Oil delivery pump state evaluation method, device and equipment and storage medium
By combining the ensemble mode decomposition method and exponential weighted moving average method, the state evaluation model of the global average pooled convolutional neural network is trained, which solves the problem of inaccurate state evaluation of oil pumps in the prior art, and achieves more accurate and real-time state evaluation.
Patent Information
- Application Number
- CN202510288815.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-12
- Publication Date
- 2025-06-27
AI Technical Summary
The existing oil pump status evaluation methods cannot capture potential failure signs in a timely and accurate manner, and the prediction of the deep network model is inaccurate, which affects the accuracy of oil pump status evaluation.
The vibration data of the oil pump is extracted by the ensemble mode decomposition method, and processed with the exponential weighted moving average method to obtain the weighted target features, and state evaluation is performed through the global average pooled convolutional neural network training state evaluation model.
By extracting accurate sample data, a complete state evaluation model is trained to achieve accurate evaluation of the oil pump status, improving the accuracy and real-time evaluation.
Smart Images
Figure CN120217095A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of oil and gas pipeline safety, and particularly relates to a method, device, equipment and storage medium for evaluating the state of an oil transfer pump. Background Art
[0002] An oil transfer pump is a crucial device in industrial production, mainly used to transport liquid media, such as products like crude oil and petroleum, from one location to another to ensure the normal operation of industrial production. Failures of the oil transfer pump may lead to production interruptions, equipment damage, and even safety accidents. Therefore, accurate evaluation and monitoring of its state are of great importance.
[0003] Existing methods for evaluating the state of an oil transfer pump usually rely on simple parameter monitoring and regular inspections. Such methods often fail to promptly and accurately detect potential fault signs. Although there are also methods using deep network models for evaluation, due to the complexity of the vibration data of the oil transfer pump, simply using manual methods to divide the state cannot obtain accurate sample data, resulting in inaccurate predictions of the trained model and affecting the accuracy of the oil transfer pump state evaluation. Summary of the Invention
[0004] The present invention provides a method for evaluating the state of an oil transfer pump to accurately evaluate the state of the oil transfer pump.
[0005] According to a first aspect of the present invention, there is provided a method for evaluating the state of an oil transfer pump, including: extracting features from the vibration data during the operation of the oil transfer pump by using the ensemble empirical mode decomposition method to obtain a feature set, where the feature set includes vibration components in different frequency dimensions;
[0006] Processing the feature set by using the exponential weighted moving average method to obtain a weighted target feature, and dividing the state of the oil transfer pump based on the weighted target feature;
[0007] Marking the state of the oil transfer pump for the vibration data according to the weighted target feature of the divided state;
[0008] Inputting the marked vibration data into a global average pooling convolutional neural network for training to obtain a state evaluation model, and evaluating the state of the currently measured vibration data of the oil transfer pump through the state evaluation model.
[0009] Optionally, extracting features from the vibration data during the operation of the oil transfer pump by using the ensemble empirical mode decomposition method to obtain a feature set includes:
[0010] Obtaining a specified number of white noise sequences, where the intensities of the white noise sequences are different;
[0011] Each time, a white noise sequence is added to the vibration data to obtain new vibration data, and the new vibration data is decomposed by the ensemble empirical mode decomposition method to obtain independent vibration components with a single frequency;
[0012] The vibration components with the same frequency obtained each time are averaged to construct the feature set.
[0013] Optionally, the exponential weighted moving average method is used to process the feature set to obtain weighted target features, including:
[0014] The target features with the most obvious waveform trend change are selected from the feature set;
[0015] The target features are processed by the exponential weighted moving average method to obtain the weighted target features.
[0016] Optionally, the exponential weighted moving average method is used to process the target features to obtain the weighted target features, including:
[0017] The preset window size and weight are obtained, and the window sliding times are determined according to the sample size of the vibration data and the window size;
[0018] Each time the window slides, the exponential weighted moving average method is used to calculate the target features within the window according to the weight to obtain weighted vibration components;
[0019] The weighted vibration components obtained each time the window moves are combined to obtain the weighted target features.
