An online liquid ultrasonic flowmeter fault diagnosis method

Abstract

The application provides an online liquid ultrasonic flowmeter fault diagnosis method, and the application is aimed at a liquid ultrasonic flowmeter, a comprehensive fault diagnosis index system is constructed, an expert system and intelligent reasoning are fused, intelligent optimization algorithms are applied, and the high consistency of diagnosis results is ensured, so that an efficient, accurate and reliable performance monitoring and fault diagnosis scheme is provided, not only the equipment mechanism analysis and multi-dimensional monitoring indexes are deeply integrated, the intelligent level of complex fault identification is improved, but also the accuracy and adaptability of the model under different working conditions are ensured through dynamic parameter optimization, so that the high reliability of the diagnosis results is finally ensured, the accuracy of flow monitoring and the efficiency of fault prevention in industrial production are greatly promoted, and the application is especially suitable for scenes with high precision requirements.

Description

Technical Field

[0001] This invention relates to the field of flow meter fault diagnosis technology, and in particular to a fault diagnosis method for online liquid ultrasonic flow meters. Background Technology

[0002] In the application of ultrasonic flow meters for liquids, their non-contact measurement characteristics and adaptability to multiple media make them the preferred solution for industrial flow monitoring. However, the accuracy decay, zero-point offset, signal attenuation or distortion, and faults caused by extreme environmental factors encountered during long-term operation, if not diagnosed and corrected in a timely manner, will directly affect production accuracy, product quality, and operating costs. Existing fault diagnosis technologies, including LSTM, SVM, CNN, and BRB, each have their limitations: LSTM can handle long sequence data, but training becomes time-consuming as the model size increases; SVM is inefficient when handling large amounts of data and relies on carefully selected kernel functions; CNN models are powerful but have poor interpretability and are highly dependent on a large number of samples; BRB and its extended form ERB, while efficient and interpretable, face challenges in model performance and adaptability when dealing with dynamically changing data and environments, especially when environmental conditions change rapidly, making it difficult to accurately identify fault types in real time.

[0003] Therefore, it is necessary to provide a fault diagnosis method for online liquid ultrasonic flow meters to solve the above-mentioned technical problems. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides an online liquid ultrasonic flow meter fault diagnosis method to solve the deficiencies in the comprehensiveness, intelligence, adaptability, and accuracy of liquid ultrasonic flow meter fault diagnosis.

[0005] This invention provides a fault diagnosis method for an online ultrasonic liquid flow meter, the fault diagnosis method comprising the following steps:

[0006] S1. Obtain the actual operational data set of the liquid ultrasonic flow meter;

[0007] S2. Identify and optimize a diagnostic evaluation system based on the key indicators affecting the performance of the ultrasonic flow meter.

[0008] S3. Construct an index reference value matrix based on the normal operation data and historical fault data of the liquid ultrasonic flow meter;

[0009] S4. Use confidence functions to evaluate the measured values ​​of each indicator, calculate the confidence distribution of each indicator, and obtain the confidence distribution results;

[0010] S5. Match the confidence distribution results with the rules in the extended confidence rule base, use decision logic to infer the state of the flow meter, and obtain the inference results;

[0011] S6. Based on the matching and inference results in step S5, diagnose the specific fault type and severity of the liquid ultrasonic flow meter.

[0012] Preferably, step S1 includes the following steps:

[0013] S1.1 Use sensor networks to monitor various operating parameters of the flow meter in real time, including flow rate, sound velocity, signal strength, temperature and pressure, and ensure the continuity and integrity of data recording;

[0014] S1.2 Ensure that the timestamps of all sensor data are synchronized to facilitate data alignment during subsequent analysis;

[0015] S1.3. Conduct preliminary screening of the collected data to remove data with obvious errors or outliers, such as abnormal values ​​caused by sensor malfunctions.

[0016] Preferably, step S2 includes the following steps:

[0017] S2.1 Based on the working principle of the flow meter, select key indicators that directly affect the measurement accuracy and reliability, such as fluid noise ratio, signal propagation time and echo intensity.

[0018] S2.2. Through expert evaluation or statistical analysis, assign weights to each indicator to reflect its impact on overall performance.

[0019] S2.3 Utilize optimization algorithms to adjust the parameters of the fault diagnosis model.

