A vehicle weighing data acquisition method and system based on an industrial control system

Abstract

The application relates to the technical field of vehicle dynamic weighing, and discloses a vehicle weighing data acquisition method and system based on an industrial control system, wherein the method comprises the following steps: synchronously collecting original load time sequence signals of a weighing sensor strip and road surface temperature values; collecting baseline temperature pairing data in idle time intervals in stages; generating an enhanced temperature zero drift function through weighted regression and performing temperature compensation; estimating an instantaneous speed sequence and marking a working condition; performing dynamic load transfer compensation on a low-speed non-uniform speed working condition; performing axle load proportion checking; performing secondary confirmation and axle weight correction on suspected abnormal axles based on waveform structure characteristic parameters; and packaging and outputting weighing data records.

Description

Technical Field

[0001] This invention relates to the field of vehicle dynamic weighing technology, specifically to a method and system for acquiring vehicle weighing data based on an industrial control system. Background Technology

[0002] Dynamic weighing stations collect vehicle axle load signals through multiple sets of weighing sensor strips arranged along the driving direction, and use road surface temperature sensors to compensate for zero-point drift. Existing methods typically rely on collecting static baseline paired data during idle time intervals between adjacent vehicles, and then calculating the zero-point compensation amount by regression fitting of a temperature zero-point drift function.

[0003] Under conditions of dense, low-speed vehicle queuing and significant ambient temperature differences, the idle time between adjacent vehicles is severely insufficient, resulting in sparse baseline pairing data that meets the requirement for complete data acquisition duration. This prevents the temperature zero-point drift function from being updated in a timely manner, causing the zero-point compensation to lag behind the actual temperature conditions. Simultaneously, existing methods use the original measurements of each axle as input for load transfer compensation. When temperature drift is not correctly compensated, temperature deviation and load transfer compensation deviation are superimposed on the same axle weight value, systematically contaminating the basis for proportional calibration. Normal axles are easily misjudged as abnormal axles, leading to a decrease in the effective data utilization rate. Summary of the Invention

[0004] This invention provides a vehicle weighing data acquisition method and system based on an industrial control system, which solves the technical problems in related technologies such as insufficient temperature drift compensation of weighing sensors, axle load error caused by dynamic load transfer when the vehicle is traveling at low speed and non-uniform speed, and difficulty in effectively identifying and correcting abnormal axle loads.

[0005] This invention discloses a vehicle weighing data acquisition method based on an industrial control system, which includes at least the following steps: synchronously acquiring the original load time sequence signal of each weighing sensor strip and the real-time temperature value at the corresponding position, and obtaining vehicle axle number information and vehicle type classification code;

[0006] Monitor the idle time interval between adjacent vehicles on each strip, classify the idle time interval according to the preset minimum baseline acquisition duration threshold, and generate a baseline temperature pairing dataset containing complete pairing data and low confidence pairing data.

[0007] Weighted regression calculation is performed on the baseline temperature paired dataset to generate enhanced temperature zero drift functions for each strip. Temperature compensation is performed on the original load time series signal based on the enhanced temperature zero drift functions and the current real-time temperature value to generate a temperature-compensated axial load signal.

[0008] The instantaneous velocity sequence is estimated based on the timestamp sequence of the load signals of each axis after full temperature compensation, and the coefficient of variation is calculated. The current data is marked with the working condition based on the comparison result of the coefficient of variation and the preset velocity fluctuation threshold.

[0009] For the collected data marked as low-speed non-uniform speed conditions, the speed change is converted into a dynamic load transfer estimate based on the instantaneous speed sequence. The corresponding dynamic load transfer estimate is subtracted from the axle load values ​​after full temperature compensation to obtain the compensated axle load value sequence.

[0010] The ratio of each axle load value after compensation to the total axle load is used to generate a measured load distribution ratio sequence. This sequence is then compared with the standard axle load distribution ratio vector retrieved based on vehicle type classification codes, and axles with ratio values ​​exceeding the corresponding standard range are marked as suspected abnormal axles.

[0011] Energy concentration parameters and peak symmetry parameters are extracted from the load signals of suspected abnormal shafts and compared with the standard range in qualified data of similar historical working conditions. Based on the comparison results, the final state of the suspected abnormal shaft is determined and the shaft load is corrected for confirmed abnormal shafts.

[0012] Furthermore, the step of classifying the idle time intervals based on a preset minimum baseline acquisition duration threshold includes:

[0013] When the idle time interval exceeds the preset minimum baseline acquisition duration threshold, the mean value of the static baseline signal and the corresponding temperature value within that interval are extracted and added to the baseline temperature pairing dataset of the corresponding strip as complete paired data.

[0014] When the idle time interval is shorter than the preset minimum baseline acquisition duration threshold, the mean value and temperature value of the instantaneous baseline signal segment within that interval are extracted and marked as low-confidence paired data, which are then added to the baseline temperature paired dataset of the corresponding strip.

[0015] Further, performing weighted regression calculations on the baseline temperature paired dataset includes:

[0016] Piecewise linear regression was performed using complete paired data as the primary regression sample, while low-confidence paired data were included in the regression calculation with reduced weights to generate enhanced temperature zero-point drift functions for each band.

[0017] The contribution coefficient of the low-confidence paired data in the regression objective function is lower than that of the complete paired data.

[0018] Furthermore, the step of performing temperature compensation on the original load time-series signal based on the enhanced temperature zero-point drift function and the current real-time temperature value includes:

[0019] The zero-point compensation amount is calculated based on the enhanced temperature zero-point drift function of each strip and the current real-time temperature value. The corresponding zero-point compensation amount is subtracted from the original load time sequence signal to obtain the load signal after zero-point correction.

