Power control method, model training method and system of inverter

Abstract

Embodiments of the present application provide a power control method, a model training method and a system of an inverter. The method is applied to a controller of an anti-backflow system, the anti-backflow system comprising a set of inverters. The method comprises: obtaining a first output total power sequence of the anti-backflow system; wherein the first output total power sequence comprises an output total power of the anti-backflow system at each time in a preset time period; inputting the first output total power sequence into a preset intelligent prediction model for processing to obtain a proportional parameter value, a differential parameter value and an integral parameter value; and controlling power of the inverters in the set of inverters according to the first output total power sequence, the proportional parameter value, the differential parameter value and the integral parameter value. The method is used to accurately control the output power of the inverter, thereby efficiently and accurately preventing the anti-backflow of the distributed photovoltaic power generation system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of inverters, in particular to a power control method, a model training method and a system of an inverter. BACKGROUND

[0002] In a distributed photovoltaic power generation system, when the direct current power output by a photovoltaic module is converted into alternating current by an inverter, if the real-time power consumption of a local load is lower than the system power generation, the excess power will be fed back to the power grid through the user-side power distribution network, forming a reverse current phenomenon. In order to prevent reverse current from occurring, the output power of the inverter needs to be precisely regulated.

[0003] In related technologies, the output power of each inverter is corrected by evenly dividing the reverse current power value. However, due to the differences in operating parameters and environmental conditions of each inverter, the uniform distribution strategy may cause some inverters to be over-adjusted and have power over-regulation, while other inverters still have reverse current risk due to insufficient adjustment, making it difficult for the overall system to effectively suppress reverse current.

[0004] Therefore, there is an urgent need for a method that can accurately control the output power of an inverter to efficiently and accurately prevent the occurrence of reverse current in a distributed photovoltaic power generation system. SUMMARY

[0005] The embodiments of the present application provide a power control method, a model training method and a system of an inverter to achieve the effect of efficiently and accurately preventing reverse current in a distributed photovoltaic power generation system.

[0006] In a first aspect, the embodiments of the present application provide a power control method of an inverter. The method is applied to a controller of an anti-reverse current system, the anti-reverse current system includes a set of inverters, and the method includes:

[0007] obtaining a first total output power sequence of the anti-reverse current system; wherein the first total output power sequence includes the output total power of the anti-reverse current system at each time in a preset time period;

[0008] inputting the first total output power sequence into a preset intelligent prediction model for processing to obtain a proportional parameter value, a differential parameter value, and an integral parameter value; wherein the proportional parameter value is used to correct the error of the output total power of the anti-reverse current system, the differential parameter value is used to correct the cumulative error of the output total power of the first total output power sequence in the preset time period, and the integral parameter value is used to predict the error of the output total power of the anti-reverse current system;

[0009] controlling the power of the inverters in the set of inverters according to the first total output power sequence, the proportional parameter value, the differential parameter value, and the integral parameter value.

[0010] In a possible implementation, the first output total power sequence is input to a preset intelligent prediction model for processing to obtain the proportional parameter value, the differential parameter value, and the integral parameter value, including:

[0011] According to the first output total power sequence and the preset power threshold, a power difference sequence is determined; the power difference sequence includes a power difference value of the anti-flow system at each moment in a preset time period; the power difference value represents a power value that needs to be corrected by the inverter;

[0012] The power difference sequence is processed based on the preset intelligent prediction model to obtain the proportional parameter value, the differential parameter value, and the integral parameter value.

[0013] In a possible implementation, the power difference sequence is processed based on the preset intelligent prediction model to obtain the proportional parameter value, the differential parameter value, and the integral parameter value, including:

[0014] The power difference sequence is processed based on the preset intelligent prediction model to obtain a first power difference vector; the first power difference vector represents a change trend of the importance of the power difference value in the power difference sequence;

[0015] The power difference sequence and the first power difference vector are processed based on the preset intelligent prediction model to obtain a second power difference vector; the second power difference vector represents a feature of the power difference value with an importance greater than a preset threshold in the power difference sequence.

[0016] The second power difference vector is processed based on the preset intelligent prediction model to obtain the proportional parameter value, the differential parameter value, and the integral parameter value.

[0017] In a possible implementation, the power difference sequence is processed based on the preset intelligent prediction model to obtain the first power difference vector, including:

[0018] The power difference sequence is processed based on the preset intelligent prediction model to obtain a third power difference vector; the third power difference vector represents the importance of the power difference value in the power difference sequence.

[0019] The power difference sequence and the third power difference vector are processed based on the preset intelligent prediction model to obtain a fourth power difference vector; the fourth power difference vector represents the power difference value with an importance greater than a preset threshold in the power difference sequence.

[0020] The power difference sequence, the third power difference vector, and the fourth power difference vector are processed based on the preset intelligent prediction model to obtain the first power difference vector.

[0021] In a possible implementation, the power difference sequence and the first power difference vector are processed according to a preset intelligent prediction model to obtain a second power difference vector, including:

[0022] The power difference sequence and the first power difference vector are processed according to a preset intelligent prediction model to obtain an attention weight; the attention weight represents an importance of the first power difference vector.

[0023] The first power difference vector and the attention weight are processed by weighted summation according to the preset intelligent prediction model to obtain the second power difference vector.

[0024] In a possible implementation, the power of the inverters in the inverter set is controlled according to the first output total power sequence, the proportional parameter value, the differential parameter value, and the integral parameter value, including:

[0025] The power correction parameter is determined according to the first output total power sequence, the proportional parameter value, the differential parameter value, and the integral parameter value; the power correction parameter represents a percentage of power that needs to be corrected by the inverters in the inverter set.

[0026] The current output power of the inverters in the inverter set is obtained; and the power value of the inverters in the inverter set is controlled according to the current output power and the power correction parameter.

[0027] In a possible implementation, the anti-flow system further includes an auxiliary power supply set; and the first output total power sequence of the anti-flow system is obtained, including:

[0028] The second output total power sequence of the inverter set and the third output total power sequence of the auxiliary power supply set are obtained; the second output total power sequence includes the output total power of the inverter set at each time in a preset time period, and the third output total power sequence includes the output total power of the auxiliary power supply set at each time in the preset time period.

[0029] The first output total power sequence is determined according to the second output total power sequence and the third output total power sequence.

[0030] In a second aspect, the embodiments of the present application provide a training method of a preset intelligent prediction model of a power control method of an inverter, the power control method of the inverter being the method in the first aspect and / or the various possible implementation manners of the first aspect; the training method of the preset intelligent prediction model is applied to a controller of an anti-flow system, the anti-flow system including an auxiliary power supply set, and the training method of the preset intelligent prediction model includes:

[0031] obtain an output total power sequence set of the anti-flow system, an access number of an auxiliary power supply in the auxiliary power supply set, and a load type of the anti-flow system; the output total power sequence set includes at least one output total power sequence, and the output total power sequence includes an output total power of the anti-flow system at each time point in a preset time period;

[0032] splice the output total power sequence in the output total power sequence set with the corresponding access number of the auxiliary power supply and the load type to obtain a to-be-processed data set;

[0033] input the to-be-processed data set into an initial model for training to obtain a preset intelligent prediction model.

