Abnormality detection method and system based on ultrasonic sensor

Abstract

The present invention innovatively provides an abnormality detection method and system based on ultrasonic sensors, including several groups of ultrasonic sensors at different installation positions, which transmit ultrasonic waves at different angles and receive echo signals to obtain waveform data; pre-process the waveform data at different angles to ensure that all ultrasonic sensors collect data at the same time point, and use the correction principle to correct the waveform to eliminate errors caused by changes in the installation position and angle of the ultrasonic sensor; extract the characteristic values ​​of the corrected data, compare the characteristic values ​​at different angles, and calculate the deviation value between them, preset the abnormality judgment principle, and obtain the deviation value, compare the deviation value with the abnormality judgment principle to obtain the abnormal state; when the abnormal state is a weak signal abnormality, compare the waveform deviations detected by the ultrasonic sensors at different positions, preliminarily judge the weak signal area, and determine the specific position of the weak signal through the feature matching method.

Description

Technical Field

[0001] The present invention relates to the technical field of abnormality detection, and more particularly to an abnormality detection method and system based on ultrasonic sensors. Background Art

[0002] The safety of rail vehicle operation has always attracted widespread attention. However, traditional rail vehicle inspection methods mainly rely on manual inspections, but they have problems such as low efficiency and susceptibility to human interference. Therefore, rail vehicle anomaly detection technology is becoming a hot topic in research and application fields. Generally, there are thousands of parts installed on the bottom of rail vehicles, and many trains need to be inspected and maintained every day. The safety status inspection of vehicle components is an extremely arduous task. Currently, it includes manual inspection and machine vision inspection. Compared with manual inspection, machine vision inspection has better accuracy and efficiency. During the visual inspection process, it is impossible to detect the slight abnormalities of some parts at the bottom of the rail vehicle in a timely manner, and there are visual blind spots, which are not easy to detect. Summary of the invention

[0003] In view of the deficiencies in the prior art, an object of the present invention is to provide an abnormality detection method and system based on an ultrasonic sensor, so as to overcome the above-mentioned defects in the prior art.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] The abnormality detection method based on ultrasonic sensor includes the following steps:

[0006] Ultrasonic waveform collection steps: including several groups of ultrasonic sensors installed at different positions, emitting ultrasonic waves at different angles and receiving echo signals to obtain waveform data;

[0007] Ultrasonic waveform correction step: pre-process the waveform data at different angles to ensure that all ultrasonic sensors collect data at the same time point, and use the correction principle to correct the waveform to eliminate the error caused by the change of the installation position and angle of the ultrasonic sensor;

[0008] Abnormal step: convert the corrected waveform data to the same reference plane, extract the characteristic values ​​of the corrected data, compare the characteristic values ​​at different angles, and calculate the deviation values ​​between them. Preset the abnormal judgment principle, obtain the deviation value, compare the deviation value with the abnormal judgment principle, and judge whether there is an abnormality according to the abnormal judgment principle. If there is an abnormality, perform abnormal analysis on the abnormality to obtain the abnormal state, which includes weak signal abnormality and foreign body abnormality.

[0009] Weak signal position determination step: when the abnormal state is a weak signal abnormality, the waveform deviations detected by ultrasonic sensors at different positions are compared, the weak signal area is preliminarily determined, and the specific position of the weak signal is determined by a feature matching method.

[0010] Preferably, the feature matching method includes: extracting feature values ​​from each ultrasonic sensor and establishing a database containing known weak signal waveform features, comparing the detected waveform features with the features in the database according to a feature matching algorithm, and if the weak signal in the feature library is highly matched with the detected waveform, it is determined that the weak signal position is located in an area with environmental conditions similar to those recorded in the feature library, and when multiple ultrasonic sensors detect similar waveform features, the weak signal position is obtained through a geometric positioning algorithm.

[0011] Preferably, the geometric positioning algorithm is: receiving waveform data detected by different ultrasonic sensors, removing the influence of the train model on the waveform data, and determining the position information of the ultrasonic sensor, obtaining the distance from the weak signal source to each sensor by calculation, and obtaining the specific position of the weak signal source in three-dimensional space by distance calculation.

[0012] Preferably, the correction principle determines the parameters that need to be corrected, such as the waveform shape, and the correction parameters based on the installation position and angle of the sensor, and processes the waveform data using a correction algorithm, the correction algorithm includes polynomial fitting waveform shape correction, and verifies the corrected waveform data.

