Photovoltaic inverter output test and capacity evaluation method

Abstract

The invention relates to the technical field of electric power measurement, in particular to a photovoltaic inverter output test and capacity evaluation method, which comprises the following steps of: synchronously acquiring current, voltage and power related data, carrying out joint mapping analysis, identifying non-periodic change, finishing structure alignment, and comprehensively comparing and processing a multi-track relation, so as to obtain a photovoltaic inverter output test and capacity evaluation result. And finally outputting a photovoltaic inverter output test and capacity evaluation result. According to the method, the expression ability of electrical parameter coupling evolution characteristics is enhanced by constructing a multi-parameter joint mapping image, aperiodic fluctuation characteristics are extracted in combination with structure alignment of an offset section, multi-track synchronous alignment analysis is realized in an image mode, and a fitting error is further mapped to a path structure to form capacity combination expression; finally, an output intensity distribution trend is presented based on a coupling coverage relation, the recognition capability of a coupling mode and continuous output characteristics is enhanced, and the stability and precision of evaluation under variable working conditions are improved.

Description

A method for testing the output and assessing the capacity of a photovoltaic inverter Technical Field

[0001] This invention relates to the field of power measurement technology, and in particular to a method for testing the output and evaluating the capacity of a photovoltaic inverter. Background Technology

[0002] The field of power measurement technology involves the measurement, monitoring, and evaluation of electricity, electrical energy, and electrical parameters, and is widely used in power systems, photovoltaic power generation, wind power generation, and other renewable energy fields. Core aspects of this technology include accurate measurement of power flow, and analysis and evaluation of power factor, voltage, and current. Power measurement technology is not only used to monitor and analyze power production and distribution, but also to optimize energy efficiency and ensure the operational stability of power equipment. Common power measurement technologies include power metering, real-time monitoring systems, power quality analysis, and various energy efficiency assessment methods. With the development of renewable energy, the application of power measurement technology in new energy fields such as photovoltaic and wind power is gradually increasing, especially in terms of equipment performance monitoring and optimization. Traditional photovoltaic inverter output testing and capacity assessment methods refer to the process of evaluating the performance and capacity of photovoltaic inverters by directly measuring and analyzing their output power, power factor, and other electrical parameters. This method typically uses a test platform based on DC power supply and load circuits, utilizing power meters and data recording equipment for real-time data acquisition and monitoring. Traditional testing methods require manual setting and adjustment of test conditions, making the testing process cumbersome and susceptible to external environmental influences. The operating status of an inverter is typically determined by measuring parameters such as its output voltage, current, and frequency, and these parameters are used to evaluate the device's capacity and efficiency. This testing process involves multiple steps, different test scenarios, and parameter settings to ensure the inverter's reliability and output stability under varying loads and operating conditions.

[0003] Existing technologies use static testing platforms for data acquisition, with fixed test scenarios. They lack the ability to dynamically capture parameter changes, and the measurement results of voltage, current, and power are fragmented in the time dimension, failing to reflect the synergistic evolution characteristics between multiple parameters. The evaluation method relies on single-point measurement results to determine equipment capacity, lacking image-based feature extraction and structured alignment analysis mechanisms. This results in evaluation results that are insensitive to output fluctuation characteristics, easily overlooking nonlinear offsets and changes in coupling relationships. Under load disturbances or unstable input conditions, misjudgments are prone to occur, affecting the stability and adaptability of capacity assessment. Summary of the Invention

[0004] To achieve the above objectives, the present invention adopts the following technical solution: a method for testing the output power and evaluating the capacity of a photovoltaic inverter, comprising the following steps: S1: synchronously measuring the AC output terminal of the photovoltaic inverter using a voltage measuring device and a current measuring device, acquiring the current change curve, voltage change curve, and instantaneous power trajectory over a continuous time period, and jointly mapping the measurement results to generate a power output image; S2: based on the power output image, analyzing the voltage change curve, extracting non-periodic offset segments, and performing structural alignment with the current change trajectory over the same time period to construct an offset feature image; S3: based on the offset feature image, comparing the intersection paths of the power change trajectory with the voltage change trajectory and the current change trajectory, extracting the coupled distribution structure with synchronous characteristics, and outputting a structural alignment image; S4: based on The structure alignment image is used to calculate the fitting difference between the power output trajectory and the coupling path. Linear expansion is performed along the coupling trajectory, and the fitted segments are arranged and combined to generate a capacity identification image. After generating the capacity identification image, the integrity of the coupling path, the trajectory fitting deviation, and the linear continuity of the fitted segments are further comprehensively calculated to form a capacity assessment parameter set. The inverter capacity assessment result is output based on the preset capacity level range. The actual capacity percentage is inferred by the coverage of the linear fitted segments and the degree of matching of the coupling path in the capacity identification image to obtain the actual output capacity of the inverter. S5: Based on the capacity identification image, the overlap relationship between the power trajectory, voltage curve, and coupling coverage path is determined, and the output intensity distribution of the photovoltaic inverter is expressed in a combined form. The photovoltaic inverter output test and capacity assessment results are output.

[0005] As a further aspect of the present invention, the power output image includes current change characteristics, voltage fluctuation characteristics, and instantaneous power characteristics; the offset feature image includes voltage aperiodic offset characteristics, current alignment structure characteristics, and time synchronization characteristics; the structure alignment image includes coupled distribution structure characteristics, path intersection position characteristics, and synchronization alignment relationship characteristics; the capacity identification image includes trajectory fitting difference characteristics, path linear expansion characteristics, and capacity combination distribution characteristics; and the photovoltaic inverter output test and capacity evaluation results include output intensity distribution characteristics, power trajectory matching characteristics, and structure path coverage characteristics.

[0006] As a further aspect of the present invention, the non-periodic offset segment refers to the continuous fluctuation segment in the voltage change curve that deviates from the normal periodic fluctuation pattern, and analyzes the impact of potential anomalies and dynamic loads.

[0007] As a further aspect of the present invention, the coupled distribution structure with synchronous characteristics refers to a cross-path structure that exhibits temporal consistency and morphological coordination characteristics in the power, voltage and current change trajectories.

[0008] As a further embodiment of the present invention, the specific steps of S1 are as follows: S101: Based on the AC output terminal of the photovoltaic inverter, the voltage measuring device and the current measuring device are invoked to synchronously detect the output signal within a continuous time period of the output terminal, extract the periodic change values ​​of the AC voltage signal and the AC current signal respectively, and perform pairing processing on the instantaneous voltage value and the instantaneous current value at the time point to obtain a voltage and current synchronization data frame; S102: According to the voltage and current synchronization data frame, the instantaneous voltage value and the instantaneous current value in each time segment are multiplied point-to-point to construct the instantaneous power value sequence of the corresponding time segment, and the value variation range between adjacent sample points in the sequence is continuously measured to generate an instantaneous power change trajectory sequence; S103: The voltage and current synchronization data frame and the instantaneous power change trajectory sequence are invoked to perform dimensional alignment and numerical normalization processing on the data, establish a two-dimensional joint matrix map within the corresponding time segment, and construct an image data frame based on the matrix map to obtain an energy output image.

