be used as the control parameter limits of the high-power DC boost simulation model and the control parameter limits of the inverter simulation model. For ease of description, this embodiment uses normalized values as the control parameter limits of the high-power DC boost simulation model and the control parameter limits of the inverter simulation model, that is, in the first simulation scenario to the 390625th simulation scenario, the value of the high-power DC boost simulation model control parameter is one of 0, 0.25, 0.5, 0.75, and 1, and the value of the inverter simulation model control parameter is one of 0, 0.25, 0.5, 0.75, and 1. All value combinations are traversed, and there are 390625 control parameter value combinations, corresponding to the first simulation scenario to the 390625th simulation scenario. The inverter simulation model control parameter value matrix obtained is a matrix with 390625 rows and 4 columns.
[0074] According to the first simulation scenario to the Nth simulation scenario, N simulation simulations are performed on the grid-connected simulation model to obtain the first waveform data of high-power DC boost output to the Nth waveform data of high-power DC boost output and the first waveform data of inverter output to the Nth waveform data of inverter output.
[0075] Exemplarily, a loop simulation program interface is compiled based on the MATLAB platform, and the loop simulation program interface is connected to the PSCAD / EMTDC simulation platform to realize automatic simulation of the grid-connected simulation model. According to the high-power DC boost simulation model control parameter values and inverter simulation model control parameter values corresponding to the first simulation scenario to the 390625th simulation scenario, a total of 390625 simulations are performed to obtain the high-power DC boost output first waveform data to the high-power DC boost output 390625th waveform data, and the inverter output first waveform data to the inverter output 390625th waveform data.
[0076] According to the inverter output first waveform data to the inverter output Nth waveform data and the inverter simulation model control parameter value matrix, the first attribution value to the Nth attribution value are obtained through the control attribution algorithm; the first attribution value to the Nth attribution value are respectively combined with the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data and the inverter output first waveform data to the inverter output Nth waveform data to obtain the first simulation data group to the Nth simulation data group.
[0077] The values of the first to Nth attribution values are one of 0 and 1.
[0078] Further, the method of obtaining the first attribution value to the Nth attribution value by controlling the attribution algorithm according to the inverter outputting the first waveform data to the inverter outputting the Nth waveform data and the inverter simulation model control parameter value matrix includes:
[0079] Fourier decomposition is performed on the inverter output first waveform data to the inverter output Nth waveform data respectively, and the fundamental wave amplitude and total harmonic distortion rate of the inverter output first waveform data to the inverter output Nth waveform data are obtained.
[0080] According to the fundamental amplitude and total harmonic distortion rate of the first waveform data output by the inverter to the Nth waveform data output by the inverter, a first abnormality detection algorithm is used to obtain the first attribution value to the Nth attribution value; the first abnormality detection algorithm is: if and ,but , otherwise, parameter attribution identification is performed; among them, is the preset lower limit of fundamental amplitude, is the preset upper limit of fundamental amplitude, is the preset upper limit of total harmonic distortion rate, Output the first i The fundamental amplitude of the waveform data, Output the first i Total harmonic distortion of waveform data, For the i Attribution value, i =1, 2...N; the parameter attribution identification steps are: if j Take [1, n 2], all integer values in the range satisfy ,but ,otherwise ;in, is the control parameter value matrix of the inverter simulation model. j Line i The elements of the column, j =1,2…… n 2, The inverter is preset j Control parameter standard lower limit, The inverter is preset j Control parameter standard upper limit.
[0081] For example, the preset =9.3kV, =10.7kV, =4%. Based on the first attribute value Take the calculation of as an example. The calculated =2.52kV, =35.96%, due to , , indicating that after high-power DC boost and inverter, the inverter outputs the first waveform data that does not meet the power quality requirements. This indicates that at least one of the high-power DC boost simulation model control parameter values and inverter simulation model control parameter values corresponding to the first simulation scenario is abnormal, so parameter attribution identification is required. Next, parameter attribution identification is performed. Since the inverter simulation model control parameter value matrix, Preset , , , , , , , , so . The principle is that in actual engineering, the sensitivity of the inverter control parameters is much higher than the sensitivity of the high-power DC boost control parameters. Therefore, in the parameter attribution identification, if it is determined that there is an abnormality in the values of the control parameters of the inverter simulation model, it is directly determined that the reason why the inverter output first waveform data does not meet the power quality requirements is due to the abnormal values of the control parameters of the inverter simulation model, that is, there is no abnormality in the values of the control parameters of the high-power DC boost simulation model. If it is determined that there is no abnormality in the values of the control parameters of the inverter simulation model, then there is an abnormality in the values of the control parameters of the high-power DC boost simulation model. Using this parameter attribution identification method, the calculation efficiency can be improved. Using the same method, it is calculated that .
