A multi-sensor based health assessment method for mobile equipment

Abstract

The present invention discloses a method for health assessment of moving equipment based on multiple sensors, including: collecting the operating parameter groups of the moving equipment at different times, including temperature data, noise data, magnetic flux data and vibration acceleration data, and calibrating the health score; determining the vibration severity at the corresponding time using the vibration acceleration data; calculating the actual speed at the corresponding time using the vibration acceleration data and the magnetic flux data; using the spectrum analysis technology to perform fault diagnosis at the corresponding time on the vibration acceleration data and the actual speed; extracting multiple measurement features from the operating parameter group, and obtaining the SPC score at the corresponding time through trend analysis; using the vibration severity factor and the fault severity factor at the corresponding time, and combining the SPC score and the health score to construct a data sample; using the data sample to construct a health assessment model and train it so as to subsequently evaluate the health score of the moving equipment. The health assessment method based on multiple sensor data has high accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis of moving equipment, and in particular to a moving equipment health assessment method based on multiple sensors. Background Art

[0002] Dynamic equipment refers to rotating equipment or reciprocating equipment (i.e. equipment that consumes energy) driven by a driver, such as pumps, compressors, fans, etc. Its energy can be electric power, pneumatic power, steam power, etc. The operating efficiency and reliability of dynamic equipment are crucial to modern industrial production. If equipment fails during the production process, it will not only affect the output and quality, but also require shutdown for maintenance, which seriously affects production efficiency.

[0003] Equipment health assessment refers to the process of analyzing, evaluating and diagnosing the equipment's operating status, performance indicators, etc. By implementing equipment health assessment, equipment failures and hidden dangers can be discovered and eliminated early, equipment reliability and service life can be improved, production costs and risks can be reduced, and production efficiency and economic benefits can be improved. In the health assessment process, traditional equipment diagnosis usually relies on a single type of sensor data, which to a certain extent limits the accuracy and reliability of fault diagnosis, and also affects the accuracy of health assessment, inevitably leading to equipment shutdowns for maintenance. Summary of the invention

[0004] One of the purposes of the present invention is to provide a multi-sensor based dynamic equipment health assessment method, which performs fault analysis and health assessment on dynamic equipment through multi-sensor data to solve the problem of low accuracy of traditional health assessment technology using a single type of sensor data proposed in the background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A multi-sensor based dynamic equipment health assessment method, the method comprising the following steps:

[0007] The operating parameter groups of the moving equipment at different times are collected through multiple sensors, and the health score is calibrated. The operating parameter groups include temperature data, noise data, magnetic flux data and vibration acceleration data;

[0008] Extract the effective value of vibration velocity from the vibration acceleration data, and query the vibration intensity at the corresponding moment in the vibration intensity index according to the power size and installation method of the moving equipment;

[0009] The actual speed of the moving equipment at the corresponding time is calculated by using the speed calculation strategy based on vibration acceleration data and magnetic flux data;

[0010] Use spectrum analysis technology to diagnose dynamic equipment faults at the corresponding moment based on vibration acceleration data and actual speed;

[0011] Extract multiple measurement features from the operating parameter group, perform SPC-based trend analysis on the measurement features at each corresponding moment, and perform data fusion to obtain the SPC score at the corresponding moment;

[0012] The vibration severity factor and fault severity factor at the corresponding moment are obtained by using the vibration severity and fault diagnosis results based on the time series, and the data sample is constructed by combining the SPC score and the health score;

[0013] Build and train a health assessment model based on data samples at different times;

[0014] Use the trained health assessment model to assess the health score of mobile devices.

[0015] Preferably, the vibration acceleration data includes vibration acceleration low-frequency data, vibration acceleration medium-frequency data, and vibration acceleration low-frequency data.

[0016] Preferably, the speed calculation strategy uses the vibration acceleration low-frequency data and the magnetic flux data as calculation parameters to calculate the speed of different moving equipment; the speed calculation strategy includes the following steps:

[0017] Determine whether the driving machine of the moving device is a motor. If so, use the vibration acceleration low-frequency data and magnetic flux data as calculation parameters to obtain the actual speed of the moving device through the first speed calculation method. Otherwise, perform the second judgment logic.

