Grid-connected control method and controller for reducing harmonic noise

Abstract

The invention provides a grid-connected control method for reducing harmonic noise and a controller, and relates to the technical field of grid-connected control, and the method comprises the steps: collecting amplitude information and phase information of each harmonic component, recording the amplitude information and phase information as first information, calculating compensation data according to the first information, building a mathematical model of an inverter, and calculating the compensation data according to the mathematical model; the method comprises the following steps: calculating compensation data corresponding to each harmonic component to obtain grid-connected parameter information of an inverter, generating a control instruction based on the compensation data, adjusting the grid-connected parameter information by the inverter according to the control instruction, generating an output current, and injecting the output current into a power grid. And the suppression effect on harmonic waves with different times is improved.

Description

Technical Field

[0001] The present application relates to the technical field of grid-connected control, and in particular to a grid-connected control method and controller for reducing harmonic noise. Background Art

[0002] In the power system, the grid-connected inverter is the interface device between the distributed power source and the power grid, and its performance directly affects the stability and power quality of the power grid. However, in the process of grid connection, the harmonic noise generated by the inverter is an issue that cannot be ignored. Harmonic noise mainly comes from the nonlinear characteristics of the inverter switching devices. These harmonic components will be injected into the power grid, causing waveform distortion of the grid voltage and current, which in turn has a negative impact on the stability and power quality of the power grid. Traditional harmonic suppression methods mostly use passive filtering, that is, adding a filter to the output end of the inverter to filter out the harmonic components. However, this method has the disadvantages of large filter size, high cost, and difficult maintenance, and the filtering effect is limited, making it difficult to meet the growing power quality requirements.

[0003] In order to overcome the shortcomings of traditional harmonic suppression methods, active harmonic suppression technology has gradually attracted attention in recent years. Active harmonic suppression technology monitors the harmonic components in the power grid in real time, calculates the corresponding compensation data, and then controls the inverter to output a compensation current opposite to the harmonic components, thereby effectively suppressing harmonic noise. However, there are still some problems with existing active harmonic suppression technologies. Some technologies can only suppress harmonics of a specific order, but are not effective for harmonics of other orders. Summary of the invention

[0004] In order to provide a certain suppression effect on harmonics of different orders, the present application provides a grid-connected control method and a controller for reducing harmonic noise.

[0005] In a first aspect, the present application provides a grid-connected control method for reducing harmonic noise, which adopts the following technical solution: A grid-connected control method for reducing harmonic noise comprises the following steps: First collection: collecting the harmonic components of the current in the power grid, obtaining the amplitude information and phase information of each harmonic component, and recording it as the first information; First calculation: According to the first information, the compensation data is calculated, and the calculation model of the compensation data is as follows: ; in, To compensate for the data; is the amplitude information of the nth harmonic component; n is the harmonic order of the harmonic component; is the fundamental angular frequency of the grid-connected current; t is the time; is the phase information of the nth harmonic component; First modeling: Establish a mathematical model of the inverter in a three-phase stationary coordinate system to obtain the grid-connected parameter information of the inverter; Generate instructions: Generate control instructions based on compensation data; Generate signal: The inverter adjusts the grid-connected parameter information according to the control instruction, generates output current, and injects the output current into the grid.

[0006] By adopting the above technical solution, the present application can fully understand the harmonic conditions in the power grid by collecting the harmonic components of the current in the power grid and obtaining the amplitude information and phase information of each harmonic component. The present application calculates compensation data through a calculation model based on the collected harmonic information. The calculation model of the present application takes into account the amplitude, harmonic order, fundamental angular frequency, time and phase information of the harmonic components, so that accurate compensation data can be generated. These compensation data will be used to guide the output of the inverter to offset the harmonics of different orders in the power grid. The controller generates control instructions based on the calculated compensation data. These control instructions will directly guide the operation of the inverter, so that the inverter can output a current with the opposite amplitude and the same phase as the harmonics in the power grid, thereby realizing harmonic compensation. The inverter adjusts its grid-connected parameter information according to the control instruction, generates an output current, and injects the current into the power grid. Through the precise control of the inverter, effective compensation for harmonics in the power grid is achieved.

[0007] Optionally, after executing the first modeling step and before executing the step of generating instructions, the method further includes: Coordinate transformation: Clarke transformation algorithm and Park transformation algorithm are used to transform the mathematical model in the three-phase stationary coordinate system into the mathematical model in the two-phase rotating coordinate system; Second calculation: Calculate the current change rate of the power grid under the action of zero voltage, recorded as the first current change rate; calculate the current change rate of the power grid under the action of non-zero voltage, recorded as the second current change rate; First prediction: Calculate the prediction data of the grid current based on the first current change rate and the second current change rate, recorded as the first data; The third calculation: calculating the difference between the first data and the real current data, and recording it as the second data; Difference judgment: judge whether the difference between the second data and the compensation data is less than the preset difference threshold, if so, execute the step of generating instructions; if not, execute the step of second calculation.

