Community LPG turnover bin equipment fault diagnosis method and system

Abstract

The invention discloses a community LPG turnover warehouse equipment fault diagnosis method and system. The method comprises the steps of collecting operation data of corresponding equipment based on a sensor deployed on community LPG turnover warehouse equipment; establishing a first association relationship between the community LPG turnover warehouse equipment; establishing a second association relationship between the community LPG turnover bin equipment and the operation scene; and determining an equipment fault diagnosis result according to the operation data, the first association relationship and the second association relationship. In a complex equipment operation environment of a community LPG turnover bin, accurate identification, positioning and early warning of equipment faults are realized, meanwhile, the problems of false alarm and missing alarm caused by single data monitoring are reduced, the comprehensiveness and timeliness of fault diagnosis are improved, and the safety and stability of community gas supply are guaranteed.

Description

Technical Field

[0001] This invention relates to the field of equipment fault diagnosis technology, and in particular to a method and system for diagnosing equipment faults in community LPG turnover warehouses. Background Technology

[0002] Community LPG distribution centers are delivery points set up within communities, capable of storing bottled gas. Their establishment reduces logistics costs while providing convenient services to community users. Current technologies, which use fixed thresholds from single sensor data to diagnose equipment anomalies, are suitable for simple, independent devices but cannot handle the interrelationships between devices in complex systems or dynamic changes in the environment. Relying on maintenance personnel to periodically inspect equipment appearance, listen for abnormal noises, or check dashboard data is inefficient and highly susceptible to subjective factors, making it difficult to detect early, hidden faults. The monitoring systems for each device operate independently, with data not integrated, making it impossible to build a global fault analysis model. Existing correlation analysis and scene awareness technologies suffer from shortcomings, including a lack of inter-device correlation modeling, insufficient scene adaptability, and a lack of in-depth analysis of historical data and operating patterns, making it difficult to identify early characteristics of equipment anomalies using machine learning algorithms. The compact spatial layout and complex physical connections of tanks, valves, and pipelines within distribution centers mean that a single device failure can quickly spread to surrounding facilities. Without integrated data between devices, it is impossible to analyze fault propagation paths and adjust fault diagnosis strategies according to the operating scenario. Summary of the Invention

[0003] This invention aims to at least partially solve one of the technical problems in the aforementioned technologies. Therefore, the purpose of this invention is to propose a method and system for fault diagnosis of community LPG storage equipment. In the complex operating environment of community LPG storage facilities, this method enables accurate identification, location, and early warning of equipment faults, while reducing false alarms and missed alarms caused by single data monitoring. This improves the comprehensiveness and timeliness of fault diagnosis, ensuring the safety and stability of community gas supply.

[0004] To achieve the above objectives, this invention proposes a method for diagnosing faults in community LPG turnover warehouse equipment, comprising:

[0005] Based on sensors deployed on community LPG turnover warehouse equipment, the corresponding equipment's operating data is collected;

[0006] Establish the primary connection between the community LPG turnover warehouse equipment;

[0007] Establish a second correlation between community LPG turnover warehouse equipment and operational scenarios;

[0008] The equipment fault diagnosis results are determined based on the operating data, the first correlation, and the second correlation.

[0009] According to some embodiments of the present invention, operational data of the corresponding equipment is collected based on sensors deployed on community LPG turnover warehouse equipment, including:

[0010] Pressure and temperature sensors are deployed in the storage tank to collect pressure and temperature data inside the tank in real time.

[0011] Gas leak detection sensors are deployed at valve connections to monitor for LPG leaks.

[0012] Deploy flow sensors on the pipeline to acquire LPG flow data;

[0013] Vibration sensors are deployed on the power components to monitor the vibration status of the power components during operation; the power components include motors and pumps.

[0014] According to some embodiments of the present invention, establishing a first association between community LPG turnover warehouse facilities includes:

[0015] Determine the first sub-association relationship based on physical connections between the community LPG turnover warehouse equipment;

[0016] Determine the second sub-association relationship between community LPG turnover warehouse equipment based on operational logic;

[0017] Determine the third sub-relationship between community LPG turnover warehouse equipment based on data association;

[0018] Establish the first association between the community LPG turnover warehouse equipment based on the first sub-association, the second sub-association, and the third sub-association.

[0019] According to some embodiments of the present invention, determining a first sub-association based on physical connections between community LPG turnover warehouse facilities includes:

[0020] Determine the pipeline information between the equipment in the community LPG turnover warehouse; draw a pipeline connection topology diagram based on the pipeline information, determine the position and connection sequence of each piece of equipment in the pipeline system, and establish the first connection relationship based on the pipeline connection;

[0021] Determine the electrical information between the equipment in the community LPG turnover warehouse; draw an electrical wiring diagram based on the electrical information, determine the connection relationship of each device in the electrical circuit, and establish a second connection relationship based on the electrical connection;

[0022] The first sub-association relationship based on physical connection between community LPG turnover warehouse equipment is determined according to the first connection relationship and the second connection relationship.