[0034] In a third aspect, an embodiment of the present application provides a power control device of an inverter, the device being applied to a controller of an anti-flow system, the anti-flow system including a set of inverters, and the device including:

[0035] an obtaining module configured to obtain a first output total power sequence of the anti-flow system; the first output total power sequence includes an output total power of the anti-flow system at each time point in a preset time period;

[0036] a processing module configured to input the first output total power sequence into a preset intelligent prediction model for processing to obtain a proportional parameter value, a differential parameter value, and an integral parameter value; the proportional parameter value is used to correct an error of the output total power of the anti-flow system, the differential parameter value is used to correct an accumulated error of the output total power of the first output total power sequence in the preset time period, and the integral parameter value is used to predict an error of the output total power of the anti-flow system;

[0037] a control module configured to control power of the inverter in the set of inverters according to the first output total power sequence, the proportional parameter value, the differential parameter value, and the integral parameter value.

[0038] In a possible implementation, the processing module includes:

[0039] a first processing module configured to determine a power difference sequence according to the first output total power sequence and a preset power threshold; the power difference sequence includes a power difference value of the anti-flow system at each time point in the preset time period; the power difference value represents a power value that needs to be corrected by the inverter;

[0040] a second processing module configured to process the power difference sequence based on the preset intelligent prediction model to obtain the proportional parameter value, the differential parameter value, and the integral parameter value.

[0041] In a possible implementation, the second processing module includes:

[0042] The third processing module is configured to process the power difference sequence based on a preset intelligent prediction model to obtain a first power difference vector; the first power difference vector represents a trend of importance of the power difference values in the power difference sequence;

[0043] The fourth processing module is configured to process the power difference sequence and the first power difference vector based on the preset intelligent prediction model to obtain a second power difference vector; the second power difference vector represents characteristics of the power difference values with importance greater than a preset threshold in the power difference sequence.

[0044] The fifth processing module is configured to process the second power difference vector based on the preset intelligent prediction model to obtain a proportional parameter value, a differential parameter value, and an integral parameter value.

[0045] In a possible implementation, the third processing module comprises:

[0046] The third processing module is configured to process the power difference sequence based on a preset intelligent prediction model to obtain a first power difference vector; the first power difference vector represents a trend of importance of the power difference values in the power difference sequence;

[0047] The third processing module is configured to process the power difference sequence based on a preset intelligent prediction model to obtain a first power difference vector; the first power difference vector represents a trend of importance of the power difference values in the power difference sequence;

[0048] The third processing module is configured to process the power difference sequence based on a preset intelligent prediction model to obtain a first power difference vector; the first power difference vector represents a trend of importance of the power difference values in the power difference sequence;

[0049] In a possible implementation, the fourth processing module comprises:

[0050] The fourth processing module is configured to process the power difference sequence and the first power difference vector based on the preset intelligent prediction model to obtain a second power difference vector; the second power difference vector represents characteristics of the power difference values with importance greater than a preset threshold in the power difference sequence.

[0051] The fourth processing module is configured to process the power difference sequence and the first power difference vector based on the preset intelligent prediction model to obtain a second power difference vector; the second power difference vector represents characteristics of the power difference values with importance greater than a preset threshold in the power difference sequence.

[0052] In a possible implementation, the control module comprises:

[0053] The control module is configured to determine a power correction parameter according to the first output total power sequence, the proportional parameter value, the differential parameter value, and the integral parameter value; the power correction parameter represents a percentage of power that needs to be corrected by the inverters in the inverter set.

[0054] obtaining a current output power of the inverter in the inverter set; and controlling a power value of the inverter in the inverter set according to the current output power and a power correction parameter.

[0055] In a possible implementation, the anti-inrush system further includes a set of auxiliary power supplies; and the obtaining module includes:

[0056] obtaining a second output total power sequence of the inverter set and a third output total power sequence of the set of auxiliary power supplies; wherein the second output total power sequence includes an output total power of the inverter set at each time point in a preset time period, and the third output total power sequence includes an output total power of the set of auxiliary power supplies at each time point in the preset time period;

[0057] determining the first output total power sequence according to the second output total power sequence and the third output total power sequence.

[0058] In a fourth aspect, an embodiment of the present application provides a device for training a preset intelligent prediction model for power control of an inverter, the device being applied to a controller of an anti-inrush system, the anti-inrush system including a set of auxiliary power supplies, and the device including:

[0059] an obtaining module, configured to obtain a set of output total power sequences of the anti-inrush system, an access number of an auxiliary power supply in the set of auxiliary power supplies, and a load type of the anti-inrush system; wherein the set of output total power sequences includes at least one output total power sequence, and the output total power sequence includes an output total power of the anti-inrush system at each time point in a preset time period;

[0060] a splicing module, configured to splice an output total power sequence in the set of output total power sequences with the access number of the auxiliary power supply corresponding to the output total power sequence and the load type, to obtain a to-be-processed data set;

[0061] a training module, configured to input the to-be-processed data set into an initial model to perform training, to obtain a preset intelligent prediction model;

[0062] The preset intelligent prediction model is the preset intelligent prediction model in the first aspect and / or various possible implementation manners of the first aspect.

[0063] In a fifth aspect, an embodiment of the present application provides an anti-inrush system, the anti-inrush system including a set of inverters, a set of auxiliary power supplies, a smart meter, a data acquisition device, and a controller; wherein the set of inverters includes at least one inverter, and the set of auxiliary power supplies includes at least one auxiliary power supply.

[0064] One end of the auxiliary power supply is connected to one end of the inverter through a load, and the other end of the inverter is connected to a power grid system through the smart meter.

[0065] The other end of the auxiliary power supply is connected with the other end of the inverter through a data acquisition device, and the other end of the data acquisition device is connected with the smart meter; the controller is connected with the data acquisition device;

[0066] The controller is configured to perform the first aspect and / or the various possible implementation manners of the first aspect.

[0067] In a sixth aspect, an embodiment of the present application provides an electronic device, comprising a memory and a processor.

[0068] The memory stores computer execution instructions.

[0069] The processor executes the computer execution instructions stored in the memory, so that the processor performs the first aspect and / or the various possible implementation manners of the first aspect.

[0070] In a seventh aspect, an embodiment of the present application provides a computer readable storage medium, which stores computer execution instructions, and the computer execution instructions are executed by a processor to implement the first aspect and / or the various possible implementation manners of the first aspect.

[0071] In an eighth aspect, an embodiment of the present application provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the first aspect and / or the various possible implementation manners of the first aspect.

[0072] The power control method, model training method and system of the inverter provided by the embodiments of the present application realize precise power management of the anti-backflow system by constructing a dynamic PID parameter regulation mechanism based on time series data driving. Specifically, the method first acquires a first output total power sequence of the anti-backflow system within a preset time period, and the sequence is essentially a power fluctuation feature set containing time dimension information. After inputting the sequence into a preset intelligent prediction model, the model dynamically generates three groups of parameter values of proportion, differential and integral through deep mining of power time series features by using a long short-term memory network. According to the first output total power sequence and the PID parameter values, the power of the inverters in the inverter set is precisely controlled; precise and dynamic regulation of the power of the inverters is realized, effectively solving the problems of inaccurate power control and untimely response in the traditional method, and significantly improving the working efficiency and stability of the inverters. BRIEF DESCRIPTION OF DRAWINGS

[0073] The accompanying drawings, which are incorporated into and form a part of the specification, illustrate an embodiment consistent with the present application and, together with the description, serve to explain the principles of the application.