[0020] The zero-point calibrated load signal is divided by the sensitivity temperature coefficient of each strip at the current temperature, determined based on historical calibration data, to generate a temperature-compensated axial load signal; wherein the sensitivity temperature coefficient is a dimensionless proportionality coefficient obtained by fitting historical calibration data.

[0021] Furthermore, the estimation of the instantaneous velocity sequence based on the timestamp sequence of each shaft load signal after full temperature compensation includes:

[0022] Calculate the peak time difference of each adjacent axis on the same strip, divide the known axis distance by the peak time difference to obtain the instantaneous velocity value at the corresponding position, and form an instantaneous velocity sequence;

[0023] Wherein, the coefficient of variation is the ratio of the standard deviation to the mean of the instantaneous velocity sequence; when the coefficient of variation exceeds the preset velocity fluctuation threshold, the current acquisition is marked as a low-speed non-uniform speed condition and dynamic load transfer compensation is performed; when the coefficient of variation does not exceed the preset velocity fluctuation threshold, the axial load value after full temperature compensation is used as the compensated axle load value sequence.

[0024] Furthermore, the step of converting the velocity change into a dynamic load transfer estimate based on the instantaneous velocity sequence includes:

[0025] Extract the velocity change between adjacent axes from the instantaneous velocity sequence;

[0026] The longitudinal dynamics simplification formula is used to convert each speed change into a dynamic load transfer estimate for each axle; wherein, the longitudinal dynamics simplification formula is: the longitudinal acceleration component is obtained by dividing the speed change by the corresponding time difference, and the load transfer between the front and rear axles is obtained by multiplying the total vehicle mass, the longitudinal acceleration component, the ratio of the vehicle's center of gravity height and the wheelbase.

[0027] Furthermore, the energy concentration parameter is extracted as follows:

[0028] The peak region is defined as the half-peak width range before and after the peak moment. The ratio of the sum of squares of the signal amplitudes of each sampling point within the peak region to the sum of squares of the signal amplitudes of each sampling point within the entire acquisition window is calculated. This ratio is the energy concentration parameter.

[0029] 8. The vehicle weighing data acquisition method based on an industrial control system according to claim 1, characterized in that the peak symmetry parameter is extracted in the following way:

[0030] Centered on the peak moment, the signal sequences of the leading edge segment and the trailing edge segment are extracted respectively. The cosine similarity is used to calculate the similarity between the signal sequences of the leading edge segment and the trailing edge segment. The obtained cosine similarity value is the peak symmetry parameter.

[0031] Furthermore, the step of determining the final state of suspected abnormal axes based on the comparison results and performing axis weight correction on confirmed abnormal axes includes:

[0032] When either the energy concentration parameter or the peak symmetry parameter exceeds the corresponding standard range, the axis is determined to be an anomalous axis; when both the energy concentration parameter and the peak symmetry parameter are within their respective standard ranges, the axis is restored from a suspected anomalous axis to a normal axis.

[0033] For the confirmed abnormal axis, the corrected axle load value is calculated based on the median of the proportion corresponding to the axis in the standard axis load distribution proportion vector and the sum of the compensated axle loads of the remaining normal axes; wherein, the corrected axle load value is equal to the sum of the compensated axle loads of the remaining normal axes multiplied by the median of the standard proportion of the confirmed abnormal axis and then divided by the sum of the median of the standard proportions of all normal axes.

[0034] This invention provides a vehicle weighing data acquisition system based on an industrial control system, comprising:

[0035] The multi-source signal synchronous acquisition module is used to synchronously acquire the original load time sequence signal of each weighing sensor strip and the real-time temperature value at the corresponding position, and obtain vehicle axle number information and vehicle type classification code;

[0036] The baseline pairing data hierarchical acquisition module is used to monitor the idle time interval between adjacent vehicles on each strip, and classifies the idle time interval according to a preset minimum baseline acquisition duration threshold to generate a baseline temperature pairing dataset containing complete pairing data and low confidence pairing data.

[0037] The temperature compensation module is used to perform weighted regression calculation on the baseline temperature pairing dataset to generate enhanced temperature zero drift functions for each strip, and to perform temperature compensation on the original load time series signal based on the enhanced temperature zero drift functions and the current real-time temperature value to generate a temperature-compensated axis-by-axis load signal.

[0038] The working condition marking module is used to estimate the instantaneous velocity sequence based on the timestamp sequence of each shaft load signal after full temperature compensation and calculate the coefficient of variation. Based on the comparison result of the coefficient of variation and the preset velocity fluctuation threshold, the working condition is marked on the currently collected data.

[0039] The dynamic load transfer compensation module is used to convert the speed change into a dynamic load transfer estimate based on the instantaneous speed sequence of the collected data marked as low-speed non-uniform speed working conditions. The corresponding dynamic load transfer estimate is subtracted from the axle load value after full temperature compensation to obtain the compensated axle load value sequence.

[0040] The axle load ratio verification module is used to generate a measured load distribution ratio sequence by the ratio of each axle load value after compensation to the total axle load. This sequence is then compared with the standard axle load distribution ratio vector retrieved based on vehicle type classification codes, and axles with ratio values ​​exceeding the corresponding standard range are marked as suspected abnormal axles.

[0041] The anomaly confirmation and axle load correction module is used to extract energy concentration parameters and peak symmetry parameters from the load signals of suspected abnormal axles, compare them with the standard range in the qualified data of similar historical working conditions, determine the final state of the suspected abnormal axle based on the comparison results, and perform axle load correction on the confirmed abnormal axle.

