Equipment fault prediction method, system and equipment based on Internet of Things, and readable medium

Abstract

The invention relates to an equipment fault prediction method and system based on the Internet of Things, equipment and a readable medium. The method comprises the following steps: acquiring real-time data of equipment; aggregating the real-time data by using an Internet of Things gateway, and carrying out preliminary cleaning on the real-time data; preprocessing the preliminarily cleaned real-time data to obtain real-time feature data; comparing the real-time feature data with historical data to obtain the similarity between the real-time feature data and the historical data; if the similarity is smaller than a similarity threshold value, constructing a fault prediction algorithm model; inputting the real-time feature data into a fault prediction algorithm model for prediction so as to obtain a health state score of the equipment; obtaining a fault risk score based on the health state score; and if the fault risk score is higher than a score threshold, the fault prediction algorithm model outputs a fault type. Potential faults of the equipment can be timely and accurately found and early warned, the maintenance cost is reduced, and the production efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, system, device and readable medium for predicting equipment failure based on the Internet of Things. Background Art

[0002] In industrial production, equipment failure can lead to production stagnation and economic losses. Currently, equipment failures are mainly detected by regular manual inspections or simple threshold alarm systems. However, the above methods cannot accurately predict equipment failures in a timely manner, affecting production efficiency and increasing maintenance costs. Summary of the invention

[0003] Based on this, it is necessary to provide an equipment failure prediction method, system, device and readable medium based on the Internet of Things to address the above technical problems.

[0004] In a first aspect, the present application provides a device failure prediction method based on the Internet of Things, and the device failure prediction method based on the Internet of Things includes:

[0005] Acquire real-time data of the device, wherein the real-time data includes sensor data collected by multiple sensors;

[0006] Aggregate the real-time data using an IoT gateway and perform preliminary cleaning on the real-time data;

[0007] Preprocessing the real-time data after preliminary cleaning to obtain real-time feature data;

[0008] Comparing the real-time feature data with the historical data to obtain the similarity between the real-time feature data and the historical data;

[0009] If the similarity is less than the similarity threshold, a fault prediction algorithm model is constructed;

[0010] Inputting the real-time feature data into the fault prediction algorithm model for prediction to obtain a health status score of the equipment;

[0011] Obtaining a failure risk score based on the health status score;

[0012] If the fault risk score is higher than the score threshold, the fault prediction algorithm model outputs the fault type.

[0013] Optionally, the preprocessing of the preliminarily cleaned real-time data to obtain real-time feature data includes:

[0014] Clean the real-time data after preliminary cleaning to remove noise and outliers;

[0015] The cleaned real-time data is subjected to feature extraction to obtain the real-time feature data; the real-time feature data includes the temperature change rate, the frequency and the fluctuation index of the vibration signal.

[0016] Optionally, before comparing the real-time feature data with the historical data to obtain the similarity between the real-time feature data and the historical data, the method further includes:

[0017] Aligning the real-time feature data in time series;

[0018] The aligned real-time feature data are fused; and the fused real-time feature data are compared with the historical data.

[0019] Optionally, the similarity between the real-time feature data and the historical data is obtained by using the following formula:

[0020]

[0021] X 实时 is the real-time feature data, X 历史 is the historical data, . represents the dot product, ||X 实时 || is the Euclidean norm of the real-time feature data, ||X 历史 || is the Euclidean norm of the historical data.

[0022] Optionally, the constructing of a fault prediction algorithm model includes:

[0023] Select the LSTM network model;

[0024] Adjusting the number of neurons in the input layer of the long short-term memory network according to the dimension of the real-time feature data;

[0025] Increasing the number of hidden layers or neurons in the long short-term memory network model;

[0026] Regularize the long short-term memory network model;

[0027] The regularized long short-term memory network model is trained, verified and optimized to obtain the fault prediction algorithm model.

[0028] Optionally, the following formula is used to obtain a fault risk score based on the health status score:

[0029]

[0030] Among them, R 风险 is the failure risk score, and HScore is the health status score.

[0031] Optionally, if the fault risk score is higher than a score threshold, the fault prediction algorithm model outputs a fault type based on a support vector machine or a decision tree.