[0008] By adopting the above technical solution, the present application adopts Clarke transformation algorithm and Park transformation algorithm to convert the mathematical model in the three-phase stationary coordinate system into the mathematical model in the two-phase rotating coordinate system, which helps to simplify the mathematical model of the inverter and makes the subsequent calculation more efficient and accurate. Calculating the current change rate of the power grid under zero voltage and non-zero voltage helps to understand the response characteristics of the power grid under different voltage conditions. The current change rate is an important indicator of the dynamic performance of the power grid and can reflect the stability and response speed of the power grid. Based on the first current change rate and the second current change rate, the predicted data of the power grid current is calculated. This step uses the known current change rate to predict the future current value, and then calculates the difference between the predicted current value and the actual current value, and compares the difference with the compensation data to determine whether the difference between the two meets expectations. If so, it means that the harmonic components in the power grid can be well suppressed, and then the step of generating instructions is executed; otherwise, it means that there is a calculation error or the harmonic components in the power grid are not well suppressed (there is a problem with the compensation data or a new harmonic component has appeared), and the second calculation step needs to be re-executed to verify whether the phenomenon is caused by the calculation error.

[0009] Optionally, after executing the step of determining the difference and before executing the step of generating the instruction, the method further includes: Fourth calculation: calculating predicted data of the grid voltage based on the first data; Inverse Park transformation: Inverse Park transformation is performed on the predicted data of the grid voltage to obtain the voltage data in the two-phase stationary coordinate system, which is recorded as the third data; Calculating a voltage phase: calculating a voltage phase based on the third data; Output control: adding the voltage phase to the original compensation data as new compensation data.

[0010] By adopting the above technical solution, the present application calculates the predicted data of the grid voltage based on the first data (the predicted data of the grid current), which can more effectively manage the stability and performance of the grid. The present application can convert the predicted grid voltage data into a format that is easier to process and apply through the inverse Park transformation, thereby improving the applicability and practicality of the data. The present application optimizes the voltage control process by calculating the voltage phase and updating the compensation data, thereby improving the stability and performance of the grid.

[0011] Optionally, after the step of performing the inverse Park transformation and before the step of calculating the voltage phase, the method further includes: Second collection: collect the active power in the power grid, recorded as the first power; collect the reactive power in the power grid, recorded as the second power; Fifth calculation: calculating active power based on the first data and the third data, recorded as the third power; calculating reactive power based on the first data and the third data, recorded as the fourth power; Sixth calculation: calculating the difference between the first power and the third power, and recording it as the fourth data; calculating the difference between the second power and the fourth power, and recording it as the fifth data; Power judgment: judge whether the fourth data and the fifth data meet expectations, if so, execute the step of calculating the voltage phase; if not, execute the second calculation step.

[0012] By adopting the above-mentioned technical scheme, in the second acquisition step, the present application collects active power and reactive power in real time, and then in the fifth calculation step, calculates the active power and reactive power under the predicted data based on the predicted current data and voltage data (after coordinate conversion), and then in the sixth calculation step, calculates the difference between the real-time data of active power and the predicted data, and then in the power judgment step, judges whether these differences meet expectations, and if so, it means that the differences are within an acceptable range, and then executes the step of calculating the voltage phase, otherwise, executes the second calculation step to verify whether there are calculation errors in the process.

[0013] Optionally, after executing the step of calculating the voltage phase and before executing the step of output control, the method further includes: First filtering: inputting the fourth data into a low-pass filter to obtain new fourth data; Seventh calculation: performing a product operation on the new fourth data and the preset adjustment coefficient to obtain an active frequency adjustment amount, converting the active frequency adjustment amount into a voltage phase adjustment amount, which is recorded as the sixth data; First summation: sum the voltage phase with the sixth data to obtain a new voltage phase.

[0014] By adopting the above-mentioned technical scheme, the present application inputs the fourth data into a low-pass filter, smoothes it, and filters out the high-frequency components. Then, the fourth data after smoothing is multiplied by a preset adjustment coefficient to obtain an active power adjustment amount, and the adjustment amount is converted to obtain a voltage phase adjustment amount. Then, the voltage phase obtained in the step of calculating the voltage phase is summed with the voltage phase adjustment amount to obtain a more accurate new voltage phase. The entire step is based on real-time data and dynamic changes of the power system, which helps to maintain the stability and power factor of the power system.

[0015] Optionally, after performing the seventh calculation step and before performing the first summation step, the method further includes: Second filtering: inputting the fifth data into a low-pass filter to obtain new fifth data; Eighth calculation: multiplying the new fifth data by the preset adjustment coefficient to obtain a reactive power adjustment amount, converting the reactive frequency adjustment amount into a voltage phase adjustment amount, which is recorded as the seventh data; Second summation: sum the sixth data and the seventh data to obtain new sixth data.

