High pour-point oil centrifugal pump fault diagnosis method based on viscosity and temperature real-time correction

Abstract

The invention discloses a high pour-point oil centrifugal pump fault diagnosis method based on viscosity and temperature real-time correction, relates to the technical field of equipment fault detection, and improves the operation safety and reliability of oil transportation equipment. According to the method, the viscosity-temperature correlation parameter equation is established, the normal value threshold interval of various data is obtained from the historical oil transportation record according to the viscosity-temperature correlation parameter equation, the fault event and the corresponding abnormal correlation data range are set, and the time-space fault event chain is set for the historical oil transportation record according to the abnormal correlation data range. The method comprises the steps of establishing an oil transportation fault diagnosis tree, matching an abnormal associated data range of each fault event according to a real-time oil transportation record, judging an abnormal type of a corresponding state monitoring area and a real-time fault response event, and matching an abnormal fault associated path from the oil transportation fault diagnosis tree according to a time-space sequence of the real-time fault response event. And a fault event expected to occur in the state monitoring area is judged according to the abnormal fault association path.

Description

Technical Field

[0001] This invention relates to the field of equipment fault detection technology, specifically a fault diagnosis method for high-viscosity oil centrifugal pumps based on real-time viscosity and temperature correction. Background Technology

[0002] In the field of oil transportation, the transportation of high-pour-point crude oil has always been an extremely challenging problem. Due to its special physical properties, high-pour-point crude oil has a high pour point and viscosity, and is prone to solidification and poor flow during transportation, which seriously affects the normal operation of oil transportation equipment and the efficiency of oil transportation.

[0003] Traditional methods for condition monitoring and fault diagnosis of oil transportation equipment have many limitations. On the one hand, condition monitoring during the transportation of high-viscosity oil is not comprehensive or accurate enough, lacking targeted monitoring of key areas, resulting in the inability to obtain accurate transportation data in a timely manner. On the other hand, in terms of fault diagnosis, previous methods often fail to accurately determine the type and probability of faults based on transportation data, and cannot effectively establish correlations between faults, thus failing to predict faults in advance and take corresponding preventive measures. This means that when oil transportation equipment malfunctions, it cannot be dealt with effectively and promptly, which can not only cause interruptions in oil transportation but also lead to equipment damage, increased maintenance costs, and production losses. Therefore, a fault diagnosis method for high-viscosity oil centrifugal pumps based on real-time viscosity and temperature correction is needed. Summary of the Invention

[0004] To address the aforementioned technical problems, the present invention aims to provide a fault diagnosis method for high-viscosity centrifugal pumps based on real-time viscosity and temperature correction.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction includes the following steps: Step 1: Set up status monitoring areas according to the flow direction of high-pour-point oil in the oil transportation equipment, set up data sensing devices in each status monitoring area, and then collect historical oil transportation records and real-time oil transportation records of each status monitoring area through the data sensing devices. Step 2: Establish viscosity-temperature correlation parameter equations, and correct the viscosity-temperature correlation parameter equations based on historical oil transportation records. Then, obtain the normal value threshold ranges of various data from historical oil transportation records based on the corrected viscosity-temperature correlation parameter equations. Step 3: Set the fault events and corresponding abnormal related data ranges, set the spatiotemporal fault event chain for historical oil transportation records based on the abnormal related data ranges, set characteristic fault events for each status monitoring area based on the spatiotemporal fault event chain, and set causal related fault events for fault time, thereby establishing an oil transportation fault diagnosis tree. Step 4: Match the abnormal correlation data range of each fault event according to the real-time oil transfer records, and then determine the abnormal type and real-time fault response event of the corresponding status monitoring area. Match the abnormal fault correlation path from the oil transfer fault diagnosis tree according to the spatiotemporal order of the real-time fault response event, and determine the fault events expected to occur in the status monitoring area according to the abnormal fault correlation path.

[0006] Furthermore, the process of collecting various data from the status monitoring area through data sensing devices includes: Based on the flow direction of the high-pour-point oil within the oil transfer equipment, the equipment is divided into m status monitoring zones. A data sensing device is installed in each monitoring zone, and these zones are numbered a1, a2, a3, ..., a... m , where m is a natural number greater than 0; Before the oil transportation equipment transports high-pour-point oil, each data sensing device synchronizes its time through the system clock to ensure that all data collected by the data sensing devices have spatiotemporal consistency. A data integration cycle is set up so that during the process of transporting high-pour-point oil in the oil transportation equipment, each data sensing device collects laser image data, vibration signals, temperature change records, pressure values, oil flow records, and oil viscosity signals of its associated status monitoring area.