[0032] In a second aspect, the present application further provides an IoT-based device failure prediction system, the IoT-based device failure prediction system comprising:

[0033] A data acquisition module, used to acquire real-time data of the device; the data acquisition module includes a plurality of sensors, and the real-time data includes sensor data collected by the plurality of sensors;

[0034] An Internet of Things gateway, used to aggregate the real-time data and perform preliminary cleaning on the real-time data;

[0035] A preprocessing module is used to preprocess the real-time data after preliminary cleaning to obtain real-time feature data;

[0036] A comparison module, used for comparing the real-time feature data with the historical data to obtain the similarity between the real-time feature data and the historical data;

[0037] A fault prediction algorithm model construction module is used to construct a fault prediction algorithm model when the similarity is less than a similarity threshold; the fault prediction algorithm model is used to obtain a health status score of the device based on the real-time feature data, obtain a fault risk score based on the health status score, and output a fault type when the fault risk score is higher than a score threshold.

[0038] In a third aspect, the present application also provides a device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of any of the above-mentioned methods for predicting equipment faults based on the Internet of Things are implemented.

[0039] In a fourth aspect, the present application also provides a readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any of the above-mentioned methods for predicting device faults based on the Internet of Things are implemented.

[0040] In the above-mentioned device fault prediction method, system, device and readable medium based on the Internet of Things, real-time data of the device is obtained, and the real-time data includes sensor data collected by multiple sensors; the real-time data is aggregated using an Internet of Things gateway, and the real-time data is preliminarily cleaned; the real-time data after preliminary cleaning is preprocessed to obtain real-time feature data; the real-time feature data is compared with historical data to obtain the similarity between the real-time feature data and the historical data; if the similarity is less than a similarity threshold, a fault prediction algorithm model is constructed; the real-time feature data is input into the fault prediction algorithm model for prediction to obtain a health status score of the device; a fault risk score is obtained based on the health status score; if the fault risk score is higher than the score threshold, the fault prediction algorithm model outputs the fault type; potential faults of the equipment can be discovered and warned in a timely and accurate manner, reducing maintenance costs and improving production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A flowchart of a device failure prediction method based on the Internet of Things provided in one embodiment of the present application;

[0042] Figure 2 A structural block diagram of an equipment failure prediction system based on the Internet of Things provided in another embodiment of the present application;

[0043] Figure 3 This is a diagram of the internal structure of a computer device in yet another embodiment of the present application. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0045] In one embodiment, see Figure 1 The present application provides a device failure prediction method based on the Internet of Things. The device failure prediction method based on the Internet of Things may include the following steps: S10 to S80.

[0046] S10: Acquire real-time data of the device, where the real-time data includes sensor data collected by multiple sensors.

[0047] S20: Aggregate the real-time data using an IoT gateway, and perform preliminary cleaning on the real-time data.

[0048] S30: pre-processing the real-time data after preliminary cleaning to obtain real-time feature data.

[0049] S40: Compare the real-time feature data with the historical data to obtain the similarity between the real-time feature data and the historical data.

[0050] S50: If the similarity is less than the similarity threshold, construct a fault prediction algorithm model.

[0051] S60: Inputting the real-time feature data into the fault prediction algorithm model for prediction to obtain a health status score of the device.

[0052] S70: Obtain a failure risk score based on the health status score.

[0053] S80: If the fault risk score is higher than the score threshold, the fault prediction algorithm model outputs the fault type.

[0054] The IoT-based equipment failure prediction method of the present application can timely and accurately discover and warn of potential failures of the equipment, reduce maintenance costs, and improve production efficiency.

[0055] In step S10, refer to Figure 1 In step S10, real-time data of the device is acquired, and the real-time data includes sensor data collected by multiple sensors.

[0056] As an example, the multiple sensors may include temperature sensors, vibration sensors, pressure sensors, etc.; each of the sensors may collect relevant sensor data from the device at fixed time intervals. The multiple sensors are precisely installed at locations where relevant sensor data can be accurately collected. Specifically, the corresponding sensors may be selected for installation and configuration according to the specific model of the device and monitoring requirements, and the comprehensiveness and accuracy of the real-time data collected may be ensured. Each of the sensors may convert the collected sensor data from analog signals to digital signals.

