Error analysis and correction method for load sensors

Abstract

The present invention discloses an error analysis and correction method for a load sensor, and relates to the technical field of load sensors. The method comprises: arranging a plurality of load sensors on a load test platform; applying a known load to a first load sensor, recording the first load sensor, and N groups of displacement data transmitted from the first load sensor to a second sensor... and an Nth sensor; performing transfer load error analysis according to the N groups of displacement data, and outputting N groups of transfer load errors; when the N groups of transfer load errors are greater than or equal to a preset error threshold, performing analysis according to a multi-bit error propagation model, and obtaining N-1 error correction indicators; performing error correction on the second sensor... and the Nth sensor according to the N-1 error correction indicators. The method solves the technical problem of inaccurate measurement caused by load transfer during the measurement process of the load sensor in the prior art, and achieves the technical effect of improving the accuracy and reliability of load measurement through error correction.

Description

Technical Field

[0001] The invention relates to the technical field of load sensors, and in particular to an error analysis and correction method for load sensors. Background Art

[0002] In modern industry and scientific research, accurate measurement technology is essential to ensure product quality, improve production efficiency, and promote scientific discovery. As a key component for measuring mechanical loads, load sensors are widely used in various testing and monitoring systems, such as structural health monitoring, material testing, and robotic arm force control. However, due to the influence of various factors, including environmental conditions, equipment aging, installation errors, etc., load sensors may produce errors in actual applications. If these errors are not controlled and corrected, they will directly affect the accuracy of the measurement results. With the development of technology and the increase in application needs, the accuracy requirements for load sensors are also getting higher and higher. Traditional error correction methods may not meet the growing accuracy requirements, especially in complex or dynamic measurement environments. Therefore, developing an effective error analysis and correction method to improve the measurement accuracy and reliability of load sensors has become a technical problem that urgently needs to be solved. Summary of the invention

[0003] The present application provides an error analysis and correction method for a load sensor, which solves the technical problem of inaccurate measurement caused by load transfer during the measurement process of the load sensor in the prior art.

[0004] In view of the above problems, the present application provides an error analysis and correction method for a load sensor.

[0005] The present application provides an error analysis correction method for a load sensor, the method comprising:

[0006] A plurality of load sensors are arranged on a load test platform, and the plurality of load sensors are distributed at equal intervals; by applying a known load to a first load sensor, the first load sensor and N groups of displacement data transmitted by the first load sensor to a second sensor, a third sensor ... and an Nth sensor are recorded, where N is the number of load sensors; transfer load error analysis is performed according to the N groups of displacement data, and N groups of transfer load errors are output; when the N groups of transfer load errors are greater than or equal to a preset error threshold, the N groups of transfer load errors are input into a multi-bit error propagation model, and analysis is performed according to the multi-bit error propagation model to obtain N-1 error correction indicators; and error correction is performed on the second sensor, the third sensor ... and the Nth sensor according to the N-1 error correction indicators.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] First, multiple load sensors are arranged on the load test platform, and the multiple load sensors are distributed at equal intervals. By applying a known load to the first load sensor, the first load sensor and N sets of displacement data transmitted from the first load sensor to the second sensor, the third sensor... and the Nth sensor are recorded, where N is the number of load sensors. Then, the transfer load error analysis is performed based on the N sets of displacement data, and N sets of transfer load errors are output. When the N sets of transfer load errors are greater than or equal to the preset error threshold, the N sets of transfer load errors are input into a multi-bit error propagation model, and analyzed according to the multi-bit error propagation model to obtain N-1 error correction indicators. Finally, the second sensor, the third sensor... and the Nth sensor are error corrected according to the N-1 error correction indicators. The technical problem of inaccurate measurement caused by load transfer during the measurement process of the load sensor in the prior art is solved, and the technical effect of improving the load measurement accuracy and reliability is achieved through error correction. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic flow chart of an error analysis and correction method for a load sensor provided in an embodiment of the present application;

[0011] Figure 2 A schematic diagram of a flow chart of error correction in an error analysis and correction method for a load sensor provided in an embodiment of the present application. DETAILED DESCRIPTION

[0012] The present application solves the technical problem of inaccurate measurement caused by load transfer during the measurement process of the load sensor in the prior art by providing an error analysis and correction method for the load sensor.