[0020] Optionally, the state of the oil transfer pump is divided according to the weighted target features, including:
[0021] The inflection points are identified from the weighted target features and displayed;
[0022] The division instruction fed back by the user based on the inflection points is received, and the state of the weighted target features is determined according to the division instruction, where the state includes normal, degradation, and rapid degradation.
[0023] Optionally, the vibration data is marked with the state of the oil transfer pump according to the weighted target features with the divided state, including:
[0024] The sample number matching the weighted target features with the divided state is determined, and the vibration data corresponding to the weighted target features is determined according to the sample number;
[0025] The vibration data corresponding to the weighted target features is marked with the state of the oil transfer pump according to the state of the weighted target features.
[0026] According to another aspect of the present invention, there is provided a state evaluation device for an oil transfer pump, including: a feature set acquisition module, configured to extract features from vibration data during the operation of the oil transfer pump by using the ensemble empirical mode decomposition method to obtain a feature set, wherein the feature set includes vibration components in different frequency dimensions;
[0027] An oil transfer pump state division module, configured to process the feature set by using the exponential weighted moving average method to obtain a weighted target feature, and divide the oil transfer pump state based on the weighted target feature;
[0028] An oil transfer pump state marking module, configured to mark the oil transfer pump state of the vibration data according to the weighted target feature of the divided state;
[0029] A state evaluation module, configured to input the marked vibration data into a global average pooling convolutional neural network for training to obtain a state evaluation model, and evaluate the state of the current vibration data to be measured of the oil transfer pump through the state evaluation model.
[0030] According to another aspect of the present invention, there is provided an electronic device, the electronic device includes:
[0031] At least one processor; and
[0032] A memory communicatively connected to the at least one processor; wherein,
[0033] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method according to any embodiment of the present invention.
[0034] According to another aspect of the present invention, there is provided a computer-readable storage medium, the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method according to any embodiment of the present invention when executed.
[0035] The technical solution of the embodiment of the present invention combines the ensemble empirical mode decomposition method and the exponential weighted moving average method to divide the state of the vibration data of the oil transfer pump, so as to obtain accurate sample data to train the global average pooling convolutional neural network to obtain a state evaluation model with perfect functions, and thus accurately evaluate the state of the oil transfer pump based on the trained state evaluation model.
[0036] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. Description of the Drawings
[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.
[0038] Figure 1 is a flowchart of a method for evaluating the state of an oil transfer pump provided in Embodiment 1 of the present invention;
[0039] Figure 2 is a waveform diagram of vibration data provided in Embodiment 1 of the present invention;
[0040] Figure 3 is a waveform diagram of a feature set provided in Embodiment 1 of the present invention;
[0041] Figure 4 is a waveform diagram of weighted target features provided in Embodiment 1 of the present invention;
[0042] Figure 5 is a flowchart of a method for evaluating the state of an oil transfer pump provided in Embodiment 2 of the present invention;
[0043] Figure 6 is a schematic structural diagram of an oil transfer pump state evaluation device provided in Embodiment 3 of the present invention;
[0044] Figure 7 is a schematic structural diagram of an electronic device provided in Embodiment 4 of the present invention. Detailed implementation manners
[0045] In order to enable those skilled in the art to better understand the solutions of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, rather than all of them. 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.
[0046] It should be noted that the terms "first", "second", etc. in the description, claims and above-mentioned drawings of the present invention are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, 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 inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.
[0047] Embodiment 1
[0048] Figure 1 FIG. is a flowchart of a method for evaluating the state of an oil transfer pump provided in Embodiment 1 of the present invention. This embodiment is applicable to the situation of evaluating the state of an oil transfer pump. This method can be executed by an oil transfer pump state evaluation device, and this device can be implemented in the form of hardware and / or software. As Figure 1 shown, the method includes:
[0049] Step S101, extracting features from the vibration data during the operation of the oil transfer pump by using the ensemble empirical mode decomposition method to obtain a feature set.