[0016] By adopting the above-mentioned technical solution, the present application inputs the fifth data into a low-pass filter and smoothes it. Then, the smoothed fifth data is multiplied by a preset adjustment coefficient to obtain a reactive power adjustment amount, and the adjustment amount is converted to obtain a voltage phase adjustment amount. Then, the sixth data and the seventh data are summed to obtain a voltage phase that is more accurate than the sixth data.

[0017] Optionally, after executing the step of generating a signal, the method further includes: The third collection: collecting the current data in the power grid after the output current is injected, which is recorded as the eighth data; Second modeling: establishing a neural network model; Second prediction: input the eighth data and the compensation data into the neural network model, and output the current regulation amount; Adjustment: adjusting the compensation data based on the current adjustment amount, and using the adjusted compensation data as new compensation data; Determine the adjustment amount: determine whether the current adjustment amount is less than a preset threshold value, if so, output the weight coefficient of the neural network model; if not, execute the second prediction step; Parameter replacement: The weight coefficients of the neural network model are used as parameters of the controller.

[0018] By adopting the above technical solution, the present application establishes a neural network model, uses the neural network model to learn and simulate the relationship between the grid current and the inverter output current, outputs the current regulation amount, so as to realize the precise control of the grid current, and adjusts the compensation data based on the current regulation amount. By adjusting the compensation data, the output current of the inverter can be further corrected to make it closer to the target value. In the whole process, the neural network model essentially replaces the controller, realizes the adjustment of the inverter output current data, and then the weight coefficient of the neural network model is combined with the parameters of the controller, and the internal parameters of the controller are optimized by optimizing the weight coefficient of the neural network model. Traditional controller parameters are often set based on experience or fixed models, which are difficult to adapt to the real-time changes of the grid and inverter load. By using the weight coefficient of the neural network model as the parameter of the controller, the learning ability of the neural network can be used to adjust the parameters of the controller in real time, thereby realizing more accurate and flexible control of the grid current or voltage.

[0019] Optionally, after the step of performing the second prediction and before the step of performing the adjustment, the method further includes: Online learning: The neural network model uses an online learning algorithm to adjust the internal weight coefficients by reducing the current regulation amount.

[0020] By adopting the above technical solution, the grid status and inverter load may change at any time during actual operation. The neural network model uses an online learning algorithm to continuously adjust the internal weight coefficients to reduce the current regulation amount until it is lower than the preset regulation amount threshold. This process is essentially a real-time optimization of the neural network model to make its predicted output more accurate, thereby generating a more accurate current regulation amount. This solution improves the performance of the inverter, allowing it to maintain excellent current control capabilities under various working conditions, providing a strong guarantee for the stable operation of the power system.

[0021] Optionally, after performing the third acquisition step and before performing the second modeling step, the method further includes: Harmonic determination: determining whether the eighth data contains even harmonics, if so, executing the checking step; if not, executing the second modeling step; Check: Check the wiring.

[0022] By adopting the above technical solution, the present application can identify in advance whether there are still even harmonic components remaining in the power grid before establishing the neural network model through the step of harmonic judgment. Since even harmonics only exist when there is a wiring error / poor contact, when even harmonics are found in the eighth data, it means that there is a wiring error and the inspection step needs to be performed. Through the step of harmonic judgment, the existence of even harmonics can be keenly perceived before modeling, thereby effectively reducing the complexity and errors that may be introduced due to harmonic interference during the modeling process, which not only improves the accuracy of the model, but also provides early warning information for subsequent harmonic governance.

[0023] In a second aspect, the present invention provides a controller, comprising a processor and a memory, wherein the memory stores machine executable instructions that can be executed by the processor, and the processor executes the machine executable instructions to implement the described method.

[0024] In summary, the present application includes at least one of the following beneficial technical effects: 1. This application can fully understand the harmonic conditions in the power grid by collecting the harmonic components of the current in the power grid and obtaining the amplitude information and phase information of each harmonic component. This application calculates compensation data through a calculation model based on the collected harmonic information. The calculation model of this application takes into account the amplitude, harmonic order, fundamental angular frequency, time and phase information of the harmonic components, so it can generate accurate compensation data. These compensation data will be used to guide the output of the inverter to offset the harmonics of different orders in the power grid. The controller generates control instructions based on the calculated compensation data. These control instructions will directly guide the operation of the inverter, so that the inverter can output a current with the opposite amplitude and the same phase as the harmonics in the power grid, thereby realizing harmonic compensation. The inverter adjusts its grid-connected parameter information according to the control instructions, generates an output current, and injects the current into the power grid. Through the precise control of the inverter, effective compensation for harmonics in the power grid is achieved.

