Calorimeter online fault diagnosis method and system

Abstract

The invention discloses an on-line fault diagnosis method and system for a heat meter, and relates to the technical field of energy metering. When the system is operated, data are collected and preprocessed in real time through an integrated sensor group, key features are extracted to form a feature vector FVEC, and an on-line anomaly detection algorithm and a fault type judgment model are utilized to determine the fault type of the heat meter; and an anomaly detection index AS and a fault feature probability index TYP are accurately calculated. The data not only can reflect the health state of the system in real time, but also can further optimize a response strategy and fault judgment by executing an iteration module, so that the system has self-diagnosis and self-adaptive adjustment capabilities. The optimization process can effectively overcome the defect that a traditional heat meter system cannot process and detect faults online in real time, the accuracy and response speed of fault recognition are improved, equipment damage or energy waste caused by delayed detection is avoided, and then continuous and stable operation of the system is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of energy metering, and particularly to an online fault diagnosis method and system for a heat meter. Background Art

[0002] Energy metering technology is a key technology, which is widely applied in multiple fields such as heating systems, industrial processes, and building energy management. As an important tool in this field, a heat meter is used to measure the energy consumption in a heating system for accurate billing and energy efficiency management. In large residential communities or industrial parks, the heat meter can monitor the heat energy consumption of each user in real time.

[0003] During the daily operation of the heat meter, due to environmental factors, sensor aging, and signal interference, data anomalies occur frequently. These anomalies may manifest as drifts, mutations, or long-term deviations in measurement data, directly affecting the accuracy of heat energy metering. However, existing fault detection methods mostly rely on regular manual calibration and offline analysis, and cannot process and detect data anomalies online in real time. This method is not only time-consuming and laborious, but also may miss early fault signs, resulting in the accumulation of problems and affecting the performance of the entire heating system. In addition, the judgment of fault types by traditional detection means mainly relies on preset rules or empirical judgments, lacking the ability of dynamic adjustment and real-time feedback, and it is difficult to cope with the complex and changeable actual operating environment. The lack of online processing and detection capabilities further limits the possibility of the system to take effective intervention at the initial stage of a fault, leading to more complex subsequent problems. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides an online fault diagnosis method and system for a heat meter, which solves the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: An online fault diagnosis system for a heat meter includes a data acquisition and preprocessing module, a feature extraction module, an anomaly detection module, a fault type judgment module, and an execution iteration module;

[0006] The data acquisition and preprocessing module collects the data of the heat meter operation area in real time by integrating a sensor group at the heat meter operation area position and performs preprocessing to form a collected data group DAT;

[0007] The feature extraction module extracts features from the preprocessed collected data group DAT to obtain a temperature difference feature TDF, a flow rate change feature FCF, a pressure fluctuation feature PWF, a vibration frequency feature VDF, and a noise intensity feature NRF, and forms a feature vector FVEC;

[0008] The anomaly detection module processes the obtained feature vector FVEC using an online anomaly detection algorithm to obtain an anomaly detection index AS, and matches it with a preset operating anomaly status threshold YZ to obtain the operating status result of the heat meter;

[0009] The fault type judgment module fits based on the anomaly detection index AS and the feature vector FVEC to obtain a fault feature probability index TYP;

[0010] The execution iteration module matches the obtained fault feature probability index TYP with a preset abnormal feature evaluation threshold TZ for the operation of the heat meter to obtain an abnormal feature response strategy plan for the operation of the heat meter, and specifically executes according to the content of the abnormal feature response strategy plan for the operation of the heat meter. By recording the execution results, the fault feature probability index TYP is iteratively adjusted.

[0011] Preferably, the data acquisition and preprocessing module includes a data acquisition unit and a data preprocessing unit;

[0012] The data acquisition unit collects real-time data of the heat meter operation area by integrating a sensor group at the heat meter operation area location, including fluid temperature TMP, flow rate FLW, pipeline pressure PRS, ambient temperature ENV, vibration intensity VIB, and operation noise NRM;

[0013] Among them, the sensor group includes a temperature sensor, an ultrasonic flow sensor, a pressure sensor, an ambient temperature sensor, a vibration sensor, and a noise sensor;

[0014] The data preprocessing unit preprocesses the collected operation area data, including noise removal processing, missing value processing, and data normalization processing, to form a collected data group DAT;

[0015] Among them, the noise removal processing includes using moving average filtering to remove noise in time series data while retaining the trend of the operation area data; the missing value processing includes using linear interpolation to fill in the missing values based on the linear relationship between adjacent data points, thereby restoring a complete data sequence; the data normalization preprocessing includes using Z-score normalization for normalization processing, thereby adjusting different parameters to the same dimension.

