A method for electrical valve condition monitoring and alarming

Abstract

The present application relates to the technical field of electric valve state monitoring, in particular to a method for electric valve state monitoring and alarm. It comprises formulating knowledge graph updating simulation schematic diagram, updating the associated state of each state feature and fault type at different time nodes in real time; according to the real-time collected state features and the updated knowledge graph, the fault type of the current electric valve is predicted. The present application updates the associated state of each state feature and fault type at different time nodes in real time through formulating knowledge graph updating simulation schematic diagram, uses the fault matching quantity and the change rate as the correlation degree evaluation project of each state feature and fault type, accurately describes the relationship between each state feature and fault type, establishes a stable knowledge graph, and at the same time, combines the predicted fault type and the dynamic correlation between the current fault type and state feature, classifies the alarm degree, carries out early warning according to the alarm degree, and improves the early warning effect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric actuator valve state monitoring, in particular to a method for electric actuator valve state monitoring and alarm. BACKGROUND

[0002] As an indispensable key control unit in industrial production process, electric actuator valve is widely used in petroleum, chemical industry, power, metallurgy and other fields. Its main functions are to connect or cut off the flow of pipeline medium, change the flow direction of medium, adjust the flow and pressure of medium, etc.

[0003] The stable and reliable operation of electric actuator valve plays a crucial role in ensuring production safety, improving production efficiency and reducing operating costs. However, due to long-term service in complex and harsh conditions such as high temperature, high pressure, corrosion and vibration, electric actuator valve is prone to various types of faults such as jamming, leakage and actuator failure. Lightly, it affects product quality and increases energy consumption, and heavily, it leads to production interruption, equipment damage, and even serious safety accidents and environmental pollution.

[0004] In order to obtain the running state of electric actuator valve in time, a method for electric actuator valve state monitoring and alarm is urgently needed. SUMMARY

[0005] The purpose of the present application is to provide a method for electric actuator valve state monitoring and alarm. By formulating knowledge graph update simulation schematic diagram, real-time updating the correlation state of each state feature and fault type at different time nodes, using fault matching quantity and change rate as the correlation degree evaluation project of each state feature and fault type, accurately describing the relationship between each state feature and fault type, establishing a stable knowledge graph, and combining the predicted fault type and the dynamic correlation between the current fault type and the state feature, the alarm degree is divided into grades, to solve the problems raised in the above background technology, that is:

[0006] How to predict and adaptively warn according to the monitoring data.

[0007] To achieve the above purpose, a method for electric actuator valve state monitoring and alarm is provided, which includes configuring monitoring sensors to monitor the state data of electric actuator valve in real time, collecting the state features of each state data, i.e. using real-time data monitored by sensors, and combining historical monitoring data to establish a knowledge graph of state features and fault types.

[0008] Further, in order to improve the accuracy of the prediction result and avoid the influence of the mutation value on the prediction result, the historical monitoring data in a unit time period is extracted, the number domain and the condition domain of different state characteristics and fault types are obtained, the dynamic association between the state characteristics and the fault types is simulated, the knowledge graph is updated to simulate the schematic diagram, the association state of each state characteristic and fault type at different time nodes is updated in real time, that is, the area formed by the number domain and the area formed by the condition domain are used to simulate the association state of the state characteristics and the fault types, the proportion of the sum of the areas to the area threshold value is used to simulate the association state of the state characteristics and the fault types at different time points of the knowledge graph, and the association state of the state characteristics and the fault types at different time points of the knowledge graph is updated in real time, real-time prediction is carried out according to the association state, and in the later warning process, the corresponding warning level is determined according to the area proportion, so that the warning effect is improved.

[0009] Compared with the prior art, the beneficial effects of the present application are:

[0010] In the method for monitoring and alarming the state of the electrodynamic valve, the knowledge graph is updated to simulate the schematic diagram, the association state of each state characteristic and fault type at different time nodes is updated in real time, the fault matching amount and the change rate are used as the association degree evaluation items of each state characteristic and fault type, the relationship between each state characteristic and fault type is accurately described, a stable knowledge graph is established, the warning degree is divided in combination with the predicted fault type and the dynamic association between the current fault type and the state characteristic, the warning is carried out according to the warning degree, and the warning effect is improved. BRIEF DESCRIPTION OF DRAWINGS

[0011] Figure 1 The overall method flowchart of the present application is shown in the figure;

[0012] Figure 2 The knowledge graph update simulation schematic diagram of the present application is shown in the figure;

[0013] Figure 3 The method step diagram for monitoring the state data of the electrodynamic valve in real time is shown in the figure;

[0014] Figure 4 The method step diagram for establishing the knowledge graph of the state characteristics and the fault types is shown in the figure;

[0015] Figure 5 The method step diagram for updating the association state of each state characteristic and fault type at different time nodes in real time is shown in the figure;