[0023] According to some embodiments of the present invention, determining a second sub-association relationship based on operational logic between community LPG turnover warehouse equipment includes:

[0024] Analyze the process flow of the LPG turnover warehouse, determine the sequence and collaborative relationship of equipment in the process flow, and obtain the first related information based on the process flow logic;

[0025] The control programs and logic algorithms of each device in the community LPG turnover warehouse were analyzed separately, and the control relationships between the devices were established.

[0026] Based on the first association information and control association, a second sub-association relationship based on operational logic is determined between the community LPG turnover warehouse equipment.

[0027] According to some embodiments of the present invention, determining a third sub-association relationship based on data association between community LPG turnover warehouse equipment includes:

[0028] The correlation of data collected by sensors deployed on various devices is calculated based on the Pearson correlation coefficient to quantitatively analyze the correlation between device data and obtain the second correlation information of the devices.

[0029] Analyze historical equipment failure records, establish a failure mode correlation matrix, analyze the correlation patterns of equipment failures, and determine the third correlation information between equipment based on failure modes;

[0030] Using time series analysis, we can obtain the fourth correlation information of equipment operation data in the time dimension;

[0031] Based on the second, third, and fourth association information, a third sub-association relationship based on data association is determined between the community LPG turnover warehouse equipment.

[0032] According to some embodiments of the present invention, establishing a second association between community LPG turnover warehouse equipment and operating scenarios includes:

[0033] Acquire scenario data for community LPG distribution warehouses; the scenario data includes meteorological data, power facility data, and transportation facility data;

[0034] The operational data and scenario data of the community LPG turnover warehouse equipment were converted into a unified format and timestamps were added to form a related scenario dataset.

[0035] The Apriori algorithm is used to mine frequent itemsets and association rules among data to determine the association information between meteorology and equipment, power facilities and equipment, and transportation facilities and equipment, and to establish a second association relationship between community LPG turnover warehouse equipment and operating scenarios.

[0036] According to some embodiments of the present invention, determining equipment fault diagnosis results based on operating data, a first correlation, and a second correlation includes:

[0037] The sample running data is cleaned and normalized to obtain the processed running data;

[0038] Feature extraction is performed on the processed data to obtain time-domain and frequency-domain features;

[0039] The time-domain features and frequency-domain features are combined to form a feature vector;

[0040] Based on the equipment's historical fault records, the feature vectors, the corresponding first and second sample association relationships are labeled to obtain labeled data.

[0041] The SVM model is trained using the feature vectors of the sample running data, the corresponding first and second sample association relationships, and the labeled data. During the training process, the radial basis function kernel function is set, and the model parameters are optimized through cross-validation to obtain the target SVM model.

[0042] Input the running data, the first correlation, and the second correlation into the target SVM model to determine the equipment fault diagnosis results.

[0043] According to some embodiments of the present invention, analyzing historical fault records of equipment and establishing a fault mode correlation matrix includes:

[0044] Analyze the equipment's historical fault records, which are derived from the equipment's operation logs, maintenance work orders, and fault alarm systems.

[0045] Clean up the equipment's historical fault records, removing duplicate records, erroneous data, and incomplete data, and standardize the description of fault types;

[0046] Identify the key equipment within the community LPG turnover warehouse. For each type of key equipment, analyze its potential failure modes by consulting equipment manuals, expert experience, and historical failure data. Summarize the failure modes of all equipment and establish a unified classification system as the failure mode classification system.

[0047] Iterate through the historical fault data of the equipment and count the number of times each of the two fault modes occurs simultaneously or occurs successively within a preset time period.

[0048] A square matrix is ​​constructed using all failure modes in the failure mode classification system as rows and columns; the matrix elements are then filled with the results of the failure mode correlation analysis to establish a failure mode correlation matrix.

[0049] According to some embodiments of the present invention, a diagnostic system applying the fault diagnosis method for community LPG turnover warehouse equipment as described above includes:

[0050] The data acquisition module is used to collect operational data of the corresponding equipment based on sensors deployed on the community LPG turnover warehouse equipment;

[0051] The first module is used to establish the first association between community LPG turnover warehouse equipment;

[0052] The second module is used to establish a second association between the community LPG turnover warehouse equipment and the operating scenario;

[0053] The determination module is used to determine the equipment fault diagnosis results based on the operating data, the first correlation, and the second correlation.

[0054] This invention proposes a fault diagnosis method and system for community LPG storage facilities. It establishes data connections between equipment and analyzes fault propagation paths. Beyond relying on fixed thresholds for fault diagnosis, it further enables the perception of dynamic changes in equipment operating modes and optimizes the diagnostic logic by incorporating community gas usage environmental factors. Through sensor data fusion, equipment association modeling, and dynamic scenario adaptation, the reliability and safety of fault diagnosis in complex scenarios are improved. In the complex operating environment of community LPG storage facilities, it achieves accurate identification, location, and early warning of equipment faults, while reducing false alarms and missed alarms caused by single data monitoring, improving the comprehensiveness and timeliness of fault diagnosis, and ensuring the safety and stability of community gas supply.