[0016] Preferably, the feature extraction module includes a feature calculation unit and a feature integration unit;

[0017] The feature calculation unit extracts features from the preprocessed collected data group DAT to obtain a temperature difference feature TDF, a flow rate change feature FCF, a pressure fluctuation feature PWF, a vibration frequency feature VDF, and a noise intensity feature NRF;

[0018] The feature integration unit marks the time t for the temperature difference feature TDF, flow change feature FCF, pressure fluctuation feature PWF, vibration frequency feature VDF, and noise intensity feature NRF, and then integrates them to form a feature vector FVEC.

[0019] Preferably, the temperature difference feature TDF is used to reflect the difference state between the fluid temperature and the ambient temperature of the heat meter; the flow change feature FCF is used to reflect the change rate and stability of the flow; the pressure fluctuation feature PWF is used to reflect the stability of the pressure inside the pipeline; the vibration frequency feature VDF is used to analyze the vibration state of the device or pipeline, specifically by extracting the frequency components of the vibration signal through Fourier transform for feedback; the noise intensity feature NRF is used to reflect the noise level of the device and pipeline during operation.

[0020] Preferably, the anomaly detection module includes an anomaly detection unit and a state matching unit;

[0021] The anomaly detection unit performs weight assignment and weighted calculation processing on the temperature difference feature TDF, flow change feature FCF, pressure fluctuation feature PWF, vibration frequency feature VDF, and noise intensity feature NRF in the obtained feature vector FVEC by using an online anomaly detection algorithm, and obtains an anomaly detection index AS by synthesizing the weighted deviations of the temperature difference feature TDF, flow change feature FCF, pressure fluctuation feature PWF, vibration frequency feature VDF, and noise intensity feature NRF, for triggering the matching of the heat meter operation state result of the state matching unit;

[0022] The state matching unit matches the preset operation anomaly state threshold YZ with the anomaly detection index AS to obtain the heat meter operation state result, and triggers the execution of the fault type judgment module according to the heat meter operation state result.

[0023] Preferably, the heat meter operation state result is marked as a return signal, and the execution of the fault type judgment module is triggered according to the return signal;

[0024] When the return signal is 1, it is obtained that the heat meter operation state is abnormal, and the execution of the fault type judgment module is triggered;

[0025] When the return signal is 0, it is obtained that the heat meter operation state is normal, and the execution of the fault type judgment module is not triggered.

[0026] Preferably, the fault type judgment module verifies the anomaly detection index AS and the feature vector FVEC through a verification mechanism to verify the integrity and consistency of the data, temporarily stores the anomaly detection index AS and the feature vector FVEC in the buffer area synchronously, processes them in sequence according to the time series, and then comprehensively processes the anomaly detection index AS and the feature vector FVEC using an online fitting algorithm to obtain the fault feature probability index TYP, so as to realize the quantitative evaluation of different fault types.

[0027] Preferably, the execution iteration module includes an anomaly evaluation unit and an iteration control unit;

[0028] The anomaly evaluation unit matches the obtained fault feature probability index TYP with the preset heat meter operation feature anomaly evaluation threshold TZ to obtain the heat meter operation anomaly feature response strategy plan;

[0029] The iteration control unit specifically executes according to the content of the heat meter operation anomaly feature response strategy plan, records the execution result, and iteratively adjusts the fault feature probability index TYP.

[0030] Preferably, the heat meter operation anomaly feature response strategy plan is obtained through the following matching method:

[0031] When the fault feature probability index TYP ≥ the heat meter operation feature anomaly evaluation threshold TZ, obtain the heat meter operation status anomaly response strategy plan, including switching the operation mode, starting the self-diagnosis function of the heat meter, issuing an alarm, recording a log, and notifying relevant inspection personnel for maintenance;

[0032] When the fault feature probability index TYP < the heat meter operation feature anomaly evaluation threshold TZ, obtain the heat meter operation status anomaly non-response strategy plan.

