Valve fault data detection method and system

Abstract

The invention discloses a valve fault data detection method and system, and relates to the technical field of industrial automation control and fault diagnose.The method comprises the steps that historical data of operation of a valve are obtained, the collected historical data are analyzed through a deep learning method, and time sequence features are extracted; a fault scoring model is established according to the extracted time sequence features, the fault of the valve is warned in advance, and meanwhile the state of the valve is monitored in real time through the latest real-time data flow; and the prediction result of the fault scoring model is verified through the real-time data flow, and the valve abnormity scoring standard is dynamically adjusted according to the real-time data and the output of the fault scoring model. According to the method, the fault prediction model is constructed by collecting the time sequence data of the valve rod displacement and combining the deep learning technology, the valve operation state can be monitored in real time, abnormal features can be accurately recognized, and delay and errors of manual detection are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial automation control and fault diagnosis, and in particular to a valve fault data detection method and system. Background Art

[0002] As a key component in industrial control systems, valves are widely used in the fields of petroleum, chemical industry, electric power, metallurgy, etc. to control the flow, pressure and direction of the medium. With the continuous improvement of industrial automation, the operating status of valves is directly related to the safety and efficiency of the entire process. However, due to the inevitable wear and aging in long-term operation and the complexity of external working conditions, valves are prone to jamming, sealing failure, abnormal drive mechanism and other faults. If these faults are not detected and handled in time, they may lead to production interruption or even safety accidents.

[0003] In the prior art, most valve fault detection relies on manual periodic inspection or simple threshold-based fault warning methods. These methods have the following defects:

[0004] Lack of real-time performance: Manual inspection cannot achieve real-time monitoring and is prone to missed detection or delayed detection of faults.

[0005] Low level of intelligence: Fault detection methods based on simple threshold judgments are difficult to cope with complex changes in operating conditions.

[0006] Low accuracy: Lack of in-depth analysis and dynamic adjustment of time series data, which can easily lead to false positives or omissions due to environmental disturbances or errors.

[0007] To solve the above problems, an intelligent fault detection method based on valve stem displacement data is proposed. Through deep learning and real-time data analysis technology, the operating status of the valve is comprehensively monitored and predicted. Summary of the invention

[0008] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0009] In view of the problems existing in the above-mentioned existing data detection method and system for valve failure, the present invention is proposed.

[0010] In order to solve the above technical problems, the present invention provides the following technical solutions: a data detection method for valve failure, the method comprising the steps of:

[0011] Step 1: Obtain the historical data of valve operation, and analyze the collected historical data through deep learning methods to extract time series features;

[0012] Step 2: Establish a fault scoring model based on the extracted time series features to warn of valve failures in advance, and use the latest real-time data stream to monitor the valve status in real time;

[0013] Step 3: Verify the prediction results of the fault scoring model through real-time data streams, and dynamically adjust the valve anomaly scoring criteria based on the real-time data and the output of the fault scoring model;

[0014] If the prediction model warns in advance that a valve will fail, the valve's abnormality score will be increased to remind staff to take preventive measures; at the same time, real-time data will adjust the scoring criteria to adapt to the actual operation of the valve.

[0015] As a preferred solution of the data detection method for valve faults described in the present invention, in step 1, the method for extracting time series features is:

[0016] S101: The historical displacement data of the valve stem during valve operation collected by the sensor device is recorded as: X(t)={x1(t), x2(t), ..., x i (t)}, where x i (t) represents the i-th time series feature at time t;

[0017] S102: Use a specific function to transform each time series feature to extract useful features: f i =ln(1+x i (t) 2 );

[0018] S103: Assign weights based on the importance of each feature Form a weighted feature combination:

[0019] As a preferred solution of the data detection method for valve faults described in the present invention, in step 2, the method for establishing the fault scoring model is:

[0020] S201: Capture the impact of historical data on the current moment, combined with the time decay factor e -α(t-τ) Perform weighted integration to obtain historical integral data I(t):

[0021]

[0022] S202: Introducing real-time data stream Y(t)={y1(t), y2(t), ..., y M(t)}, combined with a specific function g j Extract real-time features: Then the overall real-time status is expressed by the mean of the real-time features:

[0023]

[0024] Among them, M represents the dimension of real-time data features;

[0025] S203: Combining the historical integral I(t) and the real-time correction G(t) to form a basic fault scoring model:

[0026] P(t)=I(t)+β·G(t)

[0027] Among them, β is a tuning parameter used to balance the impact of historical features and real-time data.