[0057] In step S20, refer to Figure 1 In step S20, the real-time data is aggregated using an IoT gateway and the real-time data is preliminarily cleaned.

[0058] As an example, the IoT gateway may be an IoT gateway deployed on the device.

[0059] As an example, the IoT gateway may perform preliminary cleaning on the real-time data to remove obvious outliers (eg, abnormal real-time data caused by sensor anomalies, etc.).

[0060] As an example, real-time data after preliminary cleaning can also be packaged and uploaded to the cloud or central data storage system using a secure data transmission protocol (such as SSL or TLS), and encryption measures (for example, encrypted channels such as VPN, etc.) are used during the transmission process to ensure data security.

[0061] In this embodiment, the real-time data is preliminarily processed through the Internet of Things gateway, which has availability and scalability, can store massive time series data, and provide support for subsequent data analysis and prediction.

[0062] In a specific embodiment, the temperature sensor can be installed at a key heat source point of the device, such as a motor housing, to monitor the surface temperature. The vibration sensor can be installed on a bearing or structural component of the device to capture abnormal vibration data. The pressure sensor can be configured in a pressure change sensitive area such as a hydraulic system of the device to record pressure data in real time. After the device is started, each sensor starts to collect data at a preset frequency (for example, once per second). The Internet of Things gateway collects the real-time data once a minute and performs a preliminary check (i.e., preliminary cleaning) on ​​the real-time data to eliminate extreme readings (i.e., sensor data) caused by sensor failure. The real-time data after preliminary cleaning is uploaded to the central data storage system through an encrypted channel (e.g., VPN) for subsequent predictive analysis.

[0063] In step S30, refer to Figure 1 In step S30, the real-time data after preliminary cleaning is preprocessed to obtain real-time feature data.

[0064] As an example, step S30 may include the following steps: S301 - S302 .

[0065] S301: Clean the real-time data after preliminary cleaning to remove noise and outliers.

[0066] S302: Extract features from the cleaned real-time data to obtain the real-time feature data; the real-time feature data includes the temperature change rate, the frequency and fluctuation index of the vibration signal

[0067] Specifically, in step S301, statistical methods or machine learning algorithms (such as box plots and cluster analysis, etc.) can be used, but not limited to, to automatically identify and remove noise and outliers that deviate significantly from normal data distribution. For example, for the temperature data collected by the temperature sensor, data points that exceed the normal operating temperature range of the device can be excluded through this step.

[0068] Specifically, in step S302, by performing feature extraction on the cleaned real-time data, key indicators related to device performance can be captured.

[0069] Specifically, the temperature change rate may be obtained by acquiring the change of temperature over time, so as to obtain real-time characteristic data reflecting the change trend of the thermal state of the device.

[0070] Specifically, the frequency of the vibration signal of the device may be analyzed by Fast Fourier Transform (FFT) to obtain real-time characteristic data that can identify potential fault frequency patterns.

[0071] Specifically, the fluctuation of the pressure data may be acquired to obtain a fluctuation index that can detect abnormal pressure changes.

[0072] Specifically, the temperature change rate can be obtained by using the key data formula expressed in LaTeX as follows:

[0073]

[0074] Among them, △T is the temperature change and △t is the time interval.

[0075] As an example, the Fourier transform of a vibration signal is a process of converting a vibration signal from the time domain to the frequency domain. This transformation can help us understand the frequency components of the vibration signal, that is, the combination of sine waves and cosine waves of different frequencies. The Fourier transform of the vibration signal in this embodiment can be obtained by the following formula:

[0076]

[0077] Where, X(f) is the Fourier transform result of the vibration signal at frequency f, which is a complex number representing the amplitude and phase of the vibration signal at frequency f; x(t) is the vibration signal, which is a function of time t; e -j2πft is the core of Fourier transform, which represents a complex wave with frequency f.

[0078] Specifically, the Fourier transform decomposes the vibration signal in the time domain into a combination of sine waves and cosine waves of different frequencies, thereby obtaining the spectrum of the vibration signal; this spectrum can tell us which frequency components in the vibration signal are significant, as well as their relative strengths and phases.