[0055] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0056] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0057] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0058] Figure 1 This is a flowchart of a method for diagnosing faults in community LPG turnover warehouse equipment according to an embodiment of the present invention;

[0059] Figure 2 This is a flowchart illustrating the establishment of a first association relationship according to an embodiment of the present invention;

[0060] Figure 3 This is a block diagram of a community LPG turnover warehouse equipment fault diagnosis system according to an embodiment of the present invention. Detailed Implementation

[0061] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0062] like Figure 1 As shown, this embodiment of the invention proposes a method for diagnosing faults in community LPG turnover warehouse equipment, including steps S1-S4:

[0063] S1. Collect operational data of the corresponding equipment based on sensors deployed on the community LPG turnover warehouse equipment;

[0064] S2. Establish the primary relationship between the community LPG turnover warehouse equipment;

[0065] S3. Establish a second correlation between community LPG turnover warehouse equipment and operating scenarios;

[0066] S4. Determine the equipment fault diagnosis results based on the operating data, the first correlation, and the second correlation.

[0067] The working principle of the above technical solution is as follows: Sensors deployed on community LPG storage facilities collect operational data from the corresponding equipment, facilitating the comprehensive acquisition of operational data from various devices. A primary correlation is established between the community LPG storage facilities, enabling data sharing and facilitating the construction of a global fault analysis model. A secondary correlation is established between the community LPG storage facilities and the operational scenario, dynamically adjusting diagnostic rules based on scenario characteristics. For example, during periods of high summer temperatures, tank pressure naturally increases, and traditional fixed thresholds may falsely report "overpressure faults." The equipment fault diagnosis result is determined based on the operational data, the primary correlation, and the secondary correlation.

[0068] The beneficial effects of the above technical solution are: it establishes data connections between devices, enabling analysis of fault propagation paths; it goes beyond fixed threshold-based fault diagnosis, further realizing the perception of dynamic changes in equipment operating modes; and it optimizes diagnostic logic by incorporating community gas usage environmental factors. Through sensor data fusion, device association modeling, and dynamic scenario adaptation, it improves the reliability and safety of fault diagnosis in complex scenarios. In the complex equipment operating environment of community LPG distribution warehouses, it achieves accurate identification, location, and early warning of equipment faults, while reducing false alarms and missed alarms caused by single data monitoring, improving the comprehensiveness and timeliness of fault diagnosis, and ensuring the safety and stability of community gas supply.

[0069] According to some embodiments of the present invention, operational data of the corresponding equipment is collected based on sensors deployed on community LPG turnover warehouse equipment, including:

[0070] Pressure and temperature sensors are deployed in the storage tank to collect pressure and temperature data inside the tank in real time.

[0071] Gas leak detection sensors are deployed at valve connections to monitor for LPG leaks.

[0072] Deploy flow sensors on the pipeline to acquire LPG flow data;

[0073] Vibration sensors are deployed on the power components to monitor the vibration status of the power components during operation; the power components include motors and pumps.

[0074] like Figure 2 As shown, according to some embodiments of the present invention, establishing a first association between community LPG turnover warehouse equipment includes steps S21-S24:

[0075] S21. Determine the first sub-association relationship based on physical connection between community LPG turnover warehouse equipment;

[0076] S22. Determine the second sub-association relationship between community LPG turnover warehouse equipment based on operational logic;

[0077] S23. Determine the third sub-relationship between community LPG turnover warehouse equipment based on data association;

[0078] S24. Establish the first association relationship between the community LPG turnover warehouse equipment based on the first sub-association relationship, the second sub-association relationship, and the third sub-association relationship.

[0079] The working principle of the above technical solution is as follows: First, the physical connections between the LPG storage facilities in the community are determined. An undirected graph G = (V, E) is constructed, where V is the set of equipment nodes and E is the set of physical connection edges. Physical connections determine the physical path of fault propagation; for example, pipeline leaks directly affect the pressure of upstream and downstream equipment. Quantifying physical connection parameters provides a foundation for fault propagation modeling. Second, the state transition diagram is used to describe the control relationships between the equipment. For example, when the tank level is <20%, a "low level alarm" state is triggered. Upon receiving the alarm, the control system automatically switches to the standby tank and records it as the "switching to standby tank" state. Production rules are used to represent the equipment linkage logic. Time series analysis is used to determine the equipment start-up and shutdown sequence; for example, before the vaporizer starts, the upstream valve must be opened first, and the heating system starts after a 30-second delay. The Granger causality test is used to verify the temporal causal relationship between the equipment. Logical relationships reflect the collaborative working methods between devices. When the state of one device changes, the response of other devices can be predicted through logical relationships. For example, valve closure will inevitably lead to a decrease in downstream flow; if the flow does not actually decrease, it indicates a fault. A third sub-relationship based on data association is determined between the community LPG storage facility devices. Data association reflects implicit relationships between devices; even without direct physical connection, data anomalies may indicate a relationship. For example, a simultaneous abnormal pressure rise in two independent storage tanks may be caused by an overall overpressure in the gas supply system. The first relationship between the community LPG storage facility devices is established based on the first, second, and third sub-relationships, including: determining the physical connection diagram, physical connection diagram, and data association diagram based on the first, second, and third sub-relationships.