[0033] An online fault diagnosis method for a heat meter includes the following steps:

[0034] Step 1: The data acquisition and preprocessing module collects the heat meter operation area data in real time by integrating a sensor group in the heat meter operation area position, and performs preprocessing to form the collected data group DAT;

[0035] Step 2: The feature extraction module extracts features from the preprocessed collected data group DAT to obtain the temperature difference feature TDF, the flow change feature FCF, the pressure fluctuation feature PWF, the vibration frequency feature VDF, and the noise intensity feature NRF, and forms the feature vector FVEC;

[0036] Step 3: The anomaly detection module processes the obtained feature vector FVEC by using an online anomaly detection algorithm, obtains an anomaly detection index AS, and matches it with a preset operating anomaly status threshold YZ to obtain the operating status result of the heat meter;

[0037] Step 4: The fault type judgment module fits according to the anomaly detection index AS and the feature vector FVEC to obtain a fault feature probability index TYP;

[0038] Step 5: The execution iteration module matches the obtained fault feature probability index TYP with a preset abnormal feature evaluation threshold TZ for the operation of the heat meter, obtains an abnormal feature response strategy plan for the operation of the heat meter, and specifically executes according to the content of the abnormal feature response strategy plan for the operation of the heat meter. By recording the execution results, the fault feature probability index TYP is iteratively adjusted.

[0039] The present invention provides a method and system for online fault diagnosis of a heat meter, having the following beneficial effects:

[0040] (1) During system operation, data is collected and preprocessed in real time through an integrated sensor group, key features are extracted to form a feature vector FVEC, and an online anomaly detection algorithm and a fault type judgment model are used to accurately calculate an anomaly detection index AS and a fault feature probability index TYP. These data can not only reflect the health status of the system in real time, but also further optimize the response strategy and fault judgment through the execution iteration module, enabling the system to have the ability of self-diagnosis and adaptive adjustment. This optimization process can effectively solve the deficiency that traditional heat meter systems cannot process and detect faults online in real time, improve the accuracy and response speed of fault identification, avoid equipment damage or energy waste caused by delayed detection, and thus ensure the continuous and stable operation of the system, as well as provide a more accurate and timely fault warning and response mechanism.

[0041] (2) In the data collection and preprocessing stage, key parameters are obtained in real time through an integrated multi-high-precision sensor and preprocessed, including moving average filtering, linear interpolation, and Z-score normalization, to ensure the high quality and consistency of the data. In the feature extraction process, the system extracts multi-dimensional features reflecting the operating status of the heat meter through complex feature calculation formulas and integrates these features into a feature vector FVEC with time stamps. This process not only ensures the integrity and accuracy of the data, but also enables the system to comprehensively capture the dynamic changes of the operating status. Compared with traditional methods, the present system has higher sensitivity and robustness in detecting minute anomalies and various complex fault modes, thus effectively avoiding misjudgments caused by inaccurate data or improper processing, and ensuring the stable operation and accurate measurement of the heat meter.

[0042] (3) Process the feature vector FVEC through an online anomaly detection algorithm, accurately calculate the anomaly detection index AS, and match it with the preset operating anomaly status threshold YZ to determine the operating status of the heat meter in real time. Through strict standardization and weight calculation, this module ensures a sensitive capture of various degrees of feature deviation, accurately identifying abnormal states. At the same time, through the return mechanism of the return signal, it realizes the intelligent triggering of the fault type judgment module, only performing further fault analysis when an anomaly is detected, thus avoiding waste of system resources. It is also applicable to continuous monitoring under complex working conditions, ensuring that the system can quickly make judgments and responses in the face of emergencies, reducing unnecessary downtime, and improving the overall operating efficiency and stability of the heat meter.

[0043] (4) Fit and calculate the anomaly detection index AS and the feature vector FVEC to accurately obtain the fault feature probability index TYP, and then match it with the preset threshold TZ to intelligently generate a highly targeted anomaly response strategy. This not only ensures that when an actual fault occurs, the system can quickly switch the operating mode, start the self-diagnosis function, and notify the maintenance personnel in a timely manner, but also reduces unnecessary intervention in non-serious abnormal situations, optimizing the utilization rate of system resources. Description of the Drawings

[0044] Figure 1 It is a schematic block diagram of an online fault diagnosis system for a heat meter according to the present invention;

[0045] Figure 2 It is a schematic diagram of the steps of an online fault diagnosis method for a heat meter according to the present invention. Detailed Embodiments