[0007] Furthermore, the generation process of the historical oil transfer records and real-time oil transfer records includes: At the end of any data integration period, the data are normalized and n dynamic sliding windows are set according to the length of the data integration period. The data are divided into n data segments by the dynamic sliding windows, and the mean and standard deviation of the data segments within the dynamic sliding windows are obtained, where n is a natural number greater than 0. The data segment is divided into several data points, and a difference threshold is set to determine whether the difference between the value of each data point and the corresponding mean and standard deviation is less than or equal to the difference threshold. Based on the judgment result, data points are verified. According to the distribution order of the dynamic sliding windows, starting from the first dynamic sliding window, it is merged with the second dynamic sliding window, the third and fourth dynamic sliding windows are merged, and so on, and the data point verification process is repeated. The dynamic sliding window merging process is repeated until only one dynamic sliding window exists. Once all data has been verified, the data integration cycle generates all historical or real-time oil transfer records, and marks the corresponding status monitoring area number and generation time.

[0008] Furthermore, the process of establishing the viscosity-temperature correlation parametric equation includes: Based on laser image data from historical oil transportation records, a visualized 3D model of the oil transportation equipment during the corresponding data integration period is established. At the same time, the oil viscosity signals in each historical oil transportation record are converted into oil viscosity change records. Temperature and viscosity changes from various historical oil transport records are mapped onto a visualized 3D model according to the spatiotemporal order of collection. Based on the temperature and viscosity changes from various monitoring areas, viscosity-temperature correlation parameter equations are established.

[0009] Furthermore, the process of correcting the viscosity-temperature correlation equation includes: Set an error threshold, and sequentially retrieve the temperature change records and oil viscosity change records from the historical oil transport records of each state monitoring area under each dynamic sliding window. Input the records into the corresponding viscosity-temperature correlation parameter equation, obtain the predicted value of the viscosity-temperature correlation parameter under the corresponding dynamic sliding window, and obtain the actual value of the viscosity-temperature correlation parameter based on the temperature change records and oil viscosity change records. Obtain the difference between the actual value of the viscosity-temperature correlation parameter and the viscosity-temperature correlation parameter equation. If the difference between three consecutive dynamic sliding windows is greater than or equal to the error threshold, trigger the parameter correction equation.

[0010] Furthermore, the process of obtaining the normal numerical threshold range includes: Based on the corrected viscosity-temperature correlation parameter equation, the theoretical viscosity threshold range under different temperature ranges is obtained, and the corresponding historical oil transfer records are matched according to each temperature range and the corresponding theoretical viscosity threshold range. Based on the vibration signals, pressure values, and oil flow rate records in the matched historical oil transportation records, normal value threshold ranges for vibration signals, pressure values, and oil flow rate records are constructed under various temperature ranges and corresponding theoretical viscosity threshold range combinations. Based on the state monitoring area and dynamic sliding window corresponding to each temperature range, theoretical viscosity range, and corresponding normal value threshold range, the temperature range, theoretical viscosity range, and corresponding normal value threshold range are mapped to the corresponding positions on the visualized 3D model.

[0011] Furthermore, the process of the spatiotemporal fault event chain includes: Several types of fault events are set, and multiple abnormal correlation data ranges are set for each fault event. Various data of historical oil transportation records generated in the same data integration cycle are mapped onto a visualized 3D model. Then, based on the abnormal correlation data range of each fault event, several spatiotemporal fault points are marked on the visualized 3D model. Based on the spatiotemporal order in which each spatiotemporal fault point appears, several spatiotemporal fault event chains are traversed on the visualized 3D model.

[0012] Furthermore, the process of establishing the oil pipeline fault diagnosis tree includes: Based on the spatiotemporal fault event chain, first-level characteristic fault events, second-level characteristic fault events, and irrelevant fault events are set for each state monitoring area. At the same time, first-level cause-related fault events, second-level cause-related fault events, first-level effect-related fault events, second-level effect-related fault events, and irrelevant related fault events are set for each fault event. Based on the distribution of the status monitoring area, establish m backbone nodes and a number of leaf nodes based on the number of fault event types. Connect the leaf nodes corresponding to the first-level characteristic fault events of each status monitoring area to the backbone nodes. Connect the second-level characteristic fault events according to the correlation of the leaf nodes connected to the backbone nodes. Connect the unrelated fault events according to the causal relationship between them and other leaf nodes, and thus obtain the oil transportation fault diagnosis tree.

