Equipment fault prediction and diagnosis system in industrial Internet of Things

Abstract

The invention provides an equipment fault prediction and diagnosis system in an industrial internet of things, which relates to the field of electric digital data processing and comprises a data acquisition module, a state sensing module, a prediction analysis module and a fault diagnosis module. The data acquisition module is used for acquiring operation data of equipment, the state sensing module is used for monitoring an operation state of the equipment, the prediction analysis module is used for predicting a future operation trend of the equipment, and the fault diagnosis module is used for identifying a fault type of the equipment; according to the system, key features can be extracted by using the operation information of the equipment, the fault trend of the equipment is predicted, and the fault reason is diagnosed in an auxiliary manner according to the historical prediction result.

Description

Technical Field

[0001] The present invention relates to the field of electrical digital data processing, and in particular to an equipment fault prediction and diagnosis system in an industrial Internet of Things. Background Art

[0002] With the rapid development of the Industrial Internet of Things (IIoT), an increasing number of industrial devices are enabling real-time data collection and remote monitoring by deploying a variety of sensors. Consequently, equipment fault prediction and diagnosis have become crucial for ensuring the continuous and safe operation of industrial systems. Currently, traditional equipment fault monitoring systems rely on manual experience or fixed rules to make simple judgments based on collected sensor data. These systems lack the ability to dynamically model equipment operating status and predict trends, making it difficult to accurately identify potential risks or provide timely warnings.

[0003] The foregoing discussion of the background art is intended only to facilitate an understanding of the present invention. This discussion does not acknowledge or admit that any of the material referred to is part of the common general knowledge.

[0004] Many authorized equipment fault analysis systems have been developed. After extensive research and reference, it was discovered that existing equipment fault analysis systems, such as the one disclosed in Publication No. CN114689351B, generally include multiple domains. Each domain contains a domain head and multiple fault data acquisition units. The domain head connects to a data transmission unit and uploads data to a fault analysis and prediction unit. The domain head also includes a built-in data preprocessing module, which executes a data preprocessing procedure. This procedure includes extracting features from the first-level data, searching for similar data in the domain head's historical transmission database, and using this data as reference data. However, this system only performs diagnosis based on data at the time of the fault, and its accuracy still needs to be improved. Summary of the Invention

[0005] The purpose of the present invention is to address the existing deficiencies and propose an equipment fault prediction and diagnosis system in the industrial Internet of Things.

[0006] The present invention adopts the following technical solutions:

[0007] An equipment fault prediction and diagnosis system in the industrial Internet of Things, including a data acquisition module, a state perception module, a prediction analysis module and a fault diagnosis module;

[0008] The data acquisition module is used to collect the operating data of the equipment, the state perception module is used to monitor the operating status of the equipment, the prediction analysis module is used to predict the future operating trend of the equipment, and the fault diagnosis module is used to identify the fault type of the equipment;

[0009] The data acquisition module includes a sensor access unit, a data preprocessing unit, and an edge cache unit. The sensor access unit is used to connect to the sensor device that detects the physical information of the device. The data preprocessing unit is used to clean the raw data transmitted by the sensor. The edge cache unit is used to temporarily store the preprocessed data information.

[0010] The state perception module includes a feature extraction unit, a state modeling unit, and a threshold determination unit. The feature extraction unit is used to extract key features from the sensor information, the state modeling unit is used to build a device operation state model, and the threshold determination unit is used to set the state threshold and identify abnormal features.

[0011] The prediction and analysis module includes a time series processing unit, a health scoring unit, and a risk warning unit. The time series processing unit is used to perform time series processing on feature information, the health scoring unit is used to evaluate the health status of the equipment, and the risk warning unit is used to predict the risk of equipment failure in the short term.

[0012] The fault diagnosis module includes a data collection unit, a fault database and a fault reasoning unit. The data collection unit is used to collect historical status data and prediction data of the faulty equipment. The fault database is used to store standard data of all faults. The fault reasoning unit is used to diagnose the fault type of the faulty equipment.

[0013] Furthermore, the feature extraction unit includes a frequency domain analysis processor, a time domain analysis processor, and a feature selection processor, wherein the frequency domain analysis processor is used to extract frequency domain features, the time domain analysis processor is used to extract time domain features, and the feature selection processor is used to screen out key features related to the device status;

[0014] The feature selection processor calculates the correlation index between the feature and the state according to the following formula:

[0015]

[0016] Where X represents the candidate feature, Y represents the candidate state, p(x) represents the probability that the candidate feature is x, p(y) represents the probability that the candidate state is y, and p(x, y) represents the probability that the candidate feature is x and the candidate state is y;

[0017] When the correlation index between a candidate feature and at least one candidate state is greater than a correlation threshold, the candidate feature is screened out as a key feature.

