A method, apparatus and medium for automatic processing and chart generation of test bench data

Abstract

The application belongs to the field of fuel cell testing, and discloses an automatic processing and chart generation method, device and medium for test bench data. The method comprises the following steps: (1) data reading and preprocessing; (2) target data positioning and extraction; (3) steady-state value calculation; (4) data arrangement and classification; and (5) automatic generation and export of charts. The application realizes one-key or script-based data extraction, calculation and chart generation, so as to improve efficiency, ensure accuracy, standardize output and liberate labor.

Description

Technical Field

[0001] This invention belongs to the field of fuel cell testing, specifically relating to a method, apparatus, and medium for automatic processing and chart generation of test bench data. Background Technology

[0002] With the development of industries such as new energy vehicles and energy storage, the performance testing requirements for core components such as batteries and motors are becoming increasingly stringent and complex. Test benches are key equipment for conducting tests such as standard operating conditions, durability, and performance mapping, and the data they generate are important bases for product development and verification.

[0003] After testing, the benchtop software typically exports the raw data as structured or semi-structured Excel files. Subsequent data analysis and visualization heavily rely on manual operation. Technicians need to open multiple Excel files, extract target data by searching, filtering, copying and pasting, and then manually plot curves in Origin, MATLAB, or Excel's built-in charting functions. The entire process is time-consuming and labor-intensive, and it is difficult to ensure the consistency of batch processing results.

[0004] The closest current implementation is a "manual operation mode" that relies entirely on manual labor. The specific steps are as follows:

[0005] a. Data location and extraction: Technicians open the Excel file exported from the test bench and, based on test logs or experience, locate the target test point (such as a discharge data block at a certain rate under a certain SOC state) in hundreds or even thousands of rows of data by visual inspection or simple Excel filtering functions.

[0006] b. Data calculation: After finding the target data column (such as voltage or current series), scroll to the bottom of the column, manually select the last 12 cells, view the average value using Excel's AVERAGE function or the status bar, and manually record the result to another summary table.

[0007] c. Chart Plotting: After calculating and summarizing the average values ​​for all required points, in Excel or professional charting software, select the "Insert Chart" function, manually select the data range (X-axis for example, "current density", Y-axis for example, "voltage"), and add data series one by one to plot, such as a "polarization curve". Plotting "sensitivity plots", "power plots", etc., requires repeating this process and may involve additional formula calculations (e.g., power = voltage × current).

[0008] d. Chart formatting: Manually adjust the generated charts, including titles, axis labels, legends, line styles, colors, etc., to make the charts meet the publication requirements of reports or papers.

[0009] In current technologies, processing a single file containing dozens of test points can take several hours. Batch processing of multiple files increases the time cost exponentially, resulting in extremely low efficiency. Manually searching, selecting, copying, pasting, and inputting data is highly prone to errors such as omissions, incorrect row / column selections, and miscalculations, affecting the accuracy of the analysis results. Furthermore, repeating almost identical procedures after each test is tedious and wastes highly skilled human resources. In addition, different personnel, or even the same person processing data at different times, may use different chart styles, color schemes, and annotation methods, leading to inconsistent chart styles in the final report, affecting professionalism and readability. Summary of the Invention

[0010] This invention overcomes the aforementioned shortcomings by providing a method, apparatus, and medium for automatic processing and chart generation of test bench data. It enables one-click or script-based data extraction, calculation, and chart generation, thereby improving efficiency, ensuring accuracy, standardizing output, and freeing up manpower.

[0011] The technical solution of this invention is as follows.

[0012] (1) Data reading and preprocessing

[0013] Read the Excel file exported from the specified test bench using Python's pandas library or openpyxl library;

[0014] By identifying the data table structure in the file, the worksheet or data area storing the original time-series data can be located.

[0015] (2) Target data location and extraction

[0016] Users can specify the "test points" to be analyzed through configuration files or predefined rules within the program. These points can be located using specific column label combinations (such as Test_ID, Step_Index) or row index patterns.

[0017] For each specified point, the program automatically locates the end of the data in its corresponding data column (such as the voltage column and the current column).

[0018] (3) Calculation of steady-state value

[0019] The program starts from the end of each target data column and extracts a fixed number of data points upwards (preferably the last 12 data points). This number (12) can be configured as a parameter.