[0050] Optionally, extracting features from the vibration data during the operation of the oil transfer pump by using the ensemble empirical mode decomposition method to obtain a feature set includes: obtaining a specified number of white noise sequences, where the intensities of the white noise sequences are different; adding one white noise sequence to the vibration data each time to obtain new vibration data, and using the ensemble empirical mode decomposition method to decompose the new vibration data to obtain independent vibration components with a single frequency; averaging the vibration components with the same frequency obtained each time to construct a feature set.
[0051] Specifically, in this embodiment, the full life cycle data of the oil transfer pump will be collected to obtain the original sample data. As Figure 2 shown is the waveform diagram of the vibration data. Due to space limitations Figure 2Only a partial display is made here. Here, the horizontal axis is the sample number, and the vertical axis is the vibration value, with the unit of m^2 / s. From the waveform diagram of the vibration data, it can be seen that it is completely impossible to distinguish the state of the oil transfer pump simply through the waveform diagram. The vibration data of the oil transfer pump usually changes with time. The features extracted from the time domain and frequency domain usually describe its state from the perspective of the overall component, and it is impossible to observe the information of non-stationary vibration data changing with time, nor can it analyze the local information of the vibration data. Since the time domain features can more obviously show the degradation trend of the vibration data of the oil transfer pump, in this embodiment, the Ensemble Empirical Mode Decomposition (EEMD) method is used to extract features from the vibration data during the operation of the oil transfer pump to obtain a feature set.
[0052] In a specific implementation, when using EEMD for feature extraction in this embodiment, specifically, a specified number of white noise sequences are obtained, and the intensities of each white noise sequence are different. For example, three white noise sequences with intensities of 10, 15, and 20 are obtained; first, a white noise sequence with an intensity of 10 is added to the vibration data to obtain new vibration data, and then the new vibration data is decomposed using the EEMD algorithm to obtain independent vibration components with a single frequency. Six frequencies are obtained according to the actual situation of the vibration data. For example, the vibration components at frequencies a, b, c, d, e, and f; then different intensities of white noise sequences are added again, and the above operations are continuously repeated. Thus, when adding a white noise sequence with an intensity of 15, the vibration components at the six frequencies decomposed are obtained, and when adding a white noise sequence with an intensity of 20, the vibration components at the six frequencies decomposed are obtained. Then, the vibration components at the same frequency obtained each time are averaged to construct a feature set. For example, for frequency a, the vibration components corresponding to frequency a in the three decomposition processes are averaged to obtain the vibration components corresponding to this frequency, and the vibration components at frequency a are used as a feature. Similarly, features at other frequencies can be obtained, and the features at all the above frequencies are combined to construct a feature set. Of course, only an example is given in this embodiment, and the dimension of the feature set is not limited, which is mainly determined by the specific situation of the vibration data of the oil transfer pump. Therefore, the feature set in this embodiment includes vibration components in different frequency dimensions.
[0053] Among them, as Figure 3 shown is the waveform diagram of the feature set. In this embodiment, by using the EEMD algorithm to obtain the features of the vibration data of the oil transfer pump, the problem that the trend of the waveform diagram of the original vibration data cannot be observed can be transformed into observable feature data and visualized. As Figure 3As shown, the horizontal axis represents the sample number, and the vertical axis represents the vibration value, with the unit of m^2 / s. By Figure 3 It can be seen that after obtaining the characteristics of the vibration data throughout the life cycle of the oil transfer pump, the vibration data begins to show a trend change from the original chaos, which facilitates the subsequent state division of the vibration data throughout the life cycle of the oil transfer pump.
[0054] Step S102: Process the feature set using the exponentially weighted moving average method to obtain the weighted target feature, and perform the state division of the oil transfer pump on the weighted target feature.
[0055] Optionally, processing the feature set using the exponentially weighted moving average method to obtain the weighted target feature includes: screening out the target feature with the most obvious waveform trend change from the feature set; processing the target feature using the exponentially weighted moving average method to obtain the weighted target feature.