[0018] Furthermore, the time series processing unit includes a time window manager, a time series modeling processor and a trend analysis processor. The time window manager is used to slide and manage the time length and step size of the input data window. The time series modeling processor is used to construct a time series analyzer. The trend analysis processor is used to judge future trends.

[0019] Furthermore, the time series modeling processor receives the integrated abnormal information and extracts time series groups containing feature items of abnormal markers, each time series group is composed of m feature values, where m is the number of state vectors contained in the abnormal information;

[0020] The timing modeling processor calculates the gradient value S of each timing group according to the following formula:

[0021]

[0022] Among them, t i represents the time point of the i-th window, represents the mean time point of m windows, v i represents the i-th eigenvalue in the time series group, Represents the mean of the time series group.

[0023] Furthermore, the trend analysis processor calculates the alienation rate according to the following formula:

[0024]

[0025] Where n is the number of key features, S i Indicates the gradient value of the time series group corresponding to the i-th key feature, k i is the alienation weight of the i-th key feature;

[0026] If the i-th key feature has no time series group, then S i Set to 0.

[0027] The beneficial effects achieved by the present invention are:

[0028] This system performs real-time predictive analysis on the equipment in the system when the system is operating normally, and extracts historical predictive data after an actual failure occurs. Compared with directly using the data at the time of the failure for diagnosis, it has higher accuracy. At the same time, the use of predictive data can also maintain the equipment in a timely manner, reducing the probability of failure.

[0029] To further understand the features and technical contents of the present invention, please refer to the following detailed description and drawings of the present invention. However, the drawings provided are only for reference and illustration and are not intended to limit the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a schematic diagram of the overall structural framework of the present invention;

[0031] Figure 2 This is a schematic diagram of the data acquisition module of the present invention;

[0032] Figure 3 This is a schematic diagram of the state perception module of the present invention;

[0033] Figure 4 This is a schematic diagram of the forecast analysis module of the present invention;

[0034] Figure 5 This is a schematic diagram of the fault diagnosis module of the present invention;

[0035] Figure 6 The following is a comparison chart of the effects after using this system for operation and maintenance. DETAILED DESCRIPTION

[0036] The following is an explanation of the embodiments of the present invention through specific embodiments. Those skilled in the art can understand the advantages and effects of the present invention from the contents disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and the details in this specification can also be modified and changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. In addition, the drawings of the present invention are only simple schematic illustrations and are not depicted in actual size. It is stated in advance. The following embodiments will further explain the relevant technical content of the present invention in detail, but the disclosed content is not intended to limit the scope of protection of the present invention.

[0037] Example 1.

[0038] This embodiment provides a device fault prediction and diagnosis system in the industrial Internet of Things, combined with Figure 1 , including data acquisition module, state perception module, prediction analysis module and fault diagnosis module;

[0039] The data acquisition module is used to collect the operating data of the equipment, the state perception module is used to monitor the operating status of the equipment, the prediction analysis module is used to predict the future operating trend of the equipment, and the fault diagnosis module is used to identify the fault type of the equipment;

[0040] The data acquisition module includes a sensor access unit, a data preprocessing unit, and an edge cache unit. The sensor access unit is used to connect to the sensor device that detects the physical information of the device. The data preprocessing unit is used to clean the raw data transmitted by the sensor. The edge cache unit is used to temporarily store the preprocessed data information.

[0041] The state perception module includes a feature extraction unit, a state modeling unit, and a threshold determination unit. The feature extraction unit is used to extract key features from the sensor information, the state modeling unit is used to build a device operation state model, and the threshold determination unit is used to set the state threshold and identify abnormal features.

[0042] The prediction and analysis module includes a time series processing unit, a health scoring unit, and a risk warning unit. The time series processing unit is used to perform time series processing on feature information, the health scoring unit is used to evaluate the health status of the equipment, and the risk warning unit is used to predict the risk of equipment failure in the short term.

[0043] The fault diagnosis module includes a data collection unit, a fault database and a fault reasoning unit. The data collection unit is used to collect historical status data and prediction data of the faulty equipment. The fault database is used to store standard data of all faults. The fault reasoning unit is used to diagnose the fault type of the faulty equipment.