[0020] The program calls the pandas or numpy mean calculation function (such as .mean()) to calculate the arithmetic mean of the 12 extracted values, and obtains an "eigenvalue" that represents the steady-state level of the point under the given parameters.

[0021] The calculation results (e.g., steady-state voltage at point A, steady-state current at point B) are automatically stored in a newly created DataFrame or dictionary data structure and organized according to data type (voltage, current, temperature, power, etc.).

[0022] (4) Data processing and classification

[0023] The program will organize all the calculated characteristic values ​​according to their physical meaning and subsequent plotting requirements. For example, it will pair the "current density" of all points with the corresponding average "voltage" to prepare for plotting polarization curves; and it will pair the "change in a certain parameter" with the corresponding average "voltage" to prepare for plotting sensitivity maps.

[0024] (5) Automatic chart generation and export

[0025] The program calls Python plotting libraries (such as matplotlib or seaborn) to automatically generate various analytical charts based on pre-designed templates.

[0026] Polarization curve: Automatically plots a curve with data points, using average current density as the X-axis and average voltage as the Y-axis.

[0027] Sensitivity plot: A curve is plotted with a certain input parameter (such as temperature) as the X-axis and a performance index (average voltage) as the Y-axis.

[0028] Power diagram: Based on the calculated average voltage and average current, calculate the average power at each point and plot the power vs. current curve.

[0029] During the chart generation process, the program automatically adds a uniformly formatted title, axis labels, legend, grid lines, and sets standardized line styles, colors, and markers.

[0030] The generated high-quality charts (such as PNG, SVG, or PDF formats) are automatically saved to a user-specified directory. Additionally, the compiled feature value summary table can be exported as a new Excel or CSV file.

[0031] A device for automatically processing and generating charts from test bench data, as shown in the attached diagram. Figure 5 As shown, it includes:

[0032] The data reading and preprocessing module 102 is used to read the Excel file exported from the test bench, identify the data table structure, and locate the worksheet or data area storing the original time series data.

[0033] The target data location and extraction module 104 is used to specify test points according to predefined rules, locate the end of the data in the corresponding data column, and obtain the row index of the steady-state end.

[0034] The steady-state value calculation module 106 is used to extract a fixed number of M data points from the end of each test point, calculate the arithmetic mean as the steady-state characteristic value of that point, and organize and store it according to data type.

[0035] The data processing and classification module 108 is used to pair and process all feature values ​​according to their physical meaning and plotting requirements, generating input data pairs suitable for different charts.

[0036] The automatic chart generation and export module 110 is used to call the plotting library, automatically generate high-quality charts such as polarization curves, sensitivity plots, and power plots according to preset templates, and save them in a specified format, while exporting a summary table of characteristic values.

[0037] A computer-readable storage medium storing computer instructions for causing a processor to execute the above-described automatic processing and chart generation method.

[0038] Compared with the prior art, the advantages of the present invention are:

[0039] 1. Significantly improved efficiency: The program automatically executes tasks that reduce manual operations of several hours to minutes or even seconds, making it especially suitable for repetitive processing of large amounts of data.

[0040] 2. Extremely high accuracy: It avoids all possible selection, input and calculation errors in manual operation, ensuring the consistency and reliability of data analysis results.

[0041] 3. Standardized output: All generated charts follow the same template, ensuring a high degree of consistency and professionalism in the format and style of the output results under multiple rounds of testing and multi-person collaboration.

[0042] 4. Flexible and configurable: Users can easily adjust data extraction rules (such as changing the value of M) and select different chart combinations to adapt to changing analysis needs by modifying configuration files or program parameters.

[0043] 5. Liberate technical personnel: Free technical personnel from tedious and repetitive data processing work, allowing them to focus more on high-value test design, results analysis, and problem discovery. Attached Figure Description

[0044] Figure 1 This is a flowchart of an automatic processing and chart generation method for test bench data according to the present invention.

[0045] Figure 2 This is a processed graph of the fuel cell stack polarization curve test data from Example 1;

[0046] Figure 3 This is a graph showing the data processing of the cathode metering ratio sensitivity test in Example 2 of the fuel cell stack.