[0009] As a further embodiment of the present invention, the specific steps of S2 are as follows: S201: Based on the power output image, extract the time axis signal sequence of the corresponding voltage change curve, and divide the voltage amplitude of the periodic segment in the sequence into intervals. Then, compare the positions of the periodic peaks and troughs in each interval. If the peak interval exceeds the preset voltage fluctuation period threshold, mark the interval as a non-periodic segment and establish a voltage offset interval sequence; S202: Call the voltage offset interval sequence, synchronously extract the continuous data segments of the current change curve within the same time period, and perform one-to-one matching processing on the voltage sequence and current sequence based on the time index. Integrate the current data in the offset interval into the voltage offset structure to generate a voltage-current aligned structure dataset; S203: According to the voltage-current aligned structure dataset, perform two-dimensional grid encoding on the voltage and current combination data of each offset interval, and set the numerical level mapping according to the data amplitude to construct an image frame sequence consistent with the original time sequence and obtain the offset feature image.

[0010] As a further aspect of the present invention, the specific steps of S3 are as follows: S301: Based on the offset feature image, extract pixel regions with power channel brightness higher than a set threshold as candidate regions for power output trajectory. Combine the image coordinate mapping relationship to obtain the voltage and current coordinates corresponding to the pixels, i.e., the VI plane position. Simultaneously extract the pixel paths of voltage trajectory and current trajectory, and match density overlapping regions in the coordinate space to identify the coupling intersection points of power trajectory and voltage and current trajectory, and obtain a set of trajectory intersection segments; S302: Based on the set of trajectory intersection segments, using image coordinates as a reference, calculate the spatial overlap between the power trajectory path and the voltage trajectory path in each intersection segment, and compare it with the overlap between the power trajectory and the current trajectory in the same region. Filter the intersection segments where the difference in spatial overlap between the power trajectory and the voltage trajectory, and between the power trajectory and the current trajectory is less than a preset threshold to obtain a set of synchronously coupled intersection segments; S303: Call the set of synchronously coupled intersection segments, perform region aggregation on the image region where the intersection segments are located, and construct a two-dimensional continuous coordinate mapping based on continuous time index. Re-encode the aggregation result into an image pixel matrix to establish a structural alignment image.

[0011] As a further embodiment of the present invention, the specific steps of S4 are as follows: S401: Based on the structure alignment image, extract the pixel path corresponding to the power output trajectory in the image, and use the coupling distribution path in the image as a reference to calculate the coordinate difference sequence between the pixel point on the power path and the corresponding coupling path point, and then perform mean and maximum difference amplitude statistics on the difference sequence to obtain a path fitting deviation index set; S402: According to the path fitting deviation index set, retrieve continuous segments in the coupling distribution path where the difference fluctuation is less than a preset difference stability threshold, and obtain a continuous pixel string along the segment direction, calculate the linear extension path in the direction of each pixel point, integrate adjacent extension path segments, and generate a linear combination fitting segment set. S403: Call the linear combination fitting segment set, take the distribution path direction as the arrangement axis, and perform pixel-aligned encoding of all fitting segments according to time order and coordinate position order. Map the combination result to a two-dimensional image frame structure to establish a capacity identification image. After generating the capacity identification image, further perform comprehensive calculations on the coupling path integrity, trajectory fitting deviation, and fitting segment linear continuity to form a capacity evaluation parameter set. Output the inverter capacity evaluation result based on the preset capacity level range. The actual capacity percentage is back-calculated by the linear fitting segment coverage and coupling path matching degree in the capacity identification image to obtain the actual output capacity of the inverter. After the structural alignment image is generated, this invention introduces a fast capacity back-calculation mechanism. The method includes the following steps: control the active power output of the photovoltaic inverter to 10% of its current maximum operating capacity; obtain the output value P1 under this power output state through a voltage / current measuring device; calculate the rated capacity of the inverter according to the following formula: Where: P1: the actual power value when the inverter is controlled at 10% of its rated output; C: the estimated rated capacity of the inverter.

[0012] Assuming stable light, temperature, and grid conditions, the output power and control percentage are approximately linearly related, which can be used for rapid on-site capacity estimation or as a reference for image-based capacity assessment. Before using the image method to assess capacity, the photovoltaic inverter control module is set to limit the output to 10% of the rated value, that is, the control system sends a command to the inverter to reduce the active power to 10%.

[0013] The output power value is recorded as 23.5kW at this time, and calculated using the following formula: Comparing this value with the evaluation results of the image method, the deviation is less than 1%, verifying the feasibility and accuracy of this method in the field evaluation scenario.

[0014] As a further aspect of the present invention, the specific steps of S5 are as follows: S501: Call the capacity identification image, extract the pixel coordinate sequence of the power output trajectory, voltage change curve and coupling structure coverage path in the image, and perform matching statistics on the overlapping pixels in the three sets of coordinate sequences to calculate the coordinate overlap rate and trajectory coverage value among the three, and obtain the trajectory overlap index set; S502: According to the trajectory overlap index set, map the index values ​​to a two-dimensional coordinate region, perform aggregation processing according to the time index of the power trajectory, and use the power, voltage and coupling path overlap level in each time period as a vector combination to construct the region label, and generate the output intensity distribution matrix; S503: Call the output intensity distribution matrix, perform overall traversal processing on the intensity value, spatial span and combination structure of the labeled region in the matrix, and perform region level evaluation and classification output according to the preset capacity level interval standard to establish the photovoltaic inverter output test and capacity evaluation results.