[0074] Figure 1 Flowchart of the power control method of the inverter provided by the embodiments of the present application Figure 1 ;

[0075] Figure 2 Flowchart of a power control method of an inverter provided for an embodiment of the present application Figure 2 ;

[0076] Figure 3 Flowchart of step S202 in a power control method of an inverter provided for an embodiment of the present application

[0077] Figure 4 Flowchart of a training method of a preset intelligent prediction model for power control of an inverter provided for an embodiment of the present application

[0078] Figure 5 Structural diagram of a power control device of an inverter provided for an embodiment of the present application

[0079] Figure 6 Structural diagram of a training device of a preset intelligent prediction model for power control of an inverter provided for an embodiment of the present application

[0080] Figure 7 Structural diagram of an anti-reverse flow system provided for an embodiment of the present application

[0081] Figure 8 Structural diagram of an electronic device provided for an embodiment of the present application

[0082] Legend of reference numerals: 70-anti-reverse flow system, 701-inverter set, 702- auxiliary power supply set, 703-smart meter, 704-data acquisition device, 705-controller, 706- load, 707-power grid

[0083] The specific embodiments of the present application have been shown through the above-described drawings, and will be described in more detail hereinafter. These drawings and written descriptions are not intended to limit the scope of the concept of the present application by any means, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0084] The exemplary embodiments will be described in detail herein with reference to the attached drawings. The following description is made with reference to the accompanying drawings in which like reference numerals represent like elements, unless the context of use indicates otherwise. The following exemplary embodiments described are not representative of all embodiments consistent with the present application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the present application as detailed in the appended claims.

[0085] In a distributed photovoltaic power generation system, photovoltaic components convert solar energy into direct current power, which is converted into alternating current power by an inverter and then supplied to local loads. When there is a dynamic imbalance between the power generation of the system and the power consumption of the local loads, if the photovoltaic power is continuously higher than the load demand, the excess power will be transmitted to the power grid through the user-side power distribution network, forming a reverse flow phenomenon. This phenomenon not only may cause power quality problems such as power grid voltage fluctuation and frequency deviation, but also may cause the system to be forced to operate at a reduced capacity due to the limitation of the power grid access specification, significantly reducing the utilization rate and economy of photovoltaic power generation.

[0086] In the prior art, to suppress reverse flow, a control strategy of equally dividing the reverse flow power value to each inverter is generally used, that is, by detecting the total reverse flow power of the system, the adjustment amount is proportionally distributed according to the number or rated power of the inverters. However, this strategy has significant limitations: due to the differences in operating parameters (such as maximum power point accuracy, conversion efficiency curve) and environmental conditions (such as light intensity, component temperature, local shading) of each inverter in the distributed photovoltaic system, the uniform distribution of the adjustment amount cannot adapt to the individual characteristics of the inverters. For example, some inverters may have power output lower than the actual adjustable range due to excessive load reduction when receiving the same adjustment amount because of higher efficiency or superior environmental conditions, resulting in power over-regulation; while other inverters may still have the probability of continuous reverse feedback due to insufficient adjustment to offset the reverse flow because of lower efficiency or environmental disadvantages. This uneven adjustment problem makes it difficult for the system as a whole to achieve accurate suppression of reverse flow, and may even cause secondary problems such as increased equipment wear and tear and shortened lifespan due to frequent over-regulation of some inverters.

[0087] Therefore, the present application provides a power control method, model training method and system for an inverter, which can solve the above problems.

[0088] The technical solutions of the present application and how the technical solutions of the present application solve the above technical problems will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of the present application will be described below with reference to the drawings.

[0089] Figure 1 A flowchart of a power control method for an inverter provided by an embodiment of the present application Figure 1 As shown in Figure 1 The method is applied to a controller of an anti-reverse flow system, the anti-reverse flow system including a set of inverters, and the method includes:

[0090] S101, obtaining a first output total power sequence of the anti-reverse flow system; wherein the first output total power sequence includes the output total power of the anti-reverse flow system at each time in a preset time period.

[0091] Exemplarily, the anti-inrush system is a power control device for distributed energy systems (such as photovoltaic power stations, energy storage systems), whose core function is to monitor the power flow between the grid and the load in real time, prevent the generated power of the inverter from flowing back into the public grid (i.e. "inrush"), and thus avoid affecting the stability of the grid. The system includes but is not limited to inverter sets, controllers, data acquisition devices, smart meters, etc.

[0092] Exemplarily, the inverter set refers to multiple inverter devices connected in parallel or series in the anti-inrush system. Each inverter is responsible for converting direct current (such as the power generated by photovoltaic panels) into alternating current and outputting it to the grid.

[0093] Exemplarily, the preset time period refers to a pre-set time range, such as 5 minutes, 1 hour, or 24 hours, for periodic monitoring and analysis of system output power. The length of the preset time period needs to be configured according to actual scene requirements (such as grid scheduling period, load fluctuation characteristics) to ensure the real-time and accuracy of power control.

[0094] Exemplarily, within the preset time period, the anti-inrush system collects and records the set of total output power of the anti-inrush system at fixed time intervals (such as every second, every minute). For example, if the preset time period is 10:00-10:10 and the sampling interval is 1 second, the sequence contains 600 data points, each corresponding to a power value at a time. This sequence is used for subsequent analysis of the power fluctuation trend and inrush risk of the system.

[0095] Exemplarily, in the power control method of the anti-inrush system, first, the access state of the auxiliary power supply is obtained through the data acquisition device (such as contactor state detection, current sensor), and the number of auxiliary power supplies accessed is counted. For example, if the system is configured with 3 diesel generators, and 2 are detected to be in operation, the access number is 2. Then, according to the rated power of each auxiliary power supply (such as the rated power of a single diesel generator is 100kW), the total output power of all auxiliary power supplies is calculated by accumulation (such as 2x100kW=200kW). At the same time, the output power of all inverters in the anti-inrush system is collected in real time through the smart meter, and the power values of each inverter are added to obtain the total output power of the inverter set (such as 10 inverters each output 50kW, so the total power is 500kW). Finally, the total output power of the inverter is added to the total output power of the auxiliary power supply to obtain the total output power of the anti-inrush system. To generate the first total output power sequence, the system repeats the above power acquisition and calculation process at fixed time intervals within the preset time period, and stores the results in chronological order. For example, if the preset time period is 1 hour and the sampling interval is 1 second, the sequence contains 3600 data points, each recording the system's total output power at that time, thus providing data support for subsequent inrush risk analysis and power regulation.

[0096] Auxiliary power refers to the backup power (such as diesel generators, energy storage batteries, etc.) connected with the inverter or load in the anti-flow system, which is used to supplement power supply when the local load demand exceeds the output capacity of the inverter. The access state of the auxiliary power can be monitored in real time by data acquisition equipment (such as contactor state detection, current sensor), and its rated power is the maximum output capacity marked on the equipment nameplate.

[0097] S102, input the first output total power sequence to a preset intelligent prediction model for processing to obtain a proportional parameter value, a differential parameter value, and an integral parameter value; wherein the proportional parameter value is used to correct the error of the output total power of the anti-flow system, the differential parameter value is used to correct the accumulated error of the output total power of the first output total power sequence within a preset time period, and the integral parameter value is used to predict the error of the output total power of the anti-flow system.

[0098] Exemplarily, the preset intelligent prediction model can be an algorithm model based on machine learning or deep learning, which is obtained by training historical data and can predict future power change trend according to the input power sequence and output proportional (P) parameter value, integral (I) parameter value and derivative (D) parameter value. The model can use long short-term memory (LSTM), convolutional neural network (CNN) or reinforcement learning architecture to capture nonlinear features in time series.

[0099] It can be understood that the proportional-integral-derivative (PID) parameter value can represent the P parameter value, the I parameter value and the D parameter value.

