Drain valve action monitoring method and device, medium and program product

Abstract

The invention discloses a trap action monitoring method and device, a medium and a program product. The method comprises the following steps: acquiring original acquisition data of a trap action monitoring associated sensor; determining the pressure wave transmission delay time of the drain valve, and determining each difference data set according to the original acquisition data and the pressure wave transmission delay time of the drain valve; extracting frequency domain characteristics of each difference data set according to a wavelet transform mode, and analyzing the frequency domain characteristics based on a long short-term memory network to obtain a preliminary judgment action state of the drain valve; and when the action state of the drain valve is preliminarily judged to be an abnormal action state, determining a data fusion result of each difference data set, and performing decision fusion and predicting the action state of the drain valve according to the convolutional neural network, the data fusion result and an evidence theory. According to the technical scheme, the working state of the steam trap can be accurately judged, and a guarantee is provided for digital modification of maintenance and overhaul work of the steam trap of the steam pipeline.

Description

Technical Field

[0001] The present invention relates to the technical field of steam pipelines, and in particular to a method, equipment, medium and program product for monitoring the operation of a steam trap. Background Art

[0002] Steam is the energy source for key equipment in tobacco industry silk-making workshops, such as tobacco drying machines and rehumidifiers. Currently, the tobacco industry's silk-making processes all utilize saturated steam. Precise control of the steam pressure at the equipment's inlet allows for precise regulation of the hot air temperature within the equipment. However, the actual steam used is affected by factors such as boiler combustion behavior, steam pressure, transportation distance, and environmental climate change. When steam is introduced into the silk-making equipment, its quality often fluctuates, resulting in reduced quality of tobacco products. For this reason, steam traps are installed along the steam pipeline network in tobacco industry silk-making workshops to promptly remove condensate from the steam transport process, ensuring increased steam dryness and smooth steam transport.

[0003] The float trap is one of the most commonly used steam traps. It uses the buoyancy of a ball or float to control the opening and closing of a drain outlet. When the water level inside the pipe rises to a certain level, the buoyancy of the ball or float causes the drain outlet to open, draining wastewater and impurities. When the water level drops to a certain level, the drain outlet automatically closes. However, due to its mechanical structure, remote access to device status information is impossible. If a mechanical failure occurs within the device, it is difficult to inspect and confirm during normal steam network operation.

[0004] Currently, the working status and fault condition prediction of steam traps mostly use the temperature data of the steam pipes before and after the steam trap as the basis for determining the working condition of the steam trap. However, this method has low judgment accuracy. For different fault conditions of steam traps with a small temperature difference before and after the steam trap, it is difficult to provide correct judgment. In other words, the existing judgment of steam trap working condition has large judgment errors and the judgment results cannot meet the needs. Summary of the Invention

[0005] The present invention provides a steam trap operation monitoring method, equipment, medium and program product, which can solve the problem that the existing steam trap working condition determination has large determination errors and the determination results cannot meet the needs.

[0006] According to one aspect of the present invention, a method for monitoring the operation of a steam trap is provided, comprising:

[0007] Acquire original data collected by steam trap operation monitoring associated sensors; wherein the steam trap operation monitoring associated sensors include pressure sensors and non-pressure associated sensors;

[0008] Determine the steam trap pressure wave transmission lag time, and determine each difference data set based on the original collected data and the steam trap pressure wave transmission lag time;

[0009] The frequency domain features of each difference data set are extracted using wavelet transform, and the frequency domain features are analyzed based on the long short-term memory network to obtain the preliminary judgment of the steam trap operation status;

[0010] When the steam trap is initially judged to be in an abnormal operating state, the data fusion results of each difference data set are determined, and decision fusion is performed based on the convolutional neural network, data fusion results and evidence theory to predict the steam trap operating state.

[0011] According to another aspect of the present invention, a steam trap operation monitoring device is provided, comprising:

[0012] A data acquisition module is used to obtain raw data collected by sensors associated with the steam trap operation monitoring; wherein the sensors associated with the steam trap operation monitoring include pressure sensors and non-pressure related sensors;

[0013] A difference data set determination module is used to determine the steam trap pressure wave transmission lag time, and determine each difference data set based on the original collected data and the steam trap pressure wave transmission lag time;

[0014] The first action determination module is used to extract the frequency domain features of each difference data set based on the wavelet transform method, and analyze the frequency domain features based on the long short-term memory network to obtain the preliminary action state of the steam trap;

[0015] The second action determination module is used to determine the data fusion results of each difference data set when the steam trap is initially determined to be in an abnormal action state, and perform decision fusion based on the convolutional neural network, data fusion results and evidence theory to predict the action state of the steam trap.

[0016] According to another aspect of the present invention, an electronic device is provided, comprising:

[0017] at least one processor; and

[0018] a memory communicatively connected to the at least one processor; wherein,

[0019] The memory stores a computer program executable by the at least one processor. The computer program is executed by the at least one processor so that the at least one processor can perform the steam trap operation monitoring method according to any embodiment of the present invention.