[0042] The present invention has the following beneficial effects:

[0043] Firstly, during the baseline pairing data acquisition phase, instantaneous baseline signal segments within idle time intervals shorter than a preset minimum baseline acquisition duration threshold are marked as low-confidence paired data. These segments are then used in weighted regression to enhance the fitting of the zero-point drift function. In situations where dense vehicle queues lead to scarce idle time intervals, the low-confidence paired data fills the data gaps during periods of rapid temperature change. This prevents the update frequency of the enhanced zero-point drift function from entirely depending on the frequency of idle time intervals that meet the full acquisition duration requirement, thus alleviating the problem of the zero-point drift function lagging behind the actual temperature state due to insufficient paired data.

[0044] Secondly, temperature compensation is performed before load transfer compensation, so that the input for load transfer estimation is the axle load value with zero temperature deviation eliminated. This avoids the situation where temperature deviation and load transfer compensation deviation are superimposed on the same axle load value, and ensures that the axle load ratios on which proportional calibration is based are not affected by the systematic offset of temperature drift.

[0045] Third, energy concentration parameters are used for suspected anomalous axes. and peak symmetry parameter Two waveform structure characteristic parameters that do not depend on speed range division are reconfirmed. This eliminates the misjudgment of proportional offset caused by speed fluctuations under low-speed non-uniform speed conditions, ensuring that normal axes are not incorrectly excluded due to operating condition fluctuations, and improving the utilization of effective data. Attached Figure Description

[0046] Figure 1 This is a flowchart of a vehicle weighing data acquisition method based on an industrial control system provided in an embodiment of the present invention;

[0047] Figure 2 This is a schematic diagram comparing the original peak loads of each strip and each axis provided in the embodiments of the present invention;

[0048] Figure 3 This is a schematic diagram of the distribution of the baseline temperature pairing dataset for strip 1 provided in an embodiment of the present invention;

[0049] Figure 4 This is a schematic diagram of the axle load variation at each stage of temperature compensation for strip 1 provided in an embodiment of the present invention;

[0050] Figure 5 This is a schematic diagram of the instantaneous velocity sequence of the inter-axle section provided in an embodiment of the present invention;

[0051] Figure 6 This is a schematic diagram comparing the axle loads of each shaft before and after dynamic load transfer compensation according to an embodiment of the present invention;

[0052] Figure 7 This is a schematic diagram illustrating the verification of the load ratio of each shaft and the standard range provided in an embodiment of the present invention;

[0053] Figure 8 This is a schematic diagram comparing the characteristic parameters of the waveform structure of axis 2 provided in this embodiment of the invention with the standard range;

[0054] Figure 9 This is a schematic diagram illustrating the cumulative impact of each compensation step on the total axle load provided in the embodiments of the present invention. Detailed Implementation

[0055] In dynamic weighing stations, multiple sets of weighing sensor strips are arranged sequentially along the vehicle's direction of travel, with road surface temperature sensors deployed at each strip location. The industrial control system collects the raw load timing signals of each strip and obtains the real-time temperature values ​​from the temperature sensors. Under conditions of dense, low-speed vehicle queuing and significant ambient temperature differences, the idle time between adjacent vehicles is severely insufficient. The static baseline pairing data used to fit the temperature zero-point drift function is sparse, the temperature zero-point drift function cannot be updated in a timely manner, and the zero-point compensation lags behind the actual temperature state. Simultaneously, load transfer compensation uses the raw measurement values ​​of each axle as input. When temperature drift is not correctly compensated, the temperature deviation and load transfer compensation deviation are superimposed, systematically contaminating the basis for standard proportional calibration. Normal axles are misjudged as abnormal, and the effective data utilization rate decreases.

[0056] A vehicle weighing data acquisition method based on an industrial control system according to an embodiment of the present invention includes the following steps:

[0057] Step 1: Synchronous acquisition of multi-source signals;

[0058] The industrial control system synchronously collects the original load timing signals of each weighing sensor strip and the real-time temperature values ​​of the road surface temperature sensors at each strip location, obtains the vehicle passage time difference between adjacent strips, and receives vehicle axle number information and vehicle type classification code provided by the vehicle identification system.

[0059] Step 2: Baseline paired data is collected in a hierarchical manner;

[0060] The idle time interval between two adjacent vehicles on each sensor strip is monitored. The idle time interval is classified based on a preset minimum baseline acquisition duration threshold, and a hierarchical labeled baseline temperature pairing dataset is generated.

[0061] It should be noted that the above-mentioned tiered acquisition specifically involves the following: when the idle time interval exceeds a preset minimum baseline acquisition duration threshold, the mean value of the static baseline signal and the corresponding temperature value within that interval are extracted and added to the baseline temperature pairing dataset of the corresponding strip as complete paired data; when the idle time interval is shorter than the preset minimum baseline acquisition duration threshold, the mean value of the instantaneous baseline signal segment and the temperature value within that interval are extracted, marked as low-confidence paired data, and added to the baseline temperature pairing dataset of the corresponding strip. Although the acquisition duration of the low-confidence paired data does not meet the requirements of static baseline acquisition, its temperature value has temporal proximity to the current strip temperature state, which can provide temperature change trend information for enhancing the update of the temperature zero-point drift function.

[0062] Step 3: Enhance the generation of the zero-point drift function and temperature compensation;

[0063] Weighted regression calculations are performed on the baseline temperature pairing datasets of each strip to generate enhanced temperature zero-point drift functions for each strip. Temperature compensation is then performed on the original load time series signal based on the enhanced temperature zero-point drift functions and the current real-time temperature value to generate a temperature-compensated axial load signal.

[0064] Step 3 may specifically include:

[0065] Step S301: Perform piecewise linear regression using the complete paired data as the main regression sample, while incorporating the low-confidence paired data into the regression calculation with reduced weights to generate the enhanced temperature zero-point drift function for each band.