[0015] As a further aspect of the present invention, the voltage fluctuation period threshold refers to the upper limit of a period of time for judging whether the voltage signal remains periodic. If the interval between adjacent peaks exceeds the threshold, it is considered that the voltage signal has serious non-periodic fluctuations.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present invention, the ability to express the electrical parameter coupling evolution characteristics is enhanced by constructing a multi-parameter joint mapping image, and the non-periodic fluctuation features are extracted by combining the structural alignment of the offset segment. The synchronous alignment analysis of multiple trajectories is realized in the form of an image. Furthermore, the fitting error is mapped to the path structure to form a capacity combination expression. Finally, the output intensity distribution trend is presented based on the coupling coverage relationship, which enhances the ability to identify coupling mode and continuous output characteristics and improves the stability and accuracy of the evaluation under variable working conditions. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 is a schematic diagram of the steps of the present invention; Figure 2 is a detailed schematic diagram of S1 of the present invention; Figure 3 is a detailed schematic diagram of S2 of the present invention; Figure 4 is a detailed schematic diagram of S3 of the present invention; Figure 5 is a detailed schematic diagram of S4 of the present invention; Figure 6 is a detailed schematic diagram of S5 of the present invention. Detailed Implementation

[0019] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0020] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0021] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0022] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0023] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0024] Please refer to Figure 1. This embodiment of the invention provides a method for testing the output and evaluating the capacity of a photovoltaic inverter, including the following steps: S1: Synchronous signal measurement is performed on the AC output terminal of the photovoltaic inverter using a voltage measurement device and a current measurement device. Current change curves, voltage change curves, and instantaneous power change trajectories within a continuous time segment are collected. The three types of measurement results are jointly mapped to output a power output image; S2: Based on the power output image, non-periodic offset segments appearing in the voltage change curve are analyzed. Combined with the current change trajectory within the same time segment, structural alignment processing is performed to construct an offset feature image; S3: The offset feature image is input, and the intersection path positions between the continuous output segments in the power change trajectory and the voltage change trajectory, as well as the current change trajectory, are compared to extract the image. The system has a continuous coupled distribution structure with synchronous characteristics, and outputs a structure alignment image; S4: Based on the structure alignment image, calculate the fitting difference between the power output trajectory and the coupled distribution path in the image, perform a linear expansion operation on the coupled trajectory, arrange and combine the fitted segments along the distribution path to generate a capacity identification image; after generating the capacity identification image, further perform comprehensive calculations on the integrity of the coupling path, trajectory fitting deviation, and linear continuity of the fitted segments to form a capacity assessment parameter set, and output the inverter capacity assessment result based on the preset capacity level range; the actual capacity percentage is back-calculated by the coverage of the linear fitted segments and the degree of matching of the coupling path in the capacity identification image to obtain the actual output capacity of the inverter; after the structure alignment image is generated, this invention introduces a fast capacity back-calculation mechanism. The method includes the following steps: controlling the active power output of the photovoltaic inverter to 10% of its current maximum operating capacity; obtaining the output value P1 under this power output state through a voltage / current measuring device; calculating the rated capacity of the inverter according to the following formula: Where: P1: the actual power value when the inverter is controlled at 10% of its rated output; C: the estimated rated capacity of the inverter.

[0025] Assuming stable light, temperature, and grid conditions, the output power and control percentage are approximately linearly related, which can be used for rapid on-site capacity estimation or as a reference for image-based capacity assessment. Before using the image method to assess capacity, the photovoltaic inverter control module is set to limit the output to 10% of the rated value, that is, the control system sends a command to the inverter to reduce the active power to 10%.

[0026] The output power value is recorded as 23.5kW at this time, and calculated using the following formula: The value was compared with the evaluation results of the image method, and the deviation was less than 1%, which verified the feasibility and accuracy of the method in the field evaluation scenario; S5: Call the capacity identification image, judge the degree of overlap between the power output trajectory, voltage change curve and the coverage path of the coupling structure, and express the output intensity distribution relationship of the photovoltaic inverter in a combined form, and output the photovoltaic inverter output test and capacity evaluation results.

[0027] The power output image includes current change characteristics, voltage fluctuation characteristics, and instantaneous power characteristics. The offset feature image includes voltage aperiodic offset characteristics, current alignment structure characteristics, and time synchronization characteristics. The structural alignment image includes coupling distribution structure characteristics, path intersection position characteristics, and synchronization alignment relationship characteristics. The capacity identification image includes trajectory fitting difference characteristics, path linear expansion characteristics, and capacity combination distribution characteristics. The photovoltaic inverter output test and capacity assessment results include output intensity distribution characteristics, power trajectory matching characteristics, and structural path coverage characteristics.

[0028] In terms of capacity assessment, this invention further introduces a dynamic power fitting and capacity mapping mechanism. Specifically, it includes converting the fitting deviation between the output power trajectory and the reference coupling path into a pixel offset index, and combining this with the distribution range of the coupling path to generate a multi-dimensional "capacity assessment label matrix". This matrix is ​​used to express the capacity performance at different times, phases, and path segments, constructing a capacity assessment index system with "linear support length, fitting stability, and coupling integrity" as core dimensions, supporting quantitative determination of inverter output capacity. To verify the capacity assessment capability of this method, two photovoltaic inverters in different states were selected for comparative analysis. Through the capacity assessment process of this invention, a capacity labeling image was constructed and an output intensity distribution matrix was output. It was found that: Inverter A: trajectory overlap rate was 96.3%, the average fitting deviation was 1.2 pixels, and the coupling path coverage was 97.5%, assessed as Grade A; Inverter B: trajectory overlap rate was 88.5%, the average fitting deviation was 3.1 pixels, and the coverage was 89.8%, assessed as Grade C.

[0029] The above data shows that the present invention can effectively identify the output capacity decay caused by loss, thermal drift or abnormal drive control, and has a high resolution capability.

[0030] Please refer to Figure 2. The specific steps of S1 are as follows: S101: Based on the AC output terminal of the photovoltaic inverter, the voltage and current measuring devices are invoked to synchronously detect the output signal during a continuous period. The periodic change values ​​of the AC voltage and AC current signals are extracted respectively, and the instantaneous voltage and current values ​​at each time point are paired to obtain a voltage and current synchronization data frame. For the AC output terminal of the photovoltaic inverter with a rated power of 50 kW, the high-frequency data acquisition process is initiated. First, Hall voltage and current sensors with an accuracy of 0.2 are selected and connected to the A-phase output circuit of the inverter to convert the high-voltage, high-current signal into a low-voltage analog signal (such as + / -10V range) suitable for processing by the acquisition card. The data acquisition card is set to a sampling frequency of 10kHz, that is, 10,000 points are collected per second, and the time interval between adjacent sampling points is precisely controlled within 0.1 milliseconds. During the set 0.1-second continuous monitoring period, trigger commands are sent to the voltage and current channels simultaneously, and the digital sequence obtained after analog-to-digital conversion of the analog signal is recorded synchronously. A dual-channel buffer is allocated in memory, consisting of a voltage sequence buffer and a current sequence buffer. For each discrete data point acquired, a microsecond-level timestamp is added based on an internal high-precision clock to ensure strict alignment of the instantaneous voltage and current values ​​on the time axis. For example, at the 5th sampling moment after monitoring starts (i.e., 0.5 milliseconds), the voltage channel acquires a value of 48.75 volts, and the current channel acquires a value of 35.56 amperes. These two values ​​are bound to their time index of 0.5 milliseconds to form a raw data pair. To extract periodic changes, a zero-crossing detection algorithm is executed: the voltage zero-crossing threshold is set to 0 volts, and the logic for determining positive crossings is configured (i.e., the current voltage is less than 0 and the next voltage is greater than or equal to 0). When the first positive zero-crossing is detected, it is marked as the start time of the current cycle; when the third positive zero-crossing is detected, the first two complete AC cycles can be determined. Linear interpolation is used to accurately calculate the specific microsecond time of the zero-crossing, thereby correcting the phase error caused by sampling rate limitations. After the above processing, all sampling points within a continuous time period are organized into a voltage and current synchronous data frame. This data frame contains a time index column, a voltage amplitude column, a current amplitude column, and a signal quality status bit, providing an unsampled full waveform data basis for subsequent power analysis.

