Gas flow abnormity monitoring system
By combining multiple prediction models and adaptability to select prediction models with high fit in the gas volume prediction technology, problems such as simple models and insufficient outlier processing in the existing technology are solved, and higher prediction accuracy and real-time abnormality monitoring are achieved.
Patent Information
- Application Number
- CN202510366791.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-26
- Publication Date
- 2025-05-23
AI Technical Summary
The existing gas volume prediction technology for gas use has problems such as simple model, insufficient outlier processing, lack of real-time performance, ignoring external factors, and low data acquisition efficiency.
A gas flow abnormality monitoring system is proposed. By combining multiple prediction models (such as moving average prediction model and fixed-period prediction model), the prediction model with high fit is adaptively selected as the prediction reference model, and data preprocessing and abnormal monitoring are performed.
It improves the accuracy of gas volume prediction, realizes real-time abnormality detection and monitoring, and enhances the adaptability and automation level of the system.
Smart Images
Figure CN120027889A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of gas monitoring, and more specifically, to a gas flow abnormality monitoring system. Background Art
[0002] With the progress of society and the development of cities, people's living standards have been significantly improved, and gas has entered thousands of households. Maintaining the normal operation of the gas metering system is an important task, and being able to timely detect the failure of the gas metering system and gas equipment is the key to this work. Among them, industrial and commercial gas users are equipped with independent gas flow meters, and the monitoring of gas flow meters has received more and more attention.
[0003] In the application of gas usage monitoring, predicting gas usage is an important part. There are many existing prediction methods, such as: prediction based on time series analysis and historical data. Time series analysis is used to obtain the historical daily gas consumption of smart gas meters and use daily or weekly cycle models for prediction; prediction based on thermal performance is used to predict gas consumption by considering parameters such as thermal conductivity, internal heat gain, and room temperature of the building; personalized prediction based on machine learning and deep learning is used to introduce machine learning and neural network technology (such as the Transform Integrated Neural Network model) and combine the data of smart gas meters to model and predict users' personalized gas usage behavior.
[0004] The existing gas consumption prediction technology has problems such as simple model, insufficient outlier processing, lack of real-time performance, neglect of external factors, and low data collection efficiency. Summary of the invention
[0005] In view of the above, the present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes a gas flow abnormality monitoring system. By combining multiple prediction models and adaptively selecting prediction results, it can better cope with complex and changeable user behaviors and improve the accuracy of prediction.
[0006] To this end, in a first aspect, an embodiment of the present invention provides a gas flow abnormality monitoring system, comprising:
[0007] A data collection module, used to collect user gas data, wherein the user gas data includes a timestamp and a corresponding accumulated gas consumption;
[0008] A data processing module, used for preprocessing the user gas data to obtain preprocessed data;
[0009] A data prediction module, configured to predict according to the preprocessed data to obtain at least two different prediction models, and determine one of the at least two prediction models with a higher fitting effect as a prediction reference model; and
[0010] The abnormality monitoring module is used to determine whether there is an abnormality based on the prediction reference model and the current user gas data.
[0011] Preferably, the prediction model comprises a moving average prediction model, and a moving average value of the user gas data in a preset time period is used as a prediction value; and / or,
[0012] The prediction model includes a fixed period prediction model, and a prediction value is obtained according to the mean, standard deviation or skewness of the user gas data in a preset period segment.
[0013] Preferably, the step of determining one of the at least two prediction models with a higher fitting effect as a prediction reference model comprises:
[0014] Draw a line graph according to the prediction model, and determine that the line with the highest fitting degree is the prediction reference model; or,
[0015] Draw a scatter plot based on the prediction model and determine the coefficient of determination R of the scatter plot 2 The higher one is the prediction reference model.
[0016] Preferably, the preprocessing includes time interval selection, data smoothing and extreme value pruning.
[0017] Preferably, the time interval selection includes: determining a preset time interval according to the integrity of the user gas data.
[0018] Preferably, determining the preset time interval according to the integrity of the user gas data includes:
[0019] determining that when the loss rate of the user gas data corresponding to the first time interval does not exceed a preset value, determining the first time interval as the preset time interval; or,
[0020] When it is determined that the loss rate of the user gas data corresponding to the first time interval exceeds a preset value, the first time interval is extended to a second time interval, the loss rate of the user gas data corresponding to the second time interval is re-determined and the preset time interval is correspondingly determined.
[0021] Preferably, the data smoothing process comprises: smoothing the missing data segments of the user gas data using a multi-period moving average method.
[0022] Preferably, the extreme value trimming includes: identifying extreme values in the user gas data according to a preset extreme threshold xσ, and replacing the corresponding extreme values with the median.