[0046] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0047] Embodiment 1

[0048] The present invention provides an online fault diagnosis system for a heat meter. Please refer to Figure 1 , which includes a data acquisition and preprocessing module, a feature extraction module, an anomaly detection module, a fault type judgment module, and an execution iteration module;

[0049] The data acquisition and preprocessing module collects real-time data in the operating area of the heat meter by integrating a sensor group in the operating area of the heat meter and performs preprocessing to form a collected data group DAT;

[0050] The feature extraction module extracts features from the preprocessed collected data group DAT to obtain the temperature difference feature TDF, the flow rate change feature FCF, the pressure fluctuation feature PWF, the vibration frequency feature VDF, and the noise intensity feature NRF, and forms a feature vector FVEC;

[0051] The anomaly detection module processes the obtained feature vector FVEC by using an online anomaly detection algorithm to obtain an anomaly detection index AS, and matches it with a preset operation anomaly status threshold YZ to obtain the operation status result of the heat meter;

[0052] The fault type judgment module fits according to the anomaly detection index AS and the feature vector FVEC to obtain a fault feature probability index TYP;

[0053] The execution iteration module matches the obtained fault feature probability index TYP with a preset heat meter operation feature anomaly evaluation threshold TZ to obtain a heat meter operation anomaly feature response strategy plan, and specifically executes according to the content of the heat meter operation anomaly feature response strategy plan. By recording the execution results, the fault feature probability index TYP is iteratively adjusted.

[0054] In this embodiment, data is collected and preprocessed in real time through an integrated sensor group, key features are extracted to form a feature vector FVEC, and an online anomaly detection algorithm and a fault type judgment model are used to accurately calculate the anomaly detection index AS and the fault feature probability index TYP. These data can not only reflect the health status of the system in real time, but also further optimize the response strategy and fault judgment through the execution iteration module, enabling the system to have the ability of self-diagnosis and adaptive adjustment. This optimization process can effectively solve the deficiency of the traditional heat meter system that cannot process and detect faults online in real time, improve the accuracy and response speed of fault identification, avoid equipment damage or energy waste caused by delayed detection, and thus ensure the continuous and stable operation of the system, as well as provide a more accurate and timely fault warning and response mechanism.

[0055] Embodiment 2

[0056] This embodiment is an explanatory description based on Embodiment 1. Please refer to Figure 1 , specifically: The data acquisition and preprocessing module includes a data acquisition unit and a data preprocessing unit;

[0057] The data acquisition unit collects the heat meter operation area data in real time by integrating a sensor group at the heat meter operation area position, including the fluid temperature TMP, the flow rate FLW, the pipeline pressure PRS, the ambient temperature ENV, the vibration intensity VIB, and the operation noise NRM;

[0058] Among them, the sensor group includes a temperature sensor, an ultrasonic flow sensor, a pressure sensor, an ambient temperature sensor, a vibration sensor, and a noise sensor;

[0059] The data preprocessing unit preprocesses the collected operation area data, including noise removal processing, missing value processing, and data normalization processing, to form the collected data group DAT, specifically DAT = {TMP, FLW, PRS, ENV, VIB, NRM};

[0060] Among them, the noise removal processing includes using moving average filtering to remove the noise in the time series data while retaining the trend of the operation area data; the missing value processing includes using linear interpolation to fill in the missing values based on the linear relationship between adjacent data points, thereby restoring the complete data sequence; the data normalization preprocessing includes using Z-score normalization for normalization processing, thereby adjusting different parameters to the same dimension.

[0061] The feature extraction module includes a feature calculation unit and a feature integration unit;

[0062] The feature calculation unit extracts features from the preprocessed collected data group DAT to obtain the temperature difference feature TDF, the flow change feature FCF, the pressure fluctuation feature PWF, the vibration frequency feature VDF, and the noise intensity feature NRF;

[0063] The feature integration unit marks the time t for the temperature difference feature TDF, the flow change feature FCF, the pressure fluctuation feature PWF, the vibration frequency feature VDF, and the noise intensity feature NRF, and then integrates them to form the feature vector FVEC, specifically FVEC = {TDF, FCF, PWF, VDF, NRF, t}.