[0031] S102: Based on the voltage and current synchronization data frame, perform point-to-point multiplication of the instantaneous voltage and current values ​​within each time segment to construct a sequence of instantaneous power values ​​for the corresponding time segment. Continuously measure the range of value variation between adjacent sample points in the sequence to generate a sequence of instantaneous power change trajectories. Perform instantaneous power calculation on each set of voltage and current data recorded in the data frame. For the 1000 sampled samples contained in the frame (based on a duration of 0.1 seconds and a sampling rate of 10kHz), read the instantaneous voltage and current values ​​I corresponding to the time index t line by line, and perform multiplication to obtain the instantaneous power P. For example, if the voltage is read as 48.75 volts and the current as 35.56 amperes at 0.5 milliseconds, the product of the two values ​​yields an instantaneous power of 1733.55 watts. Store the calculated power values ​​in chronological order into the instantaneous power value sequence. Subsequently, perform a first-order difference operation on this sequence to measure the range of power variation between adjacent sample points. Specifically, the power values ​​at the nth and (n+1)th points are read, and the absolute value of the difference between them is calculated. Taking the power calculated at 0.4 milliseconds as 1110.51 watts (39.02 volts multiplied by 28.46 amperes) and at 0.5 milliseconds as an example, the difference is 623.04 watts. To evaluate the reasonableness of this difference and generate a trajectory sequence, a maximum power variation baseline value needs to be set. This baseline value is set based on the physical limits of the inverter: for a 50 kW inverter at a 50 Hz grid frequency, the maximum rate of change of the sinusoidal power occurs near the zero-crossing point of the voltage and current (where the slope of change is largest). By calculating the product of the peak power corresponding to the rated power (approximately 100 kW) and twice the angular frequency (100 pi), and combining this with a sampling time interval of 0.1 milliseconds, the theoretical maximum single-step power increment is approximately 3140 watts. Considering harmonic interference and measurement noise in actual operation, a tolerance factor of 1.2 is introduced, setting the final instantaneous power variation threshold to 3768 watts. The actual variation value calculated at each moment is compared with this threshold (only recorded here, not discarded), and the absolute power value sequence and the variation value sequence are merged to generate an instantaneous power change trajectory sequence containing dual attributes. This sequence fully depicts the dynamic rise and fluctuation path of the inverter output power on a microsecond time scale.

[0032] S103: The system calls upon the voltage and current synchronization data frames and the instantaneous power change trajectory sequence, performs dimensional alignment and numerical normalization on the data, establishes a two-dimensional joint matrix map within the corresponding time segment, and constructs image data frames based on the matrix map to obtain the power output image. First, to eliminate the order-of-magnitude differences between different physical dimensions (voltage of several hundred volts, current of several hundred amperes, power of several tens of kilowatts), the system performs extreme value normalization on the voltage sequence, current sequence, power sequence, and power change rate sequence respectively. It iterates through the voltage sequence within the current time segment, finds the maximum value of 311 volts and the minimum value of -311 volts, and uses the formula of subtracting the minimum value from the current value and dividing by the range to map all voltage data to the interval between 0 and 1. Similarly, the same normalization operation is performed on the current sequence (peak 320 amperes) and the power sequence (peak 100 kilowatts). Next, a two-dimensional joint matrix map with a resolution of 1024 by 1024 pixels is established. The horizontal axis of this map is defined as the normalized voltage axis, and the vertical axis is defined as the normalized current axis. The data is scanned point by point, with the normalized voltage value at each moment used as the x-axis and the normalized current value as the y-axis to locate the specific pixel coordinates in the matrix. For example, if the normalized voltage is 0.75 and the normalized current is 0.65 at a certain moment, the pixel is located at (768, 665) in the matrix. After location, the color attribute of the pixel is modulated based on the normalized instantaneous power value at that moment, and a color image data frame is constructed using RGB encoding format. The specific rules are as follows: the normalized power value is multiplied by 255 and rounded to obtain the component value of the red channel (R); the normalized power change rate at that moment is multiplied by 255 and rounded to obtain the component value of the green channel (G); and the synchronization phase difference index of voltage and current is mapped to the component value of the blue channel (B). For blank areas in the matrix that are not matched by data points, a black background is filled (RGB value 0, 0, 0). Through this point-by-point mapping and color encoding, the originally abstract time series data is transformed into an image of power output with specific texture and color distribution. Each colored track in the image represents the inverter's operating state within one or more cycles. The smoothness of the track reflects the waveform quality, and the intensity of the color directly characterizes the distribution of power output strength.

[0033] Please refer to Figure 3. The specific steps of S2 are as follows: S201: Based on the power output image, extract the time axis signal sequence of the corresponding voltage change curve, and divide the voltage amplitude of the periodic segment in the sequence into intervals. Then, compare the positions of the periodic peaks and troughs in each interval. If the peak interval exceeds the preset voltage fluctuation period threshold, the interval is marked as a non-periodic segment, and a voltage offset interval sequence is established. First, parse the image metadata to backtrack and extract the original voltage change curve time axis signal sequence. Perform periodic analysis on this sequence, and cut the continuous time axis signal into several independent periodic intervals based on the positive zero crossing point of the voltage signal. Then, measure the time position of the voltage waveform peak (maximum value of the positive half-cycle) and trough (minimum value of the negative half-cycle) in each interval. To determine whether the periodicity of the interval is normal, calculate the time difference between the occurrence times of two adjacent periodic peaks. Set the voltage fluctuation period threshold according to the power grid operation standard: the standard power grid frequency is 50 Hz, corresponding to a standard period of 20 milliseconds. Considering that the allowable frequency deviation of grid-connected inverters is typically within ±0.5 Hz, i.e., a cycle deviation within ±0.2 milliseconds, a more lenient monitoring threshold range is set to detect potential islanding precursors or grid transient disturbances. The threshold is set to ±5% of the standard cycle, meaning the cycle length is limited to between 19 and 21 milliseconds. The calculated peak intervals are compared one by one. For example, if the time interval between two adjacent peaks is found to be 21.5 milliseconds, exceeding the preset 21-millisecond limit, it is immediately determined that the voltage waveform within that time period has experienced a frequency shift or phase jump. The start and end points of the time period to which this abnormal waveform belongs (e.g., 0.500 seconds to 0.522 seconds) are marked and stored as an independent record in the voltage offset interval sequence. This sequence not only records the time window of the anomaly but also includes the specific value of the offset, providing a precise time index for subsequent current behavior analysis of specific fault segments.