[0025] 2. This application calculates the predicted data of the grid voltage based on the first data (the predicted data of the grid current), which can more effectively manage the stability and performance of the grid. This application can convert the predicted grid voltage data into a format that is easier to process and apply through the inverse Park transformation, thereby improving the applicability and practicality of the data. This application optimizes the voltage control process by calculating the voltage phase and updating the compensation data, and improves the stability and performance of the grid.

[0026] 3. In the second acquisition step, the present application acquires active power and reactive power in real time. Then, in the fifth calculation step, the active power and reactive power under the predicted data are calculated based on the predicted current data and voltage data (after coordinate conversion). Then, in the sixth calculation step, the difference between the real-time active power data and the predicted data is calculated respectively. Then, in the power judgment step, it is judged whether these differences meet expectations. If so, it means that the differences are within an acceptable range. Then, the step of calculating the voltage phase is executed. Otherwise, the second calculation step is executed to verify whether there are calculation errors in the process. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a flow chart from S1 first acquisition to S5 signal generation in Example 1 of the present application; Figure 2 is a circuit topology diagram of the inverter of the present application; Figure 3 This is a flowchart of the first modeling from S3 to S4 generating instructions in Example 2 of the present application; Figure 4 This is a flow chart from S68 calculating the voltage phase to S69 output control in Example 2 of the present application; Figure 5 This is a flow chart of Example 3 of the present application. DETAILED DESCRIPTION

[0028] The following combination Figures 1 to 5 This application is described in further detail.

[0029] Embodiment 1: This embodiment discloses a grid-connected control method for reducing harmonic noise, referring to Figure 1 The method includes: S1 first acquisition, S2 first calculation, S3 first modeling, S4 instruction generation and S5 signal generation. First, the amplitude information and phase information of each harmonic component are collected and recorded as the first information. Then, the compensation data is calculated according to the first information. Then, a mathematical model of the inverter is established to obtain the grid-connected parameter information of the inverter. Then, a control instruction is generated based on the compensation data. Then, the inverter adjusts the grid-connected parameter information according to the control instruction, generates an output current, and injects the output current into the power grid. This embodiment includes the following steps: S1 first acquisition, collect the harmonic components of the current in the power grid, obtain the amplitude information and phase information of each harmonic component, and record it as the first information. The harmonics are all other frequency components in the power grid except the fundamental wave, including odd harmonics and even harmonics.

[0030] S2 first calculation, based on the first information, calculates compensation data to offset the harmonic components in the power grid, and the calculation model of the compensation data is as follows: ; in, is the compensation data (i.e. the compensation current at time t); is the amplitude information of the nth harmonic component; n is the harmonic order of the harmonic component (starting from 2, because n=1 corresponds to the fundamental wave, the fundamental wave is not considered as a harmonic, when n is an odd number, The corresponding harmonics are odd harmonics. When n is an even number, corresponds to even harmonics); is the fundamental angular frequency of the grid-connected current; t is the time; is the phase information of the nth harmonic component.

[0031] S3 is the first modeling, which establishes the mathematical model of the inverter in the three-phase stationary coordinate system (a, b, c) to obtain the grid-connected parameter information of the inverter.

[0032] The mathematical model of the inverter in the three-phase stationary coordinate system is derived based on Kirchhoff's law. In this model, the inverter is regarded as a three-phase AC time-varying system, and its switching function pattern reflects the working state of the inverter.

[0033] Reference Figure 2 , taking the direction of the grid flowing into the inverter as the positive direction, the mathematical model of the inverter is as follows: ; Among them, L is the equivalent inductance; R is the equivalent resistance; C is the equivalent capacitance; is the current of phase a in the three-phase power grid; is the voltage of the three-phase network on phase a; is the output voltage generated by the three-phase grid-side inverter on phase a; is the current of the three-phase power grid on phase b; is the voltage of the three-phase network on phase b; is the output voltage generated by the three-phase grid-side inverter on phase b; is the current of the three-phase power grid on phase c; is the voltage of the three-phase network on phase c; is the output voltage generated by the three-phase grid-side inverter on phase c; is the DC bus capacitor voltage; is the DC side output current.

[0034] is the switching function of the a-phase bridge arm; is the switching function of the b-phase bridge arm; is the switching function of the c-phase bridge arm; the calculation model of the three is: .

[0035] An inverter is a power electronic device that can convert direct current into alternating current and can adjust the frequency, amplitude and phase of the output current by controlling its switching state. In this embodiment, by establishing a mathematical model, grid-connected parameter information of the inverter, such as output voltage, current, power, etc., can be obtained, wherein the power is further calculated by voltage and current.

[0036] S4 generates instructions, and generates control instructions based on the compensation data, wherein the control instructions are used to adjust the switching state of the inverter so that the inverter outputs a compensation current that can offset the harmonic components in the power grid.

[0037] The control instruction described in this embodiment is a PWM signal generated by using pulse width modulation technology. The PWM signal accurately controls the current waveform and amplitude output by the inverter by adjusting the switching frequency and duty cycle of the inverter.