[0020] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the steam trap operation monitoring method according to any embodiment of the present invention when executed.

[0021] According to another aspect of the present invention, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the method for monitoring the operation of a steam trap according to any embodiment of the present invention is implemented.

[0022] The technical solution of the embodiment of the present invention obtains the original collected data of the steam trap action monitoring associated sensor to determine the steam trap pressure wave transmission lag time, and determines each difference data set based on the original collected data and the steam trap pressure wave transmission lag time, and then extracts the frequency domain characteristics of each difference data set according to the wavelet transform method, and analyzes the frequency domain characteristics based on the long short-term memory network to obtain the preliminary judgment of the steam trap action state. Further, when the steam trap action state is preliminarily judged to be an abnormal action state, the data fusion results of each difference data set are determined, and decision fusion is performed based on the convolutional neural network, the data fusion results and the evidence theory to predict the steam trap action state. In this solution, the working status of the steam trap is preliminarily determined by collecting data from non-single sensors and using a long-short-term memory network. When the working status of the steam trap is abnormal, the model determination source data and the decision results are fused to accurately determine the abnormal status of the steam trap at a finer granularity. This solves the problem of large errors in the existing steam trap working condition determination and the inability of the determination results to meet the needs. It can accurately determine the working status of the steam trap and provide guarantees for the digital transformation of steam pipeline steam trap maintenance and repair work.

[0023] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of 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.

[0025] Figure 1 A flow chart of a method for monitoring steam trap operation provided in Example 1 of the present invention;

[0026] Figure 2 A flow chart of a steam trap operation monitoring method provided in the second embodiment of the present invention;

[0027] Figure 3 A logical diagram of multi-layer data fusion and decision result fusion for determining abnormal status of a steam trap provided by the second embodiment of the present invention;

[0028] Figure 4 A schematic structural diagram of a steam trap operation monitoring device provided in a third embodiment of the present invention;

[0029] Figure 5 A schematic structural diagram of an electronic device that can be used to implement an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "original", "target", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] Example 1

[0033] Figure 1 This is a flow chart of a steam trap operation monitoring method provided in the first embodiment of the present invention. This embodiment is applicable to accurately determine the working condition of the steam trap. The method can be executed by a steam trap operation monitoring device. The steam trap operation monitoring device can be implemented in the form of hardware and / or software. The steam trap operation monitoring device can be configured in an electronic device. Figure 1 As shown, the method includes:

[0034] Step 110: Acquire the original data collected by the steam trap operation monitoring associated sensor.

[0035] The steam trap operation monitoring sensors must be arranged one before and one after the other, with no steam connections, steam traps, or other structures or equipment between the steam trap operation monitoring sensors and the steam traps. The steam trap operation monitoring sensors can include pressure sensors and non-pressure-related sensors. The non-pressure-related sensors can be sensors other than the pressure sensors in the steam trap operation monitoring sensors. The non-pressure-related sensors can include at least one of a temperature sensor and a mass flow sensor, i.e., the non-pressure-related sensors include a temperature sensor and / or a mass flow sensor. The raw sampled data can be sampled data from the steam trap operation monitoring sensors.

[0036] For example, a temperature sensor can be installed at the bottom of the steam pipe near the steam trap. A pressure sensor can be installed at the top of the steam pipe. A mass flow sensor can be installed between the steam trap and the next steam-consuming device, or in the middle of a tee.

[0037] In the embodiment of the present invention, the steam trap disposed on the steam pipeline may be used to monitor the associated sensors for the steam trap action, thereby obtaining raw data collected over a period of time, namely, sensor sampling data.

[0038] Step 120 : Determine the steam trap pressure wave transmission lag time, and determine each difference data set based on the original collected data and the steam trap pressure wave transmission lag time.

[0039] The steam trap pressure wave transmission lag time can be the time delay required for the pressure wave to travel from the steam trap inlet to the steam trap outlet. The difference data set can be the difference between data collected by a steam trap operation monitoring associated sensor before the steam trap and data collected by the same type of sensor after the steam trap with the same pressure wave transmission lag time.

[0040] In an embodiment of the present invention, the steam trap pressure wave transmission lag time can be determined by a direct measurement method or an indirect measurement method, and the steam trap pressure wave transmission lag time can be used as the acquisition time difference of the difference element in the difference data set to calculate the data. In this way, the data collected by the steam trap operation monitoring associated sensors of the same type arranged before and after the same steam trap at a time interval equal to the steam trap pressure wave transmission lag time can be subtracted to obtain a difference data set that matches the steam trap operation monitoring associated sensors.

[0041] Step 130: extract frequency domain features of each difference data set using wavelet transform, and analyze the frequency domain features based on a long short-term memory network to obtain a preliminary determination of the steam trap's operating state.