[0034] S202: The voltage offset interval sequence is invoked, and continuous data segments of the current change curve within the same time period are extracted synchronously. Using the time index as a reference, a one-to-one matching process is performed between the voltage and current sequences. The current data within the offset interval is integrated into the voltage offset structure, generating a voltage-current aligned structure dataset. All time segments marked as non-periodic or frequency-offset are locked. For each marked interval (e.g., 0.500 seconds to 0.522 seconds), the original high-frequency sampling database is returned to, and the corresponding current change curve data segment within that time period is retrieved and extracted using the time index. At this point, a strict time alignment operation is performed, using the timestamp of the voltage sequence as a reference, and each sampling point in the current sequence is matched one-to-one with the voltage point. Since the voltage has shifted (e.g., the period has lengthened), the current waveform usually changes accordingly. The extracted current data (including the instantaneous current value sequence, current peak value, and current phase angle) are fully integrated into the voltage offset structure. This process constructs a voltage-current aligned structure dataset containing three-dimensional information: "abnormal time window - distorted voltage waveform - response current waveform." In this dataset, each set of data fully preserves the voltage-current coupling characteristics at the fault moment. For example, for the aforementioned 21.5 millisecond voltage offset period, the dataset clearly records whether the current waveform elongates synchronously with the voltage during this extended period, or whether it maintains its original frequency, resulting in phase disorder. This structured alignment process simplifies massive amounts of continuous monitoring data into a dataset containing only key anomaly features, effectively isolating interference from normal operating data and ensuring that subsequent image encoding and feature extraction can focus on the inverter's dynamic response behavior under grid fluctuations.

[0035] S203: Based on the voltage-current aligned structure dataset, perform two-dimensional grid encoding on the voltage and current combination data for each offset interval, and set numerical level mapping according to the data amplitude to construct an image frame sequence consistent with the original time order, obtaining the offset feature image; initiate the two-dimensional grid encoding process for the voltage and current combination data of each offset interval. Construct an independent local coordinate grid, setting the grid resolution to 256x256, with the horizontal axis representing the normalized voltage within the interval and the vertical axis representing the normalized current. Traverse each pair of voltage and current values ​​within the offset interval, project them onto the grid coordinate system, and count the number of data points falling into each grid cell. To generate the offset feature image, set grayscale and color mapping rules based on the distribution density and amplitude characteristics of the data points. First, calculate the grayscale value based on the point density within the grid cell; the higher the density, the larger the grayscale (upper limit 255), which makes areas with longer waveform dwell times appear brighter in the image, intuitively reflecting the main trajectory of the waveform. Secondly, color levels are assigned based on the power factor characteristics of the data within the specified range: if the voltage and current phases are essentially synchronized within the offset range (power factor close to 1), it is mapped to white; if there is significant inductive lag or capacitive lead, it is mapped to red or blue, respectively. Following the chronological order of the original offsets, these encoded two-dimensional grid images are stitched together into a continuous sequence of image frames. For example, a 0.1-second grid voltage dip event may contain five consecutive offset cycles, and the data from these five cycles are used to generate five offset feature images. These image frames not only geometrically display the VI trajectory (such as hysteresis loops, ellipses, or chaotic scatter plots), but also reveal the power flow and impedance properties during the fault through color coding, ultimately obtaining a complete set of offset feature images.

[0036] Please refer to Figure 4. The specific steps of S3 are as follows: S301: Based on the offset feature image, extract pixel regions with power channel brightness higher than a set threshold as candidate regions for power output trajectories. Combined with the image coordinate mapping relationship, obtain the voltage and current coordinates corresponding to the pixels, i.e., the VI plane position. Simultaneously extract the pixel paths of voltage and current trajectories, and identify the coupling intersection points of power trajectories and voltage and current trajectories in the coordinate space to obtain a set of trajectory intersection segments. First, use the Canny edge detection operator to identify the pixel regions with the highest brightness in the image. These regions represent power output trajectories that are likely to occur. According to the coordinate mapping rules (horizontal axis voltage, vertical axis current) when the image is built, reversely analyze the coordinate meaning of the bright pixels to obtain the power index position path in each frame of the image. Next, in the same frame of the image, perform projection analysis along the horizontal and vertical axes respectively to extract the pixel index paths of voltage change trajectories (voltage amplitude distribution characteristics) and current change trajectories (current amplitude distribution characteristics). Call these three sets of path data (power path, voltage path, current path) to perform overlapping interval retrieval in the normalized coordinate space. Specifically, the algorithm identifies regions where the three paths are spatially close (e.g., the Euclidean distance is less than 5 pixels). These overlapping regions typically correspond to phase points with high voltage and current amplitude matching and superior energy transfer efficiency. The set of pixel coordinates satisfying the overlap condition is recorded to form a set of trajectory intersection segments. For example, in the upper right quadrant of the image (near the peak of the positive half-cycle), the power trajectory often intersects with the voltage and current trajectories. The starting and ending coordinates of this intersection region are precisely recorded as a spatial window for subsequent symmetry analysis.

[0037] S302: Based on the set of trajectory intersection segments, and using image coordinates as a reference, calculate the spatial overlap between the power trajectory path and the voltage trajectory path within each intersection segment. Compare this overlap with the overlap between the power trajectory and the current trajectory within the same region. Select intersection segments where the difference in spatial overlap between the power trajectory and the voltage trajectory, and between the power trajectory and the current trajectory, is less than a preset threshold, thus obtaining a set of synchronously coupled intersection segments. Using image pixel coordinates as a reference, quantitatively analyze the spatial overlap between each trajectory. For each intersection segment, first calculate the Jaccard similarity coefficient between the pixel set of the power trajectory path and the pixel set of the voltage trajectory path, i.e., the number of pixels in the intersection of the two sets divided by the number of pixels in the union, to obtain the power-voltage overlap index. Similarly, calculate the power-current overlap index between the power trajectory and the current trajectory. To select segments with excellent synchronous coupling characteristics, a strict symmetry condition is set: calculate the absolute value of the difference between the above two overlap indices and compare it with a preset symmetry threshold. The threshold was set based on statistical data from a large number of normal operating conditions: under ideal sinusoidal conditions, the overlap between PV and PI should be highly consistent with minimal difference. The overlap difference of 1000 sets of normal waveform data was statistically analyzed, yielding an average difference of 0.02 and a standard deviation of 0.01. Based on the principle of three times the standard deviation, the symmetry threshold was set to 0.05 (i.e., 0.02 plus 3 multiplied by 0.01). If the overlap difference of a certain cross section is less than 0.05, it is determined that the voltage and current in that section maintain a high degree of symmetry and synchronization, and it is considered a valid coupled output section. Conversely, if the difference exceeds 0.05, it is considered that the section has significant nonlinear distortion or impedance mismatch and is discarded. All sections that passed the screening were summarized to obtain the set of synchronously coupled cross sections. Table 1 shows the calculation results for some sections.