[0028] As a preferred solution of the data detection method for valve faults described in the present invention, the real-time prediction error is calculated based on the output P(t) of the fault scoring model and the verification feature G(t) generated by real-time data: E(t) = |P(t)-G(t)|. The larger the deviation, the lower the accuracy of the model prediction.

[0029] As a preferred solution of the valve fault data detection method described in the present invention, the method of dynamically adjusting the scoring criteria is:

[0030] Based on historical prediction P(t) and real-time rate of change Dynamically adjust the valve abnormality score: S(t) = P(t) + γ·ΔG(t);

[0031] Among them, γ is the adjustment parameter that controls the impact of real-time changes on the final score. If S(t) is greater than the warning threshold, the staff is prompted to take preventive measures.

[0032] As a preferred solution of the data detection method for valve failure described in the present invention, a dynamic adjustment coefficient η is set, η = 1-e -λ·E(t)

[0033] When the prediction error E(t) is large, η approaches 1, indicating that the score adjustment needs to be strengthened; when the prediction error E(t) is close to 0, η is close to 0, indicating that the model prediction effect is good.

[0034] As a preferred solution of the data detection method for valve fault described in the present invention, wherein: the prediction error E(t) is greater than the set error threshold, η is introduced into the final scoring formula, and the prediction result and real-time data stream are combined to verify: S(t)=(1+η)·P(t)+γ·ΔG(t).

[0035] The present invention also discloses a system for the data detection method for the valve failure, the system comprising:

[0036] The data acquisition module is responsible for collecting the valve stem displacement operation data of the valve, and the data preprocessing module cleans, denoises and formats the collected raw data to ensure data quality;

[0037] The time series feature extraction module uses deep learning methods to extract the key time series features F(t) of valve stem displacement and other data as the basis for modeling the subsequent fault scoring model;

[0038] The fault scoring model module builds a fault scoring model based on the time series feature F(t), dynamically adjusts the abnormality score of the valve, and generates a comprehensive score based on historical data and real-time verification results;

[0039] The real-time verification module verifies the prediction results of the fault scoring model and adjusts the accuracy of the model based on real-time data;

[0040] As well as the alarm and notification module, when the comprehensive score exceeds the set threshold, a fault alarm is triggered and maintenance personnel are notified.

[0041] The present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and is characterized in that the processor implements the steps of the above-mentioned valve fault data detection method when executing the computer program.

[0042] The present invention also discloses a computer-readable storage medium on which a computer program is stored, characterized in that when the computer program is executed by a processor, the steps of the above-mentioned valve fault data detection method are implemented.

[0043] Beneficial effects of the present invention:

[0044] 1. The present invention collects the time series data of valve stem displacement and combines deep learning technology to build a fault prediction model, which can monitor the valve operation status in real time, accurately identify abnormal characteristics, and avoid delays and errors in manual detection;

[0045] 2. The present invention designs a dynamically adjusted fault scoring model, which combines historical integral features and real-time correction features to enable the scoring criteria to be dynamically adjusted according to changes in actual working conditions, thereby improving the adaptability and reliability of the model;

[0046] 3. By verifying the output of the fault prediction model in real time and calculating the real-time prediction error, this method can warn of potential faults in advance, prompting staff to take preventive measures and reduce the probability of sudden failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0048] Figure 1 The figure is a schematic diagram of the overall process of a valve failure data detection method proposed by the present invention. DETAILED DESCRIPTION

[0049] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0050] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0051] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0052] Example 1, reference Figure 1 , as an embodiment of the present invention, provides a valve fault data detection method and system, the method comprising:

[0053] Step 1: Obtain the historical data of valve operation, and analyze the collected historical data through deep learning methods to extract time series features.

[0054] Specifically, the method for extracting time series features is:

[0055] S101: The historical displacement data of the valve stem during valve operation collected by the sensor device is recorded as: X(t)=(x1(t), x2(t), ..., x i (t)}, where x i (t) represents the i-th time series feature at time t;

[0056] The reason why the historical displacement data of the valve stem is used here is that due to internal medium deposition, foreign matter blockage or seal wear, the valve movement is not smooth, the valve stem displacement curve will show abnormal hysteresis or stagnation, and the displacement curve may suddenly stagnate or change nonlinearly during operation, or even remain fixed for a long time. In addition, the aging or deformation of the seal will cause the valve to be unable to be completely closed or opened, and the travel range of the valve stem will be reduced or extended, so the valve stem displacement data can be used as a good basis for valve failure prediction.