[0038] S5 generates a signal, and the inverter adjusts the grid-connected parameter information according to the control instruction, that is, the inverter quickly switches its switch state according to the received PWM signal. When the PWM signal is at a high level, the inverter outputs current; when the PWM signal is at a low level, the inverter stops outputting current. By adjusting the duty cycle of the PWM signal, the average value of the inverter output current can be controlled to accurately match the required compensation current. The output current will contain compensation data for offsetting the harmonic components in the power grid, and then the output current will be injected into the power grid.

[0039] This embodiment first collects the harmonic components of the current from the power grid and obtains its amplitude and phase information, then calculates compensation data for offsetting the harmonics based on this information, then establishes a mathematical model of the inverter in a three-phase stationary coordinate system to obtain grid-connected parameters, then generates control instructions based on the compensation data, and then the inverter adjusts the grid-connected parameters and generates output current based on these instructions to accurately offset the harmonic components in the power grid, thereby improving the quality of power.

[0040] Example 2, reference Figure 3 The difference between this embodiment and embodiment 1 is that after executing S3 first modeling and before executing S4 generating instructions, it also includes: S6 voltage control, including S61 coordinate transformation, S62 second calculation, S63 first prediction, S64 third calculation, S65 difference judgment, S66 fourth calculation, S67 inverse Park transformation, S7 power verification, S68 voltage phase calculation and S69 output control.

[0041] S61 coordinate transformation, using Clarke transformation algorithm to transform the mathematical model in the three-phase stationary coordinate system into the mathematical model in the two-phase stationary coordinate system, and then using Park transformation algorithm to transform the mathematical model in the two-phase stationary coordinate system into the mathematical model in the two-phase rotating coordinate system. The mathematical model in the two-phase rotating coordinate system is as follows: ; Among them, L is the equivalent inductance; R is the equivalent resistance; C is the equivalent capacitance; is the grid angular frequency; is the d-axis component of the grid voltage; is the q-axis component of the grid voltage; is the d-axis component of the grid-side current; is the q-axis component of the grid-side current; is the DC side output current.

[0042] is the switching function of the bridge arm on the k-axis, and the calculation model is: .

[0043] S62 is a second calculation, which is to calculate the current change rate of the power grid under the action of zero voltage, recorded as the first current change rate. The calculation model of the first current change rate is as follows: ; ; in, It represents the rate of change of current under the action of zero voltage on the d-axis; Indicates the rate of change of current under the action of zero voltage on the q-axis The current change rate of the power grid under the action of non-zero voltage is calculated and recorded as the second current change rate. The calculation model of the second current change rate is as follows: ; ; ; ; in, Indicates that at non-zero voltage The rate of change of current under action; Indicates that at non-zero voltage The rate of change of current under action; Indicates that at non-zero voltage The rate of change of current under action; Indicates that at non-zero voltage The rate of change of current under action.

[0044] S63 first prediction, based on the first current change rate and the second current change rate, calculate the predicted data of the grid current, recorded as the first data, and the calculation model of the first data is as follows: ; ; in, is the duration of zero voltage action; , is the corresponding non-zero voltage ( )、 ( ) Duration of effect; is the d-axis reference current data; is the q-axis reference current data. For non-zero voltage Current data at the mth iteration under the action of ; For non-zero voltage Current data at the m+1th iteration under the action of ; For non-zero voltage Current data at the mth iteration under the action of ; For non-zero voltage Current data at the m+1th iteration under the action of .

[0045] Afterwards and To perform the inverse Park transform, the process is as follows: ; in, is the rotor position angle.

[0046] Then the two-phase stationary coordinate system is converted to the three-phase stationary coordinate system. The process is as follows: .

[0047] S64 is a third calculation, which is to calculate the difference between the first data and the real current data, and record it as the second data.

[0048] S65 difference judgment, judge whether the difference between the second data and the compensation data is less than the preset difference threshold. If so, it means that the predicted data is close enough to the real data, and then execute S66 fourth calculation; if not, it means that there is an error in the calculation process, or the harmonic components in the power grid are not well suppressed (there is a problem with the compensation data or new harmonic components have appeared), and it is necessary to re-execute S62 second calculation to verify whether the phenomenon is caused by calculation error.

[0049] After re-executing the second calculation in S62, if it is found that the difference between the second data and the compensation data is still greater than the preset difference threshold value when executing this step again, it means that new harmonics appear in the power grid, and the first acquisition in S1 can be executed.

[0050] S66 fourth calculation, calculating the predicted data of the grid voltage based on the first data, the calculation process is as follows: ; ; ; Where T is the control period of the controller; is the duration of zero voltage action; , is the corresponding non-zero voltage ( )、 ( ) Duration of effect; is the d-axis component of the grid voltage; is the q-axis component of the grid voltage.