[0044] The feature extraction unit includes a frequency domain analysis processor, a time domain analysis processor, and a feature selection processor, wherein the frequency domain analysis processor is used to extract frequency domain features, the time domain analysis processor is used to extract time domain features, and the feature selection processor is used to screen out key features related to the device status;

[0045] The feature selection processor calculates the correlation index between the feature and the state according to the following formula:

[0046]

[0047] Where X represents the candidate feature, Y represents the candidate state, p(x) represents the probability that the candidate feature is x, p(y) represents the probability that the candidate state is y, and p(x, y) represents the probability that the candidate feature is x and the candidate state is y;

[0048] When the correlation index between a candidate feature and at least one candidate state is greater than a correlation threshold, the candidate feature is screened out as a key feature.

[0049] The time series processing unit includes a time window manager, a time series modeling processor and a trend analysis processor. The time window manager is used to slide and manage the time length and step size of the input data window. The time series modeling processor is used to build a time series analyzer. The trend analysis processor is used to judge future trends.

[0050] The time series modeling processor receives the integrated abnormal information and extracts time series groups containing feature items of abnormal markers, each time series group is composed of m feature values, where m is the number of state vectors contained in the abnormal information;

[0051] The timing modeling processor calculates the gradient value S of each timing group according to the following formula:

[0052]

[0053] Among them, t i represents the time point of the i-th window, represents the mean time point of m windows, v i represents the i-th eigenvalue in the time series group, Represents the mean of the time series group.

[0054] The trend analysis processor calculates the alienation rate according to the following formula:

[0055]

[0056] Where n is the number of key features, S i Indicates the gradient value of the time series group corresponding to the i-th key feature, k i is the alienation weight of the i-th key feature;

[0057] If the i-th key feature has no time series group, then S i Set to 0.

[0058] Example 2.

[0059] This embodiment includes all the contents of the first embodiment and provides an equipment fault prediction and diagnosis system in the industrial Internet of Things, including a data acquisition module, a state perception module, a prediction and analysis module, and a fault diagnosis module;

[0060] The data acquisition module is used to collect the operating data of the equipment, the state perception module is used to monitor the operating status of the equipment, the prediction analysis module is used to predict the future operating trend of the equipment, and the fault diagnosis module is used to identify the fault type of the equipment;

[0061] Combine Figure 2 The data acquisition module includes a sensor access unit, a data preprocessing unit and an edge cache unit. The sensor access unit is used to connect to the sensor device for detecting physical information of the device. The data preprocessing unit is used to clean the raw data transmitted by the sensor. The edge cache unit is used to temporarily store the preprocessed data information.

[0062] Combine Figure 3 The state perception module includes a feature extraction unit, a state modeling unit and a threshold judgment unit. The feature extraction unit is used to extract key features from the sensor information, the state modeling unit is used to build a device operation state model, and the threshold judgment unit is used to set the state threshold and identify abnormal features;

[0063] Combine Figure 4 The prediction and analysis module includes a time series processing unit, a health scoring unit, and a risk warning unit. The time series processing unit is used to perform time series processing on feature information, the health scoring unit is used to evaluate the health status of the equipment, and the risk warning unit is used to predict the risk of equipment failure in the short term.

[0064] Combine Figure 5 The fault diagnosis module includes a data collection unit, a fault database and a fault reasoning unit. The data collection unit is used to collect historical status data and prediction data of the faulty device. The fault database is used to store standard data of all faults. The fault reasoning unit is used to diagnose the fault type of the faulty device.

[0065] The sensor access unit includes a sensor interface processor, a protocol conversion processor, and a real-time communication controller. The sensor interface processor is used to identify and access a variety of industrial sensors. The protocol conversion processor is used to uniformly convert the signals output by different sensors into a data format supported by the system. The real-time communication controller is used to control the stability and real-time performance of data transmission between the sensor and the system.

[0066] The data preprocessing unit includes a data cleaning processor, a data standardization processor, and a time alignment processor. The data cleaning processor is used to filter outliers and duplicate data. The data standardization processor is used to unify data of different dimensions into a standard range. The time alignment processor is used to perform time stamp synchronization processing on the collected data.