[0100] Exemplarily, the proportional parameter value, the differential parameter value and the integral parameter value are key parameters output by the preset intelligent prediction model, which correspond to different correction functions respectively. The P parameter value is used to adjust the power error at the current time, the I parameter value focuses on correcting the change rate of the accumulated power error, and the D parameter value is used for predictive adjustment of the future power error. The three parameters work together to ensure that the corrected power of the anti-flow system is stable and reliable.

[0101] Exemplarily, in the power control method of the anti-backflow system, first, the first output total power sequence is input into a preset intelligent prediction model (such as LSTM, temporal convolutional network (TCN) or reinforcement learning controller). The model infers the PID parameter values suitable for the current system state by analyzing the time sequence dependence and fluctuation mode in the sequence. These parameters will be passed to the downstream PID controller to adjust the inverter output or load distribution strategy of the anti-backflow system in real time, and finally realize dynamic correction of power error and system optimization.

[0102] In a possible implementation, a power difference sequence is determined according to the first output total power sequence and a preset power threshold; wherein the power difference sequence includes a power difference value of the anti-backflow system at each moment in a preset time period; the power difference value represents a power value that needs to be corrected by the inverter.

[0103] In a possible implementation, the preset intelligent prediction model LSTM is fused with an attention mechanism model, the input layer of which receives the normalized power difference sequence (such as scaling the power difference value to the range of 0-1), the hidden layer thereof captures the long-term dependence in the time sequence through a gating mechanism, and the output layer generates the PID parameters. In order to improve real-time performance, the model uses a sliding window technology, and only uses the data in the last 1 hour for inference (the window length can be adjusted according to the response speed of the system). In the training stage, a transfer learning strategy is adopted, the model is first pre-trained on a public photovoltaic data set, and then the parameters are fine-tuned through a small amount of field data. In the inference process, the model updates the parameters every 1 second, and avoids frequent adjustment through a dynamic threshold mechanism: only when the prediction error exceeds 5% does the parameter update trigger. The application of the PID parameters adopts a soft start strategy, and the parameter values take effect gradually within 10 seconds to prevent system shock caused by sudden changes. In addition, the model continuously optimizes through online learning, and automatically adjusts the network weights to adapt to the new mode if it is detected that the power sequence distribution deviates (such as seasonal light changes). Finally, the PID parameters and the control logic of the anti-backflow system work together to realize millisecond-level power regulation.

[0104] S103、According to the first output total power sequence, the proportional parameter value, the differential parameter value, and the integral parameter value, the power of the inverter in the inverter set is controlled.

[0105] Exemplarily, a power difference sequence is determined according to the first output total power sequence and a preset power threshold; wherein the power difference sequence includes a power difference value of the anti-backflow system at each moment in a preset time period; the power difference value represents a power value that needs to be corrected by the inverter.

[0106] Exemplarily, the output total power and the PID parameter value generated by the preset intelligent prediction model are calculated by a PID control algorithm to obtain the power adjustment amount of the inverter set. The specific process is as follows: the controller obtains the instantaneous adjustment amount by multiplying the error between the current output total power and the preset power threshold by the P parameter value; at the same time, the trend adjustment amount is obtained by multiplying the error trend predicted based on the change rate (derivative) of the power sequence by the I parameter value; in addition, the long-term deviation is eliminated by accumulating the historical error through the D parameter value. After the total adjustment amount is obtained by adding the three amounts, the controller sends the adjustment instruction to each inverter through broadcasting to dynamically adjust the output power thereof. For example, if the current power is higher than the preset power threshold and shows an upward trend, the P parameter value will reduce the output, the D parameter value will suppress the upward trend in advance, and the I parameter value will correct the possible accumulated deviation, so that the system output is finally stabilized in the target range.

[0107] The power control method of the inverter provided in the embodiments of the present application realizes accurate power management of the anti-backflow system by constructing a dynamic PID parameter regulation mechanism based on time series data driving. Specifically, the method first obtains a first output total power sequence of the anti-backflow system in a preset time period, which is essentially a power fluctuation feature set containing time dimension information. After the sequence is input into a preset intelligent prediction model, the model dynamically generates three groups of parameter values of proportion, derivative and integral by deeply mining the power time series features through a long short-term memory network. According to the first output total power sequence and the PID parameter values, the power of the inverters in the inverter set is accurately controlled, which realizes accurate and dynamic regulation of the inverter power, effectively solves the problems of inaccurate power control and untimely response in the traditional method, and significantly improves the working efficiency and stability of the inverter.

[0108] Figure 2 Flowchart of the power control method of the inverter provided in the embodiments of the present application Figure 2 As shown in Figure 2 the embodiments of the present application, the power control method of the inverter is described in detail on the basis of the Figure 1 The power control method of the inverter is applied to the controller of the anti-backflow system, and the anti-backflow system includes an inverter set and an auxiliary power supply set. The method includes the following steps.

[0109] S201, a second output total power sequence of the inverter set and a third output total power sequence of the auxiliary power supply set are obtained; the second output total power sequence includes the output total power of the inverter set at each time point in a preset time period, and the third output total power sequence includes the output total power of the auxiliary power supply set at each time point in the preset time period; the first output total power sequence is determined according to the second output total power sequence and the third output total power sequence.

[0110] Optionally, the anti-inrush system further comprises a smart meter, a data acquisition device, etc.

[0111] Exemplarily, the smart meter is a high-precision electric energy metering device deployed at the output end of the inverter, which supports real-time acquisition of data such as voltage, current, power factor, and transmits the data to the controller through a communication protocol (such as Modbus, DL / T645) for generating the second total output power sequence.

[0112] Exemplarily, the data acquisition device is a hardware device integrating various input / output (I / O) interfaces, which detects the access state of the auxiliary power supply (such as a switching signal) through a digital input (Digital Input, DI) channel, and calculates the real-time output power in combination with the rated power. The DI channel is a digital input interface in the data acquisition device, which is used to receive the switching state signal of the auxiliary power supply (such as the relay contact closure indicating access and the disconnection indicating non-access), so as to determine whether the auxiliary power supply participates in power supply.

[0113] In the power control method of the anti-inrush system, the output power of each inverter in the inverter set is first acquired in real time by the smart meter, and the second total output power sequence is accumulated at a preset time interval (such as every second), which records the total output power of the inverter at each time in the preset time period. At the same time, the data acquisition device detects the access state of the auxiliary power supply (such as whether it is started) through the DI channel, counts the number of auxiliary power supplies accessing the system, and calculates the actual output power of each auxiliary power supply in combination with the rated power of each auxiliary power supply, and then generates the third total output power sequence. Finally, the second total output power sequence and the third total output power sequence are aligned by time and added to obtain the first total output power sequence of the anti-inrush system, which reflects the total output power of all power supplies (inverters + auxiliary power supplies) in the system in the preset time period.

[0114] S202, determining a power difference sequence according to the first total output power sequence and a preset power threshold; wherein the power difference sequence includes a power difference value of the anti-inrush system at each time in the preset time period; the power difference value represents the power value that needs to be corrected by the inverter; based on a preset intelligent prediction model, the power difference sequence is processed to obtain a proportional parameter value, a differential parameter value, and an integral parameter value.

[0115] Exemplarily, in the power control method of the anti-flow system, first, the first output total power sequence is compared with the preset power threshold at each time, and a power difference sequence is generated. The sequence reflects the deviation of the output power of the system from the target threshold at each time in the preset period. Subsequently, the power difference sequence is input into a preset intelligent prediction model. The model analyzes the periodicity, trend and randomness of the sequence, predicts the future error change trend, and outputs the PID parameter value. Among them, the P parameter value directly responds to the current error, the I parameter value is based on the error change rate to predict the future trend, and the D parameter value eliminates the long-term deviation by integrating the historical error. Finally, the PID parameter value dynamically adjusts the output power of the inverter set through the closed-loop control mechanism, so that the system output total power is always stable around the preset threshold, avoiding reverse flow.