[0042] The preliminary determined operating state of the steam trap may be determined by a long short-term memory network based on frequency domain features of a difference data set. The preliminary determined operating state of the steam trap may include one of a steam trap open fault state, a steam trap normally closed state, a steam trap oscillating open and close state, a steam trap normally open state, and a steam trap closed fault state.

[0043] Correspondingly, the wavelet transform method can be used to extract the frequency domain features of each difference data set, and based on the long short-term memory network, the frequency domain features corresponding to each difference data set can be analyzed to preliminarily determine the action state of the steam trap, that is, to obtain the preliminary judgment of the action state of the steam trap.

[0044] Step 140: When the steam trap is initially determined to be in an abnormal operating state, the data fusion results of the difference data sets are determined, and decision fusion is performed based on the convolutional neural network, the data fusion results, and evidence theory to predict the steam trap operating state.

[0045] The abnormal operation state may include at least one of a steam trap opening failure, a steam trap oscillating opening and closing, and a steam trap closing failure. The data fusion result may be a result of cross-fusion between different difference data sets.

[0046] In an embodiment of the present invention, if it is determined that the preliminary action state of the steam trap is an abnormal action state, it indicates that the working condition of the steam trap needs to be further accurately determined. Therefore, the pairwise difference data sets can be cross-fused to obtain a data fusion result, and the data fusion result is subjected to feature extraction through the convolution layer of the trained convolutional neural network. Then, the output layer of the convolutional neural network outputs the multiple action states of the steam trap after data fusion. Then, the output results of the convolutional neural network are subjected to decision fusion using evidence theory to obtain the final action state of the steam trap.

[0047] Optionally, the cross-fusion method of the difference data set may include but is not limited to statistical methods (such as Kalman filtering, DS evidence theory, etc.), weighted average methods, and cognitive models (neural networks and knowledge systems), etc.

[0048] The technical solution of the embodiment of the present invention obtains the original collected data of the steam trap action monitoring associated sensor to determine the steam trap pressure wave transmission lag time, and determines each difference data set based on the original collected data and the steam trap pressure wave transmission lag time, and then extracts the frequency domain characteristics of each difference data set according to the wavelet transform method, and analyzes the frequency domain characteristics based on the long short-term memory network to obtain the preliminary judgment of the steam trap action state. Further, when the steam trap action state is preliminarily judged to be an abnormal action state, the data fusion results of each difference data set are determined, and decision fusion is performed based on the convolutional neural network, the data fusion results and the evidence theory to predict the steam trap action state. In this solution, the working status of the steam trap is preliminarily determined by collecting data from non-single sensors and using a long-short-term memory network. When the working status of the steam trap is abnormal, the model determination source data and the decision results are fused to accurately determine the abnormal status of the steam trap at a finer granularity. This solves the problem of large errors in the existing steam trap working condition determination and the inability of the determination results to meet the needs. It can accurately determine the working status of the steam trap and provide guarantees for the digital transformation of steam pipeline steam trap maintenance and repair work.

[0049] Example 2

[0050] Figure 2 This is a flow chart of a method for monitoring the operation of a steam trap provided by the second embodiment of the present invention. This embodiment is based on the above embodiment and is specifically refined to determine the steam trap pressure wave transmission lag time. Specifically, it may include: determining the data acquisition frequency of the pressure sensor; calculating the steam trap pressure wave transmission lag time based on the layout distance of adjacent pressure sensors, the data acquisition frequency of the pressure sensor, and the pressure data cross-correlation function. Figure 2 As shown, the method includes:

[0051] Step 210: Acquire the original data collected by the steam trap operation monitoring associated sensor.

[0052] Step 220: Determine the data acquisition frequency of the pressure sensor.

[0053] The data collection frequency refers to the number of times data is collected per unit time.

[0054] Specifically, after selecting the steam trap operation monitoring associated sensor, the data collection frequency of the pressure sensor in the steam trap operation monitoring associated sensor may be further determined.

[0055] Step 230 : Calculate the steam trap pressure wave transmission lag time based on the layout distance between adjacent target pressure sensors, the data acquisition frequency of the pressure sensors, and the pressure data cross-correlation function, and determine each difference data set based on the original acquired data and the steam trap pressure wave transmission lag time.

[0056] The target pressure sensor can be a pressure sensor arranged in front of and behind a steam trap. The pressure data cross-correlation function can be used to describe the similarity of the pressure signals collected by the pressure sensors under different time delays.

[0057] In an embodiment of the present invention, the layout distance between adjacent target pressure sensors can be obtained, and then the pressure data cross-correlation function of the target pressure sensors can be determined. The sampling time of the original sampled data is divided by the data acquisition frequency of the pressure sensor to obtain the number of original acquired data of a single pressure sensor. The steam trap pressure wave transmission lag time is further calculated based on the layout distance between adjacent target pressure sensors, the number of original acquired data of a single pressure sensor, and the pressure data cross-correlation function of the pressure sensor. The steam trap pressure wave transmission lag time is then subtracted from the acquired data of similar steam trap operation monitoring associated sensors arranged before and after the same steam trap at a time interval equal to the steam trap pressure wave transmission lag time, thereby obtaining a difference data set matching the steam trap operation monitoring associated sensor.