[0056] Optionally, processing the target feature using the exponentially weighted moving average method to obtain the weighted target feature includes: obtaining the pre-set window size and weight, determining the window sliding times according to the sample size of the vibration data and the window size; calculating the weighted vibration component of the target feature within the window according to the weight using the exponentially weighted moving average method during each window sliding; combining the weighted vibration components obtained from each window movement to obtain the weighted target feature.
[0057] Specifically, in this embodiment, after obtaining the feature set, the target feature with the most obvious waveform trend change will be screened out from it, such as Figure 3As shown, the trend change of the feature at frequency f is the most obvious. Therefore, the feature at this frequency is used as the target feature. Although the target feature has a relatively obvious trend compared to the remaining features in the feature set, it is still somewhat difficult to divide its state due to the large number of samples. In order to make the trend change more obvious, in this embodiment, the target feature is processed by the Exponentially Weighted Moving-Average (EWMA) method to obtain the weighted target feature. In this embodiment, EWMA is used to divide the state of the target feature. Specifically, the average value of the target feature is obtained by using the EWMA method, and the obtained result is used as the element within the current window, thereby converting the vibration data whose state interval was originally difficult to determine into data that is easy to distinguish states. Among them, when using EWMA for processing, specifically, the preset window size and weight are obtained. For example, the window size is 10. Since the number of samples included in the target feature is 1000, the window sliding times can be determined to be 100. And each time the window slides, the weighted vibration component is calculated for the target feature within the window by the exponentially weighted moving average method according to the weight. For example, the average value of the target feature within the window is calculated every 10 data. And in this embodiment, different weights are corresponding to the vibration components within the window, so as to calculate the average value of the vibration components within the fixed window by moving according to the set weight. Among them, EWMA is mainly calculated by formula (1):
[0058] V t =βV t-1 +(1 - β)θ t (1)
[0059] Among them, t is the current moment, t - 1 is the previous moment, β is the weighting coefficient, V t is the vibration component at the current moment, V t-1 is the vibration component at the previous moment. Therefore, in this embodiment, different weights are set for the vibration components by EWMA, and the average value of the data within the fixed window is calculated by moving according to the set weight. Therefore, adjacent data can be better obtained. The weights show a state of increasing from small to large according to the distance of the data from the current one. There is no need to save historical values, and the calculation amount is relatively small. In addition, in this embodiment, after obtaining the calculation results of 100 average values, the weighted vibration components obtained by each window movement are combined to obtain the weighted target feature. As Figure 4 shown is the waveform diagram of the weighted target feature. According to Figure 4It can be seen that the upper part is the target feature at frequency f, and the lower part is the weighted target feature calculated by the EWMA method for it. After the EWMA calculation, an obvious trend appears in the data. By observing, it can be seen that between 0 and 48 samples, although the data fluctuates as a whole, there is no upward trend. After 48 samples, the vibration value of the data gradually begins to increase. When the vibration data increases to a certain extent, attenuation occurs, and degradation occurs continuously in this trend until final failure. In this embodiment, inflection points will be obtained based on the waveform diagram with an obvious trend, that is, the 48th sample is the first inflection point M, and the 95th sample is the second inflection point N, and the inflection points will be displayed. The weighted target feature is marked with a status according to the user's instruction. The marked statuses include normal, degradation, and rapid degradation, that is, 0-48 is normal, 48-95 is degradation, and above 95 is rapid degradation.
[0060] Step S103, perform a status mark on the vibration data of the oil transfer pump according to the weighted target feature with the divided status.
[0061] Optionally, performing a status mark on the vibration data of the oil transfer pump according to the weighted target feature with the divided status includes: determining the sample number matched by the weighted target feature with the divided status, and determining the vibration data corresponding to the weighted target feature according to the sample number; performing a status mark on the corresponding vibration data according to the status of the weighted target feature.