[0066] It should be noted that the aforementioned reduction in weight refers to the contribution coefficient of low-confidence paired data in the regression objective function being lower than that of the complete paired data. Let the contribution coefficient of the complete paired data be... The contribution coefficient of low-confidence paired data is Then there is ,in The contribution coefficient for complete paired data. This represents the contribution coefficient for low-confidence paired data. Because the baseline signal acquisition time for low-confidence paired data is insufficient, the stability of its mean is weaker than that of complete paired data. Therefore, a lower contribution coefficient is assigned in the regression calculation to constrain the influence of low-confidence paired data on the regression results, while preserving the coverage of the temperature variation range by the low-confidence paired data.

[0067] Step S302: Calculate the zero-point compensation amount based on the enhanced temperature zero-point drift function of each strip and the current real-time temperature value, and subtract the corresponding zero-point compensation amount from the original load time sequence signal to obtain the load signal after zero-point correction.

[0068] Step S303: Divide the zero-point calibrated load signal by the sensitivity temperature coefficient of each strip at the current temperature, determined based on historical calibration data, to generate a temperature-compensated axial load signal. The sensitivity temperature coefficient is a dimensionless proportionality coefficient obtained by fitting historical calibration data; dividing by the sensitivity temperature coefficient does not change the load dimension of the result.

[0069] Step 4: Instantaneous velocity sequence estimation and operating condition labeling;

[0070] Based on the timestamp sequence of load signals of each axle after full temperature compensation, the time difference of the peak occurrence time of adjacent axles is calculated. The instantaneous speed sequence is estimated by combining the vehicle wheelbase information. The coefficient of variation is calculated for the instantaneous speed sequence. Based on the comparison result of the coefficient of variation and the preset speed fluctuation threshold, the current data is marked with the working condition.

[0071] It should be noted that the estimation method for the above instantaneous velocity sequence is as follows: the peak time difference between adjacent axes on the same strip reflects the time interval between adjacent axes passing through that strip. Dividing the known axis distance by this time interval yields the instantaneous velocity value at the corresponding position. The coefficient of variation is the ratio of the standard deviation to the mean of the instantaneous velocity sequence. Both have the same dimension, and the ratio is dimensionless. When the coefficient of variation exceeds the preset velocity fluctuation threshold, the current acquisition is marked as a low-speed non-uniform speed condition, and the process proceeds to step 5 to perform dynamic load transfer compensation; conversely, if the coefficient of variation does not exceed the preset velocity fluctuation threshold, the axis-by-axis load value after full temperature compensation is directly used as the compensated axis load value sequence, and the process proceeds to step 6.

[0072] Step 5, Dynamic load transfer compensation;

[0073] For the collected data marked as low-speed non-uniform speed conditions, the velocity change between adjacent axes is calculated based on the instantaneous velocity sequence. The velocity change is converted into a dynamic load transfer estimate for each axis. The corresponding dynamic load transfer estimate is subtracted from the load value of each axis after full temperature compensation to obtain the axle load value sequence after double compensation.

[0074] Step 5 may specifically include:

[0075] Step S501: Extract the velocity change between adjacent axes from the instantaneous velocity sequence.

[0076] Step S502: Use the longitudinal dynamic simplified relationship to convert each velocity change into an axis-by-axis dynamic load transfer estimate.

[0077] It should be noted that the above simplified longitudinal dynamics equations are based on a linear approximation of the relationship between vehicle longitudinal acceleration and inter-axle load transfer. Let the velocity change between adjacent axles be... (Unit: m / s), the corresponding transit time difference is (Unit: s), then the longitudinal acceleration component is (Unit: m / s²); Assume the ratio of the vehicle's center of gravity height to its wheelbase is... (Dimensionless), the total mass of the vehicle is (Unit: kg), then the load transfer between the front and rear axles is (Unit: N), where For the change in velocity, In order to take advantage of the time difference, For the longitudinal acceleration component, It is the ratio of the height of the center of gravity to the wheelbase. For the total mass of the vehicle. This represents the load transfer between the front and rear axles. The simplified longitudinal dynamics equation is suitable for estimating the impact of inertial forces on axle load distribution under low-speed conditions. Before performing addition or subtraction operations with the temperature-compensated shaft load value, both must be unified to the same unit of measurement.

[0078] Step S503: Subtract the corresponding dynamic load transfer estimate from the temperature-compensated axle load values ​​to obtain the double-compensated axle load value sequence.

[0079] Step 6: Shaft load ratio verification and abnormal shaft marking;

[0080] Calculate the ratio of each axle load value to the total axle load after compensation, generate a measured load distribution ratio sequence, retrieve the corresponding standard axle load distribution ratio vector based on vehicle type classification code, compare the measured load distribution ratio sequence with the standard axle load distribution ratio vector axle by axle, and mark axles whose ratio values ​​exceed the corresponding standard range as suspected abnormal axles.

[0081] It should be noted that each element in the above measured load distribution ratio sequence is a dimensionless quantity, representing the ratio of each axle load to the total axle load. The standard axle load distribution ratio vector represents the range of proportions of each axle load to the total load for each vehicle model under normal load conditions. This vector is also dimensionless and is obtained from historical qualified weighing records. Each axle corresponds to a ratio interval. When comparing axles one by one, if the measured ratio value falls within this interval, the axle is considered normal; otherwise, it is marked as a suspected abnormal axle.

[0082] In this embodiment of the application, in order to reduce the false positive rate of suspected abnormal shafts under low-speed non-uniform speed conditions, the following steps are also included in step 6:

[0083] Step 7: Anomaly confirmation and axle load correction based on waveform structure characteristics;

[0084] Two waveform structural feature parameters, energy concentration parameter and peak symmetry parameter, are extracted from the original load signal of the suspected abnormal shaft. The energy concentration parameter and peak symmetry parameter are compared with the standard range in the qualified data of similar historical working conditions. Based on the comparison results, the final state of the suspected abnormal shaft is determined, and the shaft load is corrected for the confirmed abnormal shaft.