[0051] S67 Inverse Park transformation, the predicted data of the grid voltage is subjected to inverse Park transformation to obtain the voltage data in the two-phase stationary coordinate system, which is recorded as the third data. The calculation process is as follows: ; in, for Voltage component on the axis; for Voltage component on the axis; is the rotor position angle.

[0052] This step converts the voltage data in the two-phase rotating coordinate system back to the two-phase stationary coordinate system so as to compare and control it with the actual output of the inverter.

[0053] S7 power verification, including S71 second acquisition, S72 fifth calculation, S73 sixth calculation and S74 power judgment.

[0054] S71 collects the active power in the power grid, which is recorded as the first power. The active power is the power that actually does work in the power grid and is the basis for the stable operation of the power grid. The reactive power in the power grid is collected, which is recorded as the second power. Although the reactive power does not do work in the power grid, it is crucial to maintain the voltage stability and current balance of the power grid.

[0055] S72 fifth calculation, calculating active power based on the first data and the third data, recorded as the third power; calculating reactive power based on the first data and the third data, recorded as the fourth power. This step calculates the predicted value of active power and the predicted value of reactive power based on the predicted value of current data and the predicted value of voltage data.

[0056] S73 Sixth calculation, calculating the difference between the first power and the third power, recorded as fourth data; calculating the difference between the second power and the fourth power, recorded as fifth data. This step calculates the difference between the real-time collected active power / reactive power and the predicted active power / reactive power.

[0057] S74 power judgment, judge whether the fourth data and the fifth data meet expectations. If so, it means that the fourth data and the fifth data are caused by measurement errors or calculation errors, and then execute S68 to calculate the voltage phase; if not, it means that there is a large gap between the actual active power / reactive power and the predicted active power / reactive power, which may be due to calculation errors or new harmonic components appearing in the power grid. It is necessary to execute the second calculation of S62 to verify whether the phenomenon is caused by calculation errors.

[0058] After re-executing the second calculation of S62, if it is found that the fourth data and the fifth data still do not meet expectations when executing this step again, it means that new harmonics have appeared in the power grid, and the first acquisition of S1 can be executed.

[0059] S68 calculates the voltage phase. The voltage phase is calculated based on the third data. The calculation model is as follows: ; in, is the voltage phase.

[0060] S69 output control, adding the voltage phase to the original compensation data as new compensation data. This step adds the voltage phase to the original compensation data in order to consider the change of the voltage phase in the compensation data, thereby more accurately controlling the output of the inverter and achieving accurate harmonic compensation.

[0061] Based on the technical solution provided in Example 1, this embodiment adds S6 voltage control and S7 power verification. Compared with the current control used alone in Example 1, it can accurately control the output of the inverter and realize effective compensation of grid harmonics.

[0062] Reference Figure 4 In other embodiments, when in S74 power judgment, one of the fourth data or the fifth data does not meet expectations, or both do not meet expectations, that is, an error occurs in the calculation and the predicted data of the current and the predicted data of the voltage are recalculated, after executing S62 second calculation until executing S68 voltage phase calculation, before executing S69 output control, it also includes: S681 first filtering, inputting the fourth data into a low-pass filter, the function of the low-pass filter is to remove the high-frequency components in the signal and retain the low-frequency components, thereby smoothing the data and reducing noise interference. After filtering, a new fourth data is obtained, which is more stable and can better reflect the actual state of the power grid.

[0063] S682 is the seventh calculation, which multiplies the new fourth data by the preset adjustment coefficient to obtain the active frequency adjustment amount, and converts the active frequency adjustment amount into a voltage phase adjustment amount, which is recorded as the sixth data. The sixth data represents the voltage phase amount that needs to be adjusted to compensate for the active power deviation.

[0064] The process of converting the active frequency adjustment amount into the voltage phase adjustment amount is as follows: 1. According to the active frequency Calculate cycle time : .

[0065] 2. Adjust the amount according to the active frequency Get the time change , calculate the time variation In cycle time The proportion of : .

[0066] 3. The proportion Multiply Get voltage phase adjustment

[0067] .

[0068] The regulation coefficient is set according to the operation characteristics and control objectives of the power grid, and is used to adjust the regulation strength of the active power. Its value range is 0-10. The closer to 0, the smaller the regulation strength, and vice versa.

[0069] S683: second filtering, inputting the fifth data into a low-pass filter to obtain new fifth data.

[0070] The eighth calculation of S684 is to multiply the new fifth data by the preset adjustment coefficient to obtain the reactive power adjustment amount, and convert the reactive frequency adjustment amount into the voltage phase adjustment amount (the same as the steps of the seventh calculation of S682), recorded as the seventh data, which represents the voltage phase amount that needs to be adjusted to compensate for the reactive power deviation.

[0071] S685: second summation, the sixth data is summed with the seventh data to obtain new sixth data. The new sixth data combines the influence of active and reactive power regulation on the voltage phase and is the ultimate basis for subsequent voltage phase adjustment.