[0013] Preferably, the abnormality judgment principle includes setting a deviation threshold. When the calculated deviation value exceeds the threshold, it is considered that an abnormal situation exists, and a change trend is set. If the change of the deviation value conforms to the change trend, it is determined to be a foreign body abnormality. If the change of the deviation value does not conform to the change trend, it is determined to be a weak signal abnormality.

[0014] Preferably, a signal enhancement step is also included, which includes temporarily enhancing the signal of the ultrasonic sensor to highlight the characteristics of the target signal, and providing a filtering algorithm to filter out non-target signals in the enhanced signal using the extracted characteristics.

[0015] Preferably, an offline data calibration step is also included, in which the ultrasonic sensor is calibrated offline before formal testing, and the algorithms for waveform correction, deviation calculation and abnormality judgment are continuously optimized based on actual test data and feedback results.

[0016] Ultrasonic sensor-based anomaly detection system, including

[0017] Ultrasonic waveform acquisition module: includes several groups of ultrasonic sensors installed in different positions, which transmit ultrasonic waves at different angles and receive echo signals to obtain waveform data;

[0018] Ultrasonic waveform correction module: pre-processes waveform data at different angles to ensure that all ultrasonic sensors collect data at the same time point, and corrects the waveform using the correction principle to eliminate errors caused by changes in the installation position and angle of the ultrasonic sensor;

[0019] Abnormal module: convert the corrected waveform data to the same reference plane, extract the characteristic values ​​of the corrected data, compare the characteristic values ​​at different angles, and calculate the deviation values ​​between them. Preset the abnormal judgment principle, obtain the deviation value, compare the deviation value with the abnormal judgment principle, and judge whether there is an abnormality according to the abnormal judgment principle. If there is an abnormality, perform abnormal analysis on the abnormality to obtain the abnormal state, which includes weak signal abnormality and foreign body abnormality.

[0020] Weak signal position judgment module: When the abnormal state is a weak signal abnormality, the waveform deviations detected by ultrasonic sensors at different positions are compared, the weak signal area is preliminarily judged, and the specific position of the weak signal is determined by a feature matching method.

[0021] The beneficial effects of the present invention are as follows: during the progress detection of the bottom of the train, ultrasonic sensors arranged at multiple angles are used to detect foreign objects at the bottom of the train and weak signals at the bottom of the train, and train abnormalities are detected through weak signals. In the detection process, the influence of weak signals on ultrasound can be used to detect abnormalities inside the train and to troubleshoot hidden dangers in the train. Multiple ultrasonic sensors with different installation positions are used to improve the detection accuracy of the ultrasonic sensors. The location where the weak signal occurs can be located by the different installation positions of the ultrasonic sensors, and the source of the weak signal can be quickly determined. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is the overall structural diagram of the present invention;

[0023] Figure 2 is a sampling diagram of the ultrasonic sensor of the present invention;

[0024] Figure 3 It is an ultrasonic waveform diagram of the present invention. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] It should be noted that when a component is referred to as being "fixed to" another component, it may be directly on the other component or there may also be a component centered. When a component is considered to be "connected to" another component, it may be directly connected to the other component or there may also be a component centered. When a component is considered to be "set on" another component, it may be directly set on the other component or there may also be a component centered. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0027] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0028] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings:

[0029] The abnormality detection method based on ultrasonic sensor includes the following steps:

[0030] like Figure 1-3 As shown, the present invention provides an ultrasonic waveform collection step: including a plurality of groups of ultrasonic sensors at different installation positions, emitting ultrasonic waves at different angles and receiving echo signals to obtain waveform data, deploying a plurality of groups of ultrasonic sensors at the bottom of a train, these ultrasonic sensors are installed at different positions and angles, and multi-angle collection is performed on the bottom of the train, and each ultrasonic sensor has the ability to emit ultrasonic waves and receive its echo signals, and by controlling each ultrasonic sensor to start emitting ultrasonic waves, waveform data from different directions and angles can be captured, and the waveform data reflects the reflection information of the train to the ultrasonic waves in the environment;

[0031] Ultrasonic waveform correction steps: pre-process the waveform data at different angles to ensure that all ultrasonic sensors collect data at the same time point, and use the correction principle to correct the waveform to eliminate the errors caused by the changes in the installation position and angle of the ultrasonic sensor; due to the different installation positions and angles of the ultrasonic sensors, the received waveform data has deviations in time and form, and the waveform data at different angles needs to be corrected to eliminate the errors caused by the physical position; synchronize the data between different sensors through timestamp calibration, and adjust the waveform data according to the installation position of the sensor using the correction principle to ensure that the data collected by all ultrasonic sensors are logically on the same reference plane;