[0085] Step 7 may specifically include:

[0086] Step S701: Calculate the energy concentration parameter for the original load signal corresponding to the suspected abnormal axis. Let the total energy of the load signal within the entire acquisition window be... The signal energy in the peak region is The peak region is defined as the range of half the peak width before and after the peak time. The energy concentration parameter is then defined as follows: ,in It is the sum of the squares of the signal amplitudes at each sampling point within the peak region. It is the sum of the squares of the signal amplitudes at all sampling points within the entire acquisition window. This is the energy concentration parameter. The ratio of two dimensionless energy quantities results in a dimensionless quantity. The energy concentration parameter of a normal shaft load signal falls within a stable statistical range. Sensor malfunctions or external interference can cause energy to diffuse into non-peak regions, making... Too low.

[0087] Step S702: Calculate the peak symmetry parameter for the original load signal corresponding to the suspected abnormal axis. Using the peak time as the center, extract the signal sequences of the leading and trailing edges respectively. Calculate the similarity between the two sequences using cosine similarity; the resulting cosine similarity value is the peak symmetry parameter. ,in This refers to the peak symmetry parameter. The load peak generated when the wheel normally passes over the sensor strip has approximately symmetrical characteristics. Abnormal forces cause significant asymmetry in the peak shape, thus affecting the peak symmetry parameter. Too low.

[0088] Step S703, the energy concentration parameter and peak symmetry parameter The values ​​were compared with the standard ranges in historical data for similar operating conditions. When the energy concentration parameter... or peak symmetry parameter If any parameter exceeds the corresponding standard range, the axis is determined to be an anomalous axis; when the energy concentration parameter With peak symmetry parameter When all parameters are within their respective standard ranges, the axis is restored from a suspected abnormal axis to a normal axis.

[0089] Step S704: For the confirmed abnormal axis, calculate the corrected axle load value based on the median of the proportion corresponding to that axis in the standard axis load distribution proportion vector and the sum of the compensated axle loads of the remaining normal axes. Let the sum of the compensated axle loads of the remaining normal axes be... The sum of the proportions of normal axle load to the total vehicle weight is Confirm that the median proportion of the abnormal axis in the standard axis load distribution ratio vector is... The corrected axle load value is ,in This represents the total axle load after compensation for the remaining normal axles. It is the sum of the median values ​​of the standard proportions of the normal axes (obtained by accumulating the median values ​​of the corresponding proportions of each normal axis in the standard axis load distribution proportion vector). To confirm the standard scale median of the abnormal axis, To correct the axle load value. and All are dimensionless proportional quantities. and They have the same dimensions.

[0090] It should be noted that the energy concentration parameter and peak symmetry parameter All methods are based on the spatial structure extraction of the load signal waveform, independent of the interval division of vehicle speed. Under low-speed, non-uniform operating conditions, changes in vehicle speed affect the time scale of the load signal, but do not alter the spatial distribution characteristics and peak symmetry of the signal energy. Therefore, the energy concentration parameter... and peak symmetry parameter It can still serve as an effective basis for anomaly detection under conditions of velocity fluctuation.

[0091] Step 8: Record, encapsulate, and output the weighing data;

[0092] The compensated axle load values ​​of all normal axles are summed with the corrected axle load values ​​of the confirmed abnormal axles to generate the verified total vehicle weight data. The verified total vehicle weight data, together with the status identifiers of each axle, temperature compensation parameters, enhanced temperature zero-point drift function update status, dynamic load transfer compensation information, and speed fluctuation markers, are encapsulated into a weighing data record and output through the industrial control system.

[0093] It should be noted that the status identifiers for each axis mentioned above include three categories: normal axes, suspected abnormal axes that have recovered to normal status after waveform confirmation, and confirmed abnormal axes. The enhanced temperature zero-point drift function update status record records the number of complete paired data and the number of low-confidence paired data used in the most recent enhanced temperature zero-point drift function update, for subsequent data quality traceability.

[0094] In this implementation, during the baseline pairing data acquisition phase, instantaneous baseline signal segments within idle time intervals shorter than a preset minimum baseline acquisition duration threshold are included in the acquisition range and marked as low-confidence paired data. These segments participate in the fitting of the enhanced temperature zero-point drift function in weighted regression with reduced weights. Under conditions where dense vehicle queues lead to scarce idle time intervals, the low-confidence paired data fills the data gaps during periods of rapid temperature change. This prevents the update frequency of the enhanced temperature zero-point drift function from entirely depending on the frequency of idle time intervals meeting the full acquisition duration requirement, thus alleviating the problem of the enhanced temperature zero-point drift function lagging behind the actual temperature state due to insufficient paired data. Furthermore, this implementation places temperature compensation before load transfer compensation, ensuring that the input for load transfer estimation is the axle load value with eliminated temperature zero-point deviation. This avoids the situation where temperature deviation and load transfer compensation deviation are superimposed on the same axle load value, thus ensuring that the axle load ratios used for proportional calibration are not affected by systematic shifts in the temperature drift direction. In addition, an energy concentration parameter is used for suspected abnormal axles. and peak symmetry parameter Two waveform structure characteristic parameters that do not depend on speed range division are reconfirmed. This eliminates the misjudgment of proportional offset caused by speed fluctuations under low-speed non-uniform speed conditions, ensuring that normal axes are not incorrectly excluded due to operating condition fluctuations, and improving the utilization of effective data.