[0047] Figure 4 This is a graph showing the data processing of the fuel cell stack cathode humidity sensitivity test in Example 3;

[0048] Figure 5 A device for automatically processing and generating charts from test bench data. Detailed Implementation

[0049] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto. For process parameters not specifically noted, conventional techniques can be referred to.

[0050] An automatic processing and chart generation method for test bench data, as described in this embodiment, includes the following steps:

[0051] (1) Data reading and preprocessing

[0052] Read the Excel file exported from the specified test bench using Python's pandas library or openpyxl library;

[0053] By identifying the data table structure in the file, the worksheet or data area storing the original time-series data can be located.

[0054] (2) Target data location and extraction

[0055] The test points to be analyzed are specified through configuration files or predefined rules within the program; the test points are located by specific combinations of column labels or row index patterns.

[0056] (3) Calculation of steady-state value

[0057] Starting from the end of each target data column, extract a fixed number of data points upwards, where the fixed number of data points is a preset positive integer M; configure these fixed number of data points as parameters, that is, use the last M data points at the end of the test point;

[0058] The arithmetic mean of the extracted M data points is calculated using the NumPy numerical computing library in Python. This yields an "eigenvalue" representing the steady-state level of that point under specific parameters. The formula for calculating the arithmetic mean is as follows:

[0059]

[0060] In the formula, is the eigenvalue; M is the number of selected steady-state data points; k is the row index of the steady-state end of the test point located in step (2); The value of the data column corresponding to the i-th row;

[0061] The calculation results, namely the steady-state voltage, steady-state current and other characteristic values ​​of each point, are automatically stored in a newly created DataFrame or dictionary data structure and organized according to data type.

[0062] (4) Data processing and classification

[0063] All calculated eigenvalues ​​are organized according to their physical meaning and subsequent drawing requirements;

[0064] (5) Automatic chart generation and export

[0065] The system calls Python's plotting libraries to automatically generate various analytical charts based on pre-designed templates. The plotting libraries include matplotlib or seaborn. The plots include polarization curves, three-chamber voltage drop plots, power curves, CVM voltage distribution plots, power curves, and sensitivity curves.

[0066] The generated charts are automatically saved to the user-specified directory. Meanwhile, the compiled feature value summary table can also be exported as a new Excel spreadsheet in XLSX or CSV format.

[0067] Preferably, in step (2), the column label combination includes the test number Test_ID and the step index Step_Index; wherein, the test number Test_ID is a globally unique identifier used to distinguish independent test items of different test types and different batches; the step index Step_Index is a local sequence identifier within a single test used to identify test steps of different parameter levels under the same test item;

[0068] The row indexing mode is as follows: First, define a function to find the row index corresponding to each preset test point in the original data. Traverse the preset test point value list (such as different currents, metering ratios, back pressures, etc.), search backwards from the last row of the data, and determine whether the key column value of the current row is within the allowable deviation range of the preset test point. If a row that meets the conditions is found, add the row index to the key_list list. After finding a test point, immediately break out of the inner loop and continue to search for the next test point.

[0069] Preferably, in step (2), the test points are combined into a voltage column (Voltage) or a current column (Current) by a specific combination of column labels or row index patterns.

[0070] Preferably, in step (3), the data types include voltage, current, temperature, and power.

[0071] Preferably, step (4) specifically involves: pairing the "current density" and the corresponding "voltage" average values ​​of all points to prepare for drawing polarization curves; pairing the "variable value of a certain parameter" and the "voltage" average value to prepare for drawing sensitivity diagrams.

[0072] Preferably, in step (5), the high-quality chart is in PNG, SVG or PDF format.

[0073] Preferably, in step (5), the polarization curve is automatically plotted with average current density as the X-axis and average voltage as the Y-axis.

[0074] Sensitivity plot: A curve is plotted with a certain input parameter (such as temperature) as the X-axis and the average change value of a performance index (such as individual unit voltage) as the Y-axis;

[0075] Power diagram: Based on the calculated average voltage and average current, calculate the average power at each point and plot the power versus current curve.

[0076] Preferably, in step (5), during the chart generation process, the program automatically adds a title, axis labels, legend, and grid lines in a uniform format, and sets standardized line styles, colors, and markers.