[0018] The second judgment logic is whether the moving device and the motor are in the same unit and the motor speed and the speed ratio of the moving device to the motor are known. If so, the moving device speed is calculated using the motor speed and the speed ratio of the moving device to the motor. Otherwise, the vibration acceleration low-frequency data is used as a calculation parameter, and the actual speed of the moving device is obtained through a third speed calculation method.

[0019] Preferably, the first speed calculation method comprises the following steps:

[0020] Performing spectrum analysis on the magnetic flux data to obtain a magnetic flux spectrum diagram, and determining the frequency corresponding to the highest peak point in the magnetic flux spectrum diagram as the current frequency;

[0021] Calculate the rated speed of the motor according to the current frequency, and determine the speed range based on the rated speed of the motor;

[0022] Perform spectrum analysis on the low-frequency data of vibration acceleration to obtain a spectrum diagram, and determine the actual speed of the moving equipment based on the highest peak within the speed range of the spectrum diagram;

[0023] The third speed calculation method is: performing spectrum analysis on the low-frequency data of vibration acceleration to obtain a vibration spectrum diagram, and determining the actual speed of the moving equipment according to the first peak in the spectrum diagram.

[0024] Preferably, in extracting the effective value of the vibration velocity, the trend of the vibration acceleration intermediate frequency data is removed and integrated to obtain the vibration velocity data in the time domain, and the effective value of the vibration velocity data is calculated to obtain the effective value of the vibration velocity.

[0025] Preferably, the multiple measurement features extracted from the operating parameter group include vibration acceleration peak value, vibration velocity effective value, displacement peak-to-peak value, temperature value, and loudness value.

[0026] Preferably, the fault diagnosis includes a first fault diagnosis and / or a second fault diagnosis, the first fault diagnosis is a structural fault diagnosis, and the second fault diagnosis diagnoses different fault types according to different transmission components.

[0027] Preferably, in the second fault diagnosis, if the moving device is a gear transmission, a gear fault classification method is executed; if the moving device is a bearing transmission, a rolling bearing fault classification method or a sliding bearing fault classification method is executed according to the bearing type of the transmission.

[0028] Preferably, the SPC-based trend analysis comprises the following steps:

[0029] Get all the measures in the historical period and calculate the mean and standard deviation;

[0030] The calculated warning value is the sum of the mean value of the measurement and 3 times the measurement standard deviation, and the alarm value is the sum of the mean value of the measurement and 6 times the measurement standard deviation;

[0031] When the current measurement value is greater than or equal to the alarm value, the SPC score of the corresponding measurement is set to 0; when the current measurement value is less than the warning value, the SPC score of the corresponding measurement is set to 100; when the current measurement value is between the warning value and the alarm value, the interpolation method is used to calculate the SPC score between 0-100.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] The present invention combines data from multiple sensors to analyze the working status of dynamic equipment from different angles, thereby enhancing the accuracy of the health assessment results of the dynamic equipment. At the same time, through continuous monitoring and health assessment, it can continuously track the performance changes of the equipment, detect abnormalities in a timely manner, and improve the reliability of the overall operation of the dynamic equipment. By setting the speed calculation strategy of the dynamic equipment, accurate calculation of the speed of the dynamic equipment can be achieved, providing accurate data support for the subsequent fault diagnosis of the dynamic equipment, ensuring the reliability of fault diagnosis, and further ensuring the reliability of the health assessment results of the dynamic equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 Flowchart of the method for evaluating the health of dynamic equipment.

[0035] Figure 2 Flowchart of the speed calculation strategy.

[0036] Figure 3 This is a second fault diagnosis flowchart in this embodiment. DETAILED DESCRIPTION

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 belong to the protection scope of the present invention.

[0038] The dynamic equipment health assessment method of the present invention sets a vibration acceleration sensor, a temperature sensor, a noise sensor, and a magnetic flux sensor on the dynamic equipment, and each sensor collects parameter data of the dynamic equipment during operation. The dynamic equipment health assessment is performed by integrating multiple parameter data, thereby improving the accuracy and reliability of the assessment.

[0039] Figure 1 A flowchart showing a method for evaluating the health of a moving device is shown in FIG. Figure 1 As shown, the specific process of the dynamic equipment health assessment method is as follows.