[0062] Specifically, in this embodiment, it is mainly necessary to obtain the status of the vibration data in the whole life cycle of the oil transfer pump. What is adopted is to select the status of the weighted target feature with an obvious trend change, and then determine the status of the vibration data based on the status of the weighted target feature. Specifically, it is to determine the sample number matched by the weighted target feature. For example, the status of the target feature corresponding to the sample numbers 0-48 is normal, multiplied by the window size of 10, so as to obtain that the status of the vibration data corresponding to 0-480 is normal; the status of the target feature corresponding to the sample numbers 48-95 is degradation, multiplied by the window size of 10, so as to obtain that the status of the vibration data corresponding to 480-950 is degradation; the status of the target feature corresponding to the sample numbers above 95 is rapid degradation, multiplied by the window size of 10, so as to obtain that the status of the vibration data corresponding to above 950 is normal. Therefore, in this embodiment, through the status division result of the feature with an obvious trend change, the status of the original data of the oil transfer pump with an unobvious trend change is indirectly marked, thereby improving the accuracy of the status mark of the oil transfer vibration data and the accuracy of the samples.
[0063] Step S104, input the marked vibration data into a global average pooling convolutional neural network for training to obtain a status evaluation model, and perform a status evaluation on the currently measured vibration data of the oil transfer pump through the status evaluation model.
[0064] Specifically, in this embodiment, after obtaining the marked vibration data, the marked vibration data can be used as a sample and input into the global average pooling convolutional neural network for training to obtain a state evaluation model. Therefore, when obtaining the current vibration data to be measured of the oil transfer pump, the state of the current vibration data to be measured can be accurately evaluated based on the state evaluation model. Among them, the reason for the accurate state evaluation result is mainly due to the following two aspects: First, due to the accuracy of the sample data, the parameters of the state evaluation model obtained by training with the sample data are more accurate; Second, the traditional convolutional neural network structure usually includes a convolutional layer for feature extraction and a fully connected layer for output. Therefore, when the input dimension of the data changes, the network structure needs to be dynamically adjusted. However, in the global average pooling convolutional neural network of this application, the fully connected layer is replaced by an average pooling layer. For the features of each channel of the output, the average value of the data it contains is calculated, and then classification is performed through the softmax function. Therefore, the model structure of this application has changed, the network parameters have decreased, and it can accept inputs of any dimension. At the same time, due to the reduction of parameters, overfitting can be avoided to a certain extent.
[0065] In the embodiment of the present application, the ensemble empirical mode decomposition method and the exponentially weighted moving average method are combined to classify the vibration data of the oil transfer pump, so as to obtain accurate sample data to train the global average pooling convolutional neural network to obtain a state evaluation model with perfect functions, and then accurately evaluate the state of the oil transfer pump based on the trained state evaluation model.
[0066] Embodiment 2
[0067] Figure 5 The flowchart of a method for evaluating the state of an oil transfer pump provided by the second embodiment of the present invention. This embodiment is based on the above embodiment. After evaluating the state of the current vibration data to be measured of the oil transfer pump through the state evaluation model, it further includes: obtaining new sample data according to the evaluation result of the vibration data to be measured; optimizing the state evaluation model with the new sample data. As Figure 5 shown, the method includes:
[0068] Step S201, using the ensemble empirical mode decomposition method to extract features from the vibration data during the operation of the oil transfer pump to obtain a feature set.
[0069] Optionally, the ensemble empirical mode decomposition method is used to extract features from the vibration data during the operation of the oil transfer pump to obtain a feature set, including: obtaining a specified number of white noise sequences, where the intensities of the white noise sequences are different; adding one white noise sequence to the vibration data each time to obtain new vibration data, and using the ensemble empirical mode decomposition method to decompose the new vibration data to obtain independent vibration components with single frequencies; averaging the vibration components with the same frequency obtained each time to construct a feature set.
[0070] Step S202: Process the feature set using the exponential weighted moving average method to obtain weighted target features, and classify the states of the oil transfer pump based on the weighted target features.