[0038] Table 1. Analysis of Overlap and Symmetry in Trajectory Intersection Sections

[0039] As shown in Table 1, Seg_001 and Seg_002 were retained because the difference was very small, while Seg_003 was removed because the difference was too large, and finally the optimized set of synchronous coupling cross segments was generated.

[0040] S303: The synchronous coupling cross-segment set is invoked to perform region aggregation on the image regions where the cross-segments are located. A two-dimensional continuous coordinate mapping is constructed based on the continuous time index. The aggregation result is re-encoded into an image pixel matrix to establish a structural alignment image. Image region aggregation processing is then performed on these discrete cross-segments. Since the original cross-segments may be discontinuous sets of pixels, morphological dilation and closing algorithms are applied to connect adjacent cross-points into continuous closed regions, forming coupling domains with a certain area. Next, a two-dimensional continuous coordinate mapping relationship is established. Using the continuous time index of the original data, these aggregated spatial regions are mapped back to the time axis to clarify the time period corresponding to each coupling domain. Finally, the image pixel matrix is ​​reconstructed based on the aggregation result. A new blank image is created, the background pixels are set to low grayscale values ​​(e.g., 30), and the pixels corresponding to the identified coupling regions are set to bright green (RGB values ​​0, 255, 0). The brightness saturation is finely adjusted according to the time sequence to establish a structural alignment image. This image visually demonstrates which phase intervals (such as peaks or zero crossings) of the inverter maintain the optimal voltage-current synchronization coupling structure throughout the entire operating cycle, providing a standard reference template for subsequent capacity assessment.

[0041] Please refer to Figure 5. The specific steps of S4 are as follows: S401: Based on the structural alignment image, extract the pixel path corresponding to the power output trajectory within the image. Using the coupling distribution path in the image as a reference, calculate the coordinate difference sequence between the pixel points on the power path and the corresponding coupling path points. Then, perform statistical analysis on the mean and maximum difference magnitude of the difference sequence to obtain the path fitting deviation index set. Extract the pixel path sequence of the actual power output trajectory within the image, and simultaneously extract the confirmed high-quality coupling distribution path (i.e., the center line of the highlighted area in the structural alignment image) as a reference. Calculate the Euclidean distance between the pixel points on the actual power path and the corresponding points on the reference coupling path point by point. For each point on the path, search for the closest point in the reference path and calculate the coordinate difference between them. After traversing the entire path, obtain a difference sequence containing all point deviation values. Perform statistical analysis on this sequence to calculate the arithmetic mean and maximum difference magnitude of the differences. For example, in the analysis of a certain trajectory segment, the calculated average deviation is 1.5 pixels, and the maximum deviation is 4.2 pixels. These two values ​​constitute the path fitting deviation index set. The average deviation reflects the overall fitting accuracy of the inverter output waveform, while the maximum deviation captures the degree of local abrupt changes or distortions in the waveform (such as glitches). This set of metrics quantifies graphical waveform differences into specific numerical indicators and is a key intermediate parameter for evaluating inverter output quality.

[0042] S402: Based on the path fitting deviation index set, retrieve continuous segments in the coupled distribution path where the difference fluctuation is less than a preset difference stability threshold, and obtain continuous pixel strings along the segment direction. Calculate the linear extension path in the direction of each pixel, integrate adjacent extension path segments, and generate a linear combination fitting segment set; perform refined screening and reconstruction of the coupled distribution path. Set the difference stability threshold to 2.0 pixels, which corresponds to the allowable waveform distortion rate at the image resolution (approximately 0.2% full-scale deviation). Traverse the difference sequence and retrieve continuous segments where the difference fluctuation is consistently less than 2.0 pixels. If more than 50 consecutive points in the sequence meet this low deviation condition, extract this string of pixels. Subsequently, perform linear extension processing on the extracted pixel string: calculate the local tangent direction at the location of each point, and linearly extend forward 5 pixel units along the tangent direction. Use a weighted average method to smoothly connect adjacent extension path segments, eliminating the bends at the connection points, and generate a linear combination fitting segment set. This step essentially removes segments with high noise and poor fit from the waveform, retains and repairs the output segment with high linearity and high stability, and constructs an ideal trajectory that represents the true effective capacity of the inverter.

[0043] S403: The linear combination fitted segment set is invoked. Using the distribution path direction as the arrangement axis, all fitted segments are pixel-aligned and encoded according to their temporal order and coordinate position. The combined result is mapped to a two-dimensional image frame structure to establish a capacity label image. The construction of the capacity label image is initiated. Using the distribution path direction (i.e., the phase angle direction in the VI phase plane) as the arrangement axis, all fitted segments are pixel-aligned and encoded according to their temporal order and coordinate position. A new image frame is created, and the fitted segments are mapped into the frame. To visually identify the capacity level, different color codes are assigned based on the length and linearity of the fitted segments: long segments exceeding 100 pixels in length and with minimal fitting residuals are marked in green, representing high capacity support capability; shorter segments or those with slightly lower linearity are marked in yellow. Finally, all encoded fitted segments are combined and mapped into a two-dimensional image structure to establish the capacity label image. The image visually reveals the inverter's load driving capability at different phase angles through color distribution: the more continuous and wider the green area, the fuller the inverter's output capacity and the stronger its waveform control capability. After generating the capacity identification image, comprehensive calculations are performed on the coupling path integrity, trajectory fitting deviation, and fitting segment linear continuity to form a capacity assessment parameter set, and the inverter capacity assessment result is output based on a preset capacity level range. The actual capacity percentage is inferred by back-calculating the coverage of the linear fitting segment and the degree of coupling path matching in the capacity identification image to obtain the inverter's actual output capacity. After the structural alignment image is generated, this invention introduces a fast capacity back-calculation mechanism. This method includes the following steps: controlling the active power output of the photovoltaic inverter to 10% of its current maximum operating capacity; obtaining the output value P1 under this power output state through a voltage / current measuring device; and calculating the inverter's rated capacity according to the following formula: Where: P1: the actual power value when the inverter is controlled at 10% of its rated output; C: the estimated rated capacity of the inverter.