[0020] Preferably, in step S3, the expression for constructing the reference value matrix is:

[0021]

[0022] Where X represents the constructed reference value matrix, A i,j This represents the i-th reference value of the j-th rule.

[0023] Preferably, in step S4, the expression for calculating the confidence level of the input indicators in the confidence distribution of each indicator is as follows:

[0024]

[0025] Among them, h i,j (i = 1, 2, ..., I; j = 1, 2, ..., J) is the characteristic index X i The reference values ​​for different assessment levels, x i,j It is indicator X i The input data, J represents the total number of reference values, and satisfies h i,j+1 ≥hi,j .

[0026] Compared with related technologies, the online liquid ultrasonic flow meter fault diagnosis method provided by the present invention has the following beneficial effects:

[0027] This invention targets liquid ultrasonic flow meters and provides an efficient, accurate, and reliable performance monitoring and fault diagnosis solution by constructing a comprehensive fault diagnosis index system, integrating expert systems and intelligent reasoning, applying intelligent optimization algorithms, and ensuring high consistency of diagnostic results. It not only deeply integrates equipment mechanism analysis and multi-dimensional monitoring indicators to improve the intelligent level of complex fault identification, but also ensures the accuracy and adaptability of the model under different operating conditions through dynamic parameter optimization, ultimately guaranteeing the high reliability of diagnostic results. This greatly promotes the accuracy of flow monitoring and the efficiency of fault prevention in industrial production, and is especially suitable for scenarios with high precision requirements. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the process of the present invention;

[0029] Figure 2 This is a schematic diagram of the reference value structure for human-in-the-loop configuration according to an embodiment of the present invention;

[0030] Figure 3 This is a schematic diagram of the reasoning mechanism of the model in an embodiment of the present invention;

[0031] Figure 4 This is a flowchart of the EBRB update process according to an embodiment of the present invention. Detailed Implementation

[0032] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0033] Example

[0034] A fault diagnosis method for an online ultrasonic liquid flow meter, the method comprising the following steps:

[0035] S1. Obtain the actual operational data set of the liquid ultrasonic flow meter.

[0036] Specifically, step S1 includes the following steps:

[0037] S1.1 Use a sensor network to monitor various operating parameters of the flow meter in real time, including flow rate, sound velocity, signal strength, temperature and pressure, and ensure the continuity and integrity of data recording.

[0038] S1.2 Ensure that the timestamps of all sensor data are synchronized to facilitate data alignment during subsequent analysis.

[0039] S1.3. Conduct preliminary screening of the collected data to remove data with obvious errors or outliers, such as abnormal values ​​caused by sensor malfunctions.

[0040] S2. Identify and optimize a diagnostic evaluation system based on the key indicators affecting the performance of the ultrasonic flow meter.

[0041] Specifically, step S2 includes the following steps:

[0042] S2.1. Based on the working principle of the flow meter, select key indicators that directly affect the measurement accuracy and reliability, such as fluid noise ratio, signal propagation time and echo intensity.

[0043] S2.2. Through expert evaluation or statistical analysis, assign weights to each indicator to reflect its impact on overall performance.

[0044] S2.3 Utilize optimization algorithms to adjust the parameters of the fault diagnosis model.

[0045] In this embodiment, a threshold-based update algorithm is used to adjust the parameters of the fault diagnosis model in order to achieve optimal diagnostic performance.

[0046] S3. Construct an index reference value matrix based on the normal operation data and historical fault data of the liquid ultrasonic flow meter.

[0047] Specifically, in this embodiment, experts determine the reference levels and reference values ​​for the indicators based on their analysis of historical engineering tests and understanding of the working principles of the ultrasonic flow meter, thereby constructing a reference value matrix X.

[0048]

[0049] Where A i,j This represents the i-th reference value of the j-th rule.

[0050] S4. Use the confidence function to evaluate the measured value of each indicator, calculate the confidence distribution of each indicator, and obtain the confidence distribution results.

[0051] Specifically, in this embodiment, each indicator in the evaluation indicator system is treated as a piece of evidence R. The weight ω and reliability θ of each evidence indicator are determined. Based on the reference value matrix constructed in step S3, a rule-based information transformation method is used to transform the indicator data into a confidence distribution. form

[0052]

[0053] Among them κ i,n This represents the confidence level of the i-th rule for the n-th result, and the expression for calculating the confidence level of the input indicator is:

[0054]

[0055] Among them, h i,j (i = 1, 2, ..., I; j = 1, 2, ..., J) is the characteristic index X i The reference values ​​for different assessment levels, x i,j It is indicator X i The input data, J represents the total number of reference values, and satisfies h i,j+1 ≥h i,j .