[0071] Optionally, the exponential weighted moving average method is used to process the feature set to obtain weighted target features, including: screening out the target features with the most obvious waveform trend changes from the feature set; processing the target features using the exponential weighted moving average method to obtain weighted target features.
[0072] Optionally, processing the target features using the exponential weighted moving average method to obtain weighted target features includes: obtaining a pre-set window size and weight, determining the number of window slides according to the sample size of the vibration data and the window size; calculating the weighted vibration components of the target features within the window according to the weight using the exponential weighted moving average method each time the window slides; combining the weighted vibration components obtained each time the window moves to obtain weighted target features.
[0073] Step S203: Mark the states of the oil transfer pump for the vibration data based on the weighted target features with classified states.
[0074] Optionally, marking the states of the oil transfer pump for the vibration data based on the weighted target features with classified states includes: determining the sample numbers matching the weighted target features with classified states, and determining the vibration data corresponding to the weighted target features according to the sample numbers; marking the states of the oil transfer pump for the corresponding vibration data according to the states of the weighted target features.
[0075] Step S204: Input the marked vibration data into a global average pooling convolutional neural network for training to obtain a state evaluation model, and evaluate the state of the current vibration data to be measured of the oil transfer pump through the state evaluation model.
[0076] Step S205: Obtain new sample data according to the evaluation results of the vibration data to be measured, and optimize the state evaluation model with the new sample data.
[0077] Specifically, in this embodiment, after the vibration data to be measured is input into the state evaluation model to obtain the evaluation result, it is detected whether there is an abnormal situation in the evaluation result, such as garbled characters, or the evaluation result is significantly incorrect. When it is determined that the evaluation result is normal, the evaluation result is marked in the vibration data to be measured to obtain new sample data.
[0078] Among them, in this embodiment, after obtaining the new sample data, the obtained new sample data is used to train and optimize the state evaluation model, and the state evaluation model is optimized based on the samples obtained from the latest vibration data of the oil transfer pump, thereby ensuring the real-time performance and accuracy of the state evaluation model.
[0079] It should be noted that in this embodiment, the state evaluation model can also be periodically detected. For example, it is detected once every other day or once every other week, and the user can set the detection period according to the accuracy of the state evaluation. Among them, when the latest obtained vibration data of the oil transfer pump is input into the state evaluation model and the evaluation result is obtained during the detection, the operator will identify whether the evaluation result is accurate according to the actual situation, and determine the accuracy rate of the state evaluation model according to the identification results of multiple vibration data. When it is determined through the detection that the accuracy rate of the state evaluation model is lower than the preset threshold, an alarm prompt will be generated, so as to facilitate the user to repair or maintain the state evaluation model according to the alarm prompt, thereby ensuring the accuracy of the oil transfer pump state evaluation.
[0080] In the embodiment of the present application, the ensemble empirical mode decomposition method and the exponentially weighted moving average method are combined to divide the state of the vibration data of the oil transfer pump, so as to obtain accurate sample data to train the global average pooling convolutional neural network to obtain a state evaluation model with perfect functions, and then accurately evaluate the state of the oil transfer pump based on the trained state evaluation model.
[0081] Embodiment III
[0082] Figure 6 A schematic structural diagram of an oil transfer pump state evaluation device provided by Embodiment III of the present invention. As Figure 6 shown, the device includes: a feature set acquisition module 310, an oil transfer pump state division module 320, an oil transfer pump state marking module 330, and a state evaluation module 340.
[0083] Among them, the feature set acquisition module 310 is used to extract features from the vibration data during the operation of the oil transfer pump by using the ensemble empirical mode decomposition method to obtain a feature set, where the feature set includes vibration components in different frequency dimensions;
[0084] The fuel transfer pump status classification module 320 is used to process the feature set by the exponentially weighted moving average method to obtain weighted target features, and classify the status of the fuel transfer pump based on the weighted target features;
[0085] The fuel transfer pump status marking module 330 is used to mark the status of the fuel transfer pump for the vibration data according to the weighted target features of the classified status;
[0086] The status evaluation module 340 is used to input the marked vibration data into the global average pooling convolutional neural network for training to obtain a status evaluation model, and evaluate the status of the current vibration data to be measured of the fuel transfer pump through the status evaluation model.