[0044] Assuming stable light, temperature, and grid conditions, the output power and control percentage are approximately linearly related, which can be used for rapid on-site capacity estimation or as a reference for image-based capacity assessment. Before using the image method to assess capacity, the photovoltaic inverter control module is set to limit the output to 10% of the rated value, that is, the control system sends a command to the inverter to reduce the active power to 10%.

[0045] The output power value is recorded as 23.5kW at this time, and calculated using the following formula: Comparing this value with the evaluation results of the image method, the deviation is less than 1%, verifying the feasibility and accuracy of this method in the field evaluation scenario.

[0046] Please refer to Figure 6. The specific steps of S5 are as follows: S501: Call the capacity identification image, extract the pixel coordinate sequences of the power output trajectory, voltage change curve, and coupling structure coverage path in the image, and perform matching statistics on the overlapping pixels in the three sets of coordinate sequences to calculate the coordinate overlap rate and trajectory coverage value among the three, and obtain the trajectory overlap index set; extract the three sets of pixel coordinate sequences of the actual power output trajectory, voltage change curve, and processed coupling structure coverage path in the image respectively, perform pixel-level matching statistics, find the pixels whose spatial positions completely overlap in these three sets of coordinate sequences, and calculate the coordinate overlap rate among the three, that is, the number of pixels with three overlapping lines divided by the total number of pixels of the actual power trajectory. For example, if the total number of pixels is 10,000, and the number of pixels with three overlapping lines is 9,440, then the coordinate overlap rate is 94.4%. At the same time, calculate the trajectory coverage value, that is, the length ratio of the coupling structure path in the effective power output area, and use the calculated overlap rate value and coverage value as core parameters to form the trajectory overlap index set. This set of indicators precisely quantifies the uniformity of voltage, current, and power in the spatiotemporal distribution at the microscopic pixel overlap level, and is the core basis for evaluating the dynamic performance and capacity effectiveness of inverters.

[0047] S502: Based on the trajectory overlap index set, the index values ​​are mapped to a two-dimensional coordinate region, aggregated according to the time index of the power trajectory, and the power, voltage, and coupling path overlap level within each time period are combined as vectors to construct region labels, generating an output intensity distribution matrix. The monitoring period is divided into several evaluation sub-intervals (e.g., every 20 milliseconds is a cycle), and the overlap rate, coverage, and stability score within each interval are constructed as performance evaluation vectors. An output intensity distribution matrix is ​​generated, with rows and columns corresponding to time slices, and cells storing the aforementioned evaluation vectors. For example, the vector data for the nth time slice is [0.944, 0.98]. This matrix digitally records the performance fluctuations of the inverter during continuous operation in detail, accurately pinpointing the specific time points of performance degradation.

[0048] S503: Calls the output intensity distribution matrix, performs a comprehensive traversal of the intensity values, spatial span, and combined structure of the marked areas in the matrix, and outputs the area level rating and classification based on the preset capacity level range standard, establishing the photovoltaic inverter output test and capacity assessment results; performs a comprehensive traversal and evaluation of the intensity values ​​(overlap rate), spatial span (coverage), and combined structure of all marked areas in the matrix. Classification is based on the preset capacity level range standard. The standard is set as follows: If the overlap rate is greater than or equal to 95% and the coverage is greater than or equal to 95%, it is rated A (Excellent), representing full output and no capacity loss; if the overlap rate is between 90% and 95% and the coverage is greater than 90%, it is rated B (Good), representing normal output with minor deviations; if the overlap rate is between 80% and 90%, it is rated C (Medium); below 80%, it is rated D (Poor). The overlap rate of 94.4% calculated in S501 was compared with the standard. Since 94.4% falls within the 90% to 95% range, the output performance for this time period was rated as B. The rating results for all time segments were statistically analyzed. If the total percentage of A and B grades exceeded 98%, the final conclusion was generated: "The photovoltaic inverter output test is qualified, and the capacity assessment is good." This result indicates that although the tested inverter did not achieve a theoretically perfect synchronization state under the current operating conditions, its 94.4% overlap rate shows that its dynamic response performance remains relatively stable, with only a very small phase deviation, fully meeting the safety standards for grid-connected operation.

[0049] To further verify the accuracy of the capacity assessment results, this method introduces a back-calculation mechanism based on 10% of the inverter's rated output power as a control. The specific operation is as follows: the active power output of the photovoltaic inverter is controlled to 10% of its current maximum operating capacity, and the actual power output value under this state is obtained through voltage / current measurement devices. Record during testing. Based on this, the estimated rated capacity of the inverter can be calculated: Where C is the estimated rated capacity. This method is based on the premise that temperature, illumination, and power grid conditions are relatively stable, and assumes that the output power and the control ratio are approximately linearly related. It can be used for rapid on-site estimation of capacity assessment and as a reference for comparison with the results of the graphical method.

[0050] The comparison showed that the deviation between this estimated value and the evaluation result of the image recognition method was less than 1%, which effectively verified the feasibility and accuracy of this method in the on-site capacity assessment scenario, and further corroborated the reliability of the level assessment based on the output intensity distribution matrix.

[0051] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for testing the output and evaluating the capacity of a photovoltaic inverter, characterized in that, Includes the following steps: S1: The voltage and current measuring devices are used to perform synchronous measurements on the AC output terminal of the photovoltaic inverter, and the current change curve, voltage change curve and instantaneous power trajectory are collected over a continuous time period. The measurement results are then combined and mapped to generate an energy output image. S2: Based on the power output image, analyze the voltage change curve, extract the non-periodic offset segment, and perform structural alignment with the current change trajectory in the same time period to construct an offset feature image; S3: Based on the offset feature image, compare the intersection paths of the power change trajectory with the voltage change trajectory and the current change trajectory, extract the coupling distribution structure with synchronous characteristics, and output the structure alignment image; S4: Based on the structure alignment image, calculate the fitting difference between the power output trajectory and the coupling path, perform linear expansion along the coupling trajectory, arrange and combine the fitting segments, and generate a capacity identification image; After generating the capacity identification image, further perform comprehensive calculations on the integrity of the coupling path, the trajectory fitting deviation, and the linear continuity of the fitting segment to form a capacity evaluation parameter set, and output the inverter capacity evaluation result based on the preset capacity level range; The actual capacity percentage is inferred by the coverage of the linear fitting segment and the degree of matching of the coupling path in the capacity identification image to obtain the actual output capacity of the inverter; S5: Based on the capacity identification image, determine the overlap relationship between the power trajectory, voltage curve and coupling coverage path, express the output intensity distribution of the photovoltaic inverter in a combined form, and output the photovoltaic inverter output test and capacity evaluation results.