[0013] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0014] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0015] Examples, such as Figure 1 As shown, an embodiment of the present application provides an error analysis and correction method for a load sensor, wherein the method comprises:

[0016] A plurality of load sensors are arranged on the load testing platform, and the plurality of load sensors are distributed at equal intervals.

[0017] On the load test platform, multiple load sensors are arranged in an equidistant distribution. The equidistant distribution means that the distance between each sensor and its adjacent sensor is equal. Such a layout helps to ensure that the monitoring and measurement of the load can maintain uniformity and consistency throughout the test area. At the same time, it avoids the measurement blind spots or error accumulation caused by uneven distribution of sensors, so that each sensor can independently and accurately reflect the load condition at its location. By arranging multiple load sensors at equal intervals, the load distribution on the test platform can be more comprehensively obtained, including the load size, direction, and the law of change over time and position.

[0018] By applying a known load to the first load sensor, the first load sensor and N sets of displacement data transmitted from the first load sensor to the second sensor, the third sensor, ... and the Nth sensor are recorded, where N is the number of the load sensors.

[0019] During the load test, in order to evaluate the mutual influence between the sensors and the propagation of the load on the entire test platform, a known and accurate load can be applied to the first load sensor. This known load is used as the test benchmark to compare and calibrate the responses of other sensors. As the first load sensor is subjected to the load, it will produce corresponding displacement or deformation, and this displacement data is recorded as the first set of data. At the same time, since the test platform is an interconnected system, this load will also affect other sensors through structural transmission or medium transmission. Therefore, it is also necessary to record these displacement data propagated to the second sensor, the third sensor... until the Nth sensor. In this process, each sensor will produce a certain displacement or deformation due to the propagation load received. These displacement data are recorded separately to form N sets of displacement data. Each set of displacement data represents the response of the corresponding sensor under the influence of the known load applied by the first load sensor.

[0020] Furthermore, by applying a known load to the first load sensor, the method further comprises:

[0021] A load test sample is set, wherein the load test sample is obtained by analyzing an application scenario of a load sensor; and a known load is applied to the first load sensor according to the load test sample.

[0022] Based on an in-depth analysis of the actual application scenarios of the load sensor, the various working conditions, environmental conditions and expected load characteristics that the sensor may face are obtained, so as to set up appropriate load test samples according to the actual situation; and apply a known load to the first load sensor according to the set load test samples. By setting reasonable load test samples and applying a known load to the first load sensor according to these samples, the performance and measurement accuracy of the sensor can be effectively evaluated, providing strong support for subsequent data analysis and error correction.

[0023] Perform transfer load error analysis based on the N groups of displacement data, and output N groups of transfer load errors.

[0024] After applying a known load to the first load sensor and recording N sets of displacement data, N sets of displacement data are needed to analyze the transfer load error. Specifically, by comparing the displacement response of each sensor with its expected response under an ideal interference-free state and calculating the difference between the actual response and the expected response, the transfer load error of each sensor can be obtained, and finally N sets of transfer load errors are obtained. Each set of errors corresponds to the deviation between the displacement response of a sensor and the expected response. These error data not only reveal the degree of influence of load propagation on the sensor measurement results, but also provide an important basis for subsequent error correction.

[0025] When the N groups of transfer load errors are greater than or equal to a preset error threshold, the N groups of transfer load errors are input into a multi-bit error propagation model, and analyzed according to the multi-bit error propagation model to obtain N-1 error correction indicators.

[0026] During the load test, if any or all of the N groups of transfer load errors obtained through displacement data analysis are greater than or equal to the preset error threshold, it means that the error caused by load propagation has exceeded the acceptable range and needs to be corrected. The N groups of transfer load errors are input into the multi-bit error propagation model for analysis to obtain N-1 error correction indicators.