[0080] The physical connection graph (undirected graph), the operational logic graph (directed graph), and the data association graph (weighted graph) are merged into a heterogeneous information network G. H :

[0081] G H =(V,E) T ,W)

[0082] Where V is the set of device nodes; E T Let W be a set of edges of various types, including physical connection edges, logical association edges, and data association edges; W is the edge weight matrix.

[0083] The beneficial effects of the above technical solution are as follows: the physical connection graph, physical connection graph and data connection graph are determined based on the first sub-association relationship, the second sub-association relationship and the third sub-association relationship; the physical connection graph (undirected graph), the operation logic graph (directed graph) and the data connection graph (weighted graph) are integrated into a heterogeneous information network, which facilitates the accurate determination of the first association relationship between the community LPG turnover warehouse equipment.

[0084] According to some embodiments of the present invention, determining a first sub-association based on physical connections between community LPG turnover warehouse facilities includes:

[0085] Determine the pipeline information between the equipment in the community LPG turnover warehouse; draw a pipeline connection topology diagram based on the pipeline information, determine the position and connection sequence of each piece of equipment in the pipeline system, and establish the first connection relationship based on the pipeline connection;

[0086] Determine the electrical information between the equipment in the community LPG turnover warehouse; draw an electrical wiring diagram based on the electrical information, determine the connection relationship of each device in the electrical circuit, and establish a second connection relationship based on the electrical connection;

[0087] The first sub-association relationship based on physical connection between community LPG turnover warehouse equipment is determined according to the first connection relationship and the second connection relationship.

[0088] The working principle and beneficial effects of the above technical solution: LPG is transported within the LPG transfer warehouse through pipelines between equipment such as storage tanks, pumps, valves, and flow meters. For example, storage tanks are connected to transfer pumps via pipelines, and LPG in the storage tanks is transported to downstream equipment by the pumps. If a pipeline becomes blocked or leaks, it will directly affect the operating status of connected storage tanks and pumps; therefore, there is a close relationship between storage tanks, pipelines, and pumps. By drawing a pipeline connection topology diagram, the position and connection sequence of each device in the pipeline system are clearly shown, thereby determining the device associations based on pipeline connections. The control systems and power systems of the equipment are electrically connected through wires and cables. For example, the power supply line of the motor is connected to the distribution cabinet, which is in turn connected to the entire power control system. Once an electrical circuit malfunctions, such as a short circuit or open circuit, it will cause the relevant equipment to be unable to supply power normally or the control signal transmission to be abnormal, thus affecting the operation of the equipment. By sorting out the electrical wiring diagram, the connection relationship of each device in the electrical circuit can be clarified, and the device associations based on electrical connections can be determined. This facilitates the accurate determination of the first sub-association relationship based on physical connections between the equipment in the community LPG transfer warehouse.

[0089] According to some embodiments of the present invention, determining a second sub-association relationship based on operational logic between community LPG turnover warehouse equipment includes:

[0090] Analyze the process flow of the LPG turnover warehouse, determine the sequence and collaborative relationship of equipment in the process flow, and obtain the first related information based on the process flow logic;

[0091] The control programs and logic algorithms of each device in the community LPG turnover warehouse were analyzed separately, and the control relationships between the devices were established.

[0092] Based on the first association information and control association, a second sub-association relationship based on operational logic is determined between the community LPG turnover warehouse equipment.

[0093] The working principle and beneficial effects of the above technical solution: The storage, transportation, and distribution processes of LPG involve the coordinated operation of multiple devices. In the LPG filling process, the storage tank valve is first opened, and LPG flows into the filling pipeline under its own pressure or by the action of a pump. After being measured by a flow meter, it enters the container to be filled. During this process, the opening status of the storage tank valve, the operating status of the pump, and the measurement accuracy of the flow meter are all interconnected. A malfunction in any link will affect the entire filling process. By analyzing the process flow of the LPG turnover warehouse, the sequence and coordinated role of the equipment in the process can be clarified, and the equipment relationships based on the process flow logic can be determined. The operation of the equipment is controlled by the instructions of the control system, and there is interaction and response of control signals between the various devices. When the storage tank pressure exceeds the set upper limit, the control system automatically sends a command to close the feed valve and activate the pressure relief device. Pressure sensors, control systems, feed valves, and pressure relief devices are associated based on control logic. By analyzing the control programs and logic algorithms of the equipment, the control relationships between the devices can be identified. This facilitates the accurate determination of the second sub-relationships between the community LPG turnover warehouse equipment based on operational logic.