[0082] Furthermore, the method of using the first to Nth simulation data groups to train a neural network model to obtain a first neural network model includes:
[0083] The average value of the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data is calculated in sequence to obtain the first DC waveform average value. To the average value of the Nth DC waveform ; Sequentially obtain the variance of the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data to obtain the variance of the first DC waveform To the Nth DC waveform variance .
[0084] Fourier decomposition is performed on the first waveform data output by the inverter to the Nth waveform data output by the inverter in sequence, thereby obtaining the total harmonic distortion rate of the first waveform output by the inverter. Total harmonic distortion rate of the inverter output Nth waveform ; Calculate the effective value of the inverter output first waveform data to the inverter output N waveform data in sequence to obtain the effective value of the inverter output first waveform Until the inverter outputs the effective value of the Nth waveform .
[0085] Establish the to , to , to , to and the mapping relationship between the first attribution value to the Nth attribution value, taking the first attribution value to the Nth attribution value as labels, to , to , to , to As a sample, the neural network model is input and trained to obtain the first neural network model.
[0086] For example, we get =9.87kV, =0.35kV 2 , =35.96%, =2.98kV, and the first attribution value So =9.87kV, =0.35kV 2 , =35.96%, =2.98kV, Input the neural network model. Using the same method, 390,625 sets of samples and labels are input into the neural network model for training to obtain the first neural network model.
[0087] Furthermore, the method of preprocessing the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data to obtain monitoring data samples, inputting the monitoring data samples into the first neural network model, and obtaining the high-power DC boost control effect identification result includes:
[0088] The high-power DC boost output waveform monitoring data is averaged to obtain the DC waveform monitoring average value. ; Obtain the variance of the high-power DC boost output waveform monitoring data to obtain the DC waveform monitoring variance .
[0089] Perform Fourier decomposition on the inverter output waveform monitoring data to obtain the inverter waveform monitoring total harmonic distortion rate ; Calculate the effective value of the inverter output waveform monitoring data to obtain the inverter waveform monitoring total harmonic distortion rate .
[0090] The , , , The monitoring data samples are combined to obtain monitoring data samples, and the monitoring data samples are input into the first neural network model to obtain monitoring attribution values; if the monitoring attribution value is 0, "the control system is normal" is output, otherwise "the control system is abnormal" is output.
[0091] For example, the operation and maintenance personnel monitor the , , , By combining and inputting into the first neural network model, the identification result of high-power DC boost control effect can be obtained.
[0092] Example 2
[0093] Based on the same inventive concept, Figure 2 As shown, this embodiment also provides a high-power DC boost control effect optimization system, the system comprising:
[0094] The simulation model building module is used to read the basic parameters of the high-power DC boost model and the inverter model, build the high-power DC boost simulation model and the inverter simulation model on the simulation platform; combine the high-power DC boost simulation model and the inverter simulation model to obtain a grid-connected simulation model.
[0095] The first neural network model construction module is connected to the simulation model construction module, and is used to read the number of control parameters in the high-power DC boost simulation model and the number of control parameters in the inverter simulation model, calculate the number of simulations N, set the first simulation scene to the Nth simulation scene, perform N simulation simulations on the high-power DC boost simulation model and the inverter simulation model, and record the first simulation data group to the Nth simulation data group obtained by the simulation; use the first simulation data group to the Nth simulation data group to train the neural network model to obtain the first neural network model.
[0096] The high-power DC boost control effect identification module is connected to the first neural network model construction module, and is used to monitor the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data within a preset first monitoring period; the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data are preprocessed to obtain monitoring data samples, and the monitoring data samples are input into the first neural network model to obtain a high-power DC boost control effect identification result.
[0097] The high-power DC boost control parameter adjustment module is connected to the high-power DC boost control effect identification module and is used to determine whether the high-power DC boost control parameter needs to be adjusted according to the high-power DC boost control effect identification result.