[0027] Furthermore, the multiple load sensors include surface distribution and line distribution; when the multiple load sensors are surface distribution, the multiple load sensors are divided into edge load sensors and center load sensors, and load simulation is performed on the edge load sensors and the center load sensors to generate a first group of training data sets; when the multiple load sensors are line distribution, load simulation is performed on the multiple load sensors to generate a second group of training data sets; based on the first group of training data sets and the second group of training data sets and identification information identifying displacement errors, a multi-bit error propagation model is constructed.

[0028] In the process of constructing the multi-bit error propagation model, due to the different distribution modes of load sensors on the test platform (such as surface distribution and line distribution), it is necessary to process and generate corresponding training data sets respectively. These training data sets will be used for model training and optimization to ensure that the model can accurately simulate and predict the load propagation errors under different distribution modes. When multiple load sensors distributed on the surface are used for testing, they can be further divided into edge load sensors and center load sensors to analyze the propagation characteristics of the load at different positions; next, by simulating different load conditions, the displacement data of these sensors under the action of loads are recorded. These simulated data, together with the identification information of the displacement error (such as the difference between the actual displacement and the theoretical displacement), together constitute the first set of training data sets. For multiple load sensors distributed on the line, the load simulation is directly performed on the entire sensor sequence. This simulation also needs to consider different load conditions and propagation characteristics to ensure the comprehensiveness and representativeness of the training data set; by recording the displacement data of each sensor under the simulated load and combining the corresponding displacement error identification information, the second set of training data sets is generated. After obtaining the first and second training data sets, these data can be used to construct a multi-bit error propagation model, which will learn the complex relationship between load propagation error and factors such as sensor distribution and load conditions through a training process based on machine learning or statistical methods; during the training process, the model will continuously adjust its internal parameters to minimize the prediction error and improve the generalization ability of the model. Ultimately, the constructed multi-bit error propagation model will be able to receive new displacement data as input and quickly and accurately predict the error caused by load propagation.

[0029] Furthermore, constructing a multi-bit error propagation model according to the first set of training data sets, the second set of training data sets and identification information identifying displacement errors includes:

[0030] Acquire a multivariate linear regression model; train the multivariate linear regression model according to the first set of training data sets, and output a surface distribution error propagation model; train the multivariate linear regression model according to the second set of training data sets, and output a line distribution error propagation model; integrate the surface distribution error propagation model and the line distribution error propagation model, and output a multi-bit error propagation model.

[0031] In the process of constructing a multi-bit error propagation model, a strategy of phased training and integration of models under different distribution modes is adopted to make full use of the characteristics of sensor data under different distribution modes and improve the adaptability and accuracy of the model. Preferably, a multivariate linear regression model is selected as the basis, the displacement error is regarded as the dependent variable, and various factors affecting the displacement error (such as sensor position, load size, propagation path, etc.) are regarded as independent variables; the multivariate linear regression model is trained using the first set of training data sets (including the displacement data and displacement error identification information of the edge load sensor and the center load sensor), and the model parameters are adjusted so that the model can better fit the relationship between the displacement error in the training data and the sensor distribution, load conditions and other factors. After the training is completed, an error propagation model for the surface distribution load sensor, namely, the surface distribution error propagation model, is obtained; similarly, the same multivariate linear regression model is trained using the second set of training data sets (including the displacement data and displacement error identification information of the line distribution load sensor), so that the model can accurately predict the displacement error between sensors under the line distribution mode. After the training is completed, the line distribution error propagation model is obtained. Finally, the surface distribution error propagation model and the line distribution error propagation model are integrated to form a unified multi-bit error propagation model.

[0032] Furthermore, the method comprises:

[0033] Construct an environmental data acquisition module, and obtain the environmental impact data of the load test platform, including test temperature, test humidity and vibration level, according to the environmental data acquisition module; generate an environmental feedback network layer according to the environmental impact data; optimize the multi-bit error propagation model according to the environmental feedback network layer, and output the optimized multi-bit error propagation model.