[0094] According to some embodiments of the present invention, determining a third sub-association relationship based on data association between community LPG turnover warehouse equipment includes:

[0095] The correlation of data collected by sensors deployed on various devices is calculated based on the Pearson correlation coefficient to quantitatively analyze the correlation between device data and obtain the second correlation information of the devices.

[0096] Analyze historical equipment failure records, establish a failure mode correlation matrix, analyze the correlation patterns of equipment failures, and determine the third correlation information between equipment based on failure modes;

[0097] Using time series analysis, we can obtain the fourth correlation information of equipment operation data in the time dimension;

[0098] Based on the second, third, and fourth association information, a third sub-association relationship based on data association is determined between the community LPG turnover warehouse equipment.

[0099] The working principle and beneficial effects of the above technical solution are as follows: Correlation calculations are performed on the data collected by the sensors of each device. The correlation coefficient between the pressure data of the storage tank and the flow data of the delivery pipeline is calculated. A high correlation coefficient indicates a strong correlation between the two, meaning that changes in tank pressure significantly affect pipeline flow, and vice versa. The correlation between device data is quantitatively analyzed using the Pearson correlation coefficient to determine device associations. Historical fault records of the equipment are analyzed to statistically determine the frequency of simultaneous or sequential faults of different devices. If the probability of a fault in the delivery pump is high after a fault in the pressure sensor of the storage tank, it indicates a correlation between the fault modes of the two devices, possibly due to abnormal pressure causing changes in pump load, which in turn leads to pump failure. By establishing a fault mode correlation matrix, the correlation patterns of device faults are analyzed to determine the correlation between devices based on fault modes. Time series analysis is used to study the correlation of equipment operating data over time. Granger causality tests are used to determine whether changes in the operating state of device A will cause changes in the state of device B in the future. If the test results show that the pressure change of device A is a Granger cause of the flow rate change of device B, then a time-series correlation between device A and device B can be determined. This facilitates the accurate determination of third sub-relationships between community LPG turnover warehouse equipment based on data association.

[0100] According to some embodiments of the present invention, establishing a second association between community LPG turnover warehouse equipment and operating scenarios includes:

[0101] Acquire scenario data for community LPG distribution warehouses; the scenario data includes meteorological data, power facility data, and transportation facility data;

[0102] The operational data and scenario data of the community LPG turnover warehouse equipment were converted into a unified format and timestamps were added to form a related scenario dataset.

[0103] The Apriori algorithm is used to mine frequent itemsets and association rules among data to determine the association information between meteorology and equipment, power facilities and equipment, and transportation facilities and equipment, and to establish a second association relationship between community LPG turnover warehouse equipment and operating scenarios.

[0104] The working principle and beneficial effects of the above technical solution are as follows: Meteorological data, such as temperature, humidity, wind speed, and lightning warnings, are collected in real time through connection to the meteorological bureau's API or meteorological monitoring stations deployed near the transit warehouse, with a collection frequency of once per minute. Power facility data includes the operating status of power facilities near the transit warehouse (e.g., transformer load, line current); traffic facility data includes traffic conditions (vehicle traffic flow, frequency of heavy vehicle passage). Power fluctuations may affect the equipment control system, and vibrations from passing heavy vehicles can interfere with equipment operation. The operating data and scene data of the community LPG transit warehouse equipment are uniformly converted and timestamped to form a related scene dataset. The Apriori algorithm is used to mine frequent itemsets and association rules between data to determine the correlation information between meteorology and equipment: a regression model of temperature and tank pressure is constructed, such as predicting the increment threshold of tank pressure for every 1°C increase in temperature. If the actual pressure increment exceeds the predicted value, it is marked as a potential risk. Association rules are established between lightning weather and the equipment's electrostatic protection system; when a lightning warning is in effect, the electrostatic grounding resistance is checked for abnormalities. Correlation information between power facilities and equipment: Analyze the relationship between power load rate and voltage stability of equipment control systems. When the transformer load rate exceeds 80%, monitor voltage fluctuations in the equipment control system. Correlation information between transportation facilities and equipment: Study the correlation between vehicle flow, heavy vehicle frequency, and equipment vibration. Set a vehicle flow threshold; when the threshold is exceeded, compare whether equipment vibration data shows an abnormal increase. Establish a second correlation between community LPG turnover warehouse equipment and operating scenarios. For example, analyze the relationship between historical meteorological data, surrounding facility data, and equipment failure data to identify correlation rules such as "high temperature weather and power voltage fluctuation → equipment pressure sensor failure." Set support and confidence thresholds (e.g., support ≥ 10%, confidence ≥ 80%) to filter effective rules. This facilitates the accurate establishment of a second correlation between community LPG turnover warehouse equipment and operating scenarios.