[0040] Step 1: Collect operating parameter groups of different moving equipment at different times and calibrate the health score. The operating parameter groups include temperature data, noise data, magnetic flux data and vibration acceleration data. The vibration acceleration data includes acceleration data of different sampling frequencies, which consists of vibration acceleration low-frequency data, vibration acceleration medium-frequency data and vibration acceleration high-frequency data.

[0041] In the present invention, the health score is calibrated by professional and technical personnel in this field according to the operating conditions of the dynamic equipment at different times; each operating parameter group has its corresponding health score.

[0042] In the present invention, the actual rotation speed of the moving equipment is extracted using low-frequency data of vibration acceleration; the effective value of the vibration velocity of the moving equipment is extracted using medium-frequency data of vibration acceleration to determine the vibration severity; the early, middle and late faults of the moving equipment are diagnosed using medium-frequency data of vibration acceleration and high-frequency data of vibration acceleration. Accurate estimation of the running status of the moving equipment is achieved through vibration acceleration data of different sampling frequencies.

[0043] In this embodiment, the vibration acceleration sensor sampling is divided into 250Hz low-frequency sampling, 4000Hz medium-frequency sampling and 25600Hz high-frequency sampling. The medium and low-frequency sampling duration is 10 seconds, and the high-frequency sampling is 1 second. The sampling rate of the flux sensor is 250Hz and the collection duration is 10 seconds. The sound data is collected at a sampling rate of 8K for 10 seconds. The sampling rate of temperature data is 1Hz and the sampling duration is 1s, which is used to monitor the temperature changes of the moving equipment.

[0044] Step 2: extract the effective value of the vibration velocity from the vibration acceleration intermediate frequency data, and query the vibration intensity at the corresponding moment in the vibration intensity index according to the power size and installation method of the moving equipment.

[0045] As one of the specific implementation methods for extracting the effective value of vibration velocity, Butterworth filtering is used to remove the trend of the intermediate frequency data of vibration acceleration, and then integration is performed to obtain a vibration velocity waveform in the time domain, that is, vibration velocity data; the root mean square value of the vibration velocity data is calculated, and the obtained value is the effective value of the vibration velocity.

[0046] In the present invention, the ISO10816 vibration severity index is an important index to judge the quality of the vibration severity of the moving equipment. The corresponding vibration severity level can be found in the ISO10816 vibration severity index for the effective value of the vibration velocity, the power size of the moving equipment and the installation method, including excellent, good, medium and poor.

[0047] Step 3, using a speed calculation strategy based on vibration acceleration data and magnetic flux data, calculate the actual speed of the moving equipment at the corresponding moment.

[0048] In the present invention, accurate speed calculation is the core of dynamic equipment fault diagnosis. A speed calculation strategy is set, and low-frequency data of vibration acceleration and magnetic flux data are used as calculation parameters. Different speed calculation methods are provided for different types of dynamic equipment to obtain more accurate speed values ​​more quickly.

[0049] Reference Figure 2 As shown, specifically, the speed calculation strategy includes the following steps:

[0050] Step 3.1, execute the first judgment logic: determine whether the driving machine of the moving device is a motor, if so, execute the first speed calculation method to calculate the actual speed of the moving device, otherwise execute step 3.2;

[0051] Step 3.2, execute the second judgment logic to determine whether the moving device and the motor are in the same unit and the motor speed and the speed ratio of the moving device and the motor are known. If so, execute the second speed calculation method to calculate the actual speed of the moving device, otherwise execute the third speed calculation method to calculate the actual speed of the moving device.

[0052] In the speed calculation of the present invention, if the moving device is a motor, the collected magnetic flux is that of the moving device, and at this time, the magnetic flux and vibration acceleration low-frequency data are used as calculation parameters; if the moving device is not a motor, but the moving device and the motor are in the same unit, then the magnetic flux is that of the motor in the same unit, and the motor speed can be calculated by the magnetic flux, and then the moving device speed can be calculated by the speed ratio of the motor and the moving device; otherwise, the actual speed of the moving device is directly extracted through the vibration acceleration data.