[0067] The edge cache unit includes a cache memory, a data compression processor and a cache refresh processor, wherein the cache memory is used to store temporary data, the data compression processor is used to compress the pre-processed data, and the cache refresh processor is used to regularly update the cache content;

[0068] The feature extraction unit includes a frequency domain analysis processor, a time domain analysis processor, and a feature selection processor, wherein the frequency domain analysis processor is used to extract frequency domain features, the time domain analysis processor is used to extract time domain features, and the feature selection processor is used to screen out key features related to the device status;

[0069] The feature selection processor calculates the correlation index between the feature and the state according to the following formula:

[0070]

[0071] Where X represents the candidate feature, Y represents the candidate state, p(x) represents the probability that the candidate feature is x, p(y) represents the probability that the candidate state is y, and p(x, y) represents the probability that the candidate feature is x and the candidate state is y;

[0072] When the correlation index between a candidate feature and at least one candidate state is greater than a correlation threshold, the candidate feature is selected as a key feature;

[0073] The state modeling unit includes a state space builder, a model training processor, and a dynamic update processor. The state space builder builds the operating state vector space of the device based on the feature dimension. The model training processor is used to learn and train the normal state model of the device. The dynamic update processor performs incremental update training on the model according to new data.

[0074] The threshold determination unit includes a threshold setting processor, an anomaly detection processor, and a feedback output processor. The threshold setting processor is used to set the threshold limit corresponding to each feature. The anomaly detection processor is used to identify and mark feature points that exceed the threshold. The feedback output processor is used to integrate the anomaly information and feed it back to the prediction analysis module.

[0075] The time series processing unit includes a time window manager, a time series modeling processor, and a trend analysis processor. The time window manager is used to slide and manage the time length and step size of the input data window. The time series modeling processor is used to build a time series analyzer. The trend analysis processor is used to judge future trends.

[0076] The feature extraction unit extracts key features within the time length of a window, and the state modeling unit constructs a state vector. The threshold judgment unit marks abnormal features in the state vector. This is called a loop operation, which is performed once every step time.

[0077] The time series modeling processor receives the integrated abnormal information and extracts time series groups containing feature items of abnormal markers, each time series group is composed of m feature values, where m is the number of state vectors contained in the abnormal information;

[0078] The timing modeling processor calculates the gradient value S of each timing group according to the following formula:

[0079]

[0080] Among them, t i represents the time point of the i-th window, represents the mean time point of m windows, v i represents the i-th eigenvalue in the time series group, represents the mean of the time series group;

[0081] The trend analysis processor calculates the alienation rate according to the following formula:

[0082]

[0083] Where n is the number of key features, S i Indicates the gradient value of the time series group corresponding to the i-th key feature, k i is the alienation weight of the i-th key feature;

[0084] If the i-th key feature has no time series group, then S i Set to 0;

[0085] The health scoring unit includes a scoring standard information library, a state scoring processor, and a score normalization processor. The scoring standard information library is used to store the health score calculation standard of each device's operating status. The state scoring processor calculates the device health score based on the current features and model output. The score normalization processor normalizes the health score.

[0086] The risk warning unit includes a risk identification processor, a grading processor, and a warning release processor. The risk identification processor determines whether the device has a risk based on the score and trend. The grading processor is used to divide the risk degree into different levels. The warning release processor is used to push the risk results to the corresponding staff.

[0087] The fault diagnosis unit includes a fault receiving processor, a history scheduling processor, and a data integration processor. The fault receiving processor is used to receive information about the device that is currently faulty. The history scheduling processor is used to call the historical status information and prediction information of the device. The data integration processor is used to integrate the historical information with the current information.

[0088] The fault database includes a fault sample memory, a label standard processor, and a structure index processor. The fault sample memory is used to store data samples of various types of equipment faults. The label standard processor is used to label data samples. The structure index processor is used to quickly query fault characteristics and related cases.

[0089] The fault inference unit includes a rule matching processor, a pattern recognition processor, and a diagnosis output processor. The rule matching processor is used to perform rule matching on the integrated data to obtain label information. The pattern recognition processor identifies the approaching fault type based on the label information. The diagnosis output processor accurately analyzes the integrated data and sample data to determine the fault type and output a diagnosis result.

[0090] The "i" in the above question is an ordinal number and has no actual meaning.

[0091] Part of the code of this system is as follows:

[0092]

[0093] self.preprocess_unit=DataPreprocessingUnit()

[0094] self.cache_unit=EdgeCacheUnit()

[0095] def collect_data(self):

[0096] raw_data=self.sensor_unit.connect_and_read()

[0097] cleaned_data=self.preprocess_unit.clean(raw_data)

[0098] self.cache_unit.cache(cleaned_data)

[0099] return cleaned_data

[0100] class SensorAccessUnit:

[0101] def connect_and_read(self):

[0102] #Simulate collecting physical information from multiple sensors

[0103] return{"temp":65,"vibration":0.02,"pressure":1.2}classDataPreprocessingUnit:

[0104] def clean(self,data):

[0105] #Data cleaning logic, such as denoising, standardization, etc.