[0016] Figure 6 The method step diagram for evaluating the association degree of the state characteristics and the fault types according to the area state is shown in the figure;

[0017] Figure 7 The method step diagram for warning according to the warning degree is shown in the figure. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0019] Please refer to Figure 1 As shown in the figure, a method for electric actuator valve state monitoring and alarm is provided, comprising the following steps:

[0020] S1, configure a monitoring sensor to monitor the state data of the electric actuator valve in real time, and collect the state characteristics of each state data;

[0021] S2, establish a knowledge graph of state characteristics and fault types in combination with historical monitoring data;

[0022] S3, extract historical monitoring data in a unit time period, obtain the quantity domain and condition domain of different state characteristics and fault types, and simulate the dynamic association of state characteristics and fault types;

[0023] S4, develop a knowledge graph update simulation diagram to update the association state of each state characteristic and fault type at different time nodes in real time;

[0024] S5, predict the fault type of the current electric actuator valve according to the real-time collected state characteristics and the updated knowledge graph;

[0025] S6, combine the predicted fault type and the dynamic association of the current fault type and state characteristics, grade the alarm degree, and perform early warning according to the alarm degree.

[0026] The specific content is as follows:

[0027] In order to realize real-time monitoring, different types of sensors are used for real-time monitoring in the present scheme, and different types of sensors are arranged at specific positions to obtain corresponding state data and collect the state characteristics of each state data. The sensors configured in the present scheme include:

[0028] Vibration sensor: IEP acceleration sensor with a wide frequency band of 0.5Hz-15kH is selected, the installation position is selected to be the key measuring points of the valve body, valve cover, actuator housing and the like which are sensitive to faults, magnetic attraction or bolt fixing method is adopted to ensure good signal coupling, and the monitoring target is to capture the structural vibration anomalies caused by impact in the valve opening and closing process, loose valve core parts, abnormal fluid excitation and the like;

[0029] Acoustic emission sensor: Choose high-sensitivity industrial microphone or contact acoustic emission sensor. The microphone is placed near the valve, with a specific range of 0.5-1m, to avoid environmental noise interference. The acoustic emission sensor is directly coupled to the surface of the valve body for detecting high-frequency elastic waves.

[0030] Pressure sensor: High-precision pressure transmitters are installed at the inlet and outlet of the valve and the gas source input of the pneumatic actuator. The working pressure of the medium, the pressure difference between the two ends of the valve, and the stability of the driving gas source pressure of the actuator are monitored. Pressure pulsation can reflect flow field anomalies or sealing problems.

[0031] Temperature sensor: Pt100 thin film platinum resistance or K-type thermocouple is installed on the surface of the valve body and the actuator, which is prone to heat. Monitor abnormal temperature rise caused by excessive friction, medium leakage or electrical failure.

[0032] Valve position sensor: Non-contact high-precision valve position transducer is used to provide continuous feedback signals of the actual position of the valve core.

[0033] Current / voltage sensor: For electric actuators, monitor the three-phase current and voltage signals of the driving motor to diagnose motor failure, power anomalies or load changes.

[0034] As shown in Figure 3 , according to the corresponding signal change condition fed back by each sensor, the time domain, frequency domain, time-frequency domain and deep features of the corresponding signal are extracted using feature extraction algorithm as state features, and the normal state features of each monitoring signal of the electric actuator under normal state are obtained. Signals exceeding normal state features are marked as abnormal signals, and corresponding state features are marked as abnormal state features.

[0035] Signals that do not exceed normal state features are marked as normal signals, and corresponding state features are marked as normal state features as a basis for comparison later.

[0036] In order to facilitate state feature comparison, as shown in Figure 4As shown, by combining historical monitoring data, a knowledge graph of state features and fault types is established. Since the same fault type is simultaneously fed back by multiple state features, during the specific establishment process, a monitoring time period needs to be formulated, and data statistics are performed within the monitoring time period, that is, according to the monitoring signals fed back by each sensor, the normal state features are combined for comparison, the monitoring signals exceeding the normal state feature range are marked as abnormal signals, and the abnormal ranges of each abnormal signal are divided, the fault causes corresponding to each abnormal range are counted, for example, the time domain features include mean, variance, standard deviation, and skewness, and the normal range of the mean is [n1, n2], the mean exceeding the normal range is an abnormal time domain feature, for example, the corresponding abnormal mean ranges are [n3, n4], [n5, n6], and [n7, n8], when the mean is in [n3, n4], the matching fault causes in historical monitoring include g1 and g2, and when the mean is in [n5, n6] or [n7, n8], the matching fault causes include g3 and g4, therefore, not only the state feature type needs to be considered, but also the range of the corresponding value.