[0034] In order to further improve the accuracy and adaptability of the multi-bit error propagation model, it is necessary to consider the impact of the environmental factors of the load test platform on the test results, including test temperature, test humidity and vibration level, etc. These factors may have a significant impact on the performance of the sensor, the propagation characteristics of the load, and the generation and accumulation of errors. Therefore, an environmental data acquisition module is constructed, and an environmental feedback network layer is generated based on it to optimize the multi-bit error propagation model. Specifically, an environmental data acquisition module can be constructed based on a variety of sensors, such as temperature sensors, humidity sensors and vibration sensors, so as to obtain environmental impact data on the load test platform, including test temperature, test humidity and vibration level; based on the collected environmental impact data, an environmental feedback network layer is constructed. The environmental feedback network layer will serve as a supplementary part of the multi-bit error propagation model to simulate and predict the impact of environmental factors on error propagation. By using machine learning or statistical methods, a mathematical model or mapping relationship between environmental parameters and error propagation can be established. This model or mapping relationship will serve as the core of the environmental feedback network layer, which is used to adjust the output of the error propagation model according to real-time or predicted environmental parameters. According to the information provided by the environmental feedback network layer, the multi-bit error propagation model is optimized. Specifically, the environmental parameter prediction value output by the environmental feedback network layer can be used as one of the inputs of the multi-bit error propagation model, so that the model can take into account the influence of environmental factors when predicting errors. At the same time, by comparing the prediction performance of the model under different environmental conditions, the parameters and structure of the model can be further adjusted and optimized to improve its accuracy and stability in various environments.

[0035] Error correction is performed on the second sensor, the third sensor, . . . , and the Nth sensor according to the N-1 error correction indicators.

[0036] After obtaining N-1 error correction indicators, the second sensor, the third sensor, and finally the Nth sensor may be error corrected according to these indicators, that is, the error correction indicator corresponding to each sensor may be used to adjust its measurement value.

[0037] Furthermore, if Figure 2 As shown, according to the N-1 error correction indicators, the second sensor, the third sensor, ... and the Nth sensor are subjected to error correction, and the method includes:

[0038] Record N-1 groups of corrected displacement data after correction by the second sensor, the third sensor... and the Nth sensor; perform transfer load error analysis on the N-1 groups of corrected displacement data, and output N groups of corrected transfer load errors; determine whether the N groups of corrected transfer load errors are less than the preset error threshold, and if the N groups of corrected transfer load errors are greater than the preset error threshold, iteratively optimize the multi-bit error propagation model.

[0039] After completing the error correction from the second sensor to the Nth sensor, record the N-1 groups of corrected displacement data from the second sensor to the Nth sensor after error correction, and perform transfer load error analysis on the N-1 groups of corrected displacement data. Through analysis, N groups of corrected transfer load errors can be obtained, which reflect the amount of error generated by load propagation between sensors after correction. Compare the N groups of corrected transfer load errors with the preset error threshold; if all corrected transfer load errors are less than the preset error threshold, it means that the current error correction and model prediction effects have achieved the expected goals, and no further iterative optimization is required; if any group of corrected transfer load errors is greater than the preset error threshold, it means that the correction effect is not ideal or there is a deviation in the model prediction, and it is necessary to iteratively optimize the multi-bit error propagation model. For example, increase the diversity and number of training data sets to better cover the error conditions under different load conditions and environmental factors; according to the error analysis results, adjust the internal parameters of the multi-bit error propagation model, such as weights, bias terms, etc., to improve the prediction accuracy of the model. Through the iterative optimization process, the performance of the multi-bit error propagation model can be continuously improved until the predetermined error threshold requirements are met.