[0013] Furthermore, the process for determining the types of anomalies in the status monitoring area and the real-time fault response events includes: The real-time temperature values ​​of each state monitoring area during the current abnormal monitoring cycle are input into the viscosity-temperature correlation parameter equation. The real-time oil viscosity values ​​of the corresponding state monitoring area are obtained based on the real-time oil viscosity signal. Then, the real-time oil viscosity values ​​are determined based on the output results of the viscosity-temperature correlation parameter equation to determine whether the real-time oil viscosity values ​​are within the corresponding theoretical viscosity range. If the real-time oil viscosity value is within the corresponding theoretical viscosity range, the current oil is judged to be normal; otherwise, the oil properties are judged to be abnormal, and a real-time fault response event is generated. At the same time, based on the theoretical viscosity range and the real-time temperature range, the associated normal value threshold range is retrieved, and then it is determined whether each real-time data is within the corresponding normal value threshold range. If not, the corresponding real-time data is recorded as abnormal real-time data; otherwise, it is recorded as normal real-time data. The abnormal data range of each fault event is matched with the values ​​of all current real-time abnormal data. If the abnormal data range of all fault events is satisfied, it is determined that there is a mechanical fault abnormality in the corresponding status monitoring area, and a real-time fault response event is generated according to the fault event name. Otherwise, no operation is performed.

[0014] Furthermore, the process of determining the expected fault events in the status monitoring area based on the abnormal fault association path includes: The latest occurrence time of each abnormal real-time data corresponding to each real-time fault response event is recorded as the occurrence time of the corresponding real-time fault response event. Then, based on the status monitoring area corresponding to each real-time fault response event, each real-time fault response event is sequentially input into the oil pipeline fault diagnosis tree according to the occurrence time. Then, based on the occurrence order of each real-time fault response event, the perfectly matching abnormal fault association path is traversed in the oil pipeline fault diagnosis tree. Then, based on the fault events in the leaf node following the last fault event in the abnormal fault association path, the type of fault expected to occur in the status monitoring area is determined. If multiple perfectly matching abnormal fault association paths exist simultaneously, the abnormal fault association paths are filtered based on subsequent fault events. At the same time, the types of faults expected to occur in the status monitoring area are determined based on each abnormal fault association path, and then a maintenance request is sent to maintenance personnel based on the predicted fault types.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention establishes and modifies the viscosity-temperature correlation parameter equation to obtain the normal value threshold range for various data. This scientific analysis method based on historical data can fully consider the complex factors in the high-pour-point oil transportation process, providing a clear standard for judging whether the oil transportation equipment is in normal operation. Furthermore, once the real-time data exceeds the threshold range, anomalies can be quickly detected, improving the accuracy of fault early warning.

[0016] 2. By setting the scope of fault events and abnormal correlation data, a spatiotemporal fault event chain, characteristic fault events, and causal correlation fault events are established, ultimately constructing an oil transportation fault diagnosis tree. This systematic construction method comprehensively considers various expected fault situations and their interrelationships during oil transportation, systematically sorting and integrating fault events, and providing a clear logical framework and diagnostic path for accurate fault diagnosis. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention.

[0018] Figure 1 This is a flowchart of the fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction as described in this invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other implementation methods obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0020] like Figure 1 As shown, the fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction includes the following steps: Step 1: Set up status monitoring areas according to the flow direction of high-pour-point oil in the oil transportation equipment, set up data sensing devices in each status monitoring area, and then collect historical oil transportation records and real-time oil transportation records of each status monitoring area through the data sensing devices. Step 2: Establish viscosity-temperature correlation parameter equations, and correct the viscosity-temperature correlation parameter equations based on historical oil transportation records. Then, obtain the normal value threshold ranges of various data from historical oil transportation records based on the corrected viscosity-temperature correlation parameter equations. Step 3: Set the fault events and corresponding abnormal related data ranges, set the spatiotemporal fault event chain for historical oil transportation records based on the abnormal related data ranges, set characteristic fault events for each status monitoring area based on the spatiotemporal fault event chain, and set causal related fault events for fault time, thereby establishing an oil transportation fault diagnosis tree. Step 4: Match the abnormal correlation data range of each fault event according to the real-time oil transfer records, and then determine the abnormal type and real-time fault response event of the corresponding status monitoring area. Match the abnormal fault correlation path from the oil transfer fault diagnosis tree according to the spatiotemporal order of the real-time fault response event, and determine the fault events expected to occur in the status monitoring area according to the abnormal fault correlation path.