[0058] In an optional embodiment of the present invention, determining each difference data set based on the original collected data and the steam trap pressure wave transmission lag time may include: determining the difference data set of the pressure sensor based on the original collected data of the pressure sensor and the steam trap pressure wave transmission lag time; and determining the difference data set of the non-pressure-related sensor based on the original collected data of the non-pressure-related sensor and the steam trap pressure wave transmission lag time.

[0059] In an embodiment of the present invention, the original data collected by pressure sensors arranged before and after the same steam trap at intervals equal to the steam trap pressure wave transmission lag time can be subtracted to obtain a pressure sensor difference data set. Furthermore, the original data collected by current non-pressure sensors (temperature sensors or mass flow sensors) arranged before and after the same steam trap at intervals equal to the steam trap pressure wave transmission lag time can be subtracted to obtain a current non-pressure sensor difference data set.

[0060] Step 240: extract frequency domain features of each difference data set using a wavelet transform method, and analyze the frequency domain features based on a long short-term memory network to obtain a preliminary determination of the steam trap's operating state.

[0061] In an optional embodiment of the present invention, before analyzing the frequency domain features based on the long short-term memory network to obtain a preliminary judgment of the steam trap operation state, the method may also include: obtaining historical acquisition data of sensors associated with the steam trap operation monitoring; normalizing the historical acquisition data to obtain feature data to be extracted, and training the long short-term memory network through the key features of the frequency domain analysis in the feature data to be extracted and the steam trap operation labeling type.

[0062] The historical data may be historical data collected by sensors associated with steam trap operation monitoring, located before and after steam traps in steam pipelines. The feature data to be extracted may be the normalized results of the historical data collected by sensors associated with steam trap operation monitoring. Key features for frequency domain analysis may be indicators that facilitate frequency domain feature analysis of signals. Examples of such key features include, but are not limited to, kurtosis, mean, root mean square value, standard deviation, and crest factor. The steam trap operation annotation type may be data that annotates the steam trap operating condition category.

[0063] In an embodiment of the present invention, historical collected data of sensors associated with steam trap action monitoring can be obtained, and then the historical collected data can be normalized. The normalized historical collected data is used as feature data to be extracted, and key features of frequency domain analysis that are helpful for frequency domain analysis are further parsed from the feature data to be extracted, and then the steam trap operating state that matches the historical collected data is determined, and the steam trap action labeling type is obtained. Based on the key features of frequency domain analysis and the steam trap action labeling type, a long short-term memory network is trained to enable the long short-term memory network to have the function of determining the steam trap action state.

[0064] Step 250: When the steam trap is initially determined to be in an abnormal operating state, the data fusion results of each difference data set are determined, and decision fusion is performed based on the convolutional neural network, the data fusion results, and evidence theory to predict the steam trap operating state.

[0065] In an optional embodiment of the present invention, when the steam trap is preliminarily determined to be in an abnormal operating state, determining the data fusion results of the difference data sets may include: obtaining a normal operating state prediction score interval of a long short-term memory network; if the quantitative assignment of the preliminarily determined operating state of the steam trap is not within the normal operating state prediction score interval, determining that the preliminarily determined operating state of the steam trap is in an abnormal operating state, and cross-fusing the difference data sets based on a Kalman filter to obtain a data fusion result.

[0066] The normal operation state prediction score interval may be a preset score interval where the quantitative assignment of the normal operation of the steam trap is located.

[0067] In an embodiment of the present invention, a normal action state prediction score interval of a long short-term memory network preset by a technician can be obtained, and a quantitative value can be assigned to the steam trap action label type. Therefore, after the long short-term memory network is trained, a quantitative value of the steam trap action type can be directly output. Therefore, it can be determined whether the quantitative assignment of the steam trap's preliminary action state, output by the trained long short-term memory network, falls within the normal action state prediction score interval. When the quantitative assignment of the steam trap's preliminary action state is not within the normal action state prediction score interval, it is characterized as an abnormal action state. Then, a Kalman filter is used to cross-fuse the difference data sets to obtain a data fusion result.

[0068] In an optional embodiment of the present invention, after predicting the operating state of the steam trap, the method may further include: generating a first type of fault alarm signal when the predicted operating state of the steam trap is that the steam trap cannot close and the duration is greater than a first duration; or generating a second type of fault alarm signal when the predicted operating state of the steam trap is that the steam trap cannot open and the duration is greater than the first duration; or generating a third type of fault alarm signal when the predicted operating state of the steam trap is that the steam trap oscillates and opens and closes and the duration is greater than a second duration.