[0077] The following example demonstrates the processing of polarization test data for a certain type of metal stamped plate fuel cell stack. The raw data was exported from the test bench as an Excel file, containing fields such as time, current, voltage of each cell, hydrogen / air / water pressure, flow rate, and temperature.

[0078] Example 1

[0079] Fuel cell stack polarization curve test data processing:

[0080] (1) Call the pandas data processing core library to read the Excel file and automatically locate the first worksheet.

[0081] (2) Configure the test points as a list of "current values":

[0082] The key column label is "Current (A)", and the search deviation is set to ±1A. The program searches upwards from the end of the file, finding the closest steady-state last row index for each current value. Tests are loaded in ascending order of current, with the steady-state data segment for each current range located at the end of that range's data. Therefore, by traversing the "Current (A)" column backwards from the last line of the data file, the program avoids the transitional data segment and quickly matches the steady-state last row corresponding to each current value. After matching all current values ​​in the list, the program obtains the set of steady-state last row indexes for each test point.

[0083] (3) Configure the number of steady-state points M=12, take the last 12 rows of data for each test point, calculate the arithmetic mean to obtain the characteristic values ​​of each point, including average voltage, average current density, anode and cathode water circuit voltage drop, deviation from the mean, etc.

[0084] (4) Call the NumPy numerical calculation library to calculate the power. Power = average voltage × average current density × activation area × number of individual chips / 1000. At the same time, pair the current density with the corresponding average voltage.

[0085] (5) Use the matplotlib plotting library to generate a comprehensive plot including transformer polarization curves, stack power curves, three-way voltage drop curves, and CVM consistency curves for different voltage densities (see...). Figure 2 The eigenvalues ​​were then saved as PNG images; the organized eigenvalue table was exported as an Excel file.

[0086] Example 2

[0087] Data processing of cathode metering ratio sensitivity test results for fuel cell stacks:

[0088] (1) Call the pandas data processing core library to read the Excel file and automatically locate the first worksheet.

[0089] (2) Configure the test points as a list of "Air side coefficients()":

[0090] This test is a cathode metering ratio sensitivity test under constant current density conditions. During the test, the current value remains constant, and the "Air side coefficient ()" is adjusted gradually. The test points are configured as a list of "Air side coefficient ()": the key column label is specified as "Air side coefficient ()", and the search allowable deviation is set to ±0.05. The program uses a reverse deviation matching mode for location: it iterates backwards from the last line of the data file through the "Air side coefficient ()" column, checking each row to see if the current metering ratio value falls within the allowable deviation range of the target point. The first matching row is recorded as the steady-state end row index for that metering ratio range, and this process is repeated for all metering ratio points.

[0091] (3) Configure the number of steady-state points M=12. For each test point, take the last 12 rows of data and calculate the arithmetic mean to obtain the characteristic value of each point. Since this test is under constant current condition, the current density remains constant. The obtained characteristic values ​​include: constant average current density, average single-cell voltage under the corresponding metering ratio, single-cell voltage deviation from the mean, cathode inlet and outlet voltage drop, etc. The calculation and processing methods of each parameter are as follows:

[0092] Cathode metering ratio: The arithmetic mean of the "Air side coefficient ()" column within the steady-state section is used as the abscissa parameter for that point. Average voltage: The cross-sectional average value is calculated row by row for all individual cell voltage columns within the steady-state section, and then the arithmetic mean of the 12 sets of cross-sectional average values ​​is taken to obtain the average individual cell voltage (CVM-mean) at that point. Individual cell deviation from the mean: The average of the differences between the maximum and minimum values ​​of all individual cell voltages within the steady-state section is converted to mV units and used to characterize the voltage consistency of individual cells in the stack. Cathode voltage drop: The arithmetic mean of the difference between the cathode inlet pressure and the outlet pressure within the steady-state section is taken. The cathode metering ratio is paired with the average voltage, deviation from the mean, and cathode voltage drop to generate a data sequence for plotting.