[0057] S102: Use a specific function to transform each time series feature to extract useful features: f i =ln(1+x i (t) 2 );

[0058] S103: Assign weights based on the importance of each feature Form a weighted feature combination:

[0059] Step 2: Establish a fault scoring model based on the extracted time series features. Specifically, the method for establishing the fault scoring model is as follows:

[0060] S201: Capture the impact of historical data on the current moment, combined with the time decay factor e -α(t-τ) Perform weighted integration to obtain historical integral data I(t):

[0061]

[0062] S202: Introducing real-time data stream Y(t)={y1(t), y2(t), ..., y M (t)}, combined with a specific function g j Extract real-time features: Then the overall real-time status is expressed by the mean of the real-time features:

[0063]

[0064] Among them, M represents the dimension of real-time data features;

[0065] S203: Combining the historical integral I(t) and the real-time correction G(t) to form a basic fault scoring model:

[0066] P(t)=I(t)+β·G(t)

[0067] Among them, β is a tuning parameter used to balance the impact of historical features and real-time data.

[0068] In summary, by collecting the time series data of valve stem displacement and combining it with deep learning technology to build a fault prediction model, it is possible to monitor the valve operating status in real time, accurately identify abnormal features, and avoid delays and errors in manual detection.

[0069] Step 3: Verify the prediction results of the fault scoring model through real-time data stream, that is, calculate the real-time prediction error based on the output P(t) of the fault scoring model and the verification feature G(t) generated by real-time data: E(t) = |P(t)-G(t)|. The larger the deviation, the lower the accuracy of the model prediction.

[0070] Dynamically adjust the criteria for valve anomaly scoring based on real-time data and the output of the fault scoring model.

[0071] The method of dynamically adjusting the scoring criteria is:

[0072] Based on historical prediction P(t) and real-time rate of change Dynamically adjust the valve abnormality score: S(t) = P(t) + γ·ΔG(t);

[0073] Among them, γ is the adjustment parameter that controls the impact of real-time changes on the final score. If S(t) is greater than the warning threshold, the staff is prompted to take preventive measures.

[0074] In addition, it should be noted that the dynamic adjustment coefficient η is set, η = 1-e -λ·E(t)

[0075] When the prediction error E(t) is large, η approaches 1, indicating that the score adjustment needs to be strengthened; when the prediction error E(t) is close to 0, η is close to 0, indicating that the model prediction effect is good.

[0076] When the prediction error E(t) is greater than the set error threshold, η is introduced into the final scoring formula, combined with the prediction results and real-time data stream verification: S(t) = (1+η)·P(t)+γ·ΔG(t).

[0077] In summary, through the dynamically adjusted fault scoring model, combined with the historical integration characteristics and real-time correction characteristics, the scoring criteria can be dynamically adjusted according to changes in actual working conditions, thereby improving the adaptability and reliability of the model.

[0078] Example 2: In order to further verify the beneficial effects of the present invention, the valve stem displacement data x was collected during the long-term operation of a control valve in an industrial pipeline system. i (t), the unit is millimeter (mm). The test data includes real-time operation data within 100 hours and simulates two situations:

[0079] Normal working condition: The valve has no faults and the operating status is stable.

[0080] Abnormal operating conditions: The valve gradually became slightly stuck after the 50th hour, resulting in abnormal displacement characteristics.

[0081] The method of the present invention is used to analyze the above data and verify the effect of fault detection.

[0082] Table 1 Calculation result data example table

[0083]

[0084] Combining the above data, it can be seen that under normal working conditions (48-50 hours), the fault score remains relatively stable and the prediction error is small, indicating that the model can accurately predict the valve state. From 51 hours, the valve stem displacement data increases abnormally, resulting in a significant increase in the fault score and prediction error, indicating that the fault state gradually appears. In summary, this method can not only detect faults in advance (abnormal increase at 51 hours), but also dynamically adjust the scoring criteria to adapt to real-time working conditions, significantly improving the real-time and accuracy of fault detection.