[0021] Furthermore, step one is achieved through the following process: Step 101: The oil transportation equipment consists of several crude oil pumps and oil pipelines, wherein the crude oil pumps consist of a pump body, an impeller and a bearing, and the oil pipelines consist of a pipeline body and a heat tracing device. Based on the flow direction of the high-pour-point oil within the oil transfer equipment, the equipment is divided into m status monitoring zones. A data sensing device is installed in each monitoring zone, and these zones are numbered a1, a2, a3, ..., a... m , where m is a natural number greater than 0; It should be noted that the oil transportation equipment included in each condition monitoring area is different. For example, some condition monitoring areas include multiple crude oil pumps, while others only include oil pipelines. The data sensing device consists of a laser photography device, a system clock, a three-axis accelerometer, a non-contact infrared thermometer, a differential pressure transmitter (inlet and outlet of the crude oil pump), an electromagnetic flowmeter, and an online viscometer.

[0022] Step 102: Before the oil transportation equipment transports high-pour-point oil, each data sensing device synchronizes its time through the system clock to ensure that all data collected by the data sensing devices have spatiotemporal consistency. A data integration cycle is set up so that during the process of transporting high-pour-point oil in the oil transportation equipment, each data sensing device collects laser image data, vibration signals, temperature change records, pressure values, oil flow records, and oil viscosity signals of its associated status monitoring area.

[0023] Step 103: At the end of any data integration period, normalize all data and set n dynamic sliding windows according to the length of the data integration period. Divide all data into n data segments through the dynamic sliding windows and obtain the mean and standard deviation of the data segments within the dynamic sliding windows. n is a natural number greater than 0. The data segment is divided into several data points, and a difference threshold is set to determine whether the difference between the value of each data point and the corresponding mean and standard deviation is less than or equal to the difference threshold. If all differences are less than or equal to the difference threshold, no action is taken. If either the difference between the data point and the corresponding mean or standard deviation is greater than the difference threshold, the corresponding data point will be removed and the gap will be filled by linear interpolation. After all the dynamic sliding windows have completed data point verification, according to the distribution order of the dynamic sliding windows, starting from the first dynamic sliding window, it is merged with the second dynamic sliding window, the third and fourth dynamic sliding windows are merged, and so on, and the data point verification process is repeated, and the dynamic sliding window merging process is repeated until only one dynamic sliding window exists. Once all data has been verified, the data integration cycle generates all historical or real-time oil transfer records, and marks the corresponding status monitoring area number and generation time.

[0024] When using this method, refer to steps 101 to 103: By setting up status monitoring zones based on the flow direction of high-pour-point oil and installing data sensing devices in each zone, historical and real-time oil transfer records for each key area can be collected comprehensively and accurately. This provides a more detailed and accurate understanding of the operating status of the oil transfer equipment, offering a rich and reliable data foundation for subsequent fault diagnosis and prediction.