[0037] At the same time, in order to improve the accuracy of the prediction result and avoid the influence of the mutation value on the prediction result, by extracting the historical monitoring data in a unit time period, the quantity domain and the condition domain of different state features and fault types are obtained, and the dynamic association between the state features and the fault types is simulated, and the specific content is as follows:

[0038] The quantity domain is the fault matching quantity in the monitoring time period, which is in the same abnormal value range and causes the same fault reason, for example, in the mean monitoring process, the total monitoring times are 20, the monitoring times in the abnormal mean range [n3, n4] are 15, and the times causing the fault reason g2 in the abnormal mean range [n3, n4] are 13, then the current corresponding fault matching quantity is 13 times;

[0039] The condition domain is the change rate of the change of the fault degree caused by the change of the value of the different state features. Since the fault reason produces different final fault degrees, the fault degree is associated with the change of the value of the state feature, therefore, the change rate of the change of the fault degree caused by the change of the value of the different state features is used as the correlation degree of the state feature and the fault reason, for the purpose of formulating a stable knowledge graph;

[0040] Finally, a knowledge graph update simulation schematic diagram is formulated, and the correlation state of each state feature and fault type at different time nodes is updated in real time, and the specific content is as follows:

[0041] As shown in Figure 5 Firstly, the unit interval is divided, as shown in Figure 2As shown, each square represents a unit interval of the corresponding feature. One vertex is taken as the starting point of the fault type, and the associated state feature is marked at the vertex position of the corresponding square. Each fault type and the corresponding state feature are connected to form a knowledge graph association line segment. The knowledge graph association line segment is used to separate the area of ​​the unit interval of the corresponding feature. The upper part is the fault matching quantity evaluation area, and the lower part is the change rate evaluation area.

[0042] The evaluation method for the fault matching quantity assessment area involves acquiring the fault matching quantity of each state characteristic within the monitoring period, such as... Figure 2 As shown, s1-s5 represent the number of fault matches for the same feature at different time points. Connecting the knowledge graph association segments with the fault match counts as vertical points forms fault match count association regions, and the area of ​​each fault match count association region is calculated, as shown below. Figure 2 As shown, S1 and S2 are the areas of the fault matching quantity associated regions at different time points, respectively.

[0043] For the rate of change assessment area, the origin of the coordinate system is marked at the relative vertex of the knowledge graph's associated line segments, such as... Figure 2 As shown, O is the origin of the coordinate system. A rate of change function y = kx is established, where k is the rate of change of the corresponding state characteristic. The higher the rate of change, the greater the slope of the function. The intersection points of the rate of change function y = kx with the knowledge graph's associated line segments are obtained at different time points. The area formed below the intersection point and the unit interval is marked as the rate of change associated area, as shown below. Figure 2 As shown, S3 is the area associated with the rate of change corresponding to this feature;

[0044] like Figure 6 As shown, the correlation between the current feature and the fault type is finally evaluated based on the area associated with the rate of change and the area associated with the fault matching quantity. This involves obtaining the sum of the areas associated with the rate of change and the fault matching quantity at different time points, marking this as the feature-related area, and setting an area threshold. In this scheme, the total area of ​​the unit interval is used as the area threshold. The ratio of the feature-related area to the area threshold is calculated. The higher the ratio, the higher the correlation between the current state feature and the fault type. When building the knowledge graph, the ratio of the feature-related area to the area threshold is used as the prediction ranking. For example, when the state features associated with a certain fault type include t1, t2, and t3, the ratios of the feature-related area to the area threshold for each state feature are t1, t2, and t3, respectively. s1 t s2 and t s3 Sort by size t s2 >t s3 >t s1 When the real-time collected status data matches the current associated status feature t2, the probability of it being an associated fault type is highest at this time, and the corresponding alarm status is matched according to the predicted probability in the later stage.

[0045] Finally, in order to perform real-time early warning processing, such as Figure 7 As shown, by combining the predicted fault type and the dynamic association between the current fault type and state features, the alarm level is classified and a warning is issued based on the alarm level. That is, the probability is ranked according to the ratio of the feature association area between each state feature and the fault type in the knowledge graph to the area threshold. The warning level is determined according to the ranking result. The higher the ratio, the higher the warning level of the state feature, and vice versa.

[0046] This invention develops a knowledge graph update simulation diagram to update the association status of various state features and fault types in real time at different time nodes. It uses the fault matching quantity and change rate as evaluation items for the degree of association between each state feature and fault type, accurately describes the relationship between each state feature and fault type, establishes a stable knowledge graph, and improves the accuracy of subsequent prediction results. At the same time, it combines the predicted fault type and the dynamic association between the current fault type and state features to classify alarm levels and issue warnings based on the alarm level, thereby improving the warning effect.