[0095] A dynamic weighing station (hereinafter referred to as "Station A") on a highway freight corridor operates during the morning rush hour at the transition from summer to autumn. During this period, traffic is heavy, the distance between adjacent vehicles is short, and the road surface temperature rises rapidly with increasing sunlight intensity, resulting in significant environmental temperature differences. Station A deploys three sets of weighing sensor strips (strip 1, strip 2, and strip 3) sequentially along the direction of travel, with a road surface temperature sensor located at each strip position. The industrial control system synchronously collects signals from each strip. The vehicle currently being measured is a four-axle heavy-duty truck (vehicle type code HV04), whose axle count and vehicle type code are provided in advance by the vehicle identification system.

[0096] like Figures 2-9 As shown, it includes the following steps:

[0097] The industrial control system synchronously initiates data acquisition when a vehicle enters the weighing area. The original load timing signals of each strip are recorded at a sampling rate of 500Hz, and the temperature sensor reports the road surface temperature at a frequency of 1Hz. The system also receives the axle number (4 axles) and vehicle type code (HV04) pushed by the vehicle identification system. The spacing between each strip is known; the spacing between strip 1 and strip 2 is 2.50m, and the spacing between strip 2 and strip 3 is also 2.50m.

[0098] Table 1. Raw input data collected synchronously in step 1:

[0099]

[0100] The industrial control system continuously monitors the idle time interval between two vehicles on strip 1. The preset minimum baseline acquisition time threshold is 3.0 seconds. During the current morning rush hour, the system recorded four idle intervals, some of which were shorter than the threshold and were marked as low-confidence paired data. Taking strip 1 as an example, the generated baseline temperature paired dataset is shown in the table below.

[0101] Table 2 shows the baseline temperature paired dataset for band 1:

[0102]

[0103] Pairs A and C, with idle intervals exceeding 3.0 seconds, were included in the dataset as complete paired data. Pairs B and D, with idle intervals less than 3.0 seconds, had their instantaneous baseline segment mean and corresponding temperature values ​​extracted and marked as low-confidence paired data. The temperature values ​​of the low-confidence paired data (34.2℃ and 38.6℃) covered the rapidly rising temperature range not covered by the complete paired data, filling the gap in temperature change trend information.

[0104] Taking strip 1 as an example, the generation process of the enhanced temperature zero-point drift function is illustrated. Let the contribution coefficient of the complete paired data be... Contribution coefficient of low-confidence paired data ,satisfy Weighted regression was performed using pairs A and C as the principal samples and pairs B and D as the constraint supplementary samples. Piecewise linear regression was performed within the temperature range of 31.4℃ to 38.6℃ to generate an enhanced zero-point drift function. .

[0105] When the current road surface temperature is 38.6℃, by The zero-point compensation was calculated to be 0.69 kN. The original load values ​​of each axis of strip 1 were subtracted by 0.69 kN to obtain the zero-point corrected load signals. Subsequently, based on historical calibration data, the sensitivity temperature coefficient of strip 1 at 38.6℃ was determined to be 1.023. The zero-point corrected load values ​​of each axis were divided by 1.023 to generate the temperature-compensated axis-by-axis load signals.

[0106] Table 3: Temperature compensation process data for strip 1:

[0107]

[0108] Based on the timestamps of the load signals of each axle after temperature full compensation, the peak time of each axle on strip 1 is extracted. The known adjacent axle distances for this vehicle model are: 3.80m from axle 1 to axle 2, 1.35m from axle 2 to axle 3, and 1.35m from axle 3 to axle 4. Dividing the axle distance by the time difference between the peak times of adjacent axles yields the instantaneous speed sequence.

[0109] coefficient of variation ,in The standard deviation of the instantaneous velocity sequence, This is the mean. Calculated... The preset speed fluctuation threshold is 0.10. Since 0.148 exceeds 0.10, the current acquisition is marked as a low-speed non-uniform speed condition, and the process proceeds to step 5 to perform dynamic load transfer compensation.

[0110] Table 4. Estimation results of instantaneous velocity sequence:

[0111]

[0112] The velocity change between adjacent axles is extracted from the instantaneous velocity sequence. As the vehicle passes from axle 1 to axle 4, its speed gradually increases, indicating that the vehicle accelerates within the weighing area. Under acceleration conditions, the vehicle's inertial force causes the front axle to be unloaded and the rear axle to be unloaded. Therefore, the load transfer amount is subtracted from the front axle compensation value and added back to the rear axle compensation value.

[0113] Taking the section from shaft 1 to shaft 2 as an example, the change in velocity m / s, corresponding to the time difference s, then longitudinal acceleration m / s². Assume the total mass of the vehicle. kg, the ratio of center of mass height to wheelbase The load transfer amount between shaft 1 and shaft 2 N, which is 1.25kN. Subtract 1.25kN from shaft 1 (front shaft), and add back 1.25kN to shaft 2 (rear shaft). The compensation results for each shaft are shown in the table below.

[0114] Table 5. Results of Dynamic Load Transfer Compensation:

[0115]

[0116] The total axle load after double compensation is kN. The measured load distribution ratios for each axle are as follows: axle 1 21.82%, axle 2 27.16%, axle 3 25.62%, and axle 4 25.40%. Searching for the standard axle load distribution ratio vector corresponding to vehicle model code HV04 and comparing it axle by axle, it was found that the measured ratio of axle 2 (27.16%) exceeded the upper limit of the standard range (25.80%), and it was marked as a suspected abnormal axle.

[0117] Table 6: Axis Load Ratio Verification Results

[0118]

[0119] Two waveform structural feature parameters are extracted from the original load signal of axis 2.

[0120] Regarding the energy concentration parameter, the total energy of the original load signal of shaft 2 within the entire acquisition window. Signal energy in the peak region Calculated Among the qualified data of similar historical operating conditions The standard range is 0.75 to 0.95, for axis 2. It falls within the standard range.