[0056] Secondly, in this step, the inconsistency between the antecedent and consequent of each piece of evidence can be used to determine the weight of each evidence indicator; the calculation process for inconsistency is as follows: First, based on the confidence distribution constructed in step S4, the antecedent similarity (SRA) and consequent similarity (SRC) between pairwise rules can be calculated, and their expressions are as follows:

[0057]

[0058] Then according to SRA(R) i ,R j ) and SRC(R i ,R j The weight of the i-th extended confidence rule can be calculated using the following expression:

[0059]

[0060] Among them, Incons(R) i ) represents the inconsistency degree of the i-th extended execution rule, and its calculation formula is:

[0061]

[0062] Among them, R i This represents the i-th extended confidence rule. This represents the confidence distribution of the u-th input value of the k-th attribute. This represents the confidence distribution generated for the nth result.

[0063] S5. Match the confidence distribution results with the rules in the extended confidence rule base, and use the ER parsing algorithm to infer the state of the flow meter and obtain the inference results.

[0064] Specifically, ultrasonic flow meters are widely used in the oil and gas industry to measure fluid volume or mass flow rate. In this embodiment, an 8-channel ultrasonic flow meter is used as an example for experimental verification. The experiment tested the ultrasonic flow meter in four different states: healthy, gas-injected, installation-affected, and waxed. To make the obtained diagnostic results more accurate, a feature extraction algorithm is first used to extract the importance score, i.e., the indicator with the highest influence on the fault state. The indicators in this data and their corresponding importance scores are shown in Table 1 below:

[0065] Table 1

[0066]

[0067]

[0068] Based on the importance scores in the table above, this scheme uses the four highest-scoring indicators for each metric to construct the model, namely, contour coefficient, symmetry, signal quality VII, and gain at both ends of the path III.

[0069] After determining the evaluation metrics used in the model, reference values ​​for different states need to be set, as shown in Table 2 below:

[0070] Table 2

[0071]

[0072] After performing the above operations, the dataset obtained in step S1 is used, with 60% allocated to the training set and 40% to the test set. The training set is partitioned using a human-in-the-loop strategy, as described below. Figure 2 As shown. The training set is input into the model. Based on the model parameters optimized in step S2, each training set rule is converted into the rule shown in step S4 and then input into the test set for inference. The model's inference mechanism is as follows. Figure 3 As shown, the ER parsing algorithm consists of the following three steps:

[0073] Step (1): Calculate the individual matching degree. Assume that the vector representation of the prerequisite attributes of each test sample is x(x... i (i = 1, ..., M) are converted into distributed confidence scores, where the distributed confidence score of the i-th attribute is represented as follows:

[0074] S(xi)={(Ai1,…ai j ); j = 1, ..., Ji}

[0075] The matching degree of the test sample relative to each premise attribute in each rule can then be calculated.

[0076]

[0077] in This represents the distance between the confidence distributions of the i-th premise attribute in the given test data and the i-th premise attribute in the k-th rule in the rule base. This represents the matching degree between the i-th premise attribute in the given test data and the i-th premise attribute in the k-th rule in the rule base. The larger the distance between two attributes, the less they match. If the distance between two attributes exceeds 1, they are considered a complete mismatch. To express.

[0078] Step (2): Calculate the activation weights. Based on step (1), calculate the matching degree of sample x for each attribute in each rule. Then use the rule weight θ k (k = 1, ..., L) and attribute weights ω i For each confidence rule R (i = 1, ..., M), the activation weight can be calculated for each sample x. The formula for calculating the activation weight of the k-th confidence rule is as follows:

[0079]

[0080] Step (3) Synthesizing Activation Rules. After calculating the activation rule weights, all rules with weights greater than 0 are extended to the new rules as activations. These activated rules are then merged into a single rule using an evidence reasoning parsing algorithm, which serves as the final result. Without loss of generality, assuming that all L extended confidence rules are activation rules, the confidence distribution of the synthesized rule is as follows:

[0081]

[0082] in, This represents the confidence level of the nth result in the kth rule. Let β represent the activation weight of the k-th rule. n Denotes the combined confidence level, and satisfies β n ≥0,

[0083] S6. Based on the matching and inference results in step S5, diagnose the specific fault type and severity of the liquid ultrasonic flow meter.