[0072] S686 first sums the voltage phase and the sixth data to obtain a new voltage phase.

[0073] This embodiment focuses on the precise adjustment of the voltage phase. First, the deviation data of active power and reactive power are smoothed by a low-pass filter to reduce noise interference. Then, the active frequency adjustment amount and reactive power adjustment amount are calculated using the preset adjustment coefficient and converted into the corresponding voltage phase adjustment amount. Then, the two voltage phase adjustment amounts are summed to obtain the comprehensive voltage phase adjustment amount. Finally, this adjustment amount is added to the original voltage phase to obtain the optimized voltage phase, which aims to achieve fine control of power balance and stability of the power grid and improve the overall performance and power supply quality of the power grid.

[0074] Example 3: Reference Figure 5 The difference between this embodiment and embodiment 1 is that after executing S5 to generate a signal, it also includes: S81 is the third collection, collecting the current data in the power grid after the output current is injected, recorded as the eighth data, and the eighth data reflects the state of the power grid after the compensation current is added.

[0075] S82 harmonic judgment, perform harmonic analysis on the eighth data to determine whether the eighth data contains even harmonics. If so, it indicates that there are wiring errors or poor contact at the wiring points in the power grid, and it is necessary to perform S83 inspection; if not, it indicates that the power grid state is relatively good, and execute S84 second modeling.

[0076] S83 checks whether there are any wiring errors or poor contact, and adjusts the parts with wiring errors or poor contact.

[0077] S84 second modeling, if the eighth data does not contain even harmonics, or after the wiring problem is checked and corrected, a neural network model is established.

[0078] S85 second prediction, input the eighth data and the compensation data into the neural network model, the neural network model will analyze and predict based on these input data (the eighth data and the compensation data), and then output a current adjustment amount.

[0079] This current regulation amount represents the amount of adjustment that needs to be made to the compensation data in order to make the power grid reach or approach the ideal state.

[0080] S86 online learning, in order to improve the prediction accuracy and adaptability of the neural network model, the weight coefficient of the neural network model is adjusted using the online learning algorithm, specifically: the weight coefficient of the neural network model is adjusted by reducing the current regulation amount until the current regulation amount is less than the preset regulation amount threshold.

[0081] S87: adjusting, adjusting the compensation data based on the current adjustment amount, and using the adjusted compensation data as new compensation data.

[0082] S88 determines the adjustment amount, and determines whether the current adjustment amount is less than a preset threshold. If so, it means that the adjustment is effective, and the weight coefficient of the neural network model is output as a parameter for subsequent control; if not, it means that the adjustment is not sufficient, and it is necessary to execute the second prediction of S84.

[0083] S89 parameter replacement, taking the weight coefficient of the neural network model as the parameter of the controller, the controller can accurately adjust and control according to the actual situation of the power grid. By taking the weight coefficient of the neural network model as the parameter of the controller, it can adapt to the dynamic changes of the power grid and further improve the harmonic compensation effect and stability of the power grid.

[0084] In this embodiment, the controller includes three parameters: proportional gain Kp, integral time Ti and differential time Td, so the neural network model uses three neurons to correspond to the three parameters of the controller.

[0085] In other embodiments, if the controller includes multiple parameters, the neural network model should use the same number of neurons as the controller parameters to achieve a correspondence in quantity.

[0086] This embodiment first collects the grid current data after the compensation current is injected, and determines whether it contains even harmonics. If it does, the wiring is checked, otherwise a neural network model is established, and then the collected current data and compensation data are input into the neural network model for prediction, and the current regulation amount is output, and the model weight coefficient is adjusted through the online learning algorithm to reduce the regulation amount to below the preset threshold. Based on the obtained current regulation amount, the compensation data is adjusted, and it is determined whether the adjusted current regulation amount meets the requirements. If it does, the weight coefficient of the neural network model is applied as a parameter of the quasi-PR controller to achieve accurate compensation and regulation of grid harmonics.

[0087] Embodiment 4: This embodiment discloses a controller, characterized in that the controller includes a processor and a memory, the memory stores machine executable instructions that can be executed by the processor, and the processor executes the machine executable instructions to implement the described method.

[0088] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A grid-connected control method for reducing harmonic noise, characterized in that: include: First collection: collecting the harmonic components of the current in the power grid, obtaining the amplitude information and phase information of each harmonic component, and recording it as the first information; First calculation: According to the first information, the compensation data is calculated, and the calculation model of the compensation data is as follows: ； in, To compensate for the data; is the amplitude information of the nth harmonic component; n is the harmonic order of the harmonic component; is the fundamental angular frequency of the grid-connected current; t is the time; is the phase information of the nth harmonic component; First modeling: Establish a mathematical model of the inverter in a three-phase stationary coordinate system to obtain the grid-connected parameter information of the inverter; Generate instructions: Generate control instructions based on compensation data; Generate signal: The inverter adjusts the grid-connected parameter information according to the control instruction, generates output current, and injects the output current into the grid.