[0079] It should be noted that in practical applications, Fourier transform is usually calculated by numerical methods. Especially when processing the vibration data actually collected, these data are usually discrete. Therefore, discrete Fourier transform (DFT) or its fast algorithm (FFT) can be used for spectrum analysis.

[0080] In one embodiment, the real-time data after preliminary cleaning can be cleaned using a moving average filter or an IQR (interquartile range) method to remove outliers. The temperature data is differentially operated to calculate the temperature change rate and record the average value per minute; the vibration data is FFT transformed to identify the frequency components and extract the frequency peaks associated with the known fault model. The pressure data is analyzed in time series to calculate the fluctuation index to reflect the pressure stability. The extracted real-time feature data is stored.

[0081] Through the above preprocessing steps, the data quality can be significantly improved, providing important feature data information for subsequent equipment status prediction.

[0082] As an example, after step S30, the following steps may also be included: aligning the real-time feature data in time series; fusing the aligned real-time feature data; and comparing the fused real-time feature data with the historical data.

[0083] Specifically, the real-time feature data from different sensors may be aligned in time series to ensure the consistency and integrity of the real-time feature data; for example, the temperature change rate, the frequency and fluctuation index of the vibration signal may be merged according to the timestamp.

[0084] Specifically, data fusion technology, such as multi-sensor data fusion algorithm, can be used to integrate multi-source real-time feature data into a unified input vector to provide a comprehensive equipment status description for subsequent fault prediction algorithm models.

[0085] In one embodiment, the real-time feature data from different sensors may be merged into a single feature vector according to the timestamp; a data fusion algorithm (eg, weighted average method) may be applied to combine the features according to their importance.

[0086] In step S40, refer to Figure 1 In step S40, the real-time feature data is compared with the historical data to obtain the similarity between the real-time feature data and the historical data.

[0087] As an example, in step S40, a historical data set may be established first, and the historical data set needs to include historical normal operation data of the equipment and historical fault data of the equipment; the historical normal operation data and the historical fault data may both be labeled data.

[0088] As an example, in step S40, the real-time feature data and the historical data may be compared using similarity to determine whether the device state of the device deviates from a normal range.

[0089] As an example, in step S40, the similarity between the real-time feature data and the historical data may be obtained by using the following formula:

[0090]

[0091] X 实时 is the real-time feature data, X 历史 is the historical data, . represents the dot product, ||X 实时 || is the Euclidean norm of the real-time feature data, ||X 历史 || is the Euclidean norm of the historical data.

[0092] In step S50, refer to Figure 1 In step S50, if the similarity is less than the similarity threshold, a fault prediction algorithm model is constructed.

[0093] As an example, in step S50, the similarity may be compared with the similarity threshold, and if the similarity is less than the similarity threshold, a further fault prediction process is triggered.

[0094] As an example, in step S50, building a fault prediction algorithm model may include the following steps: S501 to S505.

[0095] S501: Select a long short-term memory network model (LSTM).

[0096] S502: Adjust the number of neurons in the input layer of the long short-term memory network according to the dimension of the real-time feature data.

[0097] S503: Increase the number of hidden layers or neurons in the long short-term memory network model.

[0098] S504: Regularize the long short-term memory network model.

[0099] S505: Training, verifying and optimizing the regularized long short-term memory network model to obtain the fault prediction algorithm model.

[0100] Specifically, in step S504, the Dropout technology can be used to regularize the long short-term memory network model to reduce the risk of overfitting.

[0101] Specifically, in step S505, the historical data of the device is input into the long short-term memory network model after regularization processing to train the long short-term memory network model; the historical data includes the historical normal operation data of the device and the historical fault data of the device; the weight of the long short-term memory network model is adjusted by using the back propagation algorithm and the gradient descent optimization algorithm. The prediction performance of the trained long short-term memory network model is verified by the cross-validation method. The model parameters of the long short-term memory network model, such as the learning rate and the hidden layer structure, are adjusted according to the verification results to improve the accuracy of subsequent fault prediction.

[0102] In one embodiment, historical data of the equipment may be collected, the historical data including historical normal operation data of the equipment and historical fault data of the equipment. The above-mentioned long short-term memory network model is trained, verified and optimized using the historical data to obtain a fault prediction algorithm model. The fault prediction algorithm model is regularly retrained and optimized using newly collected data to adapt to possible changes in the equipment.