[0023] Preferably, the anomaly monitoring module includes:
[0024] A leakage warning unit, configured to generate a leakage warning and send a warning signal when it is determined that the current user gas data exceeds the predicted value corresponding to the prediction model by a first warning value.
[0025] Preferably, the anomaly monitoring module further includes:
[0026] A theft warning unit, configured to generate a theft warning and send a warning signal when it is determined that the current user gas data exceeds the predicted value corresponding to the prediction model by a second warning value, where the second warning value is greater than the first warning value.
[0027] The gas flow anomaly monitoring system provided by the embodiment of the present invention combines multiple prediction models and adaptively selects a prediction model with a high fitting degree as a prediction reference model to better cope with complex and variable user behaviors, improving the accuracy of prediction. Further anomaly detection based on the prediction reference model can also achieve accurate anomaly monitoring, which is beneficial to improving work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Figure 1 It is a schematic diagram of the framework structure of the gas flow anomaly monitoring system provided by the embodiment of the present invention;
[0029] Figure 2 It is a line graph of the prediction structure of the data prediction module of the gas flow anomaly monitoring system provided by the embodiment of the present invention;
[0030] Figure 3 It is a scatter diagram of the prediction structure of the data prediction module of the gas flow anomaly monitoring system provided by the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0031] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and are intended to explain the present invention and should not be construed as limiting the present invention.
[0032] The disclosure below provides many different embodiments or examples to implement different structures of the present invention. In order to simplify the disclosure of the present invention, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present invention. In addition, the present invention can repeat reference numbers and / or letters in different examples. This repetition is for the purpose of simplicity and clarity, and does not itself indicate the relationship between the various embodiments and / or settings discussed. In addition, the present invention provides various specific processes and examples of materials, but those of ordinary skill in the art can be aware of the applicability of other processes and / or the use of other materials.
[0033] Please refer to Figure 1 The purpose of the present invention is to provide a gas flow abnormality monitoring system 100 for real-time prediction and abnormal monitoring of user gas usage. The gas flow abnormality monitoring system 100 includes:
[0034] The data collection module 10 is used to collect user gas data, wherein the user gas data includes a timestamp and a corresponding accumulated gas consumption;
[0035] The data processing module 20 is used to pre-process the user gas data to obtain pre-processed data;
[0036] A data prediction module 30 is used to predict according to the preprocessed data to obtain at least two different prediction models, and determine one of the at least two prediction models with a higher fitting effect as a prediction reference model; and
[0037] The abnormality monitoring module 40 is used to determine whether there is an abnormality based on the prediction reference model and the current user gas data.
[0038] Among them, the data acquisition module specifically obtains the user's original gas usage data through remote collection equipment (smart gas meter), including timestamp, cumulative gas usage and other information.
[0039] Further, the prediction model includes a moving average prediction model, which uses a moving average value of the user gas data in a preset time period as a prediction value; and / or,
[0040] The prediction model includes a fixed period prediction model, and a prediction value is obtained according to the mean, standard deviation or skewness of the user gas data in a preset period segment.
[0041] Specifically, the moving average prediction model uses the moving average of the previous n days as the new prediction value, which is suitable for users with relatively stable gas usage patterns. The actual gas usage curve and the predicted gas usage curve have the same trend, which indicates that the prediction effect is good. Fixed period prediction model: Based on the data statistics of the past few months (such as mean, standard deviation, skewness, etc.), it is suitable for gas usage behaviors with obvious periodicity. By adjusting the cycle length, it can adapt to the gas usage characteristics of different users.
[0042] In this embodiment, at least two methods, moving average prediction and fixed period prediction, are integrated, and the optimal model is dynamically selected according to different application scenarios. This dual-model architecture can cope with a variety of complex situations, especially when the user's gas usage pattern changes, it can quickly adapt and adjust the prediction results. The moving average method is suitable for situations where the gas consumption fluctuates slightly, while the fixed period prediction rule can handle gas usage patterns with periodic characteristics. Through automated model selection, human intervention is avoided, and the automation level and adaptability of the system are improved.
[0043] It is understandable that the prediction model may also include other prediction models, such as thermal performance prediction model, machine learning prediction model, etc. By comparing the fit of multiple prediction models, the one with the highest fit is selected as the prediction reference model, thereby improving the prediction accuracy of different user behaviors.
[0044] Further, the step of determining one of the at least two prediction models with a higher fitting effect as a prediction reference model includes:
[0045] Draw a line graph according to the prediction model, and determine that the line with the highest fitting degree is the prediction reference model; or,
[0046] Draw a scatter plot based on the prediction model and determine the coefficient of determination R of the scatter plot 2 The higher one is the prediction reference model.