[0032] Abnormal step: convert the corrected waveform data to the same reference plane and extract the eigenvalues ​​of the corrected data. The eigenvalues ​​include amplitude, frequency, phase, waveform shape parameters, etc. These eigenvalues ​​can reflect the essential attributes of the waveform data. Compare the eigenvalues ​​at different angles and calculate the deviation between them. The size and change trend of the deviation value reflect whether there is an abnormality in the environment. Preset the abnormal judgment principle and obtain the deviation value. Compare the deviation value with the abnormal judgment principle. Determine whether there is an abnormality according to the abnormal judgment principle. If there is an abnormality, perform an abnormal analysis on the abnormality to obtain the abnormal state. The abnormal state includes The abnormality judgment principle is to set a deviation threshold. When the calculated deviation value exceeds the threshold, it is considered that an abnormal situation exists, and a change trend is set. If the change of the deviation value conforms to the change trend, it is considered to be a foreign body abnormality. If the change of the deviation value does not conform to the change trend, it is considered to be a weak signal abnormality. A deviation threshold of one or more characteristic values ​​is set. When the characteristic value exceeds the deviation threshold, it is considered to be an abnormal situation. When analyzing the change trend of the characteristic, if the characteristic value suddenly changes, it is considered to be a foreign body abnormality. If the deviation value changes continuously, it is considered to be a weak signal abnormality.

[0033] Weak signal position judgment steps: When the abnormal state is a weak signal abnormality, determine the specific position of the weak signal through further judgment; compare the waveform deviations detected by ultrasonic sensors at different positions. Since weak signals will be attenuated and interfered during propagation, the waveform deviation detected by the ultrasonic sensor closer to the signal source is more obvious. Through the comparison results of the waveform deviations, preliminarily judge the weak signal area, and determine the specific position of the weak signal through the feature matching method. Through the waveform deviations detected by multiple sensors, the unknown occurrence of weak signals is matched.

[0034] The feature matching method includes: extracting feature values ​​from each ultrasonic sensor, the feature values ​​can distinguish different fixed weak signals or signal types, and establishing a database containing known weak signal waveform features, the waveform features correspond to weak signals in different environments, collecting sample data of various weak signals in different environments, and extracting features from the sample data, extracting features from each sample data, and storing the extracted features and the corresponding feature extraction and environmental condition information in the database; comparing the detected waveform features with the features in the database according to the feature matching algorithm, if the weak signal in the feature library is highly matched with the detected waveform, it is judged that the weak signal position is located in an area similar to the environmental conditions recorded in the feature library, comparing the detected waveform features with the waveform features in the database, finding the matching weak signal position area, and identifying the features of different weak signals through a training model; and when multiple ultrasonic sensors detect similar waveform features, obtaining the weak signal position through a geometric positioning algorithm, estimating the position of the weak signal through the geometric positioning algorithm, and further determining the position of the weak signal.

[0035] The geometric positioning algorithm is as follows: receiving the waveform data detected by different ultrasonic sensors, removing the influence of the train model on the waveform data, and determining the position information of the ultrasonic sensor, obtaining the distance from the weak signal source to each sensor by calculation, and obtaining the specific position of the weak signal source in three-dimensional space by distance calculation; marking the sensor position as , , , the time difference (or phase difference) of the signal received by the sensor has been converted into a distance difference, recorded as wait, Indicates a weak signal source to the sensor , The distance difference; Assuming the signal source position is , and the distances to the sensors are , , ,

[0036]

[0037]

[0038]

[0039] The distance difference equation is

[0040] Through the above formula, the three-dimensional coordinates of the sensor can be calculated.

[0041] Correction principle: According to the installation position and angle of the sensor, the parameters that need to be corrected, such as the waveform shape, are determined, and the correction parameters are determined, and the waveform data is processed using a correction algorithm, wherein the correction algorithm includes polynomial fitting waveform shape correction, and the corrected waveform data is verified; according to the installation position and angle of the sensor, as well as the characteristics of the waveform data, the parameters that need to be corrected, such as the waveform shape, are determined, and according to the determined correction parameters, a suitable correction algorithm is selected to process the waveform data, and the corrected waveform data is verified to ensure that it can accurately reflect the actual environment; in the waveform correction process summary, the collected waveform data can be processed using polynomial fitting to eliminate errors caused by factors such as sensor position and angle; waveform data at different installation positions and angles are obtained, and a polynomial curve is mathematically fitted using the collected data points, and the fitted polynomial curve is applied to the original waveform data to correct it.

[0042] The correction principle also includes phase correction. During the signal processing process, due to the influence of factors such as the installation position and angle of the ultrasonic sensor, the signal may have a phase deviation. Phase correction is used to eliminate the deviation and ensure the accuracy of the signal.