[0064] Among them, the temperature difference feature TDF is used to reflect the difference state between the fluid temperature of the heat meter and the ambient temperature; the flow change feature FCF is used to reflect the change rate and stability of the flow; the pressure fluctuation feature PWF is used to reflect the stability of the pressure inside the pipeline; the vibration frequency feature VDF is used to analyze the vibration state of the equipment or pipeline, specifically by extracting the frequency components of the vibration signal through Fourier transform for feedback; the noise intensity feature NRF is used to reflect the noise level of the equipment and pipeline during operation;

[0065] The temperature difference feature TDF is obtained through the following feature calculation formula:

[0066]

[0067] Wherein, TMPi represents the fluid temperature at the i-th time point, ENVi represents the ambient temperature at the i-th time point, △ti represents the time interval at the i-th time point, log represents the logarithmic function, TMPmax represents the peak value of the fluid temperature, and TMPmin represents the valley value of the fluid temperature;

[0068] The flow rate change feature FCF is obtained through the following feature calculation formula:

[0069]

[0070] Wherein, T represents the total observation time length, specifically multiple time intervals △t, the time interval △t includes multiple time points i, FLWi represents the flow rate at the time point i, d represents the derivative, represents the change rate of the flow rate FLW with respect to the time point i, FLWmax represents the peak value of the flow rate, and FLWmin represents the valley value of the flow rate;

[0071] The pressure fluctuation feature PWF is obtained through the following feature calculation formula:

[0072]

[0073] Wherein, n represents the total number of time points, PRSi represents the pipeline pressure at the i-th time point, represents the average value of the pipeline pressure, and PRSprev represents the pipeline pressure at the (i - 1)-th time point;

[0074] The vibration frequency feature VDF is obtained through the following feature calculation formula:

[0075]

[0076] Wherein, w represents the total number of frequency components after Fourier transform, VIBk represents the k-th frequency component of the vibration intensity signal, and F(VIBk) represents the Fourier transform result of the vibration intensity VIB signal at the k-th frequency component, specifically representing the amplitude of the vibration intensity VIB;

[0077] The noise intensity feature NRF is obtained through the following feature calculation formula:

[0078]

[0079] Wherein, NRMi represents the operating noise at the i-th time point, represents the average value of the operating noise.

[0080] In this embodiment, in the data acquisition and preprocessing stage, key parameters are obtained in real time through integrating multiple high-precision sensors and preprocessed, including moving average filtering, linear interpolation, and Z-score standardization, to ensure the high quality and consistency of the data. In the feature extraction process, the system extracts multi-dimensional features reflecting the operating state of the heat meter through complex feature calculation formulas and integrates these features into a feature vector FVEC with time stamps. This process not only ensures the integrity and accuracy of the data but also enables the system to comprehensively capture the dynamic changes in the operating state. Compared with traditional methods, this system has higher sensitivity and robustness in detecting minor anomalies and various complex fault modes, thus effectively avoiding misjudgments caused by inaccurate data or improper processing and ensuring the stable operation and accurate measurement of the heat meter.

[0081] Embodiment 3

[0082] This embodiment is an explanatory description based on Embodiment 2. Please refer to Figure 1 , specifically: The anomaly detection module includes an anomaly detection unit and a state matching unit;

[0083] The anomaly detection unit assigns weights and performs weighted calculation processing on the temperature difference feature TDF, flow change feature FCF, pressure fluctuation feature PWF, vibration frequency feature VDF, and noise intensity feature NRF in the obtained feature vector FVEC by using an online anomaly detection algorithm, and obtains an anomaly detection index AS by synthesizing the weighted deviations of the temperature difference feature TDF, flow change feature FCF, pressure fluctuation feature PWF, vibration frequency feature VDF, and noise intensity feature NRF, for triggering the matching of the operating state result of the heat meter by the state matching unit.

[0084] The anomaly detection index AS is obtained through the following calculation formula:

[0085]

[0086] In the formula, m represents the total number of features in the feature vector FVEC, FVECj represents the j-th feature in the feature vector FVEC, represents the mean of the j-th feature in the feature vector FVEC, σFVECj represents the standard deviation of the j-th feature in the feature vector FVEC, specifically representing the fluctuation range of the feature value, represents the standardized deviation of the j-th feature in the feature vector FVEC, specifically representing the degree of deviation of the feature, and uj represents the preset weight value of the j-th feature;

[0087] The state matching unit matches the preset operating anomaly state threshold YZ with the anomaly detection index AS to obtain the operating state result of the heat meter, and triggers the execution of the fault type judgment module according to the operating state result of the heat meter.