[0116] Figure 3 A flowchart of step S202 in the power control method of the inverter provided in the embodiments of the present application is shown as Figure 3 indicated, step S202 includes:

[0117] S2021, processing the power difference sequence based on a preset intelligent prediction model to obtain a first power difference vector; wherein the first power difference vector represents the change trend of the importance of the power difference value in the power difference sequence.

[0118] Optionally, the preset intelligent prediction model can be an LSTM model or a TCN model, and the embodiments of the present application do not make specific limitations. It can be understood that the preset intelligent prediction model of the present embodiment is an initial model based on an LSTM model, which is obtained by training based on a large amount of historical data.

[0119] Exemplarily, the processing flow of the preset intelligent prediction model based on LSTM on the power difference sequence is as follows: first, the power difference sequence is input into the input layer of the LSTM model, and the layer normalizes the data (such as scaling to the range of 0-1) to eliminate the dimension difference. Subsequently, the data enters the hidden layer of the LSTM, and the forgotten gate filters the historical information (such as the power difference value trend in the past 10 minutes) that needs to be retained, and the input gate updates the cell state by combining the power difference value at the current time to generate a new hidden state. The first power difference vector output by the hidden layer can capture the key features (such as the mutation point of the power difference value and the periodic fluctuation frequency) in the sequence, and map the PID parameter value through the fully connected layer. For example, if the vector shows that the power difference value in a certain period shows an upward trend, the LSTM will generate a larger differential parameter to suppress the upward trend in advance; if the vector shows that there is a static error for a long time, a larger integral parameter will be generated to gradually eliminate the deviation. Finally, the first power difference vector as the core output of the model provides a dynamic adjustment basis for subsequent PID control.

[0120] In one example, the power difference sequence is processed based on a preset intelligent prediction model to obtain a third power difference vector, where the third power difference vector represents the importance of the power difference values in the power difference sequence. The power difference sequence and the third power difference vector are processed based on the preset intelligent prediction model to obtain a fourth power difference vector, where the fourth power difference vector represents the power difference values whose importance is greater than a preset threshold in the power difference sequence. The power difference sequence, the third power difference vector, and the fourth power difference vector are processed based on the preset intelligent prediction model to obtain the first power difference vector.

[0121] In one example, the power difference sequence is processed based on a preset intelligent prediction model to obtain a third power difference vector, where the third power difference vector represents the importance of the power difference values in the power difference sequence. The power difference sequence is input into an LSTM model, and the historical information (such as the power difference value trend in the past 10 minutes) that needs to be retained is screened through a forgetting gate. At the same time, the cell state is updated by combining the power difference value at the current time through an input gate to generate the third power difference vector.

[0122] In one example, the power difference sequence and the third power difference vector are processed based on a preset intelligent prediction model to obtain a fourth power difference vector, where the fourth power difference vector represents the power difference values whose importance is greater than a preset threshold in the power difference sequence. The third power difference vector and the original power difference sequence are input into another layer of the LSTM, and the weight of each time feature is calculated through a gating mechanism (such as an input gate) to retain only the features whose weight exceeds a preset threshold (such as 0.8) to generate the fourth power difference vector.

[0123] In one example, the third layer of the LSTM takes the original power difference sequence, the third power difference vector, and the fourth power difference vector as input to generate the first power difference vector through an output gate.

[0124] In one example, the power difference sequence and the first power difference vector are processed based on a preset intelligent prediction model to obtain a second power difference vector, where the second power difference vector represents the features whose importance is greater than a preset threshold in the power difference sequence.

[0125] Exemplarily, the hidden state outputs of all time steps of the LSTM (i.e., the set of first power difference vectors) are sent into an attention mechanism model for weighted fusion. The attention mechanism automatically calculates the attention weight of each time step hidden state, and the weight depends on the degree of influence of the power difference at this moment on the overall system state - the more critical the influence (such as large deviation or dramatic change), the higher the weight. Finally, all hidden states are weighted and summed according to their weights to generate the final second power difference vector. For example, if the power difference value at a certain moment corresponds to a weight of 0.9 (more than the threshold 0.8), the feature at this moment will be highlighted, while the feature with a weight of 0.5 will be suppressed.

[0126] In an example, the power difference sequence and the first power difference vector are processed according to a preset intelligent prediction model to obtain an attention weight; wherein the attention weight represents the importance of the first power difference vector; the first power difference vector and the attention weight are weighted and summed according to the preset intelligent prediction model to obtain a second power difference vector.

[0127] Exemplarily, the preset intelligent prediction model receives the first power difference vector and the original power difference sequence as input, and calculates the weight of each dimension feature using an attention mechanism. Specifically, the model generates attention weights by matching the similarity between the query vector and the key vector, for example, using dot product attention to calculate the weight distribution. Finally, the first power difference vector and the attention weight are multiplied element by element and summed to generate the second power difference vector. For example, if the power difference value at a certain moment corresponds to a weight of 0.8 in the first power difference vector, the feature at this moment will be significantly strengthened, while the feature with a weight of 0.2 will be suppressed. Finally, the second power difference vector is the core output of the model, providing a more focused feature representation for subsequent PID parameter generation.

[0128] S2023, processing the second power difference vector based on a preset intelligent prediction model to obtain a proportional parameter value, a differential parameter value, and an integral parameter value.

[0129] Exemplarily, the second power difference vector inputs a preset intelligent prediction model of a full connection layer, and maps high-order features to initial values of PID parameters through linear transformation. Subsequently, an activation function performs nonlinear adjustment on the initial values, for example, a Sigmoid function restricts the range of proportional parameters, and a ReLU function retains the positive change trend of differential parameters. Finally, a model output layer performs normalization processing on the parameters to ensure that the PID parameter values meet the input requirements of the PID controller. For example, if the second power difference vector shows that the feature weight of the power difference value at a certain moment is 0.95, the model will generate a larger P parameter value to quickly respond to the current error, generate a moderate I parameter value to suppress future fluctuations, and gradually eliminate long-term deviations. Finally, the PID parameter values dynamically adjust the output power of the inverter set through a closed-loop control mechanism, so that the system output total power is stabilized near the preset threshold.

[0130] S203, determining a power correction parameter according to the first output total power sequence, the proportional parameter value, the differential parameter value, and the integral parameter value; wherein the power correction parameter represents the power percentage of the inverter in the inverter set that needs to be corrected; obtaining the current output power of the inverter in the inverter set; and controlling the power value of the inverter in the inverter set according to the current output power and the power correction parameter.

[0131] Exemplarily, the first output total power sequence is compared with a preset power threshold at each moment to generate a power difference sequence. Based on the power difference sequence, the output power of the inverter is adjusted.

[0132] Exemplarily,

[0133]

[0134] wherein, is the integral term calculation value; is the I parameter value; is the power difference sequence; is the time interval within a preset time period.

[0135]

[0136] wherein, is the differential term calculation value; is the D parameter value; is the current power difference value of the power difference sequence; is the previous moment power difference value of the power difference sequence; is the time interval within a preset time period.

[0137]

[0138] wherein, is the PID adjustment amount; is a P parameter value; is a current power difference value of the power difference sequence; is an integral term calculation value; is a differential term calculation value.