[0106] return data class EdgeCacheUnit:

[0107] def cache(self,data):

[0108] # Temporary storage data logic

[0109] pass

[0110] #State perception module

[0111]

[0112]

[0113]

[0114]

[0115] The downtime for maintenance before and after the adoption of this system is statistically analyzed and the results are as follows: Figure 6 The effect comparison chart shown shows that the use of this system can diagnose faults more quickly and reduce maintenance time.

[0116] The contents disclosed above are only preferred feasible embodiments of the present invention and do not limit the scope of protection of the present invention. Therefore, all equivalent technical changes made using the contents of the present invention description and drawings are included in the scope of protection of the present invention. In addition, the elements therein can be updated as technology develops.

Claims

1. A device fault prediction and diagnosis system in the industrial Internet of Things, characterized by: It includes data acquisition module, state perception module, prediction analysis module and fault diagnosis module; The data acquisition module is used to collect the operating data of the equipment, the state perception module is used to monitor the operating status of the equipment, the prediction analysis module is used to predict the future operating trend of the equipment, and the fault diagnosis module is used to identify the fault type of the equipment; The data acquisition module includes a sensor access unit, a data preprocessing unit, and an edge cache unit. The sensor access unit is used to connect to the sensor device that detects the physical information of the device. The data preprocessing unit is used to clean the raw data transmitted by the sensor. The edge cache unit is used to temporarily store the preprocessed data information. The state perception module includes a feature extraction unit, a state modeling unit, and a threshold determination unit. The feature extraction unit is used to extract key features from the sensor information, the state modeling unit is used to build a device operation state model, and the threshold determination unit is used to set the state threshold and identify abnormal features. The prediction and analysis module includes a time series processing unit, a health scoring unit, and a risk warning unit. The time series processing unit is used to perform time series processing on feature information, the health scoring unit is used to evaluate the health status of the equipment, and the risk warning unit is used to predict the risk of equipment failure in the short term. The fault diagnosis module includes a data collection unit, a fault database and a fault reasoning unit. The data collection unit is used to collect historical status data and prediction data of the faulty equipment. The fault database is used to store standard data of all faults. The fault reasoning unit is used to diagnose the fault type of the faulty equipment.

2. The equipment fault prediction and diagnosis system in the industrial Internet of Things according to claim 1, characterized in that: The feature extraction unit includes a frequency domain analysis processor, a time domain analysis processor, and a feature selection processor, wherein the frequency domain analysis processor is used to extract frequency domain features, the time domain analysis processor is used to extract time domain features, and the feature selection processor is used to screen out key features related to the device status; The feature selection processor calculates the correlation index between the feature and the state according to the following formula: Where X represents the candidate feature, Y represents the candidate state, p(x) represents the probability that the candidate feature is x, p(y) represents the probability that the candidate state is y, and p(x, y) represents the probability that the candidate feature is x and the candidate state is y; When the correlation index between a candidate feature and at least one candidate state is greater than a correlation threshold, the candidate feature is screened out as a key feature.

3. The equipment fault prediction and diagnosis system in the industrial Internet of Things according to claim 1, characterized in that: The time series processing unit includes a time window manager, a time series modeling processor and a trend analysis processor. The time window manager is used to slide and manage the time length and step size of the input data window. The time series modeling processor is used to build a time series analyzer. The trend analysis processor is used to judge future trends.

4. The equipment fault prediction and diagnosis system in the industrial Internet of Things according to claim 3, characterized in that: The time series modeling processor receives the integrated abnormal information and extracts time series groups containing feature items of abnormal markers, each time series group is composed of m feature values, where m is the number of state vectors contained in the abnormal information; The timing modeling processor calculates the gradient value S of each timing group according to the following formula: Among them, t i represents the time point of the i-th window, represents the mean time point of m windows, v i represents the i-th eigenvalue in the time series group, Represents the mean of the time series group.

5. The equipment fault prediction and diagnosis system in the industrial Internet of Things according to claim 4, characterized in that: The trend analysis processor calculates the alienation rate according to the following formula: Where n is the number of key features, S i Indicates the gradient value of the time series group corresponding to the i-th key feature, k i is the alienation weight of the i-th key feature; If the i-th key feature has no time series group, then S i Set to 0.

Images