[0093] (4) Use the matplotlib plotting library to generate dual cathode metrological ratio characteristic plots (see...). Figure 3 The first chart is a sensitivity plot, with the cathode metering ratio as the X-axis, the average voltage as the left Y-axis, and the deviation from the mean as the right Y-axis. The second chart is a cathode voltage drop characteristic plot, with the cathode metering ratio as the X-axis and the cathode voltage drop as the main Y-axis. The charts are automatically updated with standardized titles, axis labels, legends, and gridlines, and saved as PNG images. The compiled characteristic value table is then exported as an Excel file.

[0094] Example 3

[0095] Data processing of fuel cell stack cathode humidity sensitivity test:

[0096] (1) Call the pandas data processing core library to read the Excel file and automatically locate the first worksheet.

[0097] (2) Configure the test points as the "AIR humidifier water temperature_SV(°C)" list and the water-cooled inlet temperature_SV(°C) list:

[0098] This test is a cathode humidity sensitivity test under constant current density and constant reactor temperature conditions. The relative humidity of the cathode gas is adjusted by changing the water temperature in the cathode humidification chamber. The test points are configured as the "AIR humidification chamber water temperature_SV (°C)" list: [37, 50, 59, 65, 70] (unit: °C), with the key column label specified as "AIR humidification chamber water temperature_SV (°C)", and the allowable deviation set to ±0.1°C. Simultaneously, the water-cooled reactor inlet temperature constraint is set to 70°C, with the key column label specified as "water-cooled reactor inlet temperature_SV (°C)", and an allowable deviation of ±0.1°C, ensuring consistent reactor temperature conditions. The program uses a reverse deviation matching mode for location: it iterates backwards from the last line of the data file, simultaneously matching the AIR humidification chamber water temperature and the water-cooled reactor inlet temperature. When both line values ​​fall within the corresponding allowable deviation range, the index of that line is recorded as the steady-state end line for that humidity level. This process is repeated to locate all humidification temperature points.

[0099] (3) Configure the number of steady-state points M=12, take the last 12 rows of data for each test point, calculate the arithmetic mean to obtain the characteristic value of each point. Since this test is a constant current and constant stack temperature operation, the current density and stack temperature remain constant. The obtained characteristic values ​​include: constant average current density, average unit voltage under the corresponding AIR humidification box water temperature, and the difference between the unit voltage and the mean.

[0100] (4) Use the NumPy numerical computing library to calculate humidity using the Magnus saturated vapor pressure formula. (Saturated vapor pressure) (Unit: hPa) Magnus's formula is: ,

[0101] in:

[0102] : Temperature, in degrees Celsius (°C);

[0103] : Natural exponential function;

[0104] The coefficients 0.61078, 17.625, and 243.04 are empirical constants verified by numerous experiments and are applicable to the calculation of the saturated vapor pressure of water.

[0105] Cathode humidity, here referring to the relative humidity of the cathode, is calculated by substituting the saturated vapor pressure formula into the definition of cathode relative humidity, resulting in the following formula:

[0106]

[0107] Simultaneously, the calculated cathode relative humidity is paired with the corresponding average voltage and deviation from the mean to generate a data sequence for plotting.

[0108] (5) Use the matplotlib plotting library to generate a cathode humidity sensitivity plot (see...). Figure 4 The charts are plotted with cathode humidity as the X-axis, average voltage as the left Y-axis, and deviation from the mean as the right Y-axis. The charts are automatically updated with a uniform title, axis labels, legend, and grid lines, and saved as PNG images. The compiled feature value table is then exported as an Excel file.