[0047] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for electrodynamic valve condition monitoring and alarming, characterized by, It comprises the following steps: S1, configure a monitoring sensor to monitor the state data of the electrodynamic valve in real time, collect the state characteristics of each state data; S2, combine historical monitoring data to establish a knowledge graph of state characteristics and fault types; S3, extract historical monitoring data in a unit time period, obtain the quantity domain and the condition domain of different state characteristics and fault types, and simulate the dynamic association between state characteristics and fault types; The quantity domain in S3 is the fault matching quantity in the same abnormal value range and the same fault reason within the monitoring time period; The condition domain in S3 is the change rate of the fault degree caused by the change of different state characteristic values; S4, develop a knowledge graph update simulation diagram to update the association state of each state characteristic and fault type at different time nodes in real time; The method for updating the association state of each state characteristic and fault type at different time nodes in real time in S4 comprises the following steps: S4.1, divide a unit interval, take one vertex of the unit interval as the starting point of the fault type, and mark the associated state characteristics at the vertex position of the corresponding grid; S4.2, connect each fault type with the corresponding state characteristics to form a knowledge graph association line segment, and use the knowledge graph association line segment to separate the area of the unit interval of the corresponding characteristics, the upper part is the fault matching quantity evaluation area, and the lower part is the change rate evaluation area; S4.3, connect the knowledge graph association line segment with the fault matching quantity as the foot point to form a fault matching quantity association area, and calculate the area of each fault matching quantity association area; S4.4, mark the coordinate origin at the relative top corner of the knowledge graph association segment, and establish a change rate function wherein is the change rate of the corresponding state feature, and the change rate function at different time points is obtained the intersection point of the knowledge graph association segment, and the change rate association area is formed by the intersection point and the unit interval below. S4.5, evaluate the association degree of the state characteristics and the fault type according to the area state; S5, predict the fault type of the current electrodynamic valve according to the real-time collected state characteristics and the updated knowledge graph; S6, combine the predicted fault type and the dynamic association between the current fault type and the state characteristics to grade the alarm degree, and perform early warning according to the alarm degree.

2. The method for electrodynamic valve condition monitoring and alarming as claimed in claim 1 wherein: The sensors configured in S1 include vibration sensors, acoustic emission sensors, pressure sensors, temperature sensors, valve position sensors, and current / voltage sensors.

3. The method for electrodynamic valve condition monitoring and alarming as claimed in claim 2 wherein: The method for monitoring the state data of the electrodynamic valve in real time in S1 comprises the following steps: S1.1, combine the monitoring methods of each sensor to preset the installation position of the corresponding sensor; S1.2, develop a monitoring feedback time to obtain the state data feedback by the corresponding sensor and perform signal conversion; S1.3, use a feature extraction algorithm to extract the state characteristics of the corresponding signal, and develop the normal state characteristics of the corresponding signal; Mark the signals exceeding the normal state characteristics as abnormal signals, and mark the corresponding state characteristics as abnormal state characteristics; Mark the signals not exceeding the normal state characteristics as normal signals, and mark the corresponding state characteristics as normal state characteristics.

4. The method for electrodynamic valve condition monitoring and alarming as claimed in claim 1 wherein: The method for establishing a knowledge graph of state characteristics and fault types in S2 comprises the following steps: S2.1, develop a monitoring time period and perform data statistics within the monitoring time period; S2.2, divide the abnormal range of each abnormal signal, and count the fault reasons corresponding to each abnormal range; S2.3, according to the correspondence between the fault cause and the abnormal signal, a knowledge graph corresponding to the state feature and the fault cause is established.

5. The method for electrodynamic valve condition monitoring and alarming as claimed in claim 1 wherein: The method for evaluating the correlation degree of the state feature and the fault type according to the area state in S4.5 comprises the following steps: S4.5.1, obtain the sum of the area of the change rate correlation area and the area of the fault matching amount correlation area at different time points, marked as the feature correlation area; S4.5.2, set an area threshold, calculate the ratio of the feature correlation area to the area threshold, the higher the ratio, the higher the correlation degree of the current state feature and the fault type, and when establishing the knowledge graph, the ratio of the feature correlation area to the area threshold is used as the prediction order.

6. The method for electrodynamic valve condition monitoring and alarming as claimed in claim 5 wherein: The area threshold in S4.5.2 is the total area of the unit interval.

7. The method for electrodynamic valve condition monitoring and alarming as claimed in claim 1 wherein: The method for prewarning according to the alarm degree in S5 comprises the following steps: S5.1, according to the ratio of the feature correlation area to the area threshold of each state feature and fault type in the knowledge graph, a probability ranking is performed; S5.2, according to the ranking result, a prewarning level is set, the higher the ratio of the state feature, the higher the prewarning level corresponding to the state feature, and vice versa, the lower the ratio of the state feature, the lower the prewarning level corresponding to the state feature.

Images