[0047] The line chart corresponding to the prediction model is as follows: Figure 2 As shown, Figure 2 : In the figure are partial enlarged views of the line graph obtained by the moving average prediction method and the fixed period method.
[0048] The prediction effect is shown by line graph and scatter plot. The closer the scatter points are to the fitting line, the better the prediction effect of the model. 2 The higher the value, the better the model fitting effect, that is, the better the prediction effect. The light line is the actual gas consumption, and the dark line is the predicted gas consumption.
[0049] The scatter plot corresponding to the prediction model is as follows: Figure 3 As shown, Figure 3 The two figures are the scatter plots obtained by the moving average prediction method and the scatter plots obtained by the fixed period prediction method. The horizontal axis is the actual gas consumption, the vertical axis is the predicted gas consumption, and the red line is the fitting line with a slope of 1. The more concentrated the scatter points are, the better the prediction effect is.
[0050] Furthermore, the preprocessing includes time interval selection, data smoothing and extreme value pruning.
[0051] Through multiple data processing technologies such as dynamic time interval selection, data smoothing and outlier pruning, the impact of data noise and outliers on model prediction is effectively reduced. Compared with traditional fixed time interval analysis, this adaptive time interval selection can automatically adjust according to data integrity, improve data utilization, and thus enhance the prediction accuracy of the model. The analysis window can be adjusted under different user behavior patterns to capture gas usage patterns more flexibly. The use of smoothing and extreme value pruning effectively reduces the sensitivity of the prediction model to abnormal data and improves the robustness of the prediction results.
[0052] Furthermore, the time interval selection includes: determining a preset time interval according to the integrity of the user gas data. Specifically, according to the integrity of the user data, dynamically selecting a suitable time interval as a time window, such as 0.5h, 1h, 2h, 12h or 24h.
[0053] Furthermore, determining the preset time interval according to the integrity of the user gas data includes:
[0054] determining that when the loss rate of the user gas data corresponding to the first time interval does not exceed a preset value, determining the first time interval as the preset time interval; or,
[0055] When it is determined that the loss rate of the user gas data corresponding to the first time interval exceeds a preset value, the first time interval is extended to a second time interval, the loss rate of the user gas data corresponding to the second time interval is re-determined and the preset time interval is correspondingly determined.
[0056] Specifically, if the loss rate exceeds 20%, a larger time interval is selected to improve the integrity of the data. For example, when the selected time interval is 0.5h, part of the relevant data is missing, and the corresponding loss rate may exceed 20%. At this time, the time interval 0.5h is replaced with 1h. If the loss rate of the relevant data with the time interval of 1h is lower than 20%, 1h is determined as the selected time interval. If the data loss rate corresponding to 1h is still higher than 20%, 2h is further replaced as an alternative time interval, and so on, until the loss rate of the data corresponding to the selected interval time does not exceed 20%. In this way, the data integrity of the prediction model is improved, which is conducive to improving the accuracy of subsequent predictions.
[0057] Furthermore, the data smoothing process includes: smoothing the missing data segments of the user gas data using a multi-period moving average method. Specifically, for the missing data segments, the multi-period moving average method is used for smoothing, the average value of the data in the same time period of the previous three days is calculated, and the amount of data in the missing segments is redistributed, thereby reducing the impact of outliers.
[0058] Furthermore, the extreme value pruning includes: identifying the extreme values in the user's gas data according to a preset extreme threshold xσ, and replacing the corresponding extreme values with the median. Specifically, x represents a multiple, which can be selected as appropriate according to different situations. σ is the standard deviation of the user's historical gas consumption data. When x is 2, it is equivalent to using 2 times the standard deviation to prune the extreme values. If the gas consumption in this period exceeds or is lower than the mean of the gas consumption in the same period within a certain period of the user's history plus or minus 2 times the standard deviation, it is considered to be an extreme value and is processed. x can be 3, 4, or 5. The higher the value, the higher the tolerance for extreme values. This can be used as an adjustable parameter according to the actual situation of different users. In this way, the robustness of the data can be effectively improved, and the interference of extreme values on the model prediction results can be avoided.
[0059] Furthermore, the abnormality monitoring module 40 includes:
[0060] The leakage warning unit is used to generate a leakage warning and send out a warning signal when it is determined that the current user gas data exceeds the prediction value corresponding to the prediction model and reaches a first warning value.
[0061] In this embodiment, by real-time monitoring of the gas usage changes of users, identifying abnormal situations and issuing warnings, the safety of the gas system is improved. When it is detected that the gas usage changes exceed a lower first preset value, it means that the actual gas usage deviates from the predicted gas usage to a small extent, and the abnormality report and warning reminder can be made accordingly.