[0087] Optionally, the feature set acquisition module is used to obtain a specified number of white noise sequences, where the intensities of the white noise sequences are different;
[0088] Each time a white noise sequence is added to the vibration data to obtain new vibration data, and the new vibration data is decomposed by the ensemble empirical mode decomposition method to obtain independent vibration components with a single frequency;
[0089] The vibration components with the same frequency obtained each time are averaged to construct a feature set.
[0090] Optionally, the fuel transfer pump status classification module includes a weighted target feature acquisition unit, which is used to screen out the target features with the most obvious waveform trend change from the feature set;
[0091] The target features are processed by the exponentially weighted moving average method to obtain weighted target features.
[0092] Optionally, the weighted target feature acquisition unit is used to obtain a pre-set window size and weight, and determine the number of window slides according to the sample size of the vibration data and the window size;
[0093] Each time the window slides, the weighted vibration components are calculated for the target features within the window by the exponentially weighted moving average method according to the weight;
[0094] The weighted vibration components obtained by each window movement are combined to obtain weighted target features.
[0095] Optionally, the fuel transfer pump status classification module includes a fuel transfer pump status classification unit, which is used to identify inflection points from the weighted target features and display the inflection points;
[0096] Receive the classification instruction fed back by the user based on the inflection point, and determine the status of the weighted target features according to the classification instruction, where the status includes normal, degradation, and rapid degradation.
[0097] Optionally, an oil transfer pump status marking module is configured to determine the sample number that matches the weighted target feature of the divided status, and determine the vibration data corresponding to the weighted target feature according to the sample number;
[0098] Mark the status of the oil transfer pump according to the vibration data corresponding to the status of the weighted target feature.
[0099] Optionally, the device further includes a status evaluation model optimization module, configured to obtain new sample data according to the evaluation result of the vibration data to be measured;
[0100] Optimize the status evaluation model with the new sample data.
[0101] The oil transfer pump status evaluation device provided by the embodiments of the present invention can execute the oil transfer pump status evaluation method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method.
[0102] Embodiment 4
[0103] Figure 7 The structural schematic diagram of the electronic device 10 that can be used to implement the 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 electronic device can also represent various forms of mobile devices, such as, personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are only examples and are not intended to limit the implementation of the present invention described and / or claimed herein.
[0104] As Figure 7 shown, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc. Among them, the memory stores a computer program executable by at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. The input / output (I / O) interface 15 is also connected to the bus 14.
[0105] 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 disk, an optical disc, 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.
[0106] The processor 11 can be various general-purpose and / or special-purpose processing components 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 dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the fuel pump status evaluation method.
[0107] In some embodiments, the fuel pump status evaluation method can be implemented as a computer program, which is tangibly contained 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 onto 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 fuel pump status evaluation method described above can be executed. Alternatively, in other embodiments, the processor 11 can be configured to execute the fuel pump status evaluation method by any other suitable means (e.g., by means of firmware).
[0108] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGA), application-specific integrated circuits (ASIC), application-specific standard products (ASSP), systems-on-chip (SOC), complex programmable logic devices (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: implemented in one or more computer programs, the one or more computer programs can be executed and / or interpreted on a programmable system including at least one programmable processor, the programmable processor can be a dedicated or general-purpose programmable processor, can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.
[0109] A computer program for implementing the method 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 apparatus, such that the computer programs, when executed by the processor, cause the functions / operations specified in the flowchart and / or block diagram to be implemented. The computer programs may be executed entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.
[0110] In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium may be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0111] In order to provide interaction with a user, the systems and techniques described herein may 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 a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).
[0112] The systems and techniques described herein can be implemented in a computing system that includes backend 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 frontend components (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected with each other by digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), blockchain network, and the Internet.