[0053] The first speed calculation method uses the vibration acceleration low-frequency data and the magnetic flux data as calculation parameters to extract the actual speed of the moving equipment, and specifically includes the following steps:

[0054] (1) Perform spectrum analysis on the magnetic flux data to obtain a magnetic flux spectrum diagram, and determine the frequency corresponding to the highest peak point in the magnetic flux spectrum diagram as the current frequency;

[0055] (2) Calculate the rated speed of the motor according to the current frequency using the formula: 120*current frequency / number of motor poles. Set the speed range [ro,r+o] according to the rated speed r of the motor, where o is a preset value set by a person skilled in the art;

[0056] (3) Perform spectrum analysis on the low-frequency data of vibration acceleration, that is, Fourier transform, to obtain a vibration acceleration spectrum diagram. In the vibration acceleration spectrum diagram, find the highest peak in the frequency range that matches the speed range. The speed corresponding to the frequency of the highest peak is the actual speed of the moving equipment.

[0057] In the first rotation speed calculation method, low-frequency data of vibration acceleration is used, which helps to reduce the frequency spectrum search range.

[0058] In the second speed calculation method, since the moving device and the motor are in the same unit, the speed ratio between the motor and the device driven by the motor is fixed. The speed of the motor can be calculated through the magnetic flux data. At this time, the speed of the moving device driven by the motor can be obtained by using the motor speed and the speed ratio.

[0059] In the third speed calculation method, the vibration acceleration low-frequency data is subjected to spectrum analysis, that is, Fourier transform, to obtain a vibration spectrum diagram, and the moving equipment speed is determined according to the frequency corresponding to the first peak in the vibration acceleration spectrum diagram.

[0060] In this embodiment, a sampling rate of 10 seconds can allow the frequency resolution to reach 0.1 Hz, so as to obtain an accurate rotation speed.

[0061] Step 4: Use spectrum analysis technology to diagnose the fault of the dynamic equipment at the corresponding moment on the vibration acceleration data and the actual rotation speed.

[0062] In step 4 of the present invention, the vibration velocity data extracted from the medium-frequency data of vibration acceleration and the actual rotational speed extracted from the low-frequency data of vibration acceleration are used to diagnose the medium- and late-stage faults of the moving equipment, and the vibration velocity data extracted from the high-frequency data of vibration acceleration and the actual rotational speed extracted from the low-frequency data of vibration acceleration are used to diagnose the early-stage faults of the moving equipment.

[0063] In the present invention, the fault diagnosis is a first fault diagnosis and / or a second fault diagnosis, the first fault diagnosis is a structural fault diagnosis, and the second fault diagnosis diagnoses different fault types according to different transmission components of the moving equipment.

[0064] Structural failures of moving equipment usually include axis misalignment, rotor imbalance, friction and structural looseness.

[0065] (1) Axis misalignment fault diagnosis conditions: The diagnosis conditions are: there is an obvious peak at 2 times the speed frequency in the speed spectrum;

[0066] (2) The judgment condition for rotor imbalance fault is that the vibration velocity spectrum diagram occupies the dominant component at 1 times the rotation frequency and the amplitude is very high. This amplitude can refer to the medium or poor level of vibration severity;

[0067] (3) The judgment condition for friction fault is that there are peaks at 0.5 times the rotation frequency, 1.5 times the rotation frequency, 2.5 times and other fractional multiples of the rotation frequency in the vibration velocity spectrum. The higher the amplitude of the fractional multiple frequency, the more serious the fault;

[0068] (4) Judgment conditions for structural looseness failure: There are peaks at 1, 2, 3, 4, 5, 6, 7, 8, etc. integer multiples of the rotation frequency in the vibration velocity spectrum, and the vibration intensity is medium or poor.

[0069] Structural faults are usually related to the rotational speed frequency, so the frequency range of the analysis is not very high. The present invention uses the speed frequency to perform structural fault analysis, which has less noise than using the acceleration frequency analysis, and can make the analysis simpler.

[0070] Reference Figure 3 In the second fault diagnosis, if the moving equipment is a gear transmission, the gear fault type is determined by executing the gear fault classification method, including wear or breakage; if the moving equipment is a bearing transmission, the rolling bearing fault classification method or the sliding bearing fault classification method is selected according to the bearing type of the transmission to determine the fault types of different bearings.