[0025] Furthermore, step two is achieved through the following process: Step 201: Based on the laser image data in the historical oil transportation records, establish a visual 3D model of the oil transportation equipment under the corresponding data integration period, and at the same time convert the oil viscosity signals in each historical oil transportation record into oil viscosity change records. The temperature change records and oil viscosity change records of each historical oil transportation record are mapped onto a visualized 3D model according to the spatiotemporal order of collection. Based on the transport properties of high-pour-point oil, it can be known that the viscosity of high-pour-point oil is mainly affected by temperature and composition. As the temperature increases, the thermal motion of molecules intensifies, the intermolecular forces weaken, and the viscosity decreases; conversely, as the temperature decreases, the viscosity increases. In addition, the content and structure of long-chain hydrocarbons, cycloalkanes and aromatic hydrocarbons in high-viscosity oils also have a significant impact on viscosity. For example, the higher the content of long-chain hydrocarbons and macromolecular compounds, the greater the viscosity of the oil. Therefore, it can be concluded that the transportation efficiency of high-pour-point oil in oil transportation equipment is mainly affected by temperature. Based on the temperature change records and oil viscosity change records of various monitoring areas, a viscosity-temperature correlation parameter equation is established: ; in The viscosity-temperature correlation parameter represents the i-th condition monitoring area, where T represents temperature in °C, 273.15 is the thermodynamic temperature conversion constant for converting Celsius temperature to Kelvin temperature, and i is a natural number less than m and greater than 0, representing the i-th condition monitoring area; A and B are empirical parameters whose values ​​are determined based on the component characteristics of high-pour-point oil and historical oil transportation data. The specific methods for determining the values ​​of empirical parameters A and B are as follows: the least squares method is used to fit historical oil transportation data, and the specific steps are as follows: Extract at least 500 sets of temperature-viscosity data from historical oil transportation records; The data is grouped according to temperature range, with no fewer than 100 data points in each group. The temperature range is divided into low temperature range, medium temperature range, and high temperature range. The low temperature range is less than 20 degrees Celsius, the medium temperature range is between 20 and 50 degrees Celsius, and the high temperature range is greater than 50 degrees Celsius. Perform nonlinear regression analysis on each set of data to calculate the optimal values ​​of A and B within that interval; Verify the fitting results and ensure that the correlation coefficient R² is greater than or equal to the set threshold. If the correlation coefficient is lower than the set threshold, the sample size is increased and the model is refitted.

[0026] Step 202: Set the error threshold, and sequentially retrieve the temperature change records and oil viscosity change records from the historical oil transfer records of each state monitoring area under each dynamic sliding window. Input them into the corresponding viscosity-temperature correlation parameter equation, obtain the predicted value of viscosity-temperature correlation parameter under the corresponding dynamic sliding window, and at the same time obtain the actual value of viscosity-temperature correlation parameter based on the temperature change records and oil viscosity change records. Obtain the difference ∆β between the actual value of the viscosity-temperature correlation parameter and the viscosity-temperature correlation parameter equation. If the difference ∆β of three consecutive dynamic sliding windows is greater than or equal to the error threshold, trigger the parameter correction equation; otherwise, do not perform any operation. The trigger parameter correction equation is as follows: ; in and The empirical parameters after the (k+1)th and kth corrections, respectively. The parameter is used to correct the operation; k is a natural number greater than 1.

[0027] Step 203: Based on the corrected viscosity-temperature correlation parameter equation, obtain the theoretical viscosity threshold range under different temperature ranges, and match the corresponding historical oil transfer records according to each temperature range and the corresponding theoretical viscosity threshold range. Based on the vibration signals, pressure values, and oil flow rate records in the matched historical oil transportation records, normal value threshold ranges for vibration signals, pressure values, and oil flow rate records are constructed under various temperature ranges and corresponding theoretical viscosity threshold range combinations. Based on the state monitoring area and dynamic sliding window corresponding to each temperature range, theoretical viscosity range, and corresponding normal value threshold range, the temperature range, theoretical viscosity range, and corresponding normal value threshold range are mapped to the corresponding positions on the visualized 3D model.

[0028] When using this method, refer to steps 201 to 203: A viscosity-temperature correlation parameter equation was established and corrected based on historical oil transportation records to obtain the normal value threshold ranges for various data. This method fully considers the special properties of high-pour-point oil and can more accurately define the normal range of oil transportation data, providing a scientific and accurate basis for judging whether equipment is abnormal.

[0029] Furthermore, step three is achieved through the following process: Step 301: Set several types of fault events and set multiple abnormal correlation data ranges for each fault event. The fault events include sudden pressure drop, sudden flow drop, and increase in oil viscosity. The various data from historical oil transport records generated in the same data integration cycle are mapped onto a visualized 3D model. Then, based on the abnormal correlation data range of each fault event, several spatiotemporal fault points are marked on the visualized 3D model. Based on the spatiotemporal order in which each spatiotemporal fault point appears, several spatiotemporal fault event chains are traversed on the visualized 3D model.