[0139] The PID adjustment amount is converted into a power correction parameter in percentage form; according to the current output power of each inverter, the power value is adjusted in proportion to the correction parameter, and the power is ensured within a safe range through clamping operation.

[0140] The power control method of the inverter provided by the embodiment of the application comprises the following steps: acquiring a second output total power sequence of an inverter set and a third output total power sequence of an auxiliary power supply set, the two sequences recording in detail the output total power of the inverter set and the auxiliary power supply set at each moment in a preset time period, providing comprehensive and accurate data support for subsequent analysis; then, determining a first output total power sequence according to the two sequences, and calculating a power difference sequence in combination with a preset power threshold, wherein the power difference value directly reflects the power value that needs to be corrected by the inverter, and the direction and amplitude of the power adjustment are clear; subsequently, processing the power difference sequence by using a preset intelligent prediction model to obtain a proportional parameter value, a differential parameter value and an integral parameter value; determining a power correction parameter based on the parameters, and clearly determining the power percentage that needs to be corrected by the inverter; finally, acquiring the current output power of the inverter, and determining the final power value in combination with the power correction parameter. The precise dynamic regulation and control of the power of the inverter are realized, the technical problems such as inaccurate power control and response lag in the traditional method are effectively solved, and the working performance and stability of the inverter are significantly improved.

[0141] On the basis of the above-mentioned embodiment, the embodiment of the application further provides an operable interface; the operable interface comprises a configuration interface, a running analysis interface, a data acquisition interface, an efficiency comparison interface, an abnormal alarm interface and a decision suggestion interface.

[0142] The user interface comprises six core modules: the configuration interface allows users to manually set the auxiliary power supply's rated power, inverter parameters, etc., and supports batch import / export and version management; the operation analysis interface displays the total power, inverter status, and auxiliary power supply load rate in real time, visually presenting the operating trend through dynamic line graphs and status indicator lights, and supports full-screen viewing and data backtracking; the data acquisition interface continuously collects information such as the system's total power and equipment status, providing custom field sorting and historical data query functions, and supporting filtering by time range or equipment ID; the efficiency comparison interface compares the auxiliary power supply's fuel efficiency and the inverter's power conversion efficiency through bar charts, generates optimization suggestion reports, and supports export; the anomaly alarm interface monitors power fluctuations, control oscillations, and other anomalies in real time, highlighting unprocessed alarms in red, and supports audible and visual prompts and log export; the decision suggestion interface provides suggestions for equipment load optimization and parameter adjustment based on historical data, marking the adoption rate and optimization effect, forming a closed-loop management process of "configuration-monitoring-analysis-optimization," and all interfaces are quickly navigated through the top tab bar to ensure operational continuity.

[0143] Figure 4 A flowchart illustrating a training method for a preset intelligent prediction model for a power control method of an inverter, as provided in this application embodiment, is shown below. Figure 4 As shown, the power control method for the inverter is the method provided in the above method embodiment. The training method of the preset intelligent prediction model is applied to the controller of the anti-reverse current system, which includes an auxiliary power supply set. The training method of the preset intelligent prediction model includes:

[0144] S401. Obtain the set of total output power sequences of the anti-reverse current system, the number of auxiliary power supplies connected in the auxiliary power supply set, and the load type of the anti-reverse current system; wherein, the set of total output power sequences includes at least one total output power sequence, and the total output power sequence includes the total output power of the anti-reverse current system at each moment in the preset time period.

[0145] For example, the total output power of the anti-reverse current system is collected in real time using smart meters or power sensors, and time-series data is generated according to preset time periods (such as per second or per minute). At the same time, the number of online auxiliary power supply devices is counted, and the load type is identified through load characteristic analysis. The collected total power sequence, auxiliary power supply quantity, and load type information are stored in a database, and archived by time or load type to form a historical data set.

[0146] Perform statistical analysis on the total output power sequence (such as mean, peak, and volatility), analyze the system power supply reliability in conjunction with the number of auxiliary power supplies, and optimize the power allocation strategy through load type data.

[0147] A report containing total power sequence characteristics, auxiliary power supply available number and load type distribution is generated to provide a basis for system operation and parameter adjustment.

[0148] S402, splice the output total power sequence in the output total power sequence set with its corresponding auxiliary power supply access number and load type to obtain a to-be-processed data set.

[0149] Exemplarily, the timestamp or sequence ID is indexed to ensure that each output total power sequence corresponds to a unique auxiliary power supply number and load type. The auxiliary power supply number (scalar value) and load type (classification label) are added as new columns to each row of data of the total power sequence. Check if there are missing values or time misalignment in the spliced data, and correct or mark the abnormal data. The verified data is stored in a preset format (such as CSV, JSON) to form a to-be-processed data set that can be directly called by a machine learning model or analysis tool.

[0150] S403, input the to-be-processed data set into an initial model for training to obtain a preset intelligent prediction model.

[0151] Exemplarily, the to-be-processed data set is normalized (such as scaling the power value to the range of 0-1), missing value filling (such as forward filling), and feature engineering (such as extracting the power fluctuation rate as a new feature).

[0152] The initial model can be an LSTM model or an RNN model, and the present application does not make specific limitations; it can be understood that the embodiment of the present application adopts an LSTM model as the initial model, and first, the hyperparameters of the LSTM model are randomly initialized.

[0153] The to-be-processed data set is divided into a training set and a validation set, the model output is calculated by forward propagation, the parameters are adjusted by back propagation to minimize the loss function (such as mean square error), and the model performance is converged until the model performance is converged. The generalization ability of the model is evaluated using the test set, and if the performance is not up to standard, the hyperparameters (such as learning rate, network layer number) are adjusted and retrained. The trained model parameters and structure are exported as a file (such as H5, ONNX format) to obtain a preset intelligent prediction model.

[0154] The embodiment of the application provides a training method of a preset intelligent prediction model for power control of an inverter. First, an output total power sequence set of an anti-flow system is acquired. The set covers the output total power of the anti-flow system at each moment within a preset period, thereby providing detailed and real-time basic data for model training. Meanwhile, the number of connected auxiliary power supplies in the auxiliary power supply set and the load type of the anti-flow system are collected. Next, the output total power sequence, the number of connected auxiliary power supplies and the load type are spliced to form a data set to be processed. This step effectively integrates multi-dimensional data and enhances the data representation capability. Finally, the data set to be processed is input into an initial model for training. Through continuous optimization of model parameters, a preset intelligent prediction model is finally obtained. The model can realize intelligent prediction and accurate control of PID parameters, thereby effectively improving the power regulation accuracy and adaptability of the inverter, realizing dynamic and stable control of the inverter power, and meeting the demand of a complex and variable power system.

[0155] Figure 5 The structure diagram of the power control device of the inverter provided by the embodiment of the application is shown as Figure 5 The device is applied to a controller of an anti-flow system. The anti-flow system includes a set of inverters. The power control device 50 of the inverter provided by the embodiment of the application includes:

[0156] The acquisition module 501 is configured to acquire a first output total power sequence of the anti-flow system. The first output total power sequence includes the output total power of the anti-flow system at each moment within a preset period.

[0157] The processing module 502 is configured to input the first output total power sequence into a preset intelligent prediction model for processing, so as to obtain a proportional parameter value, a differential parameter value and an integral parameter value. The proportional parameter value is used to correct the error of the output total power of the anti-flow system. The differential parameter value is used to correct the cumulative error of the output total power of the first output total power sequence within the preset period. The integral parameter value is used to predict the error of the output total power of the anti-flow system.