[0040] Furthermore, according to the N-1 error correction indicators, error correction is performed on the second sensor, the third sensor, ... and the Nth sensor, and the method further includes:

[0041] Obtain an error correction index corresponding to the second sensor, correct the second sensor according to the error correction index, and record a correction transfer influence index caused by correcting the second sensor; optimize the error correction index corresponding to the third sensor according to the correction transfer influence index, and correct the third sensor according to the optimized error correction index.

[0042] The second sensor is corrected based on the error correction index corresponding to the second sensor, and the correction transfer influence index caused by the correction of the second sensor is recorded. The correction transfer influence index represents the additional influence of the second sensor on the measurement value of the third sensor after correction. The error correction index of the third sensor is optimized by using the recorded correction transfer influence index. Since the correction of the second sensor has had a certain influence on the third sensor, it is necessary to re-evaluate the error of the third sensor and consider this transfer influence. By adjusting the error correction index of the third sensor, its measurement error can be further reduced, and the indirect influence caused by the correction of the second sensor is taken into account. The third sensor is corrected according to the optimized error correction index. Through this method, the measurement accuracy of the entire sensor network can be significantly improved, and various complex factors in the load propagation and correction process can be taken into account.

[0043] Furthermore, the third sensor is corrected according to the optimized error correction index, and the correction transfer influence index caused by correcting the third sensor is recorded; the error correction index corresponding to the fourth sensor is optimized according to the correction transfer influence index, and the fourth sensor is corrected according to the optimized error correction index, and so on, to obtain optimized N-1 error correction indicators.

[0044] After completing the correction of the second sensor and recording the correction transfer impact index, the error correction index of the third sensor is optimized according to this index. The optimized error correction index not only considers the error characteristics of the third sensor itself, but also considers the transfer impact caused by the correction of the second sensor. Then, the third sensor is corrected by applying this optimized error correction index, and the correction transfer impact index that may be generated on the subsequent sensor (i.e., the fourth sensor) during the correction process is recorded. The error correction index of the fourth sensor is optimized using the correction transfer impact index caused by the correction of the third sensor. This process is similar to the processing of the third sensor, that is, the transfer impact of the previous sensor correction is considered, and the error correction index of the current sensor is adjusted accordingly. The optimized error correction index is applied to the correction of the fourth sensor, and the correction transfer impact that may be generated on the subsequent sensors is continued to be recorded. This process is continued until the last sensor (i.e., the Nth sensor), that is, for each sensor, its error correction index is optimized according to the transfer impact of its previous sensor correction, and this optimized index is applied for correction. At the same time, the correction transfer impact index is recorded and transmitted to ensure that the errors in the entire sensor network are fully considered and corrected. Through the above-mentioned chain correction process, the optimized N-1 error correction indicators will eventually be obtained. This method significantly improves the measurement accuracy and reliability of the sensor network, allowing the entire system to more accurately reflect the actual load conditions.

[0045] In summary, the embodiments of the present application have at least the following technical effects:

[0046] First, multiple load sensors are arranged on the load test platform, and the multiple load sensors are distributed at equal intervals. By applying a known load to the first load sensor, the first load sensor and N sets of displacement data transmitted from the first load sensor to the second sensor, the third sensor... and the Nth sensor are recorded, where N is the number of load sensors. Then, the transfer load error analysis is performed based on the N sets of displacement data, and N sets of transfer load errors are output. When the N sets of transfer load errors are greater than or equal to the preset error threshold, the N sets of transfer load errors are input into a multi-bit error propagation model, and analyzed according to the multi-bit error propagation model to obtain N-1 error correction indicators. Finally, the second sensor, the third sensor... and the Nth sensor are error corrected according to the N-1 error correction indicators. The technical problem of inaccurate measurement caused by load transfer during the measurement process of the load sensor in the prior art is solved, and the technical effect of improving the load measurement accuracy and reliability is achieved through error correction.