[0105] According to some embodiments of the present invention, determining equipment fault diagnosis results based on operating data, a first correlation, and a second correlation includes:

[0106] The sample running data is cleaned and normalized to obtain the processed running data;

[0107] Feature extraction is performed on the processed data to obtain time-domain and frequency-domain features;

[0108] The time-domain features and frequency-domain features are combined to form a feature vector;

[0109] Based on the equipment's historical fault records, the feature vectors, the corresponding first and second sample association relationships are labeled to obtain labeled data.

[0110] The SVM model is trained using the feature vectors of the sample running data, the corresponding first and second sample association relationships, and the labeled data. During the training process, the radial basis function kernel function is set, and the model parameters are optimized through cross-validation to obtain the target SVM model.

[0111] Input the running data, the first correlation, and the second correlation into the target SVM model to determine the equipment fault diagnosis results.

[0112] The working principle and beneficial effects of the above technical solution are as follows: Data cleaning is performed on the sample operation data. For pressure data, a reasonable pressure range is set. If the sample operation data exceeds this range and three consecutive data collection anomalies occur, it is judged as abnormal data and removed. Temperature data is judged for outliers using the 3σ principle. For gas leak detection sensor data, if there are drastic fluctuations within a short period (e.g., within 1 minute) that do not conform to normal fluctuation patterns, they are marked and further checked to see if they are false alarms. If confirmed as false alarms, they are deleted. Flow data is checked for zero values ​​or negative data. If such anomalies occur without a reasonable explanation (e.g., the equipment is shut down normally), the data is removed. Vibration data uses the sliding window method to calculate the mean and variance of the data within the window. If the variance is too large (more than three times the variance during normal operation), abnormal vibration data is considered to exist within the window and is removed. Physical quantities such as temperature and pressure are standardized using Z-score; dimensionless features such as vibration spectra are scaled using Min-Max.

[0113] Feature extraction is performed on the processed operational data to obtain time-domain features, including: calculating the mean, variance, and peak value of pressure data. The mean reflects the average pressure level, the variance reflects the degree of pressure fluctuation, and the peak value can be used to detect sudden pressure changes. The mean, variance, and rate of temperature change are extracted from the temperature data. The mean, root mean square (RMS) value, and kurtosis of vibration data are calculated. The RMS value reflects the energy magnitude of the vibration, and the kurtosis can be used to detect the presence of impact vibrations.

[0114] Feature extraction is performed on the processed operational data, including frequency domain features. This involves performing Fourier transforms on data such as pressure, temperature, and vibration to convert time-domain signals into frequency-domain signals. The frequency-domain signals are then analyzed to extract the main frequency components and their corresponding amplitudes. By observing the spectrum, the characteristic frequency range of the equipment during normal operation is determined. An abnormal increase in amplitude within a specific frequency range may indicate a corresponding fault. For example, in motor bearing failures, abnormal amplitudes typically appear at specific fault characteristic frequencies (such as the rolling element passage frequency).

[0115] The time-domain and frequency-domain features are combined to form a feature vector; the first association (equipment topology) is converted into an adjacency matrix feature; and the second association (scene influence) is quantified into environmental factor weights. Based on the equipment's historical fault records, the feature vector, the corresponding sample first association, and sample second association are labeled to obtain labeled data; for example, a three-level labeling system: 0: normal; 1: early fault (requires planned maintenance); 2: severe fault (immediate shutdown). Example: the initial stage of bearing cage wear is labeled as 1, and upgraded to 2 when regular pulses appear. The feature vector of the sample operating data, the corresponding sample first association, sample second association, and labeled data are used as training samples to train the SVM model; during training, a radial basis function kernel function is set, and the model parameters are optimized through cross-validation to obtain the target SVM model.

[0116] Inputting the running data, the first correlation, and the second correlation into the target SVM model facilitates accurate determination of equipment fault diagnosis results.

[0117] According to some embodiments of the present invention, analyzing historical fault records of equipment and establishing a fault mode correlation matrix includes:

[0118] Analyze the equipment's historical fault records, which are derived from the equipment's operation logs, maintenance work orders, and fault alarm systems.

[0119] Clean up the equipment's historical fault records, removing duplicate records, erroneous data, and incomplete data, and standardize the description of fault types;

[0120] Identify the key equipment within the community LPG turnover warehouse. For each type of key equipment, analyze its potential failure modes by consulting equipment manuals, expert experience, and historical failure data. Summarize the failure modes of all equipment and establish a unified classification system as the failure mode classification system.

[0121] Iterate through the historical fault data of the equipment and count the number of times each of the two fault modes occurs simultaneously or occurs successively within a preset time period.

[0122] A square matrix is ​​constructed using all failure modes in the failure mode classification system as rows and columns; the matrix elements are then filled with the results of the failure mode correlation analysis to establish a failure mode correlation matrix.