Claims

1. A method for automatically processing and generating charts of test bench data, characterized in that, Includes the following steps: (1) Data reading and preprocessing Read the Excel file exported from the specified test bench using Python's pandas library or openpyxl library; By identifying the data table structure in the file, the worksheet or data area storing the original time-series data can be located. (2) Target data location and extraction The test points to be analyzed are specified through configuration files or predefined rules within the program; the test points are located by specific combinations of column labels or row index patterns. (3) Calculation of steady-state value Starting from the end of each target data column, extract a fixed number of data points upwards, where the fixed number of data points is a preset positive integer M; Configure these fixed number of data points as parameters, that is, use the last M data points at the end of the test point; The arithmetic mean of the extracted M data points is calculated using the NumPy numerical computing library in Python. This yields an "eigenvalue" representing the steady-state level of that point under specific parameters. The formula for calculating the arithmetic mean is as follows: In the formula, X is the eigenvalue; M is the number of selected steady-state data points; k is the row index of the end of the steady-state position of the test point located in step (2); X i The value of the data column corresponding to the i-th row; The calculation results, namely the steady-state voltage, steady-state current and other characteristic values ​​of each point, are automatically stored in a newly created DataFrame or dictionary data structure and organized according to data type. (4) Data processing and classification All calculated eigenvalues ​​are organized according to their physical meaning and subsequent drawing requirements; (5) Automatic chart generation and export The system calls Python's plotting libraries to automatically generate various analytical charts based on pre-designed templates. The plotting libraries include matplotlib or seaborn. The plots include polarization curves, three-chamber voltage drop plots, power curves, CVM voltage distribution plots, and sensitivity curves. The generated charts are automatically saved to the user-specified directory. Meanwhile, the compiled feature value summary table can also be exported as a new Excel spreadsheet in XLSX or CSV format.

2. The method for automatic processing and chart generation of test bench data as described in claim 1, characterized in that, In step (2), the column label combination includes the test number Test_ID and the step index Step_Index; wherein, the test number Test_ID is a globally unique identifier used to distinguish independent test items of different test types and different batches; the step index Step_Index is a local sequence identifier within a single test used to identify test steps of different parameter levels under the same test item; The row indexing pattern is as follows: First, define a function to find the row index corresponding to each preset test point in the original data. Iterate through the preset test point value list, searching backwards from the last row of the data. Determine whether the key column value of the current row is within the allowable deviation range of the preset test point. If a row that meets the conditions is found, add the row index to the key_list list. After finding a test point, immediately exit the inner loop and continue searching for the next test point.

3. The method for automatic processing and chart generation of test bench data as described in claim 1, characterized in that, In step (2), the test points are combined into a voltage column (Voltage) or a current column (Current) through specific column label combinations or row index patterns.

4. The method for automatic processing and chart generation of test bench data as described in claim 1, characterized in that, In step (3), the data types include voltage, current, temperature, and power.

5. The method for automatic processing and chart generation of test bench data as described in claim 1, characterized in that, Step (4) specifically involves pairing the "current density" and the corresponding "voltage" average values ​​at all points to prepare for drawing polarization curves; and pairing the "variation of a certain parameter" and the "voltage" average value to prepare for drawing sensitivity diagrams.

6. The method for automatic processing and chart generation of test bench data as described in claim 1, characterized in that, In step (5), the high-quality chart is in PNG, SVG or PDF format.

7. The method for automatic processing and chart generation of test bench data as described in claim 1, characterized in that, In step (5), the polarization curve is automatically plotted with the average current density as the X-axis and the average voltage as the Y-axis. Sensitivity plot: A curve is plotted with a certain input parameter as the X-axis and the average change of the performance index as the Y-axis; Power diagram: Based on the calculated average voltage and average current, calculate the average power at each point and plot the power versus current curve.

8. The method for automatic processing and chart generation of test bench data as described in claim 1, characterized in that, In step (5), during the chart generation process, the program automatically adds a title, axis labels, legend, and grid lines in a uniform format, and sets standardized line styles, colors, and markers.

9. An apparatus for automatically processing test bench data and generating charts, characterized in that, include: The data reading and preprocessing module is used to read the Excel file exported from the test bench, identify the data table structure, and locate the worksheet or data area storing the raw time series data. The target data location and extraction module is used to specify test points according to predefined rules, locate the end of the data in the corresponding data column, and obtain the row index of the steady-state end. The steady-state value calculation module is used to extract a fixed number of M data points from the end of each test point, calculate the arithmetic mean as the steady-state characteristic value of that point, and organize and store them according to data type. The data processing and classification module is used to pair and organize all feature values ​​according to their physical meaning and plotting requirements, generating input data pairs suitable for different charts; The automatic chart generation and export module is used to call the plotting library and automatically generate high-quality charts such as polarization curves, sensitivity plots, and power plots based on preset templates, and save them in a specified format, while exporting a summary table of characteristic values.

10. A computer-readable storage medium storing computer instructions for causing a processor to execute the automatic processing and chart generation method according to any one of claims 1 to 8.

Images