[0098] Furthermore, the first neural network model building module includes:
[0099] A data reading module is used to read the control parameter quantity in the high-power DC boost simulation model. n 1, n 1>1; Read the number of control parameters in the inverter simulation model n 2, n2>1; according to the preset number of segments D, the simulation number N is obtained by the simulation number calculation formula; the simulation number calculation formula is:
[0100] .
[0101] The simulation scene construction module is connected to the data reading module and is used to 、 The number of control parameters in the high-power DC boost simulation model, the number of control parameters in the inverter simulation model, and the pre-set control parameter limits of the high-power DC boost simulation model and the control parameter limits of the inverter simulation model are used to construct the first simulation scenario to the Nth simulation scenario, and obtain the control parameter value matrix of the inverter simulation model.
[0102] The simulation module is connected to the simulation scenario construction module, and is used to perform N simulations on the grid-connected simulation model according to the first simulation scenario to the Nth simulation scenario, to obtain the first waveform data of the high-power DC boost output to the Nth waveform data of the high-power DC boost output, and the first waveform data of the inverter output to the Nth waveform data of the inverter output.
[0103] The simulation array calculation module is connected to the simulation simulation module, and is used to obtain the first attribution value to the Nth attribution value through the control attribution algorithm according to the inverter output first waveform data to the inverter output Nth waveform data and the inverter simulation model control parameter value matrix; the first attribution value to the Nth attribution value are respectively combined with the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data and the inverter output first waveform data to the inverter output Nth waveform data to obtain the first simulation data group to the Nth simulation data group.
[0104] Furthermore, the simulation array calculation module includes:
[0105] The harmonic calculation module is used to perform Fourier decomposition on the inverter output first waveform data to the inverter output Nth waveform data, and obtain the fundamental amplitude and total harmonic distortion rate of the inverter output first waveform data to the inverter output Nth waveform data.
[0106] The attribution value calculation module is connected to the harmonic calculation module and is used to obtain the first attribution value to the Nth attribution value according to the fundamental amplitude and total harmonic distortion rate of the inverter outputting the first waveform data to the inverter outputting the Nth waveform data by using the first abnormality detection algorithm; the first abnormality detection algorithm is: if and ,but , otherwise, parameter attribution identification is performed; among them, is the preset lower limit of fundamental amplitude, is the preset upper limit of fundamental amplitude, is the preset upper limit of total harmonic distortion rate, Output the first i The fundamental amplitude of the waveform data, Output the first i Total harmonic distortion of waveform data, For the i Attribution value, i =1, 2...N; the parameter attribution identification steps are: if j Take [1, n 2], all integer values in the range satisfy ,but ,otherwise ;in, is the control parameter value matrix of the inverter simulation model. j Line i The elements of the column, j =1,2…… n 2, The inverter is preset j Control parameter standard lower limit, The inverter is preset j Control parameter standard upper limit.
[0107] Furthermore, the first neural network model building module also includes:
[0108] A high-power DC boost output processing module is used to sequentially calculate the average value of the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data to obtain the first DC waveform average value To the average value of the Nth DC waveform ; Sequentially obtain the variance of the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data to obtain the variance of the first DC waveform To the Nth DC waveform variance .
[0109] The inverter output processing module is connected to the high-power DC boost output processing module, and is used to sequentially perform Fourier decomposition on the inverter output first waveform data to the inverter output Nth waveform data, thereby obtaining the total harmonic distortion rate of the inverter output first waveform. Total harmonic distortion rate of the inverter output Nth waveform ; Calculate the effective value of the inverter output first waveform data to the inverter output N waveform data in sequence to obtain the effective value of the inverter output first waveform Until the inverter outputs the effective value of the Nth waveform .
[0110] The neural network training module is connected to the inverter output processing module to establish the to , to , to , to and the mapping relationship between the first attribution value to the Nth attribution value, taking the first attribution value to the Nth attribution value as labels, to , to , to , to As a sample, the neural network model is input and trained to obtain the first neural network model.
[0111] Furthermore, the high-power DC boost control effect identification module includes:
[0112] A high-power DC boost output monitoring processing module is used to calculate the average value of the high-power DC boost output waveform monitoring data to obtain the DC waveform monitoring average value ; Obtain the variance of the high-power DC boost output waveform monitoring data to obtain the DC waveform monitoring variance .