[0071] The gear fault classification method uses a time domain synchronous averaging algorithm to reduce the noise of the medium-frequency or high-frequency data of vibration acceleration to improve the signal-to-noise ratio, and then calculates the envelope value. The envelope value is subjected to spectral analysis to obtain the envelope spectrum. The peak frequency of the envelope spectrum, the drive shaft gear fault frequency, and the output shaft gear fault frequency are used to monitor the wear or breakage of the gear.

[0072] Rolling bearing faults are mainly four types of faults: inner ring fault BPFI, outer ring fault BPFO, cage fault FTF and ball fault BSF. Rolling bearing faults have specific fault frequencies for different bearing models, but these fault frequencies are generally modulated together with the speed frequency. The rolling bearing fault classification method of the present invention calculates the bandpass filter through the optimal spectrum kurtosis algorithm, and then performs envelope demodulation to highlight the impact signal, and then analyzes the fault frequency characteristics of the bearing. Specifically, the fault frequencies of the four different faults are determined by the bearing model, and the current rolling bearing fault frequency = fault frequency × speed frequency; the bandpass filter parameters are calculated for the vibration acceleration intermediate frequency or high frequency data through the optimal spectrum kurtosis algorithm, and the vibration acceleration intermediate frequency or high frequency data is bandpass filtered and then envelope demodulated to obtain envelope demodulation data, that is, the bearing fault map; the fault type is determined according to the envelope demodulation data and the fault frequency: by accumulating the bearing fault frequency and its multiples to the bearing fault map, it can be checked whether the fault frequency line coincides with the peak on the actual spectrum.

[0073] Sliding bearing faults mainly include bearing wear and oil film vortex faults. The specific method for diagnosing sliding bearings is as follows: Use Butterworth filtering to remove the trend of the medium-frequency or high-frequency data of vibration acceleration, integrate it, and obtain vibration velocity data. Perform spectrum analysis on the vibration velocity data to obtain a vibration velocity spectrum diagram. If there are peaks at 2 times, 3 times, 4 times, 5 times, and 6 times the rotation frequency in the vibration velocity spectrum diagram, and the peak value is large, it is determined that there is a bearing wear fault. The oil film vortex fault is generally called a half-frequency fault, which means that there is a peak at 0.45 times-0.58 times the rotation frequency in the vibration velocity spectrum diagram.

[0074] In step 4 of the present invention, the specific fault diagnosis method for gears and bearings is a conventional technical means in the field and will not be described in detail here.

[0075] The present invention identifies and locates the specific fault mode of the bearing or gear through the second fault diagnosis, realizes accurate classification of the fault, and facilitates the subsequent maintenance of the moving equipment.

[0076] In the present invention, when the fault diagnosis is the first fault diagnosis or the second fault diagnosis, the corresponding fault diagnosis result is directly output, and the fault diagnosis result is normal or a specific fault type; when the fault diagnosis includes the first fault diagnosis and the second fault diagnosis, it is necessary to perform decision fusion on the first fault diagnosis result and the second fault diagnosis result, and when at least one of the fault diagnosis results is a specific fault type, the final fault diagnosis result is "fault", otherwise it is "normal".

[0077] Step 5: extract multiple measurement features from the operating parameter group, perform SPC-based trend analysis on the measurement features at each corresponding moment, perform data fusion on all analysis results, and obtain the SPC score at the corresponding moment.

[0078] In the present invention, trend analysis is based on the analysis of the measurement statistical process control (SPC). The trend analysis of each measurement feature is performed using two SPC statistics, the mean and the standard deviation. The decision-making fusion of the analysis results is a conventional technical means in the field, and those skilled in the art can set it according to the actual situation.

[0079] The specific process of the trend analysis based on SPC is as follows: based on the mean and standard deviation of the measurement values ​​in a historical period, set the warning value and alarm value. The warning value is the sum of the mean and 3 times of the standard deviation, and the alarm value is the sum of the mean and 6 times of the standard deviation; when the measurement value at the current time t is greater than or equal to the alarm value, the SPC score of the corresponding measurement is set to 0; when the measurement value at the current time t is less than the warning value, the SPC score of the corresponding measurement is set to 100; when the measurement value at the current time t is between the warning value and the alarm value, use the interpolation method to calculate the SPC score between 0 and 100. The historical period can be set according to the actual situation and is not limited here.