[0069] Among them, the first time length and the second time length can be two pre-set time lengths, and the first time length is greater than the second time length. For example, the first time length can be 30 seconds, and the second time length can be 15 seconds. The embodiment of the present invention does not limit the specific time lengths of the first time length and the second time length. The first type of fault alarm signal can be an alarm signal triggered when the steam trap cannot be closed and the duration is greater than the first time length. The second type of fault alarm signal can be an alarm signal triggered when the steam trap cannot be opened and the duration is greater than the first time length. The third type of fault alarm signal can be an alarm signal triggered when the steam trap exhibits oscillating motion and the duration is greater than the second time length.

[0070] In an embodiment of the present invention, when the predicted action state of the steam trap is that the steam trap cannot close and the duration is greater than the first duration, a first type of fault alarm signal is generated. When the predicted action state of the steam trap is that the steam trap cannot open and the duration is greater than the first duration, a second type of fault alarm signal is generated. When the predicted action state of the steam trap is that the steam trap oscillates and opens and closes and the duration is greater than the second duration, a third type of fault alarm signal is generated to differentiate and warn of different types of faults of the steam trap, so as to provide targeted prompts to operation and maintenance personnel to perform system maintenance.

[0071] In an optional embodiment of the present invention, the inability of the steam trap to close may include the steam trap being normally open or the steam trap closing failure; the inability of the steam trap to open may include the steam trap being normally closed or the steam trap opening failure.

[0072] In a specific example, the pressure data collected by the pressure sensors before and after the steam trap and the layout distance of the pressure sensors before and after the steam trap are collected, and then one or more temperature and mass flow data before and after the steam trap are collected, and the layout distance of the corresponding sensors is obtained at the same time. The original collected data is synchronized with the sensor synchronization controller to complete the data time. Set the pressure sensor data collection frequency to F P , construct the cross-correlation function of the pressure data before and after the steam trap, based on N=τ / F collected at the selected time interval τ P Data points are discretely solved to obtain the steam trap pressure wave transmission lag time T. For example, at a certain time point t, the data collected by the pressure sensors before and after the steam trap are P1|t and P2|t, respectively. The raw data collected by the pressure sensors at the selected time interval τ corresponds to {(P1|t, P2|t+T),…,(P1|t+τ-T, P2|t+τ)}. Processing this data set yields the difference data set of the pressure sensors before and after the valve within the time period (t, t+τ): {P1|t-P2|t+T,…,P1|t+τ-T-P2|t+τ}. Wavelet transform is used to analyze the difference data set of the pressure sensors, obtaining the detailed coefficient set and the approximate coefficient set of the data set. Frequency domain features such as key characteristic frequencies and pressure propagation states are extracted from these detailed coefficient sets and the approximate coefficient set. The same processing method is used to obtain the difference data sets and their frequency domain features for other sensors.

[0073] Alternatively, the steam trap pressure wave transmission lag time T can be calculated according to the following formula: n represents the ordinal number of the collected data point, f(t) and g(t+τ) are the cross-correlation functions of the pressure data, and sqrt is the solution of the square root of a non-negative number.

[0074] Specifically, assuming the pressure sensor has a data sampling frequency of 10 Hz, and the pressure data before and after the steam trap are collected within 1 minute, and the data processing interval is 1 second, then a total of 60 difference data sets are collected within this 1 minute, each data set including 9 data points.

[0075] The long short-term memory network is used to analyze and match the frequency domain characteristics of the difference data set to obtain the preliminary judgment of the steam trap action state. When the preliminary judgment of the steam trap action state meets the preset conditions (that is, the preliminary judgment of the steam trap action state is an abnormal action state), each difference data set is coupled separately and input into the Kalman filter algorithm for data fusion to obtain the data fusion result. The data fusion data is then input into the preset convolutional neural network for steam trap action state analysis to predict the steam trap action state.

[0076] Initially, a test bench was constructed to collect data on the changing characteristics of parameters such as pressure, temperature, mass flow rate, and flow rate when steam with different parameters (temperature, pressure, and dryness) flowed through the steam trap. Sample data was then constructed based on this data for training the long-term short-term memory network. The specific process involved normalizing the data to extract features such as kurtosis, mean, root mean square value, standard deviation, and crest factor. The five operating states of the steam trap were ranked according to valve opening (trap open failure, trap normal closed, trap oscillating opening and closing, trap normal open, and trap closed failure), and quantized and assigned values of 1, 2, 3, 4, and 5, respectively. The long-term short-term memory network training was used to output a predicted score for the steam trap's operating state. If the quantitative assignment of the initial trap operating state was not within the range [1.5-2.5, 3.5-4.5], subsequent data fusion and re-prediction of the steam trap's operation were performed. The data fusion process involves inputting data such as pressure difference, temperature difference, mass flow difference, and flow velocity difference into a pre-set Kalman filter to construct an initial state vector, obtaining an initial state vector set. This initial state vector set is then converted into a state transition matrix. This state transition matrix is then used to perform cross-fusion on the sensor data, resulting in multiple sets of fused data. After the fused data is fed into a convolutional neural network, the convolutional layer extracts convolutional features, obtaining the corresponding fused features. Based on these features, the convolutional neural network output layer outputs the fused steam trap actuation state prediction result.