[0030] Step 302: For any fault event, count the number of times it appears in each state monitoring area and the name and number of fault events corresponding to the spatiotemporal fault points connected in the spatiotemporal fault event chain. First, for any status monitoring area, construct a fault distribution pie chart based on the frequency of occurrence of each fault event within it, and set a frequency threshold and a distribution percentage threshold. If a fault event has a frequency distribution percentage in the fault distribution pie chart that is greater than or equal to both the frequency threshold and the distribution percentage threshold, then the fault event is recorded as a first-level characteristic fault event in the corresponding status monitoring area. If a fault event has a frequency distribution percentage in the fault distribution pie chart that is greater than or equal to both the frequency threshold and the distribution percentage threshold, then the fault event is recorded as a secondary characteristic fault event in the corresponding status monitoring area. If a fault event has a frequency distribution percentage in the fault distribution pie chart that is less than both the frequency threshold and the distribution percentage threshold, then the fault event is recorded as an irrelevant fault event in the corresponding status monitoring area. For any fault event, based on the fault events corresponding to the spatiotemporal fault points connected in the spatiotemporal fault event chain, a causal fault distribution pie chart is established. Then, the process of setting characteristic fault events in the state monitoring area is adopted to set first-level causal related fault events, second-level causal related fault events, first-level effect related fault events, second-level effect related fault events, and unrelated related fault events for each fault event.

[0031] Step 303: Establish m backbone nodes according to the distribution of the status monitoring area, and establish several leaf nodes according to the number of fault event types; The leaf nodes corresponding to the first-level characteristic fault events in each status monitoring area are connected to the trunk node. The second-level characteristic fault events are connected according to the correlation of the leaf nodes connected to the trunk node. The unrelated fault events are connected according to the causal relationship between them and other leaf nodes, thus obtaining the oil pipeline fault diagnosis tree. For example, if multiple secondary fault events and a primary fault event are both primary result-related fault events and primary cause-related fault events, then the leaf nodes corresponding to each secondary fault event are directly connected to the leaf nodes corresponding to the primary fault events.

[0032] When using this method, refer to steps 301 to 303: By defining fault events and corresponding anomaly-related data ranges, a spatiotemporal fault event chain and an oil pipeline fault diagnosis tree are established. This method can clearly present the causal relationships and correlation paths between faults, making fault diagnosis more systematic and comprehensive. By setting characteristic fault events and causally related fault events for each condition monitoring area, the source and development trend of faults can be located more accurately.

[0033] Furthermore, step four is achieved through the following process: Step 401: At the beginning of each data integration cycle, the anomaly monitoring cycle is set according to the time length of the dynamic sliding window. Then, at the end of each anomaly monitoring cycle, the three-dimensional visualization model is updated according to the laser image data in the real-time oil transfer record, and other real-time data of the real-time oil transfer record are marked on the three-dimensional visualization model. Step 402: Input the real-time temperature values ​​of each state monitoring area in the current abnormal monitoring cycle into the viscosity-temperature correlation parameter equation, and obtain the real-time oil viscosity values ​​of the corresponding state monitoring area according to the real-time oil viscosity signal. Then, determine whether the real-time oil viscosity values ​​are within the corresponding theoretical viscosity range based on the output results of the viscosity-temperature correlation parameter equation. If the real-time oil viscosity value is within the corresponding theoretical viscosity range, the current oil is judged to be normal; otherwise, the oil properties are judged to be abnormal, and a real-time fault response event is generated. At the same time, based on the theoretical viscosity range and the real-time temperature range, the associated normal value threshold range is retrieved, and then it is determined whether each real-time data is within the corresponding normal value threshold range. If not, the corresponding real-time data is recorded as abnormal real-time data; otherwise, it is recorded as normal real-time data. The abnormal data range of each fault event is matched with the values ​​of all current real-time abnormal data. If the abnormal data range of all fault events is satisfied, it is determined that there is a mechanical fault abnormality in the corresponding status monitoring area, and a real-time fault response event is generated according to the fault event name. Otherwise, no operation is performed.

[0034] Step 403: Record the latest occurrence time of each abnormal real-time data corresponding to each real-time fault response event as the occurrence time of the corresponding real-time fault response event; Then, based on the status monitoring area corresponding to each real-time fault response event, each real-time fault response event is sequentially input into the oil pipeline fault diagnosis tree according to the occurrence time. Then, based on the occurrence order of each real-time fault response event, the perfectly matching abnormal fault association path is traversed in the oil pipeline fault diagnosis tree. Then, based on the fault events in the leaf node following the last fault event in the abnormal fault association path, the type of fault expected to occur in the status monitoring area is determined. If multiple perfectly matching abnormal fault association paths exist simultaneously, the abnormal fault association paths are filtered based on subsequent fault events. At the same time, the types of faults expected to occur in the status monitoring area are determined based on each abnormal fault association path, and then a maintenance request is sent to maintenance personnel based on the predicted fault types.