[0043] The correction principles include shape correction. Shape correction includes establishing a waveform shape change model based on the geometric shape of the bottom of the train. The waveform shape change model reflects the changing law of the waveform shape under different installation positions. The established correction model is corresponded to the waveform data under different angles, and the shape of the waveform is adjusted.

[0044] The signal enhancement step includes temporarily enhancing the signal of the ultrasonic sensor to highlight the characteristics of the target signal, and providing a filtering algorithm to filter out non-target signals in the enhanced signal using the extracted characteristics; adjusting the gain setting of the ultrasonic sensor for the noise source of the ultrasonic sensor signal in the train detection environment, such as internal noise of the sensor, electromagnetic interference, etc., increasing the amplitude of the target signal, pre-processing the signal by using advanced signal processing techniques such as adaptive filtering and wavelet transform, analyzing the time domain characteristics of the signal, such as root mean square value, energy, slope, etc., which can reflect the strength and change trend of the signal, and extracting the characteristics that can represent the target signal. Characteristic parameters; Perform Fourier transform or other frequency domain transform on the signal, analyze the spectral distribution of the signal, and extract key features in the spectrum, such as resonant frequency, bandwidth, center frequency, etc. These features help to identify the characteristics of the target signal; design a filtering algorithm, input the enhanced signal into the filter, and filter out non-target signals. The filtering effect can be optimized by adjusting the filter parameters, such as cutoff frequency, bandwidth, etc.; use multi-sensor data fusion technology to comprehensively process data from different sensors. Commonly used data fusion methods include weighted averaging, Kalman filtering, neural network method, etc. Through data fusion, the reliability and accuracy of detection can be improved, and interference signals can be further filtered out.

[0045] It also includes an offline data calibration step. Before the formal detection, the ultrasonic sensor is calibrated offline, and the waveform correction, deviation calculation and abnormality judgment algorithms are continuously optimized according to the actual detection data and feedback results; a standard object with known size, shape and surface features is selected as a calibration reference, the standard object is placed within the detection range, and data collection begins to record the ultrasonic sensor's measurement of the standard object and the actual size value of the standard object, and the measurement error is calculated, and the cause of the error is analyzed, such as the ultrasonic sensor's poor offset, angle offset, operating frequency, etc., and the calibration parameters of the ultrasonic sensor are adjusted according to the analysis results to reduce the measurement error. After completing the preliminary calibration, the standard object is measured again to verify the calibration results, and the calibration is repeated until the requirements are met. According to the actual detection data and feedback results, the waveform correction, deviation calculation and abnormality judgment algorithms are continuously optimized, and the measurement accuracy and stability of the ultrasonic sensor are improved through algorithm optimization.

[0046] Ultrasonic sensor-based anomaly detection system, including

[0047] Ultrasonic waveform acquisition module: includes several groups of ultrasonic sensors installed in different positions, which transmit ultrasonic waves at different angles and receive echo signals to obtain waveform data;

[0048] Ultrasonic waveform correction module: pre-processes waveform data at different angles to ensure that all ultrasonic sensors collect data at the same time point, and corrects the waveform using the correction principle to eliminate errors caused by changes in the installation position and angle of the ultrasonic sensor;

[0049] Abnormal module: convert the corrected waveform data to the same reference plane, extract the eigenvalues ​​of the corrected data, compare the eigenvalues ​​at different angles, and calculate the deviation between them. Preset the abnormal judgment principle, obtain the deviation value, compare the deviation value with the abnormal judgment principle, and judge whether there is an abnormality according to the abnormal judgment principle. If there is an abnormality, perform abnormal analysis on the abnormality to obtain the abnormal state, which includes weak signal abnormality and foreign body abnormality.

[0050] Weak signal position judgment module: When the abnormal state is a weak signal abnormality, the waveform deviations detected by ultrasonic sensors at different positions are compared, the weak signal area is preliminarily judged, and the specific position of the weak signal is determined by the feature matching method.