[0077] After obtaining the trap operation state prediction results, the improved evidence theory is used to complete the decision-making layer fusion and output the final trap operation state prediction. Among them, the improved evidence theory is a method for evaluating and quantifying the value and credibility of evidence. Fusion of the decision-making layer based on this method can weaken the impact of data conflicts and improve the reliability of the prediction results.

[0078] Assume that the target recognition framework U for the steam trap action state includes {A (trap open failure), B (trap normal closing), C (trap oscillation opening and closing), D (trap normal opening), E (trap closing failure)}, and the supporting evidence m is {pressure-temperature data (m1), pressure-mass flow data (m2), pressure-flow rate data (m3)...}. Calculating the support of different pieces of evidence for the steam trap action state can achieve decision optimization. The improved evidence theory fusion rule is:

[0079]

[0080] Specifically, K is the normalization coefficient, k is the conflict factor between pieces of evidence, and the value of k indicates the degree of conflict between pieces of evidence. Q is the weighted average of the combined evidence, which is used to adjust for imbalances caused by intense evidence collisions, thereby reducing the significant impact of severe conflict on decision-making outcomes. Q = R1*m1+R2*m2+R3*m3, where R1, R2, and R3 are weighting coefficients, determined by recursive calculation and expert scoring.

[0081] Based on the controller recording and outputting the steam trap action status signal, when the duration of the abnormal working state of the steam trap exceeds the set value, the fault alarm signal matching the abnormal state of the steam trap is output synchronously. This solution solves the problems of low accuracy of single signal judgment, poor reliability of results, and difficulty in migrating and transforming judgment parameters. Figure 3 The scheme shown uses multi-layer data fusion and decision result fusion to accurately determine the abnormal state of the steam trap with finer granularity, and can realize the online display of the operating status of the conventional float-type steam trap, thereby improving the degree of digitization of the steam pipeline network in the silk-making workshop, solving the problems faced by simple data fusion such as high uncertainty and easy overfitting of the model, realizing high-precision prediction of the operating state of the steam trap, achieving good overlap with the pipeline monitoring system, and being able to directly call a large number of pipeline monitoring point data, thereby improving the scalability of the scheme. Figure 3 In the example, Difference Dataset 1 is a difference dataset of data collected by the pressure sensor, Difference Dataset 2 is a difference dataset of data collected by the temperature sensor, and Difference Dataset 3 is a difference dataset of data collected by the mass flow sensor. Prediction Result 1 is the predicted steam trap operating state by fusing the convolutional neural network based on Difference Dataset 1 and Difference Dataset 2. Prediction Result 2 is the predicted steam trap operating state by fusing the convolutional neural network based on Difference Dataset 2 and Difference Dataset 3. Prediction Result 3 is the predicted steam trap operating state by fusing the convolutional neural network based on Difference Dataset 1 and Difference Dataset 3. The final prediction of the steam trap operating state is obtained by fusing the decision results of Prediction Result 1, Prediction Result 2, and Prediction Result 3 using the improved evidence theory.

[0082] The technical solution of the embodiment of the present invention obtains the original collected data of the associated sensors for monitoring the operation of the steam trap, thereby determining the data collection frequency of the pressure sensor, and then determining the steam trap pressure wave transmission lag time based on the layout distance of adjacent target pressure sensors, the data collection frequency of the pressure sensor, and the pressure data cross-correlation function. Based on the original collected data and the steam trap pressure wave transmission lag time, each difference data set is determined, and the frequency domain features of each difference data set are further extracted according to the wavelet transform method. The frequency domain features are analyzed based on the long short-term memory network to obtain a preliminary judgment of the operation state of the steam trap. When the operation state of the steam trap is preliminarily judged to be an abnormal operation state, the data fusion results of each difference data set are determined, and decision fusion is performed based on the convolutional neural network, the data fusion results, and the evidence theory to predict the operation state of the steam trap. In this solution, the working status of the steam trap is preliminarily determined by collecting data from non-single sensors and using a long-short-term memory network. When the working status of the steam trap is abnormal, the model determination source data and the decision results are fused to accurately determine the abnormal status of the steam trap at a finer granularity. This solves the problem of large errors in the existing steam trap working condition determination and the inability of the determination results to meet the needs. It can accurately determine the working status of the steam trap and provide guarantees for the digital transformation of steam pipeline steam trap maintenance and repair work.