[0085] This embodiment also provides a computer device, which is applicable to a data detection method and system for valve failure, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement a data detection method and system for valve failure as proposed in the above embodiment.

[0086] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0087] The present embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, a data detection method and system for a valve fault as proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0088] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A valve failure data detection method, characterized in that: The method comprises the steps of: Step 1: Obtain the historical data of valve operation, and analyze the collected historical data through deep learning methods to extract time series features; Step 2: Establish a fault scoring model based on the extracted time series features to warn of valve failures in advance, and use the latest real-time data stream to monitor the valve status in real time; Step 3: Verify the prediction results of the fault scoring model through real-time data streams, and dynamically adjust the valve anomaly scoring criteria based on real-time data and the output of the fault scoring model; If the prediction model warns in advance that a valve will fail, the valve's abnormality score will be increased to remind staff to take preventive measures; at the same time, real-time data will adjust the scoring criteria to adapt to the actual operation of the valve.

2. A valve failure data detection method according to claim 1, characterized in that: In the step 1, the method for extracting the time series features is: S101: The historical displacement data of the valve stem during valve operation collected by the sensor device is recorded as: X(t) = {x1(t), x2(t), ..., x i (t)}, where x i (t) represents the i-th time series feature at time t; S102: Use a specific function to transform each time series feature to extract useful features: f i =ln(1+x i (t) 2 ); S103: Assign weights based on the importance of each feature Form a weighted feature combination:

3. A valve failure data detection method according to claim 2, characterized in that: In the step 2, the method for establishing the fault scoring model is: S201: Capture the impact of historical data on the current moment, combined with the time decay factor e -α(t-τ) Perform weighted integration to obtain historical integral data I(t): S202: Introduce real-time data stream Y(t)={y1(t),y2(t),...,y M (t)}, combined with a specific function g j Extract real-time features: Then the overall real-time status is expressed by the mean of the real-time features: Where M represents the dimension of real-time data features; S203: Combining historical integral data I(t) and real-time correction data G(t) to form a basic fault scoring model: P(t)=I(t)+β·G(t) Among them, β is a tuning parameter used to balance the impact of historical features and real-time data.

4. A valve failure data detection method according to claim 3, characterized in that: Based on the output P(t) of the fault scoring model and the verification feature G(t) generated by real-time data, the real-time prediction error is calculated: E(t) = |P(t)-G(t)|. The larger the deviation, the lower the accuracy of the model prediction.

5. A valve failure data detection method according to claim 4, characterized in that: The method of dynamically adjusting the scoring criteria is: Based on historical prediction P(t) and real-time rate of change Dynamically adjust the valve abnormality score: S(t) = P(t) + γ·ΔG(t); Among them, γ is the adjustment parameter that controls the impact of real-time changes on the final score. If S(t) is greater than the warning threshold, the staff is prompted to take preventive measures.

6. A valve failure data detection method according to claim 5, characterized in that: Establish a dynamic adjustment coefficient η, η = 1-e -λ·E(t) When the prediction error E(t) is large, η approaches 1, indicating that the score adjustment needs to be strengthened; when the prediction error E(t) is close to 0, η is close to 0, indicating that the model prediction effect is good.

7. A valve failure data detection method according to claim 6, characterized in that: The prediction error E(t) is greater than the set error threshold, and η is introduced into the final scoring formula, combined with the prediction results and real-time data stream verification: S(t) = (1+η)·P(t)+γ·ΔG(t).

8. The detection system of the valve fault data detection method according to claim 7, characterized in that: The system includes: The data acquisition module is responsible for collecting the valve stem displacement operation data of the valve, and the data preprocessing module cleans, denoises and formats the collected raw data to ensure data quality; The time series feature extraction module uses deep learning methods to extract the key time series features F(t) of valve stem displacement and other data as the basis for modeling the subsequent fault scoring model; The fault scoring model module builds a fault scoring model based on the time series feature F(t), dynamically adjusts the abnormality score of the valve, and generates a comprehensive score based on historical data and real-time verification results; The real-time verification module verifies the prediction results of the fault scoring model and adjusts the accuracy of the model based on real-time data; As well as the alarm and notification module, when the comprehensive score exceeds the set threshold, a fault alarm is triggered and maintenance personnel are notified.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the valve fault data detection method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a valve fault data detection method according to any one of claims 1 to 7 are implemented.

Images