[0113] The inverter output monitoring data processing module is connected to the high-power DC boost output monitoring processing module and is used to perform Fourier decomposition on the inverter output waveform monitoring data to obtain the inverter waveform monitoring total harmonic distortion rate. ; Calculate the effective value of the inverter output waveform monitoring data to obtain the inverter waveform monitoring total harmonic distortion rate .
[0114] The identification result output module is connected to the inverter output monitoring data processing module and is used to , , , The monitoring data samples are combined to obtain monitoring data samples, and the monitoring data samples are input into the first neural network model to obtain monitoring attribution values; if the monitoring attribution value is 0, "the control system is normal" is output, otherwise "the control system is abnormal" is output.
[0115] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.
[0116] Finally, it should be noted that: Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A high-power DC boost control effect optimization method, characterized in that: The method comprises the following steps: S1, reading the basic parameters of the high-power DC boost model and the inverter model, building the high-power DC boost simulation model and the inverter simulation model on the simulation platform; combining the high-power DC boost simulation model and the inverter simulation model to obtain a grid-connected simulation model; S2, reading the number of control parameters in the high-power DC boost simulation model and the number of control parameters in the inverter simulation model, calculating the number of simulations N, setting the first simulation scenario to the Nth simulation scenario, performing N simulation simulations on the high-power DC boost simulation model and the inverter simulation model, and recording the first simulation data group to the Nth simulation data group obtained by the simulation; using the first simulation data group to the Nth simulation data group to train the neural network model to obtain a first neural network model; S3, monitoring the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data within a preset first monitoring period; performing data preprocessing on the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data to obtain monitoring data samples, and inputting the monitoring data samples into the first neural network model to obtain a high-power DC boost control effect identification result; S4, judging whether a high-power DC boost control parameter needs to be adjusted according to the high-power DC boost control effect identification result; The method of reading the number of control parameters in the high-power DC boost simulation model and the number of control parameters in the inverter simulation model, calculating the number of simulations N, setting the first simulation scenario to the Nth simulation scenario, performing N simulation simulations on the high-power DC boost simulation model and the inverter simulation model, and recording the first simulation data group to the Nth simulation data group obtained by the simulation includes: Read the number of control parameters in the high-power DC boost simulation model n 1, n 1>1; Read the number of control parameters in the inverter simulation model n 2, n 2>1; according to the preset number of segments D, the simulation number N is obtained by the simulation number calculation formula; the simulation number calculation formula is: ; According to the number of segments 、 The number of control parameters in the high-power DC boost simulation model, the number of control parameters in the inverter simulation model, and the pre-set control parameter limits of the high-power DC boost simulation model and the control parameter limits of the inverter simulation model are used to construct the first simulation scenario to the Nth simulation scenario, and obtain the control parameter value matrix of the inverter simulation model; According to the first simulation scenario to the Nth simulation scenario, N simulations are performed on the grid-connected simulation model to obtain high-power DC boost output first waveform data to high-power DC boost output Nth waveform data, and inverter output first waveform data to inverter output Nth waveform data; According to the inverter output first waveform data to the inverter output Nth waveform data and the inverter simulation model control parameter value matrix, the first attribution value to the Nth attribution value are obtained through the control attribution algorithm; the first attribution value to the Nth attribution value are respectively combined with the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data and the inverter output first waveform data to the inverter output Nth waveform data to obtain the first simulation data group to the Nth simulation data group.
2. The high-power DC boost control effect optimization method according to claim 1, characterized in that: The method of obtaining the first attribution value to the Nth attribution value by controlling the attribution algorithm according to the inverter outputting the first waveform data to the inverter outputting the Nth waveform data and the inverter simulation model control parameter value matrix includes: Performing Fourier decomposition on the inverter output first waveform data to the inverter output Nth waveform data respectively, and obtaining the fundamental wave amplitude and total harmonic distortion rate of the inverter output first waveform data to the inverter output Nth waveform data; According to the fundamental amplitude and total harmonic distortion rate of the first waveform data output by the inverter to the Nth waveform data output by the inverter, a first abnormality detection algorithm is used to obtain the first attribution value to the Nth attribution value; the first abnormality detection algorithm is: if and ,but , otherwise, parameter attribution identification is performed; among them, is the preset lower limit of fundamental amplitude, is the preset upper limit of fundamental amplitude, is the preset upper limit of total harmonic distortion rate, Output the first i The fundamental amplitude of the waveform data, Output the first i Total harmonic distortion of waveform data, For the i Attribution value, i =1, 2...N; the parameter attribution identification steps are: if j Take [1, n 2], all integer values in the range satisfy ,but ,otherwise ;in, is the control parameter value matrix of the inverter simulation model. j Line i The elements of the column, j =1,2…… n 2, The inverter is preset j Control parameter standard lower limit, The inverter is preset j Control parameter standard upper limit.