[0083] Example 3

[0084] Figure 4 This is a schematic diagram of the structure of a steam trap operation monitoring device provided by the third embodiment of the present invention. Figure 4 As shown, the device includes:

[0085] The data acquisition module 310 is used to acquire the original data collected by the steam trap operation monitoring associated sensors; wherein the steam trap operation monitoring associated sensors include pressure sensors and non-pressure associated sensors;

[0086] The difference data set determination module 320 is used to determine the steam trap pressure wave transmission lag time and determine each difference data set based on the original collected data and the steam trap pressure wave transmission lag time;

[0087] The first action determination module 330 is used to extract the frequency domain features of each difference data set using a wavelet transform method, and analyze the frequency domain features based on a long short-term memory network to obtain a preliminary determination of the steam trap action state;

[0088] The second action determination module 340 is used to determine the data fusion results of each difference data set when the steam trap is initially determined to be in an abnormal action state, and perform decision fusion based on the convolutional neural network, the data fusion results, and evidence theory to predict the action state of the steam trap.

[0089] The technical solution of the embodiment of the present invention obtains the original collected data of the steam trap action monitoring associated sensor to determine the steam trap pressure wave transmission lag time, and determines each difference data set based on the original collected data and the steam trap pressure wave transmission lag time, and then extracts the frequency domain characteristics of each difference data set according to the wavelet transform method, and analyzes the frequency domain characteristics based on the long short-term memory network to obtain the preliminary judgment of the steam trap action state. Further, when the steam trap action state is preliminarily judged to be an abnormal action state, the data fusion results of each difference data set are determined, and decision fusion is performed based on the convolutional neural network, the data fusion results and the evidence theory to predict the steam trap action state. In this solution, the working status of the steam trap is preliminarily determined by collecting data from non-single sensors and using a long-short-term memory network. When the working status of the steam trap is abnormal, the model determination source data and the decision results are fused to accurately determine the abnormal status of the steam trap at a finer granularity. This solves the problem of large errors in the existing steam trap working condition determination and the inability of the determination results to meet the needs. It can accurately determine the working status of the steam trap and provide guarantees for the digital transformation of steam pipeline steam trap maintenance and repair work.

[0090] Optionally, the difference data set determination module 320 includes a lag time determination unit for determining the data acquisition frequency of the pressure sensor; and calculating the steam trap pressure wave transmission lag time based on the layout distance between adjacent target pressure sensors, the data acquisition frequency of the pressure sensor, and the pressure data cross-correlation function.

[0091] Optionally, the difference data set determination module 320 includes a difference data set determination unit, which is used to determine the difference data set of the pressure sensor based on the original collected data of the pressure sensor and the pressure wave transmission lag time of the steam trap; and determine the difference data set of the non-pressure-related sensor based on the original collected data of the non-pressure-related sensor and the pressure wave transmission lag time of the steam trap; wherein the non-pressure-related sensor includes a temperature sensor and / or a mass flow sensor.

[0092] Optionally, the steam trap action monitoring device also includes a long short-term memory network training module, which is used to obtain historical collection data of sensors associated with steam trap action monitoring; normalize the historical collection data to obtain feature data to be extracted, and train the long short-term memory network through frequency domain analysis of key features in the feature data to be extracted and the steam trap action labeling type.

[0093] Optionally, the second action determination module 340 further includes a data fusion unit for obtaining a normal action state prediction score interval of the long short-term memory network; if the quantitative assignment of the preliminary action state of the steam trap is not within the normal action state prediction score interval, the preliminary action state of the steam trap is determined to be an abnormal action state, and the difference data sets are cross-fused based on the Kalman filter to obtain the data fusion result.

[0094] Optionally, the steam trap operation monitoring device also includes an early warning module, which is used to generate a first type of fault alarm signal when the predicted steam trap operation state is that the steam trap cannot be closed and the duration is greater than a first duration; or, when the predicted steam trap operation state is that the steam trap cannot be opened and the duration is greater than the first duration, generate a second type of fault alarm signal; or, when the predicted steam trap operation state is that the steam trap oscillates and opens and closes and the duration is greater than the second duration, generate a third type of fault alarm signal.

[0095] Optionally, the inability of the steam trap to close includes the steam trap opening normally or the steam trap closing failure; the inability of the steam trap to open includes the steam trap closing normally or the steam trap opening failure.

[0096] The steam trap operation monitoring device provided in the embodiment of the present invention can execute the steam trap operation monitoring method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0097] Example 4

[0098] Figure 5 A schematic diagram of an electronic device that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The components shown herein, their connections and relationships, and their functions are provided for example only and are not intended to limit the implementation of the present inventions described and / or claimed herein.