2. The photovoltaic inverter output testing and capacity assessment method according to claim 1, characterized in that, The power output image includes current change characteristics, voltage fluctuation characteristics, and instantaneous power characteristics. The offset feature image includes voltage aperiodic offset characteristics, current alignment structure characteristics, and time synchronization characteristics. The structure alignment image includes coupling distribution structure characteristics, path intersection position characteristics, and synchronization alignment relationship characteristics. The capacity identification image includes trajectory fitting difference characteristics, path linear expansion characteristics, and capacity combination distribution characteristics. The photovoltaic inverter output test and capacity evaluation results include output intensity distribution characteristics, power trajectory matching characteristics, and structure path coverage characteristics.

3. The photovoltaic inverter output testing and capacity assessment method according to claim 1, characterized in that, The non-periodic offset segment refers to the continuous fluctuation section in the voltage change curve that deviates from the normal periodic fluctuation pattern, and analyzes the impact of potential anomalies and dynamic loads.

4. The photovoltaic inverter output testing and capacity assessment method according to claim 1, characterized in that, The coupled distribution structure with synchronous characteristics refers to the cross-path structure that exhibits temporal consistency and morphological coordination characteristics in the power, voltage and current change trajectories.

5. The photovoltaic inverter output testing and capacity assessment method according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Based on the AC output terminal of the photovoltaic inverter, the voltage measurement device and the current measurement device are called to synchronously detect the output signal within a continuous time period of the output terminal, extract the periodic change values ​​of the AC voltage signal and the AC current signal respectively, and pair the instantaneous voltage value and the instantaneous current value at the time point to obtain a voltage and current synchronization data frame; S102: According to the voltage and current synchronization data frame, the instantaneous voltage value and the instantaneous current value in each time segment are multiplied point by point to construct the instantaneous power value sequence of the corresponding time segment, and the value variation range between adjacent sample points in the sequence is continuously measured to generate an instantaneous power change trajectory sequence; S103: The voltage and current synchronization data frame and the instantaneous power change trajectory sequence are called to perform dimensional alignment and numerical normalization processing on the data, establish a two-dimensional joint matrix map in the corresponding time segment, and construct an image data frame based on the matrix map to obtain the power output image.

6. The photovoltaic inverter output testing and capacity assessment method according to claim 1, characterized in that, The specific steps of S2 are as follows: S201: Based on the power output image, extract the time axis signal sequence of the corresponding voltage change curve, and divide the voltage amplitude of the periodic segment in the sequence into intervals. Then, compare the positions of the periodic peaks and troughs in each interval. If the peak interval exceeds the preset voltage fluctuation period threshold, mark the interval as a non-periodic segment and establish a voltage offset interval sequence; S202: Call the voltage offset interval sequence, synchronously extract the continuous data segments of the current change curve within the same time period, and perform one-to-one matching processing on the voltage sequence and current sequence based on the time index. Integrate the current data in the offset interval into the voltage offset structure to generate a voltage-current aligned structure dataset; S203: According to the voltage-current aligned structure dataset, perform two-dimensional grid encoding on the voltage and current combination data of each offset interval, and set the numerical level mapping according to the data amplitude to construct an image frame sequence consistent with the original time order and obtain the offset feature image.

7. The photovoltaic inverter output testing and capacity assessment method according to claim 1, characterized in that, The specific steps of S3 are as follows: S301: Based on the offset feature image, extract pixel regions with power channel brightness higher than a set threshold as candidate regions for power output trajectory. Combine the image coordinate mapping relationship to obtain the voltage and current coordinates corresponding to the pixels, i.e., the VI plane position. Simultaneously extract the pixel paths of voltage trajectory and current trajectory, and match density overlap regions in the coordinate space to identify the coupling intersection points of power trajectory and voltage and current trajectory, and obtain a set of trajectory intersection segments; S302: Based on the set of trajectory intersection segments, using image coordinates as a reference, calculate the spatial overlap between power trajectory path and voltage trajectory path in each intersection segment, and compare it with the overlap between power trajectory and current trajectory in the same region. Filter the intersection segments where the difference in spatial overlap between power trajectory and voltage trajectory, and between power trajectory and current trajectory is less than a preset threshold to obtain a set of synchronously coupled intersection segments; S303: Call the set of synchronously coupled intersection segments, perform region aggregation on the image region where the intersection segment is located, and construct a two-dimensional continuous coordinate mapping based on continuous time index. Re-encode the aggregation result into an image pixel matrix to establish a structural alignment image.

8. The photovoltaic inverter output testing and capacity assessment method according to claim 1, characterized in that, The specific steps of S4 are as follows: S401: Based on the structure alignment image, extract the pixel path corresponding to the power output trajectory in the image, and use the coupling distribution path in the image as a reference to calculate the coordinate difference sequence between the pixel point on the power path and the corresponding coupling path point. Then, perform mean and maximum difference amplitude statistics on the difference sequence to obtain the path fitting deviation index set; S402: According to the path fitting deviation index set, retrieve continuous segments in the coupling distribution path where the difference fluctuation is less than the preset difference stability threshold, and obtain continuous pixel point strings along the segment direction. Calculate the linear extension path in the direction of each pixel point, integrate adjacent extension path segments, and generate a linear combination fitting segment set; S403: Call the linear combination fitting segment set, use the distribution path direction as the arrangement axis, perform pixel alignment encoding on all fitting segments according to time order and coordinate position order, and map the combination result to the two-dimensional image frame structure to establish a capacity identification image.

9. The photovoltaic inverter output testing and capacity assessment method according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: Call the capacity identification image, extract the pixel coordinate sequence of the power output trajectory, voltage change curve and coupling structure coverage path in the image, and perform matching statistics on the overlapping pixels in the three sets of coordinate sequences to calculate the coordinate overlap rate and trajectory coverage value among the three, and obtain the trajectory overlap index set; S502: According to the trajectory overlap index set, map the index values ​​to the two-dimensional coordinate region, perform aggregation processing according to the time index of the power trajectory, and use the power, voltage and coupling path overlap level in each time period as a vector combination to construct the region label and generate the output intensity distribution matrix; S503: Call the output intensity distribution matrix, perform overall traversal processing on the intensity value, spatial span and combination structure of the labeled region in the matrix, and perform region level evaluation and classification output according to the preset capacity level interval standard to establish the photovoltaic inverter output test and capacity evaluation results.

10. The photovoltaic inverter output testing and capacity assessment method according to claim 6, characterized in that, The voltage fluctuation period threshold refers to the upper limit of a period of time for judging whether the voltage signal remains periodic. If the interval between adjacent peaks exceeds this threshold, it is considered that the voltage signal has serious non-periodic fluctuations.

Images