[0121] Regarding the peak symmetry parameter, taking the peak time of axis 2 as the center, the signal sequences of the leading edge and trailing edge segments are extracted, and the cosine similarity is calculated. Among the qualified data of similar historical operating conditions The standard range is 0.82 to 0.98, for axis 2. It also falls within the standard range.

[0122] because and Both are within their respective standard ranges. Axis 2 has been restored from a suspected abnormal axis to a normal axis, and no axis weight correction is required.

[0123] Table 7 Comparison results of waveform structure characteristic parameters of axis 2:

[0124]

[0125] All four axles are normal. After double compensation, the axle load values ​​of each axle are summed, and the total vehicle weight after verification is 297.81 kN. The industrial control system encapsulates this total weight data, along with the status indicators of each axle (all four axles are marked as normal), temperature compensation parameters (zero-point compensation amount 0.69 kN, sensitivity temperature coefficient 1.023), enhanced temperature zero-point drift function update status (this update uses 2 complete paired data and 2 low-confidence paired data), dynamic load transfer compensation information, and speed fluctuation markers (low-speed non-uniform speed condition, coefficient of variation 0.148), into a weighing data record and outputs it.

[0126] Throughout the data flow, the original load signal, after being synchronously acquired in step 1, enters the baseline paired data hierarchical acquisition in step 2. The introduction of low-confidence paired data enables the weighted regression in step 3 to cover the rapid temperature rise range, thereby generating an enhanced temperature zero-point drift function that more closely reflects the current temperature state. After completing full temperature compensation in step 3, step 4 assesses the speed fluctuation state, identifying low-speed non-uniform speed conditions and triggering dynamic load transfer compensation in step 5. The compensation direction of front axle load reduction and rear axle load increase under acceleration conditions ensures that the axle load value entering step 6 for proportional verification has eliminated both temperature deviation and dynamic load transfer errors. After marking axle 2 as a suspected anomaly in step 6, step 7 performs a secondary confirmation based on waveform structure characteristic parameters, eliminating misjudgments of proportional offset caused by speed fluctuations and restoring axle 2 to a normal axle. Step 8 encapsulates and outputs the final result, fully recording the parameter status of each compensation stage, ensuring the traceability of the weighing data.

[0127] The embodiments of the present invention have been described above. However, the embodiments are not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make more equivalent embodiments under the guidance of the present embodiments, and all of them are within the protection scope of the present embodiments.

Claims

1. A method for acquiring vehicle weighing data based on an industrial control system, characterized in that, At least the following steps are included: The original load time sequence signal of each weighing sensor strip and the real-time temperature value at the corresponding position are collected synchronously to obtain vehicle axle number information and vehicle type classification code; Monitor the idle time interval between adjacent vehicles on each strip, classify the idle time interval according to the preset minimum baseline acquisition duration threshold, and generate a baseline temperature pairing dataset containing complete pairing data and low confidence pairing data. Weighted regression calculation is performed on the baseline temperature paired dataset to generate enhanced temperature zero drift functions for each strip. Temperature compensation is performed on the original load time series signal based on the enhanced temperature zero drift functions and the current real-time temperature value to generate a temperature-compensated axial load signal. The instantaneous velocity sequence is estimated based on the timestamp sequence of the load signals of each axis after full temperature compensation, and the coefficient of variation is calculated. The current data is marked with the working condition based on the comparison result of the coefficient of variation and the preset velocity fluctuation threshold. For the collected data marked as low-speed non-uniform speed conditions, the speed change is converted into a dynamic load transfer estimate based on the instantaneous speed sequence. The corresponding dynamic load transfer estimate is subtracted from the axle load values ​​after full temperature compensation to obtain the compensated axle load value sequence. The ratio of each axle load value after compensation to the total axle load is used to generate a measured load distribution ratio sequence. This sequence is then compared with the standard axle load distribution ratio vector retrieved based on vehicle type classification codes, and axles with ratio values ​​exceeding the corresponding standard range are marked as suspected abnormal axles. Energy concentration parameters and peak symmetry parameters are extracted from the load signals of suspected abnormal shafts and compared with the standard range in qualified data of similar historical working conditions. Based on the comparison results, the final state of the suspected abnormal shaft is determined and the shaft load is corrected for confirmed abnormal shafts.

2. The vehicle weighing data acquisition method based on an industrial control system according to claim 1, characterized in that, The step of classifying the idle time interval based on a preset minimum baseline acquisition duration threshold includes: When the idle time interval exceeds the preset minimum baseline acquisition duration threshold, the mean value of the static baseline signal and the corresponding temperature value within that interval are extracted and added to the baseline temperature pairing dataset of the corresponding strip as complete paired data. When the idle time interval is shorter than the preset minimum baseline acquisition duration threshold, the mean value and temperature value of the instantaneous baseline signal segment within that interval are extracted and marked as low-confidence paired data, which are then added to the baseline temperature paired dataset of the corresponding strip.

3. The vehicle weighing data acquisition method based on an industrial control system according to claim 1, characterized in that, Performing weighted regression calculations on the baseline temperature paired dataset includes: Piecewise linear regression was performed using complete paired data as the primary regression sample, while low-confidence paired data were included in the regression calculation with reduced weights to generate enhanced temperature zero-point drift functions for each band. The contribution coefficient of the low-confidence paired data in the regression objective function is lower than that of the complete paired data.

4. The vehicle weighing data acquisition method based on an industrial control system according to claim 1, characterized in that, The temperature compensation of the original load time-series signal based on the enhanced temperature zero-point drift function and the current real-time temperature value includes: The zero-point compensation amount is calculated based on the enhanced temperature zero-point drift function of each strip and the current real-time temperature value. The corresponding zero-point compensation amount is subtracted from the original load time sequence signal to obtain the load signal after zero-point correction. The zero-point calibrated load signal is divided by the sensitivity temperature coefficient of each strip at the current temperature, determined based on historical calibration data, to generate a temperature-compensated axial load signal; wherein the sensitivity temperature coefficient is a dimensionless proportionality coefficient obtained by fitting historical calibration data.