[0051] The above are only preferred embodiments of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. An abnormality detection method based on an ultrasonic sensor, characterized in that: The steps include: Ultrasonic waveform collection steps: including several groups of ultrasonic sensors installed at different positions, emitting ultrasonic waves at different angles and receiving echo signals to obtain waveform data; Ultrasonic waveform correction step: pre-process the waveform data at different angles to ensure that all ultrasonic sensors collect data at the same time point, and use the correction principle to correct the waveform to eliminate the error caused by the change of the installation position and angle of the ultrasonic sensor; Abnormal judgment step: convert the corrected waveform data to the same reference plane, extract the characteristic value of the corrected data, compare the characteristic values ​​at different angles, and calculate the deviation value between them, preset the abnormal judgment principle, and obtain the deviation value, compare the deviation value with the abnormal judgment principle, and judge whether there is an abnormality according to the abnormal judgment principle. If there is an abnormality, perform abnormal analysis on the abnormality to obtain an abnormal state, which includes weak signal abnormality and foreign body abnormality; Weak signal position determination step: when the abnormal state is a weak signal abnormality, the waveform deviations detected by ultrasonic sensors at different positions are compared, the weak signal area is preliminarily determined, and the specific position of the weak signal is determined by a feature matching method.

2. The abnormality detection method based on ultrasonic sensor according to claim 1, characterized in that: The feature matching method includes: extracting feature values ​​from each ultrasonic sensor and establishing a database containing known weak signal waveform features, comparing the detected waveform features with the features in the database according to a feature matching algorithm, and if the weak signal in the feature library is highly matched with the detected waveform, it is determined that the weak signal position is located in an area with similar environmental conditions recorded in the feature library, and when multiple ultrasonic sensors detect similar waveform features, the weak signal position is obtained through a geometric positioning algorithm.

3. The abnormality detection method based on ultrasonic sensor according to claim 2, characterized in that: The geometric positioning algorithm is as follows: receiving waveform data detected by different ultrasonic sensors, removing the influence of the train model on the waveform data, and determining the position information of the ultrasonic sensor, obtaining the distance from the weak signal source to each sensor by calculation, and obtaining the specific position of the weak signal source in three-dimensional space by distance calculation.

4. The abnormality detection method based on ultrasonic sensor according to claim 1, characterized in that: The correction principle determines the parameters that need to be corrected, such as the waveform shape, according to the installation position and angle of the sensor, and determines the correction parameters, and uses the correction algorithm to process the waveform data. The correction algorithm includes polynomial fitting waveform shape correction, and verifies the corrected waveform data.

5. The abnormality detection method based on ultrasonic sensor according to claim 1, characterized in that: The correction principle includes shape correction, which includes establishing a waveform shape change model based on the geometric shape of the bottom of the train. The waveform shape change model reflects the change law of the waveform shape under different installation positions, and the established correction model is corresponded to the waveform data under different angles, and the shape of the waveform is adjusted.

6. The abnormality detection method based on ultrasonic sensor according to claim 1, characterized in that: The abnormality judgment principle includes setting a deviation threshold. When the calculated deviation value exceeds the threshold, it is considered that an abnormal situation exists, and a change trend is set. If the change of the deviation value conforms to the change trend, it is determined to be a foreign body abnormality. If the change of the deviation value does not conform to the change trend, it is determined to be a weak signal abnormality.

7. The abnormality detection method based on ultrasonic sensor according to claim 1, characterized in that: The method also includes a signal enhancement step, which includes temporarily enhancing the signal of the ultrasonic sensor to highlight the characteristics of the target signal, and providing a filtering algorithm to filter out non-target signals in the enhanced signal using the extracted characteristics.

8. The abnormality detection method based on ultrasonic sensor according to claim 1, characterized in that: It also includes an offline data calibration step. Before the formal test, the ultrasonic sensor is calibrated offline. According to the actual test data and feedback results, the waveform correction, deviation calculation and abnormality judgment algorithms are continuously optimized.

9. An abnormality detection system based on ultrasonic sensors, characterized in that: include Ultrasonic waveform acquisition module: includes several groups of ultrasonic sensors installed in different positions, which transmit ultrasonic waves at different angles and receive echo signals to obtain waveform data; Ultrasonic waveform correction module: pre-processes waveform data at different angles to ensure that all ultrasonic sensors collect data at the same time point, and corrects the waveform using the correction principle to eliminate errors caused by changes in the installation position and angle of the ultrasonic sensor; Abnormal judgment module: convert the corrected waveform data to the same reference plane, extract the characteristic values ​​of the corrected data, compare the characteristic values ​​at different angles, and calculate the deviation values ​​between them. Preset the abnormal judgment principle, obtain the deviation value, compare the deviation value with the abnormal judgment principle, and judge whether there is an abnormality according to the abnormal judgment principle. If there is an abnormality, perform abnormal analysis on the abnormality to obtain the abnormal state, which includes weak signal abnormality and foreign body abnormality. Weak signal position judgment module: When the abnormal state is a weak signal abnormality, the waveform deviations detected by ultrasonic sensors at different positions are compared, the weak signal area is preliminarily judged, and the specific position of the weak signal is determined by a feature matching method.

Images