[0080] In this embodiment, the multiple measurement features extracted include vibration acceleration peak value, vibration velocity effective value, displacement peak-to-peak value, temperature value, and loudness value. The vibration acceleration peak value here is obtained based on the vibration acceleration intermediate frequency data. Usually, temperature is the last indicator reflecting the severity of the fault. At the same time, the loudness value obtained by the noise data is used to assist in vibration fault diagnosis. By performing spectrum analysis on the temperature data and the noise data, the corresponding temperature value and loudness value are obtained.

[0081] Step 6: Use the vibration intensity and fault diagnosis results based on the time series to obtain the vibration intensity factor and fault severity factor at the corresponding moment, and construct a data sample in combination with the SPC score and health score.

[0082] In the present invention, a time series refers to a sequence formed by arranging the values ​​of a certain statistical indicator of a certain phenomenon at different times in chronological order. For example, the vibration intensity based on the time series includes the vibration intensity at the current time t and the historical vibration intensity before the current time t, and they are arranged in chronological order.

[0083] In this embodiment, a fusion result is obtained by decision fusion for the vibration intensity level based on the time series, and the fusion result is used to find the specific score in the preset vibration intensity score evaluation table. For example, the vibration intensity based on the time series includes excellent, good, good, medium, and excellent, which are set in sequence. Excellent appears for the first time and has the highest frequency, so excellent is used as the fusion result at the current time t. In the preset vibration intensity score table, specific scores are set for different vibration intensity levels, and the score of excellent vibration intensity is found as the vibration intensity factor.

[0084] In this embodiment, the fault severity factor is obtained by calculating the frequency of occurrence of the fault in the fault diagnosis result based on the time series. For example, the fault results based on the time series include normal, normal, axis misalignment, rotor imbalance, and normal. Then, the fault severity factor at this time is (2 / 5)*100%=40.

[0085] Step 7: Build and train a health assessment model based on data samples at different times.

[0086] In this embodiment, a health assessment model H=a+bx+cy+dz+exy+fxz+gyz is constructed using a polynomial, where a is a constant term, b, c, d, e, f, and g are all coefficients, x is the SPC score, y is the vibration intensity factor, and z is the fault severity factor; as for how to use a data set composed of data samples to train the model, this is a conventional technical means in this field, and those skilled in the art can set it according to actual conditions. Of course, the health assessment model can also be a deep learning neural network model, or other linear network models, and those skilled in the art can set it according to actual conditions.

[0087] In the present invention, the health score prediction is achieved by constructing a mapping relationship between the SPC score, vibration severity and fault type and the health score. The health score can intuitively indicate the degree of intervention required by the operation and maintenance personnel. The intervention degree is divided into: normal, attention, key attention, recommended inspection and shutdown inspection; the corresponding health score intervals are 80-100, 60-80, 40-60, 20-40, 0-20; the health score and the intervention degree are combined to determine whether to perform maintenance operations on the moving equipment.

[0088] Step 8: Use the trained health assessment model to assess the health score of the mobile device.

[0089] In the present invention, while obtaining the health score through the health assessment model, the first fault diagnosis result and / or the second fault diagnosis result in step 4 can also be output simultaneously to provide a reference for fault location and maintenance by technicians.

[0090] The dynamic equipment health assessment method of the present invention can demonstrate good performance in actual application scenarios, and data calibration is simple and easy.

[0091] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-sensor based dynamic equipment health assessment method, characterized in that: The method comprises the following steps: The operating parameter groups of the moving equipment at different times are collected through multiple sensors, and the health score is calibrated. The operating parameter groups include temperature data, noise data, magnetic flux data and vibration acceleration data; Extract the effective value of vibration velocity from the vibration acceleration data, and query the vibration intensity at the corresponding moment in the vibration intensity index according to the power size and installation method of the moving equipment; The actual speed of the moving equipment at the corresponding time is calculated by using the speed calculation strategy based on vibration acceleration data and magnetic flux data; Use spectrum analysis technology to diagnose dynamic equipment faults at the corresponding moment based on vibration acceleration data and actual speed; Extract multiple measurement features from the operating parameter group, perform SPC-based trend analysis on the measurement features at each corresponding moment, and perform data fusion to obtain the SPC score at the corresponding moment; The vibration severity factor and fault severity factor at the corresponding moment are obtained by using the vibration severity and fault diagnosis results based on the time series, and the data sample is constructed by combining the SPC score and the health score; Build and train a health assessment model based on data samples at different times; Use the trained health assessment model to assess the health score of mobile devices.