[0088] The operation status result of the heat meter is marked as the return signal, and the execution of the fault type judgment module is triggered according to the return signal;

[0089] The return signal is obtained through a marking method;

[0090] When the return signal is 1, it indicates that the operation status of the heat meter is abnormal, and the execution of the fault type judgment module is triggered;

[0091] When the return signal is 0, it indicates that the operation status of the heat meter is normal, and the execution of the fault type judgment module is not triggered.

[0092] In this embodiment, the feature vector FVEC is processed by an online anomaly detection algorithm to accurately calculate the anomaly detection index AS, and by matching with the preset operation anomaly status threshold YZ, the operation status of the heat meter is judged in real time. This module ensures the sensitive capture of various feature deviation degrees through strict standardization and weight calculation, and can accurately identify the abnormal status; at the same time, through the return mechanism of the return signal, the intelligent trigger of the fault type judgment module is realized, and further fault analysis is only carried out when an anomaly is detected, thus avoiding the waste of system resources; it is also applicable to continuous monitoring under complex working conditions, ensuring that the system can quickly make judgments and responses in the face of emergencies, reducing unnecessary downtime, and improving the overall operation efficiency and stability of the heat meter.

[0093] Embodiment 4

[0094] This embodiment is an explanatory description based on Embodiment 3. Please refer to Figure 1 , specifically: The fault type judgment module verifies the anomaly detection index AS and the feature vector FVEC through a verification mechanism to verify the integrity and consistency of the data, synchronously stores the anomaly detection index AS and the feature vector FVEC in the buffer area, and processes them sequentially according to the time series. Then, the anomaly detection index AS and the feature vector FVEC are comprehensively processed using an online fitting algorithm to obtain the fault feature probability index TYP, realizing the quantitative evaluation of different fault types;

[0095] The fault feature probability index TYP is obtained through the following calculation formula:

[0096]

[0097] In the formula, TYPk represents the fault feature probability index TYP of the k-th fault type, m represents the total number of features in the feature vector FVEC, Pk,j represents the preset weight value between the k-th fault type and the j-th feature, FVECj represents the j-th feature in the feature vector FVEC, σFVECj represents the standard deviation of the j-th feature in the feature vector FVEC, and αk represents the sensitivity coefficient of the k-th fault type;

[0098] Among them, the fault types include temperature anomaly fault type, flow anomaly fault type, pressure anomaly fault type, vibration anomaly fault type, and noise anomaly fault type.

[0099] The execution iteration module includes an anomaly evaluation unit and an iteration regulation unit;

[0100] The anomaly evaluation unit matches the obtained fault feature probability index TYP with the preset heat meter operation feature anomaly evaluation threshold TZ to obtain a heat meter operation anomaly feature response strategy plan;

[0101] The iteration regulation unit specifically executes according to the content of the heat meter operation anomaly feature response strategy plan, and iteratively adjusts the fault feature probability index TYP by recording the execution results.

[0102] The heat meter operation anomaly feature response strategy plan is obtained through the following matching method:

[0103] When the fault feature probability index TYP ≥ the heat meter operation feature anomaly evaluation threshold TZ, obtain a heat meter operation status anomaly response strategy plan, including switching the operation mode, starting the self-diagnosis function of the heat meter, issuing an alarm, recording a log, and notifying relevant inspection personnel for maintenance;

[0104] When the fault feature probability index TYP < the heat meter operation feature anomaly evaluation threshold TZ, obtain a heat meter operation status anomaly non-response strategy plan.

[0105] In this embodiment, the anomaly detection index AS and the feature vector FVEC are fitted and calculated to accurately obtain the fault feature probability index TYP, and then by matching with the preset threshold TZ, an anomaly response strategy with strong pertinence is intelligently generated. This not only ensures that when an actual fault occurs, the system can quickly switch the operation mode, start the self-diagnosis function, and timely notify the maintenance personnel, but also can reduce unnecessary intervention in non-serious anomaly situations and optimize the system resource utilization rate.