[0158] The control module 503 is configured to control the power of the inverters in the set of inverters according to the first output total power sequence, the proportional parameter value, the differential parameter value and the integral parameter value.

[0159] In a possible implementation, the processing module 502 includes:

[0160] The first processing module 5021 is configured to determine a power difference sequence according to the first output total power sequence and a preset power threshold. The power difference sequence includes a power difference value of the anti-flow system at each moment within the preset period. The power difference value represents the power value that needs to be corrected by the inverter.

[0161] The second processing module 5022 is configured to process the power difference sequence based on a preset intelligent prediction model to obtain a proportional parameter value, a differential parameter value, and an integral parameter value.

[0162] In a possible implementation, the second processing module 5022 includes:

[0163] The third processing module 50221 is configured to process the power difference sequence based on a preset intelligent prediction model to obtain a first power difference vector; the first power difference vector represents a variation trend of the importance of the power difference values in the power difference sequence.

[0164] The fourth processing module 50222 is configured to process the power difference sequence and the first power difference vector based on a preset intelligent prediction model to obtain a second power difference vector; the second power difference vector represents a feature of the power difference values with an importance greater than a preset threshold in the power difference sequence.

[0165] The fifth processing module 50223 is configured to process the second power difference vector based on a preset intelligent prediction model to obtain the proportional parameter value, the differential parameter value, and the integral parameter value.

[0166] In a possible implementation, the third processing module 50221 includes:

[0167] The third processing module 50221 is configured to process the power difference sequence based on a preset intelligent prediction model to obtain a third power difference vector; the third power difference vector represents the importance of the power difference values in the power difference sequence.

[0168] The fourth processing module 50222 is configured to process the power difference sequence and the third power difference vector based on a preset intelligent prediction model to obtain a fourth power difference vector; the fourth power difference vector represents the power difference values with an importance greater than a preset threshold in the power difference sequence.

[0169] The fifth processing module 50223 is configured to process the power difference sequence, the third power difference vector, and the fourth power difference vector based on a preset intelligent prediction model to obtain the first power difference vector.

[0170] In a possible implementation, the fourth processing module 50222 includes:

[0171] The fourth processing module 50222 is configured to process the power difference sequence and the first power difference vector based on a preset intelligent prediction model to obtain an attention weight; the attention weight represents the importance of the first power difference vector.

[0172] The fifth processing module 50223 is configured to perform weighted sum processing on the first power difference vector and the attention weight based on a preset intelligent prediction model to obtain the second power difference vector.

[0173] In a possible implementation, the control module 503 comprises:

[0174] determine a power correction parameter according to the first output total power sequence, the proportional parameter value, the differential parameter value, and the integral parameter value; wherein the power correction parameter represents a power percentage that needs to be corrected by the inverters in the inverter set;

[0175] obtain a current output power of the inverters in the inverter set; and control the power value of the inverters in the inverter set according to the current output power and the power correction parameter.

[0176] In a possible implementation, the anti-flow system further comprises an auxiliary power supply set; the obtaining module 501 comprises:

[0177] obtain a second output total power sequence of the inverter set and a third output total power sequence of the auxiliary power supply set; wherein the second output total power sequence comprises an output total power of the inverter set at each time point in a preset time period, and the third output total power sequence comprises an output total power of the auxiliary power supply set at each time point in the preset time period;

[0178] determine the first output total power sequence according to the second output total power sequence and the third output total power sequence.

[0179] The power control device for inverters provided in this embodiment can execute the method provided in the method embodiments, and has similar implementation principles and technical effects, which will not be described here.

[0180] Figure 6 A structure diagram of a training device for a preset intelligent prediction model for power control of inverters provided in this embodiment is shown in FIG. 6, which is applied to a controller of an anti-flow system, and the anti-flow system comprises an auxiliary power supply set. The device 60 comprises: Figure 6

[0181] The obtaining module 601 is configured to obtain an output total power sequence set of the anti-flow system, an access number of auxiliary power supplies in the auxiliary power supply set, and a load type of the anti-flow system; wherein the output total power sequence set comprises at least one output total power sequence, and the output total power sequence comprises an output total power of the anti-flow system at each time point in a preset time period;

[0182] The splicing module 602 is configured to splice the output total power sequence in the output total power sequence set with the corresponding access number of auxiliary power supplies and the load type, to obtain a to-be-processed data set;

[0183] The training module 603 is configured to input the to-be-processed data set into an initial model for training, to obtain a preset intelligent prediction model. ​

[0184] The training device for the preset intelligent prediction model of power control of the inverter provided in this embodiment can execute the method provided in the method embodiments, and has similar implementation principles and technical effects. Details are not described herein again.

[0185] Figure 7 A structural schematic diagram of an anti-reflux system provided in this embodiment is shown in FIG. 7. The anti-reflux system 70 includes an inverter set 701, an auxiliary power supply set 702, a smart meter 703, a data acquisition device 704, and a controller 705. The inverter set 701 includes at least one inverter, and the auxiliary power supply set 702 includes at least one auxiliary power supply.

[0186] One end of the auxiliary power supply is connected to one end of the inverter through a load 706, and the other end of the inverter is connected to a power grid 707 system through the smart meter 703.

[0187] The other end of the auxiliary power supply is connected to the other end of the inverter through the data acquisition device 704, the other end of the data acquisition device 704 is connected to the smart meter 703, and the controller 705 is connected to the data acquisition device 704.

[0188] The specific implementation process of the controller can be referred to the method embodiments, and has similar implementation principles and technical effects. Details are not described herein again.

[0189] Figure 8 A structural schematic diagram of an electronic device provided in this embodiment is shown in FIG. 8. As shown in FIG. 8, the electronic device 80 provided in this embodiment includes at least one processor 801 and a memory 802. Optionally, the device 80 further includes a communication component 803. The processor 801, the memory 802, and the communication component 803 are connected through a bus 804. Figure 8

[0190] In the specific implementation process, the at least one processor 801 executes the computer execution instructions stored in the memory 802, so that the at least one processor 801 executes the method described above.

[0191] The specific implementation process of the processor 801 can be referred to the method embodiments, and has similar implementation principles and technical effects. Details are not described herein again.

[0192] ​In the above embodiments, it should be understood that the processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in combination with the application can be directly embodied as hardware processor execution, or executed by a combination of hardware and software modules in the processor.

[0193] The memory can include a random access memory (RAM), and can also include a non-volatile memory (NVM), such as at least one disk memory.

[0194] The bus can be an industry standard architecture (ISA) bus, a peripheral component (PCI) bus, or an extended industry standard architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus in the drawings of the present application does not limit to only one bus or one type of bus.

[0195] The present application also provides a computer program product, comprising a computer program, which, when executed by a processor, implements the above method.

[0196] The present application also provides a computer readable storage medium, which stores computer execution instructions, and when a processor executes the computer execution instructions, the above method is implemented.

[0197] The above readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0198] An example readable storage medium is coupled to the processor such that the processor can read information from the readable storage medium and can write information to the readable storage medium. Of course, the readable storage medium can also be a part of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the readable storage medium can also exist as discrete components in the device.

[0199] The division of units is only a logical functional division, and in actual implementation, there can be another division manner, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0200] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, can be located in one place, or can be distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0201] In addition, the functional units in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit.

[0202] If the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the part of the present application that essentially contributes to the prior art or the part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method of each embodiment of the present application. The foregoing storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program code storage media.