2. A multi-sensor based mobile equipment health assessment method as claimed in claim 1, characterized in that: The vibration acceleration data includes vibration acceleration low-frequency data, vibration acceleration medium-frequency data, and vibration acceleration low-frequency data.

3. A multi-sensor based moving equipment health assessment method as claimed in claim 2, characterized in that: The speed calculation strategy uses the vibration acceleration low-frequency data and the magnetic flux data as calculation parameters to calculate the speed of different moving equipment; the speed calculation strategy includes the following steps: Determine whether the driving machine of the moving device is a motor. If so, use the vibration acceleration low-frequency data and magnetic flux data as calculation parameters to obtain the actual speed of the moving device through the first speed calculation method. Otherwise, perform the second judgment logic. The second judgment logic is whether the moving device and the motor are in the same unit and the motor speed and the speed ratio of the moving device to the motor are known. If so, the moving device speed is calculated using the motor speed and the speed ratio of the moving device to the motor. Otherwise, the vibration acceleration low-frequency data is used as a calculation parameter, and the actual speed of the moving device is obtained through a third speed calculation method.

4. A multi-sensor based moving equipment health assessment method as claimed in claim 3, characterized in that: The first speed calculation method comprises the following steps: Performing spectrum analysis on the magnetic flux data to obtain a magnetic flux spectrum diagram, and determining the frequency corresponding to the highest peak point in the magnetic flux spectrum diagram as the current frequency; Calculate the rated speed of the motor according to the current frequency, and determine the speed range based on the rated speed of the motor; Perform spectrum analysis on the low-frequency data of vibration acceleration to obtain a spectrum diagram, and determine the actual speed of the moving equipment based on the highest peak within the speed range of the spectrum diagram; The third speed calculation method is: performing spectrum analysis on the low-frequency data of vibration acceleration to obtain a vibration spectrum diagram, and determining the actual speed of the moving equipment according to the first peak in the spectrum diagram.

5. A multi-sensor based moving equipment health assessment method as claimed in claim 2, characterized in that: In the extraction of the effective value of the vibration velocity, the trend of the intermediate frequency data of the vibration acceleration is removed and integrated to obtain the vibration velocity data in the time domain, and the effective value of the vibration velocity data is calculated to obtain the effective value of the vibration velocity.

6. A multi-sensor based moving equipment health assessment method as claimed in claim 1, characterized in that: The multiple measurement features extracted from the operating parameter group include vibration acceleration peak value, vibration velocity effective value, displacement peak-to-peak value, temperature value, and loudness value.

7. A multi-sensor based mobile equipment health assessment method as claimed in claim 1, characterized in that: The fault diagnosis includes a first fault diagnosis and / or a second fault diagnosis, wherein the first fault diagnosis is a structural fault diagnosis, and the second fault diagnosis diagnoses different fault types according to different transmission components.

8. A multi-sensor based moving equipment health assessment method as claimed in claim 7, characterized in that: In the second fault diagnosis, if the moving device is a gear transmission, the gear fault classification method is executed; if the moving device is a bearing transmission, the rolling bearing fault classification method or the sliding bearing fault classification method is executed according to the bearing type of the transmission.

9. The multi-sensor based moving equipment health assessment method according to claim 1, characterized in that: The SPC-based trend analysis includes the following steps: Get all the measures in the historical period and calculate the mean and standard deviation; The calculated warning value is the sum of the mean value of the measurement and 3 times the measurement standard deviation, and the alarm value is the sum of the mean value of the measurement and 6 times the measurement standard deviation; When the current measurement value is greater than or equal to the alarm value, the SPC score of the corresponding measurement is set to 0; when the current measurement value is less than the warning value, the SPC score of the corresponding measurement is set to 100; when the current measurement value is between the warning value and the alarm value, the interpolation method is used to calculate the SPC score between 0-100.

Images