[0123] The working principle of the above technical solution is as follows: Analyze historical equipment fault records, which are derived from equipment operation logs, maintenance work orders, and fault alarm systems; clean the historical fault records, removing duplicate records, erroneous data, and incomplete data, and standardize the description method of fault types; identify key equipment within the community LPG turnover warehouse, and for each type of key equipment, analyze its possible fault modes by consulting equipment manuals, expert experience, and historical fault data; summarize the fault modes of all equipment and establish a unified classification system as the fault mode classification system; ensure that each fault mode has a unique identifier and clear definition. Traverse the historical equipment fault data, counting the number of times each pair of fault modes occurs simultaneously or successively within a preset time period; construct a square matrix using all fault modes in the fault mode classification system as rows and columns. Assuming there are n fault modes, the matrix dimension is n×n. For each element M in the matrix... ij (where i represents the row and j represents the column), and is filled based on the correlation analysis results of fault mode i and fault mode j. For example, it is normalized to the [0,1] interval, for example using the formula... Where, N ij Let M be the co-occurrence count of fault mode i and fault mode j, and max(N) be the maximum co-occurrence count among all fault modes. When i = j, M ij It can be set to 1, indicating that it is fully self-associated. Example:

[0124] Assuming the community LPG storage facility mainly experiences three failure modes (A), insufficient pump flow (B), and valve sealing failure (C), the following correlation matrix is ​​obtained through statistical analysis:

[0125]

[0126] The matrix shows a high correlation between abnormal tank pressure and insufficient flow of the transfer pump (correlation value of 0.6), indicating that these two failure modes often occur simultaneously or successively.

[0127] The beneficial effects of the above technical solution are as follows: A fault mode correlation matrix is ​​established by filling matrix elements with the results of fault mode correlation analysis. This enables comprehensive processing and correlation of various fault modes, facilitating accurate analysis of the correlation patterns of equipment faults and determining the third-party correlation information between equipment based on fault modes.

[0128] like Figure 3 As shown, according to some embodiments of the present invention, a diagnostic system applying the community LPG turnover warehouse equipment fault diagnosis method described above includes:

[0129] The data acquisition module is used to collect operational data of the corresponding equipment based on sensors deployed on the community LPG turnover warehouse equipment;

[0130] The first module is used to establish the first association between community LPG turnover warehouse equipment;

[0131] The second module is used to establish a second association between the community LPG turnover warehouse equipment and the operating scenario;

[0132] The determination module is used to determine the equipment fault diagnosis results based on the operating data, the first correlation, and the second correlation.

[0133] The beneficial effects of the above technical solution are: it establishes data connections between devices, enabling analysis of fault propagation paths; it goes beyond fixed threshold-based fault diagnosis, further realizing the perception of dynamic changes in equipment operating modes; and it optimizes diagnostic logic by incorporating community gas usage environmental factors. Through sensor data fusion, device association modeling, and dynamic scenario adaptation, it improves the reliability and safety of fault diagnosis in complex scenarios. In the complex equipment operating environment of community LPG distribution warehouses, it achieves accurate identification, location, and early warning of equipment faults, while reducing false alarms and missed alarms caused by single data monitoring, improving the comprehensiveness and timeliness of fault diagnosis, and ensuring the safety and stability of community gas supply.

[0134] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for diagnosing faults in community LPG turnover warehouse equipment, characterized in that, include: Based on sensors deployed on community LPG turnover warehouse equipment, the corresponding equipment's operating data is collected; Establish the primary connection between the community LPG turnover warehouse equipment; Establish a second correlation between community LPG turnover warehouse equipment and operational scenarios; The equipment fault diagnosis results are determined based on the operating data, the first correlation, and the second correlation.

2. The method for diagnosing faults in community LPG turnover warehouse equipment as described in claim 1, characterized in that, Based on sensors deployed on community LPG turnover warehouse equipment, operational data of the corresponding equipment is collected, including: Pressure and temperature sensors are deployed in the storage tank to collect pressure and temperature data inside the tank in real time. Gas leak detection sensors are deployed at valve connections to monitor for LPG leaks. Deploy flow sensors on the pipeline to acquire LPG flow data; Vibration sensors are deployed on the power components to monitor the vibration status of the power components during operation; the power components include motors and pumps.

3. The method for diagnosing faults in community LPG turnover warehouse equipment as described in claim 1, characterized in that, Establish the primary connections between community LPG distribution warehouse facilities, including: Determine the first sub-association relationship based on physical connections between the community LPG turnover warehouse equipment; Determine the second sub-association relationship between community LPG turnover warehouse equipment based on operational logic; Determine the third sub-relationship between community LPG turnover warehouse equipment based on data association; Establish the first association between the community LPG turnover warehouse equipment based on the first sub-association, the second sub-association, and the third sub-association.