[0113] A computing system can include a client and a server. The client and the server are generally remote from each other and typically interact via a communication network. The client-server relationship is created by computer programs running on respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, and solves the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.
[0114] It should be understood that various forms of the processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in the present invention can be executed 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, and no limitation is imposed herein.
[0115] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.
Claims
1. A method for evaluating the state of an oil pump, characterized in that: include: The vibration data of the oil pump operation process is subjected to feature extraction by using the set mode decomposition method to obtain a feature set, wherein the feature set includes vibration components in different frequency dimensions; The feature set is processed by exponential weighted moving average method to obtain weighted target features, and the weighted target features are divided into oil pump states; Marking the vibration data as an oil pump state according to the weighted target feature of the divided state; The marked vibration data is input into a global average pooling convolutional neural network for training to obtain a state assessment model, and the state assessment model is used to perform state assessment on the vibration data to be tested currently on the oil pump.
2. The method according to claim 1, characterized in that The method of extracting features from the vibration data of the oil pump operation process using the set mode decomposition method to obtain a feature set includes: Obtaining a specified number of white noise sequences, wherein the strengths of the white noise sequences are different; Each time, a white noise sequence is added to the vibration data to obtain new vibration data, and the new vibration data is decomposed by using a collective mode decomposition method to obtain independent vibration components with a single frequency; The vibration components of the same frequency acquired each time are averaged to construct the feature set.
3. The method according to claim 1, characterized in that: The step of using the exponentially weighted moving average method to process the feature set to obtain the weighted target feature includes: Filter out the target feature with the most obvious waveform trend change from the feature set; The target feature is processed by the exponentially weighted moving average method to obtain the weighted target feature.
4. The method according to claim 3, characterized in that: The step of processing the target feature by using the exponentially weighted moving average method to obtain the weighted target feature includes: Obtaining a preset window size and weight, and determining the number of window sliding times according to the sample size of the vibration data and the window size; When the window slides each time, the target feature located in the window is calculated according to the weight by the exponential weighted moving average method to obtain the weighted vibration component; The weighted vibration components obtained by each window movement are combined to obtain the weighted target feature.
5. The method according to claim 3, characterized in that: The step of dividing the oil pump status according to the weighted target feature includes: Identify and obtain an inflection point from the weighted target feature, and display the inflection point; A division instruction fed back by a user based on the inflection point is received, and a state of the weighted target feature is determined according to the division instruction, wherein the state includes normal, degraded, and rapidly degraded.
6. The method according to claim 1, characterized in that The step of marking the oil pump status of the vibration data according to the weighted target feature of the divided status includes: Determine a sample number that matches the weighted target feature of the divided state, and determine vibration data corresponding to the weighted target feature according to the sample number; The corresponding vibration data is marked with an oil pump state according to the state of the weighted target feature.
7. The method according to claim 1, characterized in that After the state evaluation model is used to evaluate the current vibration data to be measured of the oil pump, the method further includes: Acquire new sample data according to the evaluation result of the vibration data to be measured; The state assessment model is optimized using the new sample data.
8. An oil pump status assessment device, characterized in that: include: A feature set acquisition module, used for extracting features from vibration data of the oil pump during operation by using a set mode decomposition method to acquire a feature set, wherein the feature set includes vibration components in different frequency dimensions; An oil pump state classification module is used to process the feature set using an exponentially weighted moving average method to obtain a weighted target feature, and to classify the oil pump state based on the weighted target feature; An oil pump status marking module, used for marking the oil pump status of the vibration data according to the weighted target feature of the divided state; The state evaluation module is used to input the marked vibration data into a global average pooling convolutional neural network for training to obtain a state evaluation model, and to perform state evaluation on the vibration data to be tested by the oil pump through the state evaluation model.
9. 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, and the computer program is executed by the at least one processor so that the at least one processor can perform the method according to any one of claims 1 to 8.
10. 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 method according to any one of claims 1 to 8 when executed.