[0035] When using this method, refer to steps 401 to 403: By matching real-time oil transfer records with abnormal correlation data ranges, the system can promptly identify the types of anomalies and real-time fault response events in the corresponding status monitoring areas. Furthermore, based on the abnormal fault correlation paths, it can predict the fault events expected to occur in the status monitoring areas, achieving real-time fault prediction. This enables staff to take appropriate preventative and handling measures in advance, avoiding fault occurrences or reducing losses caused by faults, thereby improving the operational safety and reliability of oil transfer equipment.

[0036] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction, characterized in that, Includes the following steps: Step 1: Set up status monitoring areas according to the flow direction of high-pour-point oil in the oil transportation equipment, set up data sensing devices in each status monitoring area, and then collect historical oil transportation records and real-time oil transportation records of each status monitoring area through the data sensing devices. Step 2: Establish viscosity-temperature correlation parameter equations, and correct the viscosity-temperature correlation parameter equations based on historical oil transportation records. Then, obtain the normal value threshold ranges of various data from historical oil transportation records based on the corrected viscosity-temperature correlation parameter equations. Step 3: Set the fault events and corresponding abnormal related data ranges, set the spatiotemporal fault event chain for historical oil transportation records based on the abnormal related data ranges, set characteristic fault events for each status monitoring area based on the spatiotemporal fault event chain, and set causal related fault events for fault time, thereby establishing an oil transportation fault diagnosis tree. Step 4: Match the abnormal correlation data range of each fault event according to the real-time oil transfer records, and then determine the abnormal type and real-time fault response event of the corresponding status monitoring area. Match the abnormal fault correlation path from the oil transfer fault diagnosis tree according to the spatiotemporal order of the real-time fault response event, and determine the fault events expected to occur in the status monitoring area according to the abnormal fault correlation path.

2. The fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction according to claim 1, characterized in that, The process of collecting various data from the status monitoring area using data sensing devices includes: The oil transportation equipment is divided into m status monitoring zones, and a data sensing device is installed in each status monitoring zone, where m is a natural number greater than 0. A data integration cycle is set up so that during the process of transporting high-pour-point oil in the oil transportation equipment, each data sensing device collects laser image data, vibration signals, temperature change records, pressure values, oil flow records, and oil viscosity signals of its associated status monitoring area.

3. The fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction according to claim 2, characterized in that, The generation process of the historical oil transfer records and real-time oil transfer records includes: At the end of any data integration period, the data are normalized and n dynamic sliding windows are set according to the length of the data integration period. The data are divided into n data segments by the dynamic sliding windows, and the mean and standard deviation of the data segments within the dynamic sliding windows are obtained, where n is a natural number greater than 0. The data segment is divided into several data points, and a difference threshold is set to determine whether the difference between the value of each data point and the corresponding mean and standard deviation is less than or equal to the difference threshold. Based on the judgment results, data points are verified, and then the dynamic sliding windows are merged according to their distribution order until only one dynamic sliding window exists, thereby generating historical oil transfer records or real-time oil transfer records.

4. The fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction according to claim 3, characterized in that, The process of establishing the viscosity-temperature correlation equation includes: Based on laser image data from historical oil transportation records, a visualized 3D model of the oil transportation equipment during the corresponding data integration period is established. At the same time, the oil viscosity signals in each historical oil transportation record are converted into oil viscosity change records. Temperature and viscosity changes from various historical oil transport records are mapped onto a visualized 3D model according to the spatiotemporal order of collection. Viscosity-temperature correlation parameter equations are then established based on the temperature and viscosity changes from various monitoring areas.

5. The fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction according to claim 4, characterized in that, The process of correcting the viscosity-temperature correlation equation includes: Set an error threshold, and sequentially retrieve the temperature change records and oil viscosity change records from the historical oil transport records of each state monitoring area under each dynamic sliding window. Input these records into the viscosity-temperature correlation parameter equation to obtain the predicted value of the viscosity-temperature correlation parameter under the corresponding dynamic sliding window. At the same time, obtain the actual value of the viscosity-temperature correlation parameter based on the temperature change records and oil viscosity change records. Obtain the difference between the actual value of the viscosity-temperature correlation parameter and the viscosity-temperature correlation parameter equation. If the difference of three consecutive dynamic sliding windows is greater than or equal to the error threshold, trigger the parameter correction equation; otherwise, do nothing.