3. The high-power DC boost control effect optimization method according to claim 1, characterized in that: The method of using the first to Nth simulation data groups to train a neural network model to obtain a first neural network model comprises: The average value of the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data is calculated in sequence to obtain the first DC waveform average value. To the average value of the Nth DC waveform ; Obtain the variance of the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data in sequence, and obtain the variance of the first DC waveform To the Nth DC waveform variance ; Fourier decomposition is performed on the first waveform data output by the inverter to the Nth waveform data output by the inverter in sequence, thereby obtaining the total harmonic distortion rate of the first waveform output by the inverter. Total harmonic distortion rate of the inverter output Nth waveform ; Calculate the effective value of the inverter output first waveform data to the inverter output N waveform data in sequence to obtain the effective value of the inverter output first waveform Until the inverter outputs the effective value of the Nth waveform ; Establish the to , to , to , to and the mapping relationship between the first attribution value to the Nth attribution value, taking the first attribution value to the Nth attribution value as labels, to , to , to , to As a sample, the neural network model is input and trained to obtain the first neural network model.
4. The high-power DC boost control effect optimization method according to claim 1, characterized in that: The method of preprocessing the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data to obtain monitoring data samples, and inputting the monitoring data samples into the first neural network model to obtain the high-power DC boost control effect identification result includes: The high-power DC boost output waveform monitoring data is averaged to obtain the DC waveform monitoring average value. ; Obtain the variance of the high-power DC boost output waveform monitoring data to obtain the DC waveform monitoring variance ; Perform Fourier decomposition on the inverter output waveform monitoring data to obtain the inverter waveform monitoring total harmonic distortion rate ; Calculate the effective value of the inverter output waveform monitoring data to obtain the inverter waveform monitoring total harmonic distortion rate ; The , , , Combine to obtain monitoring data samples, input the monitoring data samples into the first neural network model, and obtain monitoring attribution values; if the monitoring attribution value is 0, output "control system is normal", otherwise output "control system is abnormal".
5. A high-power DC boost control effect optimization system, characterized in that: The system comprises: A simulation model building module is used to read the basic parameters of the high-power DC boost model and the inverter model, build the high-power DC boost simulation model and the inverter simulation model on the simulation platform; combine the high-power DC boost simulation model and the inverter simulation model to obtain a grid-connected simulation model; A first neural network model construction module is connected to the simulation model construction module, and is used to read the number of control parameters in the high-power DC boost simulation model and the number of control parameters in the inverter simulation model, calculate the number of simulations N, set the first simulation scene to the Nth simulation scene, perform N simulation simulations on the high-power DC boost simulation model and the inverter simulation model, and record the first simulation data group to the Nth simulation data group obtained by the simulation; use the first simulation data group to the Nth simulation data group to train the neural network model to obtain a first neural network model; A high-power DC boost control effect identification module is connected to the first neural network model building module and is used to monitor the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data within a preset first monitoring period; pre-process the high-power DC boost output waveform monitoring data and the inverter output waveform monitoring data to obtain monitoring data samples, and input the monitoring data samples into the first neural network model to obtain a high-power DC boost control effect identification result; A high-power DC boost control parameter adjustment module is connected to the high-power DC boost control effect identification module and is used to determine whether the high-power DC boost control parameter needs to be adjusted according to the high-power DC boost control effect identification result; The first neural network model building module includes: A data reading module is used to read the control parameter quantity in the high-power DC boost simulation model. n 1, n 1>1; Read the number of control parameters in the inverter simulation model n 2, n 2>1; according to the preset number of segments D, the simulation number N is obtained by the simulation number calculation formula; the simulation number calculation formula is: ; The simulation scene construction module is connected to the data reading module and is used to 、 The number of control parameters in the high-power DC boost simulation model, the number of control parameters in the inverter simulation model, and the pre-set control parameter limits of the high-power DC boost simulation model and the control parameter limits of the inverter simulation model are used to construct the first simulation scenario to the Nth simulation scenario, and obtain the control parameter value matrix of the inverter simulation model; A simulation module, connected to the simulation scenario construction module, is used to perform N simulations on the grid-connected simulation model according to the first simulation scenario to the Nth simulation scenario, to obtain the first waveform data of the high-power DC boost output to the Nth waveform data of the high-power DC boost output, and the first waveform data of the inverter output to the Nth waveform data of the inverter output; The simulation array calculation module is connected to the simulation simulation module, and is used to obtain the first attribution value to the Nth attribution value through the control attribution algorithm according to the inverter output first waveform data to the inverter output Nth waveform data and the inverter simulation model control parameter value matrix; the first attribution value to the Nth attribution value are respectively combined with the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data and the inverter output first waveform data to the inverter output Nth waveform data to obtain the first simulation data group to the Nth simulation data group.