2. The grid-connected control method for reducing harmonic noise according to claim 1, characterized in that: After executing the first modeling step and before executing the step of generating instructions, the method further includes: Coordinate transformation: Clarke transformation algorithm and Park transformation algorithm are used to transform the mathematical model in the three-phase stationary coordinate system into the mathematical model in the two-phase rotating coordinate system; Second calculation: Calculate the current change rate of the power grid under the action of zero voltage, recorded as the first current change rate; calculate the current change rate of the power grid under the action of non-zero voltage, recorded as the second current change rate; First prediction: Calculate the prediction data of the grid current based on the first current change rate and the second current change rate, recorded as the first data; The third calculation: calculating the difference between the first data and the real current data, and recording it as the second data; Difference judgment: judge whether the difference between the second data and the compensation data is less than the preset difference threshold, if so, execute the step of generating instructions; if not, execute the step of second calculation.

3. The grid-connected control method for reducing harmonic noise according to claim 2, characterized in that: After executing the step of determining the difference and before executing the step of generating the instruction, the method further includes: Fourth calculation: calculating predicted data of the grid voltage based on the first data; Inverse Park transformation: Inverse Park transformation is performed on the predicted data of the grid voltage to obtain the voltage data in the two-phase stationary coordinate system, which is recorded as the third data; Calculating a voltage phase: calculating a voltage phase based on the third data; Output control: adding the voltage phase to the original compensation data as new compensation data.

4. The grid-connected control method for reducing harmonic noise according to claim 3, characterized in that: After the step of performing the inverse Park transformation and before the step of calculating the voltage phase, the method further includes: Second collection: collect the active power in the power grid, recorded as the first power; collect the reactive power in the power grid, recorded as the second power; Fifth calculation: calculating active power based on the first data and the third data, recorded as the third power; calculating reactive power based on the first data and the third data, recorded as the fourth power; Sixth calculation: calculating the difference between the first power and the third power, and recording it as the fourth data; calculating the difference between the second power and the fourth power, and recording it as the fifth data; Power judgment: judge whether the fourth data and the fifth data meet expectations, if so, execute the step of calculating the voltage phase; if not, execute the second calculation step.

5. The grid-connected control method for reducing harmonic noise according to claim 4, characterized in that: After executing the step of calculating the voltage phase and before executing the step of output control, the method further includes: First filtering: inputting the fourth data into a low-pass filter to obtain new fourth data; Seventh calculation: performing a product operation on the new fourth data and the preset adjustment coefficient to obtain an active frequency adjustment amount, converting the active frequency adjustment amount into a voltage phase adjustment amount, which is recorded as the sixth data; First summation: sum the voltage phase with the sixth data to obtain a new voltage phase.

6. The grid-connected control method for reducing harmonic noise according to claim 5, characterized in that: After performing the seventh calculation step and before performing the first summation step, the method further includes: Second filtering: inputting the fifth data into a low-pass filter to obtain new fifth data; Eighth calculation: multiplying the new fifth data by the preset adjustment coefficient to obtain a reactive power adjustment amount, converting the reactive frequency adjustment amount into a voltage phase adjustment amount, which is recorded as the seventh data; Second summation: sum the sixth data and the seventh data to obtain new sixth data.

7. The grid-connected control method for reducing harmonic noise according to any one of claims 1 to 6, characterized in that: After executing the step of generating a signal, the step further includes: The third collection: collecting the current data in the power grid after the output current is injected, which is recorded as the eighth data; Second modeling: establishing a neural network model; Second prediction: input the eighth data and the compensation data into the neural network model, and output the current regulation amount; Adjustment: adjusting the compensation data based on the current adjustment amount, and using the adjusted compensation data as new compensation data; Determine the adjustment amount: determine whether the current adjustment amount is less than a preset threshold value, if so, output the weight coefficient of the neural network model; if not, execute the second prediction step; Parameter replacement: The weight coefficients of the neural network model are used as parameters of the controller.

8. The grid-connected control method for reducing harmonic noise according to claim 7, characterized in that: After the step of performing the second prediction and before the step of performing the adjustment, the method further includes: Online learning: The neural network model uses an online learning algorithm to adjust the internal weight coefficients by reducing the current regulation amount.

9. The grid-connected control method for reducing harmonic noise according to claim 8, characterized in that: After performing the third acquisition step and before performing the second modeling step, the method further includes: Harmonic determination: determining whether the eighth data contains even harmonics, if so, executing the checking step; if not, executing the second modeling step; Check: Check the wiring.

10. A controller, characterized in that: The controller includes a processor and a memory, wherein the memory stores machine executable instructions that can be executed by the processor, and the processor executes the machine executable instructions to implement the method according to any one of claims 1 to 9.

Images