[0099] like Figure 5As shown, electronic device 10 includes at least one processor 11 and memory, such as ROM 12 and RAM 13, communicatively connected to at least one processor 11. The memory stores computer programs executable by the at least one processor, and processor 11 can perform various appropriate actions and processes based on the computer programs stored in ROM 12 or loaded from storage unit 18 into RAM 13. RAM 13 can also store various programs and data required for programming controller 10 operations. Processor 11, ROM 12, and RAM 13 are interconnected via bus 14. An I / O interface 15 is also connected to bus 14. ROM 12 is a read-only memory, RAM 13 is a random access memory, and I / O interface 15 is an input / output interface.

[0100] Multiple components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0101] The processor 11 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various processors that run machine learning model algorithms, a digital signal processor (DSP), and any other suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the steam trap operation monitoring method.

[0102] In some embodiments, the steam trap operation monitoring method can be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the steam trap operation monitoring method described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to execute the steam trap operation monitoring method in any other suitable manner (e.g., via firmware).

[0103] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0104] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0105] In the context of the present invention, computer-readable storage medium can be a tangible medium that can contain or store a computer program for use with an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. Computer-readable storage medium can include but is not limited to electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, computer-readable storage medium can be a machine-readable signal medium. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device or any suitable combination of the foregoing.

[0106] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0107] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0108] A computing system may include clients and servers. The clients and servers are generally remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosts and VPS servers.

[0109] The present application also discloses a computer program product comprising a computer program that, when executed by a processor, implements the steam trap operation monitoring method provided in any of the embodiments of the present application. This program product shares the same inventive concept as the steam trap operation monitoring method disclosed in each embodiment of the present application and is therefore not further described here.

[0110] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved. This is not limited herein.

[0111] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A method for monitoring the operation of a steam trap, characterized in that: include: Acquire original data collected by steam trap operation monitoring associated sensors; wherein the steam trap operation monitoring associated sensors include pressure sensors and non-pressure associated sensors; Determining a steam trap pressure wave transmission lag time, and determining each difference data set based on the original collected data and the steam trap pressure wave transmission lag time; Extracting frequency domain features of each of the difference data sets using a wavelet transform method, and analyzing the frequency domain features based on a long short-term memory network to obtain a preliminary determination of the steam trap's operating state; When the steam trap is initially determined to be in an abnormal operating state, data fusion results of the difference data sets are determined, and decision fusion is performed based on a convolutional neural network, the data fusion results, and evidence theory to predict the operating state of the steam trap.

2. The method according to claim 1, characterized in that Determine the steam trap pressure wave transmission delay time, including: determining a data acquisition frequency of the pressure sensor; The steam trap pressure wave transmission lag time is calculated based on the layout distance between adjacent target pressure sensors, the data acquisition frequency of the pressure sensors, and the pressure data cross-correlation function.

3. The method according to claim 1, characterized in that Determining each difference data set based on the original collected data and the steam trap pressure wave transmission lag time includes: Determining a difference data set of the pressure sensor according to the original collected data of the pressure sensor and the pressure wave transmission lag time of the steam trap; A difference data set of the non-pressure-related sensor is determined based on the original collected data of the non-pressure-related sensor and the steam trap pressure wave transmission lag time; wherein the non-pressure-related sensor includes a temperature sensor and / or a mass flow sensor.

4. The method according to claim 1, wherein Before analyzing the frequency domain characteristics based on the long short-term memory network to obtain a preliminary judgment of the steam trap operation state, the method further includes: Obtain historical data collected from sensors associated with steam trap operation monitoring; The historical collected data is normalized to obtain feature data to be extracted, and the long short-term memory network is trained by analyzing key features in the frequency domain of the feature data and the steam trap action labeling type.

5. The method according to claim 1, characterized in that When the steam trap is initially determined to be in an abnormal operating state, determining the data fusion results of the difference data sets includes: Obtaining a normal action state prediction score interval of the long short-term memory network; If the quantitative assignment of the preliminary determined action state of the steam trap is not within the normal action state prediction score interval, the preliminary determined action state of the steam trap is determined to be an abnormal action state, and the difference data sets are cross-fused based on the Kalman filter to obtain the data fusion result.

6. The method according to claim 1, characterized in that After predicting the steam trap operation status, it also includes: When the predicted trap valve operation state is that the trap valve cannot be closed and the duration is longer than a first duration, a first type fault alarm signal is generated; or When the predicted trap valve operation state is that the trap valve cannot be opened and the duration is longer than the first duration, a second type fault alarm signal is generated; or, When the predicted steam trap operation state is that the steam trap opens and closes in an oscillating manner, and the duration thereof is greater than the second time period, a third type fault alarm signal is generated.

7. The method according to claim 6, characterized in that The steam trap valve cannot be closed, which includes the steam trap valve being normally open or the steam trap valve closing failure; The steam trap cannot be opened includes the steam trap being normally closed or the steam trap opening failure.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor. The computer program is executed by the at least one processor so that the at least one processor can perform the steam trap operation monitoring method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the steam trap operation monitoring method according to any one of claims 1 to 7 when executed.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the computer program implements the steam trap operation monitoring method according to any one of claims 1 to 7.

Images