6. The high-power DC boost control effect optimization system as claimed in claim 5, characterized in that: The simulation array calculation module includes: A harmonic calculation module, used to perform Fourier decomposition on the inverter output first waveform data to the inverter output Nth waveform data, and obtain the fundamental amplitude and total harmonic distortion rate of the inverter output first waveform data to the inverter output Nth waveform data; The attribution value calculation module is connected to the harmonic calculation module and is used to obtain the first attribution value to the Nth attribution value according to the fundamental amplitude and total harmonic distortion rate of the inverter outputting the first waveform data to the inverter outputting the Nth waveform data by using the first abnormality detection algorithm; the first abnormality detection algorithm is: if and ,but , otherwise, parameter attribution identification is performed; among them, is the preset lower limit of fundamental amplitude, is the preset upper limit of fundamental amplitude, is the preset upper limit of total harmonic distortion rate, Output the first i The fundamental amplitude of the waveform data, Output the first i Total harmonic distortion of waveform data, For the i Attribution value, i =1, 2...N; the parameter attribution identification steps are: if j Take [1, n 2], all integer values in the range satisfy ,but ,otherwise ;in, is the control parameter value matrix of the inverter simulation model. j Line i The elements of the column, j =1,2…… n 2, The inverter is preset j Control parameter standard lower limit, The inverter is preset j Control parameter standard upper limit.
7. The high-power DC boost control effect optimization system as claimed in claim 5, characterized in that: The first neural network model building module also includes: A high-power DC boost output processing module is used to sequentially calculate the average value of the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data to obtain the first DC waveform average value To the average value of the Nth DC waveform ; Obtain the variance of the high-power DC boost output first waveform data to the high-power DC boost output Nth waveform data in sequence, and obtain the variance of the first DC waveform To the Nth DC waveform variance ; The inverter output processing module is connected to the high-power DC boost output processing module, and is used to sequentially perform Fourier decomposition on the inverter output first waveform data to the inverter output Nth waveform data, thereby obtaining the total harmonic distortion rate of the inverter output first waveform. Total harmonic distortion rate of the inverter output Nth waveform ; Calculate the effective value of the inverter output first waveform data to the inverter output N waveform data in sequence to obtain the effective value of the inverter output first waveform Until the inverter outputs the effective value of the Nth waveform ; The neural network training module is connected to the inverter output processing module to establish the to , to , to , to and the mapping relationship between the first attribution value to the Nth attribution value, taking the first attribution value to the Nth attribution value as labels, to , to , to , to As a sample, the neural network model is input and trained to obtain the first neural network model.
8. The high-power DC boost control effect optimization system as claimed in claim 5, characterized in that: The high-power DC boost control effect identification module includes: A high-power DC boost output monitoring processing module is used to calculate the average value of the high-power DC boost output waveform monitoring data to obtain the DC waveform monitoring average value ; Obtain the variance of the high-power DC boost output waveform monitoring data to obtain the DC waveform monitoring variance ; The inverter output monitoring data processing module is connected to the high-power DC boost output monitoring processing module and is used to perform Fourier decomposition on the inverter output waveform monitoring data to obtain the inverter waveform monitoring total harmonic distortion rate. ; Calculate the effective value of the inverter output waveform monitoring data to obtain the inverter waveform monitoring total harmonic distortion rate ; The identification result output module is connected to the inverter output monitoring data processing module and is used to , , , Combine to obtain monitoring data samples, input the monitoring data samples into the first neural network model, and obtain monitoring attribution values; if the monitoring attribution value is 0, output "control system is normal", otherwise output "control system is abnormal".