5. The vehicle weighing data acquisition method based on an industrial control system according to claim 1, characterized in that, The instantaneous velocity sequence estimated based on the timestamp sequence of each shaft load signal after full temperature compensation includes: Calculate the peak time difference of each adjacent axis on the same strip, divide the known axis distance by the peak time difference to obtain the instantaneous velocity value at the corresponding position, and form an instantaneous velocity sequence; Wherein, the coefficient of variation is the ratio of the standard deviation to the mean of the instantaneous velocity sequence; when the coefficient of variation exceeds the preset velocity fluctuation threshold, the current acquisition is marked as a low-speed non-uniform speed condition and dynamic load transfer compensation is performed; when the coefficient of variation does not exceed the preset velocity fluctuation threshold, the axial load value after full temperature compensation is used as the compensated axle load value sequence.

6. The vehicle weighing data acquisition method based on an industrial control system according to claim 1, characterized in that, The process of converting velocity changes into dynamic load transfer estimates based on the instantaneous velocity sequence includes: Extract the velocity change between adjacent axes from the instantaneous velocity sequence; The longitudinal dynamics simplification formula is used to convert each speed change into a dynamic load transfer estimate for each axle; wherein, the longitudinal dynamics simplification formula is: the longitudinal acceleration component is obtained by dividing the speed change by the corresponding time difference, and the load transfer between the front and rear axles is obtained by multiplying the total vehicle mass, the longitudinal acceleration component, the ratio of the vehicle's center of gravity height and the wheelbase.

7. The vehicle weighing data acquisition method based on an industrial control system according to claim 1, characterized in that, The energy concentration parameter is extracted as follows: The peak region is defined as the half-peak width range before and after the peak moment. The ratio of the sum of squares of the signal amplitudes of each sampling point within the peak region to the sum of squares of the signal amplitudes of each sampling point within the entire acquisition window is calculated. This ratio is the energy concentration parameter.

8. The vehicle weighing data acquisition method based on an industrial control system according to claim 1, characterized in that, The peak symmetry parameter is extracted as follows: Centered on the peak moment, the signal sequences of the leading edge segment and the trailing edge segment are extracted respectively. The cosine similarity is used to calculate the similarity between the signal sequences of the leading edge segment and the trailing edge segment. The obtained cosine similarity value is the peak symmetry parameter.

9. The vehicle weighing data acquisition method based on an industrial control system according to claim 1, characterized in that, The process of determining the final state of suspected abnormal axes based on comparison results and performing axis weight correction on confirmed abnormal axes includes: When either the energy concentration parameter or the peak symmetry parameter exceeds the corresponding standard range, the axis is determined to be an anomalous axis; when both the energy concentration parameter and the peak symmetry parameter are within their respective standard ranges, the axis is restored from a suspected anomalous axis to a normal axis. For the confirmed abnormal axis, the corrected axis weight value is calculated based on the median of the proportion corresponding to the axis in the standard axis load distribution proportion vector and the sum of the compensated axis weights of the remaining normal axes; wherein, the corrected axis weight value is equal to the sum of the compensated axis weights of the remaining normal axes multiplied by the median of the standard proportion of the confirmed abnormal axis and then divided by the sum of the median of the standard proportions of all normal axes.

10. A vehicle weighing data acquisition system based on an industrial control system, used to execute the vehicle weighing data acquisition method based on an industrial control system as described in any one of claims 1 to 9, characterized in that, include: The multi-source signal synchronous acquisition module is used to synchronously acquire the original load time sequence signal of each weighing sensor strip and the real-time temperature value at the corresponding position, and obtain vehicle axle number information and vehicle type classification code; The baseline pairing data hierarchical acquisition module is used to monitor the idle time interval between adjacent vehicles on each strip, and classifies the idle time interval according to a preset minimum baseline acquisition duration threshold to generate a baseline temperature pairing dataset containing complete pairing data and low confidence pairing data. The temperature compensation module is used to perform weighted regression calculation on the baseline temperature pairing dataset to generate enhanced temperature zero drift functions for each strip, and to perform temperature compensation on the original load time series signal based on the enhanced temperature zero drift functions and the current real-time temperature value to generate a temperature-compensated axis-by-axis load signal. The working condition marking module is used to estimate the instantaneous velocity sequence based on the timestamp sequence of each shaft load signal after full temperature compensation and calculate the coefficient of variation. Based on the comparison result of the coefficient of variation and the preset velocity fluctuation threshold, the working condition is marked on the currently collected data. The dynamic load transfer compensation module is used to convert the speed change into a dynamic load transfer estimate based on the instantaneous speed sequence of the collected data marked as low-speed non-uniform speed working conditions. The corresponding dynamic load transfer estimate is subtracted from the axle load value after full temperature compensation to obtain the compensated axle load value sequence. The axle load ratio verification module is used to generate a measured load distribution ratio sequence by the ratio of each axle load value after compensation to the total axle load. This sequence is then compared with the standard axle load distribution ratio vector retrieved based on vehicle type classification codes, and axles with ratio values ​​exceeding the corresponding standard range are marked as suspected abnormal axles. The anomaly confirmation and axle load correction module is used to extract energy concentration parameters and peak symmetry parameters from the load signals of suspected abnormal axles, compare them with the standard range in qualified data of similar historical working conditions, determine the final state of suspected abnormal axles based on the comparison results, and perform axle load correction on confirmed abnormal axles.

Images