[0103] In step S60, refer to Figure 1 In step S60, the real-time feature data is input into the fault prediction algorithm model for prediction to obtain a health status score of the equipment.

[0104] As an example, the real-time feature data is input into the fault prediction algorithm model for processing to obtain the health status score (HScore)

[0105] As an example, the health status score may be a value from 0 to 100, which is used to reflect the current health status of the device.

[0106] In step S70, refer to Figure 1 In step S70, a fault risk score is obtained based on the health status score.

[0107] As an example, the fault prediction algorithm model may use the following formula to obtain a fault risk score based on the health status score:

[0108]

[0109] Among them, R 风险 is the failure risk score, and HScore is the health status score.

[0110] As an example, the failure risk score may be a value between 0 and 1, used to indicate the possibility of failure of the device.

[0111] In step S80, refer to Figure 1In S80, if the fault risk score is higher than the score threshold, the fault prediction algorithm model outputs the fault type.

[0112] As an example, in step S80, the fault risk score may be compared with the score threshold. The score threshold may be a predefined threshold, and the specific value of the score threshold may be set according to actual needs. For example, the fault risk score may be a value between 0 and 1, and the score threshold may be, but is not limited to, 0.5.

[0113] As an example, if the fault risk score is higher than the score threshold, the fault prediction algorithm model enters the fault type prediction process.

[0114] As an example, if the fault risk score is higher than a score threshold, the fault prediction algorithm model outputs a fault type based on a support vector machine (SVM) or a decision tree.

[0115] In one embodiment, real-time operating data of equipment (e.g., wind turbines) is collected in real time by sensors, and the real-time operating data is preprocessed to obtain real-time feature data. The real-time feature data is input into the fault prediction algorithm model to obtain the health status score. The fault risk score is obtained based on the health status score; if the fault risk score exceeds a preset score threshold, the fault type determination is continued. The fault prediction algorithm model can output the most likely fault type, such as blade damage or generator overheating, etc. According to the output health status score, the fault risk score and the fault type, corresponding maintenance strategies and preventive measures can be formulated.

[0116] Through the above real-time method, accurate decision support can be provided for equipment maintenance management, thereby reducing the risk of equipment failure and improving equipment operation efficiency.

[0117] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0118] In another embodiment, see Figure 2 The present application also provides an IoT-based device fault prediction system, which may include: a data acquisition module 10, an IoT gateway 20, a preprocessing module 30, a comparison module 40, and a fault prediction algorithm model construction module 50; wherein the data acquisition module 10 is used to acquire real-time data of the device; the data acquisition module includes multiple sensors, and the real-time data includes sensor data collected by multiple sensors; the IoT gateway 20 is used to aggregate the real-time data and perform preliminary cleaning on the real-time data; the preprocessing module 30 is used to preprocess the real-time data after preliminary cleaning to obtain real-time feature data; the comparison module 40 is used to compare the real-time feature data with historical data to obtain the similarity between the real-time feature data and the historical data; the fault prediction algorithm model construction module 50 is used to construct a fault prediction algorithm model when the similarity is less than a similarity threshold; the fault prediction algorithm model is used to obtain a health status score of the device based on the real-time feature data, obtain a fault risk score based on the health status score, and output a fault type when the fault risk score is higher than the score threshold.

[0119] As an example, the IoT-based equipment failure prediction system can be used to perform the following Figure 1 and the steps of the device fault prediction method based on the Internet of Things mentioned above; for the specific limitations of the device fault prediction system based on the Internet of Things, please refer to the limitations of the device fault prediction method based on the Internet of Things mentioned above, which will not be repeated here. Each module in the above-mentioned device fault prediction system based on the Internet of Things can be implemented in whole or in part through software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0120] In another embodiment, a device is provided. The device may be a computer device, and the computer device may be a server. The internal structure diagram thereof may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store data such as deployment strategies. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it is implemented as follows Figure 1And the device failure prediction method based on the Internet of Things described in the related embodiments.