[0106] Embodiment 5

[0107] A heat meter online fault diagnosis method, please refer to Figure 2 , specifically: including the following steps:

[0108] Step 1: The data acquisition and preprocessing module collects the data in the operating area of the heat meter in real time by integrating a sensor group in the operating area of the heat meter, and performs preprocessing to form the collected data group DAT;

[0109] Step 2: The feature extraction module extracts features from the preprocessed collected data group DAT to obtain the temperature difference feature TDF, the flow rate change feature FCF, the pressure fluctuation feature PWF, the vibration frequency feature VDF, and the noise intensity feature NRF, and forms the feature vector FVEC;

[0110] Step 3: The anomaly detection module processes the obtained feature vector FVEC by using an online anomaly detection algorithm to obtain the anomaly detection index AS, and matches it with the preset operating anomaly state threshold YZ to obtain the operating state result of the heat meter;

[0111] Step 4: The fault type judgment module fits according to the anomaly detection index AS and the feature vector FVEC to obtain the fault feature probability index TYP;

[0112] Step 5: The execution iteration module matches the obtained fault feature probability index TYP with the preset abnormal evaluation threshold TZ of the operating characteristics of the heat meter to obtain the response strategy plan for the abnormal characteristics of the heat meter operation, and specifically executes according to the content of the response strategy plan for the abnormal characteristics of the heat meter operation. By recording the execution results, the fault feature probability index TYP is iteratively adjusted.

[0113] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A heat meter online fault diagnosis system, characterized by: It includes data collection and preprocessing module, feature extraction module, anomaly detection module, fault type judgment module and execution iteration module; The data collection and preprocessing module collects the data of the heat meter operation area in real time by integrating a sensor group at the heat meter operation area, and performs preprocessing to form a collection data group DAT; The feature extraction module extracts features from the preprocessed collected data group DAT, obtains the temperature difference feature TDF, the flow change feature FCF, the pressure fluctuation feature PWF, the vibration frequency feature VDF and the noise intensity feature NRF, and forms a feature vector FVEC; The anomaly detection module processes the acquired feature vector FVEC by using an online anomaly detection algorithm, obtains the anomaly detection index AS, and matches it with the preset operation abnormal state threshold YZ to obtain the operation state result of the heat meter; The fault type judgment module performs fitting based on the abnormal detection index AS and the feature vector FVEC to obtain the fault feature probability index TYP; The execution iteration module matches the acquired fault feature probability index TYP with the preset heat meter operation feature abnormality assessment threshold TZ, obtains the heat meter operation abnormality feature response strategy plan, and performs specific execution according to the content of the heat meter operation abnormality feature response strategy plan. By recording the execution results, the fault feature probability index TYP is iteratively adjusted.

2. A heat meter online fault diagnosis system according to claim 1, characterized in that: The data acquisition and preprocessing module includes a data acquisition unit and a data preprocessing unit; The data acquisition unit integrates a sensor group at the heat meter operation area to collect real-time data of the heat meter operation area, including fluid temperature TMP, flow FLW, pipeline pressure PRS, ambient temperature ENV, vibration intensity VIB and operation noise NRM; Among them, the sensor group includes a temperature sensor, an ultrasonic flow sensor, a pressure sensor, an ambient temperature sensor, a vibration sensor and a noise sensor; The data preprocessing unit preprocesses the collected operating area data, including noise removal, missing value processing and data standardization, to form a collection data group DAT; Among them, noise removal processing includes using moving average filtering to remove noise in time series data while retaining the trend of operating area data; missing value processing includes using linear interpolation to fill missing values ​​with the linear relationship between adjacent data points, thereby restoring the complete data sequence; data standardization preprocessing includes using Z-score standardization for standardization processing, thereby adjusting different parameters to the same dimension.

3. A heat meter online fault diagnosis system according to claim 1, characterized in that: The feature extraction module includes a feature calculation unit and a feature integration unit; The feature calculation unit extracts features from the preprocessed collected data group DAT to obtain temperature difference features TDF, flow change features FCF, pressure fluctuation features PWF, vibration frequency features VDF and noise intensity features NRF; The feature integration unit marks the temperature difference feature TDF, flow change feature FCF, pressure fluctuation feature PWF, vibration frequency feature VDF and noise intensity feature NRF with time t, and then integrates them. Constitute the feature vector FVEC.

4. A heat meter online fault diagnosis system according to claim 1, characterized in that: in, The temperature difference feature TDF is used to reflect the difference between the fluid temperature of the heat meter and the ambient temperature; the flow change feature FCF is used to reflect the rate of change of the flow and its stability; the pressure fluctuation feature PWF is used to reflect the stability of the pressure in the pipeline; the vibration frequency feature VDF is used to analyze the vibration state of the equipment or pipeline, specifically by extracting the frequency component of the vibration signal through Fourier transform for feedback; the noise intensity feature NRF is used to reflect the noise level of the equipment and pipeline during operation.