[0047] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0048] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0049] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. An error analysis and correction method for a load sensor, characterized in that: The method comprises: Arrange a plurality of load sensors on the load test platform, wherein the plurality of load sensors are distributed at equal intervals; By applying a known load to a first load sensor, recording the first load sensor and N sets of displacement data transmitted from the first load sensor to a second sensor, a third sensor, ..., and an Nth sensor, where N is the number of load sensors; Performing transfer load error analysis according to the N groups of displacement data, and outputting N groups of transfer load errors; When the N groups of transfer load errors are greater than or equal to a preset error threshold, inputting the N groups of transfer load errors into a multi-bit error propagation model, performing analysis according to the multi-bit error propagation model, and obtaining N-1 error correction indicators; Error correction is performed on the second sensor, the third sensor, . . . , and the Nth sensor according to the N-1 error correction indicators.

2. The error analysis and correction method for a load sensor according to claim 1, characterized in that: According to the N-1 error correction indicators, the second sensor, the third sensor, ... and the Nth sensor are subjected to error correction, the method comprising: Recording N-1 groups of corrected displacement data after correction by the second sensor, the third sensor, ... and the Nth sensor; Performing transfer load error analysis on the N-1 groups of corrected displacement data, and outputting N groups of corrected transfer load errors; It is determined whether the N groups of corrected transfer load errors are less than the preset error threshold; if the N groups of corrected transfer load errors are greater than the preset error threshold, iteratively optimizing the multi-bit error propagation model.

3. The error analysis and correction method for a load sensor according to claim 1, characterized in that: According to the N-1 error correction indicators, the second sensor, the third sensor, ... and the Nth sensor are subjected to error correction, and the method further comprises: Acquire an error correction index corresponding to the second sensor, correct the second sensor according to the error correction index, and record a correction transfer influence index caused by correcting the second sensor; The error correction index corresponding to the third sensor is optimized according to the corrected transfer influence index, and the third sensor is corrected according to the optimized error correction index.

4. The error analysis and correction method for a load sensor according to claim 3, characterized in that: Correcting the third sensor according to the optimized error correction index, and recording the correction transfer influence index caused by correcting the third sensor; The error correction index corresponding to the fourth sensor is optimized according to the corrected transfer influence index, and the fourth sensor is corrected according to the optimized error correction index, and so on, to obtain optimized N-1 error correction indexes.

5. The method according to claim 1, characterized in that By applying a known load to the first load sensor, the method further comprises: Setting a load test sample, wherein the load test sample is obtained by analyzing an application scenario of the load sensor; A known load is applied to the first load sensor according to the load test sample.

6. The error analysis and correction method for a load sensor according to claim 1, characterized in that: The plurality of load sensors include surface distribution and line distribution; When the multiple load sensors are distributed on a surface, the multiple load sensors are divided into edge load sensors and center load sensors, and load simulation is performed with the edge load sensors and the center load sensors to generate a first set of training data sets; When the plurality of load sensors are distributed in a line, load simulation is performed on the plurality of load sensors to generate a second set of training data sets; A multi-bit error propagation model is constructed according to the first set of training data sets, the second set of training data sets, and identification information identifying displacement errors.

7. The error analysis and correction method for a load sensor according to claim 6, characterized in that: Constructing a multi-bit error propagation model according to the first set of training data sets, the second set of training data sets and identification information identifying displacement errors, including: Obtain a multiple linear regression model; Training the multivariate linear regression model according to the first set of training data sets, and outputting a surface distribution error propagation model; Training the multivariate linear regression model according to the second set of training data sets, and outputting a line distribution error propagation model; The surface distribution error propagation model and the line distribution error propagation model are integrated to output a multi-bit error propagation model.

8. The error analysis and correction method for a load sensor according to claim 1, characterized in that: The method comprises: Constructing an environmental data acquisition module, and obtaining environmental impact data of the load test platform according to the environmental data acquisition module, including test temperature, test humidity and vibration level; generating an environmental feedback network layer according to the environmental impact data; The multi-bit error propagation model is optimized according to the environmental feedback network layer, and the optimized multi-bit error propagation model is output.

Images