[0121] In another embodiment, a device is provided. The device may be a computer device, and the computer device may be a terminal. The internal structure diagram thereof may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it is implemented as follows Figure 1 The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0122] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0123] In another embodiment, the present application further provides a device, which may include a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following is implemented: Figure 1 And the steps of the device failure prediction method based on the Internet of Things described in the related embodiments. The device can be a computer device.

[0124] In another embodiment, the present application further provides a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the following Figure 1 And the steps of the device failure prediction method based on the Internet of Things described in the related embodiments.

[0125] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0126] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0127] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A device failure prediction method based on the Internet of Things, characterized in that: include: Acquire real-time data of the device, wherein the real-time data includes sensor data collected by multiple sensors; Aggregate the real-time data using an IoT gateway and perform preliminary cleaning on the real-time data; Preprocessing the real-time data after preliminary cleaning to obtain real-time feature data; Comparing the real-time feature data with the historical data to obtain the similarity between the real-time feature data and the historical data; If the similarity is less than the similarity threshold, a fault prediction algorithm model is constructed; Inputting the real-time feature data into the fault prediction algorithm model for prediction to obtain a health status score of the equipment; Obtaining a failure risk score based on the health status score; If the fault risk score is higher than the score threshold, the fault prediction algorithm model outputs the fault type.

2. The device failure prediction method based on the Internet of Things according to claim 1 is characterized in that: The pre-processing of the real-time data after preliminary cleaning to obtain real-time feature data includes: Clean the real-time data after preliminary cleaning to remove noise and outliers; The cleaned real-time data is subjected to feature extraction to obtain the real-time feature data; the real-time feature data includes the temperature change rate, the frequency and the fluctuation index of the vibration signal.

3. The device failure prediction method based on the Internet of Things according to claim 1 is characterized in that: Before comparing the real-time feature data with the historical data to obtain the similarity between the real-time feature data and the historical data, the method further includes: Aligning the real-time feature data in time series; The aligned real-time feature data are fused; and the fused real-time feature data are compared with the historical data.

4. The device failure prediction method based on the Internet of Things according to claim 1 is characterized in that: The similarity between the real-time feature data and the historical data is obtained using the following formula: X 实时 is the real-time feature data, X 历史 is the historical data, . represents the dot product, ||X 实时 || is the Euclidean norm of the real-time feature data, ||X 历史 || is the Euclidean norm of the historical data.

5. The device failure prediction method based on the Internet of Things according to claim 1 is characterized in that: The constructing of the fault prediction algorithm model comprises: Select the LSTM network model; Adjusting the number of neurons in the input layer of the long short-term memory network according to the dimension of the real-time feature data; Increasing the number of hidden layers or neurons in the long short-term memory network model; Regularize the long short-term memory network model; The regularized long short-term memory network model is trained, verified and optimized to obtain the fault prediction algorithm model.

6. The device failure prediction method based on the Internet of Things according to claim 1 is characterized in that: The following formula is used to obtain the failure risk score based on the health status score: Among them, R 风险 is the failure risk score, and HScore is the health status score.

7. The device failure prediction method based on the Internet of Things according to claim 1 is characterized in that: If the fault risk score is higher than the score threshold, the fault prediction algorithm model outputs the fault type based on a support vector machine or a decision tree.

8. An equipment failure prediction system based on the Internet of Things, characterized in that: include: A data acquisition module, used to acquire real-time data of the device; the data acquisition module includes a plurality of sensors, and the real-time data includes sensor data collected by the plurality of sensors; An Internet of Things gateway, used to aggregate the real-time data and perform preliminary cleaning on the real-time data; A preprocessing module is used to preprocess the real-time data after preliminary cleaning to obtain real-time feature data; A comparison module, used for comparing the real-time feature data with the historical data to obtain the similarity between the real-time feature data and the historical data; A fault prediction algorithm model construction module is used to construct a fault prediction algorithm model when the similarity is less than a similarity threshold; the fault prediction algorithm model is used to obtain a health status score of the device based on the real-time feature data, obtain a fault risk score based on the health status score, and output a fault type when the fault risk score is higher than a score threshold.

9. A device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the device failure prediction method based on the Internet of Things described in any one of claims 1 to 7 are implemented.

10. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the device failure prediction method based on the Internet of Things described in any one of claims 1 to 7 are implemented.

Images