5. A heat meter online fault diagnosis system according to claim 1, characterized in that: The anomaly detection module includes an anomaly detection unit and a state matching unit; The anomaly detection unit uses an online anomaly detection algorithm to perform weight assignment and weighted calculation processing on the temperature difference feature TDF, flow change feature FCF, pressure fluctuation feature PWF, vibration frequency feature VDF and noise intensity feature NRF in the acquired feature vector FVEC, and obtains the anomaly detection index AS by comprehensively calculating the weighted deviation of the temperature difference feature TDF, flow change feature FCF, pressure fluctuation feature PWF, vibration frequency feature VDF and noise intensity feature NRF, which is used to trigger the heat meter operation status result matching of the state matching unit; The state matching unit matches the preset abnormal operation state threshold YZ with the abnormal detection index AS to obtain the operation state result of the heat meter, and triggers the execution of the fault type judgment module according to the operation state result of the heat meter.

6. A heat meter online fault diagnosis system according to claim 5, characterized in that: The heat meter operation status result is marked as a return signal, and the execution of the fault type judgment module is triggered according to the return signal; When the return signal is 1, the abnormal operation status of the heat meter is obtained, triggering the execution of the fault type judgment module; When the return signal is 0, the heat meter operation status is normal and the execution of the fault type judgment module is not triggered.

7. A heat meter online fault diagnosis system according to claim 6, characterized in that: The fault type judgment module verifies the anomaly detection index AS and the feature vector FVEC through a verification mechanism to verify the integrity and consistency of the data, and simultaneously temporarily stores the anomaly detection index AS and the feature vector FVEC in the cache area, and processes them in sequence according to the time series. Then, the anomaly detection index AS and the feature vector FVEC are comprehensively processed using an online fitting algorithm to obtain the fault feature probability index TYP, thereby realizing quantitative evaluation of different fault types.

8. A heat meter online fault diagnosis system according to claim 7, characterized in that: The execution iteration module includes an abnormality assessment unit and an iteration control unit; The abnormality assessment unit matches the acquired fault characteristic probability index TYP with the preset heat meter operation characteristic abnormality assessment threshold TZ to obtain the heat meter operation abnormality characteristic response strategy scheme; The iterative control unit performs specific execution according to the content of the abnormal operation characteristic response strategy plan of the heat meter, records the execution results, and iteratively adjusts the fault characteristic probability index TYP.

9. A heat meter online fault diagnosis system according to claim 8, characterized in that: The heat meter operation abnormality characteristic response strategy solution is obtained through the following matching methods: When the fault characteristic probability index TYP ≥ the heat meter operation characteristic abnormality assessment threshold TZ, the heat meter operation status abnormality response strategy is obtained, including switching the operation mode, starting the heat meter's self-diagnosis function, issuing an alarm, recording a log, and notifying relevant inspection personnel to perform maintenance; When the fault feature probability index TYP is less than the heat meter operation feature abnormality assessment threshold TZ, the heat meter operation state abnormal non-response strategy solution is obtained.

10. A heat meter online fault diagnosis method, applied to a heat meter online fault diagnosis system according to any one of claims 1 to 9, characterized in that: The following steps are involved: Step 1: The data collection and preprocessing module collects the data of the heat meter operation area in real time by integrating a sensor group at the heat meter operation area, and performs preprocessing to form a collection data group DAT; Step 2: The feature extraction module extracts features from the preprocessed collected data group DAT, obtains the temperature difference feature TDF, the flow change feature FCF, the pressure fluctuation feature PWF, the vibration frequency feature VDF and the noise intensity feature NRF, and forms a feature vector FVEC; Step 3: The anomaly detection module processes the acquired feature vector FVEC by using an online anomaly detection algorithm to obtain the anomaly detection index AS, and matches it with the preset operation abnormal state threshold YZ to obtain the operation state result of the heat meter; Step 4: The fault type judgment module performs fitting based on the abnormal detection index AS and the feature vector FVEC to obtain the fault feature probability index TYP; Step 5: The execution iteration module matches the acquired fault feature probability index TYP with the preset heat meter operation feature abnormality assessment threshold TZ, obtains the heat meter operation abnormality feature response strategy plan, and performs specific execution according to the content of the heat meter operation abnormality feature response strategy plan. By recording the execution results, the fault feature probability index TYP is iteratively adjusted.

Images