6. The fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction according to claim 5, characterized in that, The process of obtaining the normal numerical threshold range includes: Based on the corrected viscosity-temperature correlation parameter equation, the theoretical viscosity threshold range under different temperature ranges is obtained, and the corresponding historical oil transfer records are matched according to each temperature range and the corresponding theoretical viscosity threshold range. Based on the vibration signals, pressure values, and oil flow rate records in the matched historical oil transportation records, normal value threshold ranges for vibration signals, pressure values, and oil flow rate records are constructed under various temperature ranges and corresponding theoretical viscosity threshold range combinations. Based on the state monitoring area and dynamic sliding window corresponding to each temperature range, theoretical viscosity range, and corresponding normal value threshold range, the temperature range, theoretical viscosity range, and corresponding normal value threshold range are mapped to the corresponding positions on the visualized 3D model.

7. The fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction according to claim 6, characterized in that, The process of the spatiotemporal fault event chain includes: Several types of fault events are set, and multiple abnormal correlation data ranges are set for each fault event. Various data of historical oil transportation records generated in the same data integration cycle are mapped onto a visualized 3D model. Then, based on the abnormal correlation data range of each fault event, several spatiotemporal fault points are marked on the visualized 3D model, and several spatiotemporal fault event chains are traversed.

8. The fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction according to claim 7, characterized in that, The process of establishing the oil pipeline fault diagnosis tree includes: Based on the spatiotemporal fault event chain, first-level characteristic fault events, second-level characteristic fault events, and irrelevant fault events are set for each state monitoring area, and causal correlation fault events are set for each fault event. Based on the distribution of the status monitoring area, establish m backbone nodes and a number of leaf nodes based on the number of fault event types. Connect the leaf nodes corresponding to the first-level characteristic fault events of each status monitoring area to the backbone nodes. Connect the second-level characteristic fault events according to the correlation of the leaf nodes connected to the backbone nodes. Connect the unrelated fault events according to the causal relationship between them and other leaf nodes, and thus obtain the oil transportation fault diagnosis tree.

9. The fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction according to claim 8, characterized in that, The process for determining the types of anomalies in the status monitoring area and the real-time fault response events includes: The real-time temperature values ​​of each state monitoring area during the current abnormal monitoring cycle are input into the viscosity-temperature correlation parameter equation. The real-time oil viscosity values ​​of the corresponding state monitoring area are obtained based on the real-time oil viscosity signal. Then, the real-time oil viscosity values ​​are determined based on the output results of the viscosity-temperature correlation parameter equation to determine whether the real-time oil viscosity values ​​are within the corresponding theoretical viscosity range. Based on the determination results, the abnormality of oil properties is judged, and a real-time fault response event is generated. At the same time, based on the theoretical viscosity range and the real-time temperature range, the associated normal value threshold range is retrieved to determine whether each real-time data is within the corresponding normal value threshold range. If not, the corresponding real-time data is recorded as abnormal real-time data; otherwise, it is recorded as normal real-time data. The abnormal data range of each fault event is matched with the values ​​of all current real-time abnormal data. If the abnormal data range of all fault events is satisfied, it is determined that there is a mechanical fault abnormality in the corresponding status monitoring area, and a real-time fault response event is generated according to the fault event name. Otherwise, no operation is performed.

10. The fault diagnosis method for high-pour-point-rate centrifugal pumps based on real-time viscosity and temperature correction according to claim 9, characterized in that, The process of determining the expected fault events in the status monitoring area based on the abnormal fault association path includes: The latest occurrence time of each abnormal real-time data corresponding to each real-time fault response event is recorded as the occurrence time of the corresponding real-time fault response event. Then, based on the status monitoring area corresponding to each real-time fault response event, each real-time fault response event is sequentially input into the oil pipeline fault diagnosis tree according to its occurrence time. Then, based on the occurrence order of each real-time fault response event, the oil pipeline fault diagnosis tree is traversed to find a completely matching abnormal fault association path. Finally, based on the fault event in the leaf node after the last occurrence fault event in the abnormal fault association path, the type of fault expected to occur in the status monitoring area is determined.

Images