4. The method for diagnosing faults in community LPG turnover warehouse equipment as described in claim 3, characterized in that, Determine the first sub-association based on physical connections between community LPG turnover warehouse equipment, including: Determine the pipeline information between the equipment in the community LPG turnover warehouse; draw a pipeline connection topology diagram based on the pipeline information, determine the position and connection sequence of each piece of equipment in the pipeline system, and establish the first connection relationship based on the pipeline connection; Determine the electrical information between the equipment in the community LPG turnover warehouse; draw an electrical wiring diagram based on the electrical information, determine the connection relationship of each device in the electrical circuit, and establish a second connection relationship based on the electrical connection; The first sub-association relationship based on physical connection between community LPG turnover warehouse equipment is determined according to the first connection relationship and the second connection relationship.

5. The method for fault diagnosis of community LPG turnover warehouse equipment as described in claim 3, characterized in that, Determine the second sub-association relationship between community LPG turnover warehouse equipment based on operational logic, including: Analyze the process flow of the LPG turnover warehouse, determine the sequence and collaborative relationship of equipment in the process flow, and obtain the first related information based on the process flow logic; The control programs and logic algorithms of each device in the community LPG turnover warehouse were analyzed separately, and the control relationships between the devices were established. Based on the first association information and control association, a second sub-association relationship based on operational logic is determined between the community LPG turnover warehouse equipment.

6. The method for fault diagnosis of community LPG turnover warehouse equipment as described in claim 3, characterized in that, Determine the third sub-relationships between community LPG turnover warehouse equipment based on data association, including: The correlation of data collected by sensors deployed on various devices is calculated based on the Pearson correlation coefficient to quantitatively analyze the correlation between device data and obtain the second correlation information of the devices. Analyze historical equipment failure records, establish a failure mode correlation matrix, analyze the correlation patterns of equipment failures, and determine the third correlation information between equipment based on failure modes; Using time series analysis, we can obtain the fourth correlation information of equipment operation data in the time dimension; Based on the second, third, and fourth association information, a third sub-association relationship based on data association is determined between the community LPG turnover warehouse equipment.

7. The method for diagnosing faults in community LPG turnover warehouse equipment as described in claim 1, characterized in that, Establish a second correlation between community LPG turnover warehouse equipment and operational scenarios, including: Acquire scenario data for community LPG distribution warehouses; the scenario data includes meteorological data, power facility data, and transportation facility data; The operational data and scenario data of the community LPG turnover warehouse equipment were converted into a unified format and timestamps were added to form a related scenario dataset. The Apriori algorithm is used to mine frequent itemsets and association rules among data to determine the association information between meteorology and equipment, power facilities and equipment, and transportation facilities and equipment, and to establish a second association relationship between community LPG turnover warehouse equipment and operating scenarios.

8. The method for diagnosing faults in community LPG turnover warehouse equipment as described in claim 1, characterized in that, The equipment fault diagnosis results are determined based on operational data, the first correlation, and the second correlation, including: The sample running data is cleaned and normalized to obtain the processed running data; Feature extraction is performed on the processed data to obtain time-domain and frequency-domain features; The time-domain features and frequency-domain features are combined to form a feature vector; Based on the equipment's historical fault records, the feature vectors, the corresponding first and second sample association relationships are labeled to obtain labeled data. The SVM model is trained using the feature vectors of the sample running data, the corresponding first and second sample association relationships, and the labeled data. During the training process, the radial basis function kernel function is set, and the model parameters are optimized through cross-validation to obtain the target SVM model. Input the running data, the first correlation, and the second correlation into the target SVM model to determine the equipment fault diagnosis results.

9. The method for diagnosing faults in community LPG turnover warehouse equipment as described in claim 6, characterized in that, Analyze historical equipment failure records and establish a failure mode correlation matrix, including: Analyze the equipment's historical fault records, which are derived from the equipment's operation logs, maintenance work orders, and fault alarm systems. Clean up the equipment's historical fault records, removing duplicate records, erroneous data, and incomplete data, and standardize the description of fault types; Identify the key equipment within the community LPG turnover warehouse. For each type of key equipment, analyze its potential failure modes by consulting equipment manuals, expert experience, and historical failure data. Summarize the failure modes of all equipment and establish a unified classification system as the failure mode classification system. Iterate through the historical fault data of the equipment and count the number of times each of the two fault modes occurs simultaneously or occurs successively within a preset time period. A square matrix is ​​constructed using all failure modes in the failure mode classification system as rows and columns; the matrix elements are then filled with the results of the failure mode correlation analysis to establish a failure mode correlation matrix.

10. A diagnostic system applying the fault diagnosis method for community LPG turnover warehouse equipment as described in any one of claims 1-9, characterized in that, include: The data acquisition module is used to collect operational data of the corresponding equipment based on sensors deployed on the community LPG turnover warehouse equipment; The first module is used to establish the first association between community LPG turnover warehouse equipment; The second module is used to establish a second association between the community LPG turnover warehouse equipment and the operating scenario; The determination module is used to determine the equipment fault diagnosis results based on the operating data, the first correlation, and the second correlation.

Images