[0062] Furthermore, the abnormal monitoring module further includes:
[0063] The theft warning unit is used to generate a theft warning and send out a warning signal when it is determined that the current user's gas data exceeds the prediction value corresponding to the prediction model and reaches a second warning value, and the second warning value is greater than the first warning value. In this embodiment, the user's historical gas usage pattern is analyzed, abnormal and large changes in gas usage are detected, and potential gas theft behaviors are identified. If such behaviors are found, the abnormal emergency report is made and an alarm message is issued. For example, the system can send a text message or make a phone call to a preset manager, or the system can send a buzzer alarm.
[0064] In this embodiment, through accurate gas usage prediction, a variety of intelligent application functions are realized, such as micro-leakage warning and gas theft detection. This not only improves the safety of the gas system, but also provides valuable user analysis data for gas companies to help optimize user services and management. Real-time monitoring and anomaly detection are realized, which can quickly identify and warn of safety issues such as gas leaks. User portraits and gas usage pattern analysis are provided to support gas companies in providing personalized services and precision marketing.
[0065] The gas flow anomaly monitoring system provided in the embodiment of the present invention combines multiple prediction models and adaptively selects a prediction model with a high degree of fit as a prediction reference model, so as to better cope with complex and changeable user behaviors and improve the accuracy of predictions. Further anomaly detection is performed according to the prediction reference model, and accurate anomaly monitoring can be achieved, which is conducive to improving work efficiency.
[0066] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.
[0067] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.
Claims
1. A gas flow abnormality monitoring system, characterized in that: include: A data collection module, used to collect user gas data, wherein the user gas data includes a timestamp and a corresponding accumulated gas consumption; A data processing module, used for preprocessing the user gas data to obtain preprocessed data; A data prediction module, configured to predict according to the preprocessed data to obtain at least two different prediction models, and determine one of the at least two prediction models with a higher fitting effect as a prediction reference model; and The abnormality monitoring module is used to determine whether there is an abnormality based on the prediction reference model and the current user gas data.
2. The gas flow abnormality monitoring system according to claim 1, characterized in that: The prediction model includes a moving average prediction model, using the moving average of the user gas data in a preset time period as a prediction value; and / or, The prediction model includes a fixed period prediction model, and a prediction value is obtained according to the mean, standard deviation or skewness of the user gas data in a preset period segment.
3. The gas flow abnormality monitoring system according to claim 1, characterized in that: The step of determining one of the at least two prediction models with a higher fitting effect as a prediction reference model comprises: Draw a line graph according to the prediction model, and determine that the line with the highest fitting degree is the prediction reference model; or, Draw a scatter plot based on the prediction model and determine the coefficient of determination R of the scatter plot 2 The higher one is the prediction reference model.
4. The gas flow abnormality monitoring system according to claim 1, characterized in that: The preprocessing includes time interval selection, data smoothing and extreme value pruning.
5. The gas flow abnormality monitoring system according to claim 4, characterized in that: The time interval selection includes: determining a preset time interval according to the integrity of the user gas data.
6. The gas flow abnormality monitoring system according to claim 5, characterized in that: The step of determining the preset time interval according to the integrity of the user gas data includes: determining that when the loss rate of the user gas data corresponding to the first time interval does not exceed a preset value, determining the first time interval as the preset time interval; or, When it is determined that the loss rate of the user gas data corresponding to the first time interval exceeds a preset value, the first time interval is extended to a second time interval, the loss rate of the user gas data corresponding to the second time interval is re-determined and the preset time interval is correspondingly determined.
7. The gas flow abnormality monitoring system according to claim 4, characterized in that: The data smoothing process includes: using a multi-period moving average method to smooth the missing data segments of the user gas data.
8. The gas flow abnormality monitoring system according to claim 4, characterized in that: The extreme value pruning includes: identifying extreme values in the user gas data according to a preset extreme threshold xσ, and replacing the corresponding extreme values with the median.
9. The gas flow abnormality monitoring system according to claim 1, characterized in that: The abnormal monitoring module includes: The leakage warning unit is used to generate a leakage warning and send out a warning signal when it is determined that the current user gas data exceeds the prediction value corresponding to the prediction model and reaches a first warning value.
10. The gas flow abnormality monitoring system according to claim 9, characterized in that: The abnormal monitoring module also includes: The theft warning unit is used to generate a theft warning and send out a warning signal when it is determined that the current user gas data exceeds the prediction value corresponding to the prediction model to reach a second warning value, and the second warning value is greater than the first warning value.