[0203] Those skilled in the art can understand that all or part of the steps of the above-mentioned method embodiments can be completed by program instruction related hardware. The foregoing program can be stored in a computer readable storage medium. The program executes to perform the steps of the above-mentioned method embodiments; and the foregoing storage medium includes various media capable of storing program codes, such as ROM, RAM, magnetic disk, or optical disk.

[0204] Finally, it should be noted that other embodiments of the present application will readily occur to those skilled in the art upon consideration of the specification and practice of the present application disclosed herein. The present application is intended to include all such variations as fall within the general scope of the application, and includes the generic principles disclosed and the best mode known to the inventors to be currently practiced as well as variations thereof, without departing from the scope of the present application as defined by the claims. The specification and examples give the best application of the present application as known to at least one of the inventors at the time of the filing of this application. It is to be understood that since numerous modifications and changes will readily occur to those skilled in the art, the application is not to be limited to the exact construction and operation as illustrated and described. Accordingly, all such variations are intended to be included within the scope of the present application as defined in the claims. The application is to be limited only by the claims.

Claims

1. A power control method for an inverter, characterized in that, The method is applied to the controller of an anti-backflow system, the anti-backflow system including an inverter assembly, the method comprising: Obtain the first total output power sequence of the anti-backflow system; wherein, the first total output power sequence includes the total output power of the anti-backflow system at each moment in a preset time period; The first total output power sequence is input into a preset intelligent prediction model for processing to obtain proportional parameter values, differential parameter values, and integral parameter values; wherein, the proportional parameter value is used to correct the error of the total output power of the anti-reverse flow system, the differential parameter value is used to correct the cumulative error of the total output power of the first total output power sequence within a preset time period, and the integral parameter value is used to predict the error of the total output power of the anti-reverse flow system. The power of the inverters in the inverter set is controlled based on the first total output power sequence, the proportional parameter value, the differential parameter value, and the integral parameter value.

2. The method according to claim 1, characterized in that, The first output total power sequence is input into a preset intelligent prediction model for processing to obtain proportional parameter values, derivative parameter values, and integral parameter values, including: A power difference sequence is determined based on the first total output power sequence and a preset power threshold; wherein, the power difference sequence includes the power difference value of the anti-reverse current system at each moment in a preset time period; the power difference value represents the power value that the inverter needs to correct; The power difference sequence is processed based on the preset intelligent prediction model to obtain the proportional parameter value, the differential parameter value, and the integral parameter value.

3. The method according to claim 2, characterized in that, The power difference sequence is processed based on the preset intelligent prediction model to obtain the proportional parameter value, the differential parameter value, and the integral parameter value, including: The power difference sequence is processed based on the preset intelligent prediction model to obtain a first power difference vector; wherein, the first power difference vector characterizes the changing trend of the importance of the power difference values ​​in the power difference sequence. The power difference sequence and the first power difference vector are processed according to the preset intelligent prediction model to obtain a second power difference vector; wherein, the second power difference vector represents the power difference value in the power difference sequence whose importance is greater than a preset threshold. The second power difference vector is processed based on the preset intelligent prediction model to obtain the proportional parameter value, the differential parameter value, and the integral parameter value.

4. The method according to claim 3, characterized in that, The power difference sequence is processed based on the preset intelligent prediction model to obtain a first power difference vector, including: The power difference sequence is processed based on the preset intelligent prediction model to obtain a third power difference vector; wherein, the third power difference vector represents the importance of the power difference values ​​in the power difference sequence. Based on the preset intelligent prediction model, the power difference sequence and the third power difference vector are processed to obtain a fourth power difference vector; wherein, the fourth power difference vector represents the power difference in the power difference sequence whose importance is greater than a preset threshold. Based on the preset intelligent prediction model, the power difference sequence, the third power difference vector, and the fourth power difference vector are processed to obtain the first power difference vector.

5. The method according to claim 3, characterized in that, The power difference sequence and the first power difference vector are processed according to the preset intelligent prediction model to obtain the second power difference vector, including: The power difference sequence and the first power difference vector are processed according to the preset intelligent prediction model to obtain attention weights; wherein, the attention weights characterize the importance of the first power difference vector. The first power difference vector and the attention weight are weighted and summed according to the preset intelligent prediction model to obtain the second power difference vector.

6. The method according to claim 1, characterized in that, Based on the first total output power sequence, the proportional parameter value, the derivative parameter value, and the integral parameter value, controlling the power of the inverters in the inverter set includes: Based on the first total output power sequence, the proportional parameter value, the differential parameter value, and the integral parameter value, a power correction parameter is determined; wherein, the power correction parameter represents the percentage of power that needs to be corrected for the inverters in the inverter set; Obtain the current output power of the inverters in the inverter set; control the power value of the inverters in the inverter set according to the current output power and the power correction parameter.

7. The method according to any one of claims 1-6, characterized in that, The anti-reverse current system also includes an auxiliary power supply set; obtaining the first total output power sequence of the anti-reverse current system includes: Obtain a second total output power sequence of the inverter set and a third total output power sequence of the auxiliary power supply set; wherein, the second total output power sequence includes the total output power of the inverter set at each moment in a preset time period, and the third total output power sequence includes the total output power of the auxiliary power supply set at each moment in the preset time period; The first total output power sequence is determined based on the second total output power sequence and the third total output power sequence.

8. A training method for a preset intelligent prediction model for a power control method of an inverter, wherein the power control method of the inverter is the method as described in any one of claims 1-7, characterized in that, The training method of the preset intelligent prediction model is applied to the controller of the anti-reverse current system, which includes an auxiliary power supply set. The training method of the preset intelligent prediction model includes: Obtain the set of total output power sequences of the anti-reverse current system, the number of auxiliary power supplies connected in the auxiliary power supply set, and the load type of the anti-reverse current system; wherein, the set of total output power sequences includes at least one total output power sequence, and the total output power sequence includes the total output power of the anti-reverse current system at each moment in a preset time period; The total output power sequence in the set of total output power sequences is concatenated with the corresponding number of auxiliary power supplies connected and the load type to obtain the dataset to be processed. The dataset to be processed is input into the initial model for training to obtain a preset intelligent prediction model.

9. A power control device for an inverter, characterized in that, The device is used in the controller of an anti-backflow system, which includes an inverter assembly. The device comprises: The acquisition module is used to acquire the first total output power sequence of the anti-backflow system; wherein, the first total output power sequence includes the total output power of the anti-backflow system at each moment in a preset time period; The processing module is used to input the first total output power sequence into a preset intelligent prediction model for processing to obtain proportional parameter values, differential parameter values, and integral parameter values; wherein, the proportional parameter value is used to correct the error of the total output power of the anti-reverse flow system, the differential parameter value is used to correct the cumulative error of the total output power of the first total output power sequence within a preset time period, and the integral parameter value is used to predict the error of the total output power of the anti-reverse flow system. The control module is used to control the power of the inverters in the inverter set according to the first total output power sequence, the proportional parameter value, the derivative parameter value, and the integral parameter value.

10. A backflow prevention system, characterized in that, The anti-backflow system includes an inverter assembly, an auxiliary power supply assembly, a smart meter, a data acquisition device, and a controller; wherein the inverter assembly includes at least one inverter, and the auxiliary power supply assembly includes at least one auxiliary power supply. One end of the auxiliary power supply is connected to one end of the inverter through a load, and the other end of the inverter is connected to the power grid system through the smart meter; The other end of the auxiliary power supply is connected to the other end of the inverter through the data acquisition device, and the other end of the data acquisition device is connected to the smart meter; the controller is connected to the data acquisition device; The controller is used to perform the method as described in any one of claims 1-7.

Images