[0084] The model update process is as follows: Figure 4 As shown, it consists of three steps:

[0085] Step 1: The model performs the first step of diagnosis on the test set, obtaining misclassified samples and correctly classified samples.

[0086] Step 2: Using a human-in-the-loop strategy, by interfering with the attribute weights by experts, samples that are similar to the misclassified, weakly classified, and minority class samples among the correctly classified and strongly classified samples are added to the samples to be updated.

[0087] Step 3: Add all samples to be updated in the above steps to the rule base for the next round of diagnosis.

[0088] Step 6: Obtain the fault diagnosis results of the liquid ultrasonic flow meter.

[0089] After the above process, the confidence distribution of the model's final results is obtained. Combined with the initial reference values ​​set for the model, the category D is determined using the following utility calculation method. i .

[0090]

[0091] Where, μ n This represents the utility value of the nth result.

[0092] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0093] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0094] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

Claims

1. A fault diagnosis method for an online ultrasonic liquid flow meter, characterized in that, The fault diagnosis method includes the following steps: S1. Obtain the actual operational data set of the liquid ultrasonic flow meter; S2. Identify and optimize a diagnostic evaluation system based on the key indicators affecting the performance of the ultrasonic flow meter. S3. Construct an index reference value matrix based on the normal operation data and historical fault data of the liquid ultrasonic flow meter; S4. Use confidence functions to evaluate the measured values ​​of each indicator, calculate the confidence distribution of each indicator, and obtain the confidence distribution results; S5. Match the confidence distribution results with the rules in the extended confidence rule base, use decision logic to infer the state of the flow meter, and obtain the inference results; S6. Based on the matching and inference results in step S5, diagnose the specific fault type and severity of the liquid ultrasonic flow meter. The update process for the fault diagnosis model is as follows: Step 1: The model performs the first step of diagnosis on the test set, obtaining misclassified samples and correctly classified samples; Step 2: Using a human-in-the-loop strategy, by interfering with the attribute weights by experts, obtain samples that are similar to the misclassified, weakly classified, and minority class samples among the correct and strong classification samples, and add these samples together to the samples to be updated. Step 3: Add all samples to be updated in the above steps to the rule base for the next round of diagnosis.

2. The online liquid ultrasonic flow meter fault diagnosis method according to claim 1, characterized in that, Step S1 includes the following steps: S1.1 Use a sensor network to monitor various operating parameters of the flow meter in real time, including flow rate, sound velocity, signal strength, temperature and pressure, and ensure the continuity and integrity of data recording; S1.2 Ensure that the timestamps of all sensor data are synchronized to facilitate data alignment during subsequent analysis; S1.

3. Conduct preliminary screening of the collected data to remove data with obvious errors or outliers, such as abnormal values ​​caused by sensor malfunctions.

3. The online liquid ultrasonic flow meter fault diagnosis method according to claim 1, characterized in that, Step S2 includes the following steps: S2.1 Based on the working principle of the flow meter, select key indicators that directly affect the measurement accuracy and reliability, such as fluid noise ratio, signal propagation time and echo intensity. S2.

2. Through expert evaluation or statistical analysis, assign weights to each indicator to reflect its impact on overall performance. S2.3 Utilize optimization algorithms to adjust the parameters of the fault diagnosis model.

4. The online liquid ultrasonic flow meter fault diagnosis method according to claim 1, characterized in that, In step S3, the expression for constructing the reference value matrix is: Where X represents the constructed reference value matrix, A i,j This represents the i-th reference value of the j-th rule.

5. The online liquid ultrasonic flow meter fault diagnosis method according to claim 1, characterized in that, In step S4, the expression for calculating the confidence level of the input indicators in the confidence distribution of each indicator is as follows: Among them, h i,j (i = 1, 2, ..., I; j = 1, 2, ..., J) is the characteristic index X i The reference values ​​for different assessment levels, x i,j It is indicator X i The input data, J represents the total number of reference values, and satisfies h i,j+1 ≥h i,j .

Images