Error calibration method and system for smart water meter

Abstract

The present application relates to the field of error correction technology, and specifically discloses an error correction method and system for a smart water meter, which uses a deep learning-based neural network model to perform fine-grained time series feature extraction on the time series data of the temperature and pressure of the water body flowing through the multifunctional smart water meter within the target time period, so as to capture the time series variation pattern of the water body temperature and pressure in each local time domain, and by performing time series integral encoding on the time series variation characteristics of the water body temperature and pressure in each local time domain, to understand the time series cumulative effect of the water body temperature and pressure on the water consumption in the global time domain, and realize intelligent prediction of the water consumption, thereby performing error correction on the smart water meter based on the difference between the prediction result and the water meter measurement value. In this way, the measurement error of the smart water meter can be reflected more accurately, thereby improving the accuracy of error correction and the intelligent level of water management.

Description

Technical Field

[0001] The present application relates to the technical field of error correction, and more specifically, to an error correction method and system for a smart water meter. Background Art

[0002] With the continuous advancement of science and technology, the application of smart water meters in modern water management is becoming increasingly widespread. Smart water meters integrate modern microelectronics, sensing, and smart IC card technologies to not only accurately measure water consumption but also enable data transmission, transaction settlement, record storage, and automatic calculation of tiered water pricing. This significantly enhances the intelligence of water management and effectively avoids the many drawbacks of traditional manual meter reading.

[0003] During the use of smart water meters, it is crucial to ensure their measurement accuracy. Error correction, as a key means to ensure the measurement accuracy of smart water meters, has received widespread attention. In the prior art, for example, the invention patent application with publication number CN118776645A proposes an error correction method for a multifunctional smart water meter, which uses a pre-established multivariate linear regression model to predict the second water consumption within the target time period based on the date information of the target time period, the temperature and pressure of the water body flowing through the multifunctional smart water meter, and combines the first water consumption within the target time period collected by the multifunctional smart water meter to achieve error correction of the multifunctional smart water meter, thereby ensuring the accurate measurement of water consumption by the multifunctional smart water meter.

[0004] However, the multiple linear regression model has some significant flaws when predicting water consumption. First, the model assumes a linear relationship between variables, but in reality, the relationship between water temperature, pressure, and water consumption is often more complex and may exist in a nonlinear manner. For example, under extreme weather conditions, such as extreme heat or cold, changes in water temperature and pressure can have a significant nonlinear effect on water consumption, and this association pattern is difficult to accurately describe using a simple linear relationship. In addition, the multiple linear regression model is highly dependent on data. If the historical data is noisy or incomplete, it will significantly affect the model's prediction accuracy.

[0005] Therefore, an optimized error checking method and system for smart water meters are desired. Summary of the Invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides an error correction method and system for a smart water meter, which uses a neural network model based on deep learning to perform fine-grained time series feature extraction on the time series data of the temperature and pressure of the water body flowing through the multifunctional smart water meter during the target time period, so as to capture the time series change pattern of the water body temperature and pressure in each local time domain, and by performing time series integral encoding on the time series change characteristics of the water body temperature and pressure in each local time domain, to understand the time series cumulative effect of the water body temperature and pressure on the water consumption in the global time domain, and realize intelligent prediction of water consumption, so as to perform error correction on the smart water meter according to the difference between the prediction result and the water meter measurement value. In this way, the measurement error of the smart water meter can be reflected more accurately, thereby improving the accuracy of error correction and the intelligent level of water management.

[0007] According to one aspect of the present application, a method for error checking of a smart water meter is provided, comprising:

[0008] Acquire monitoring parameters of the multifunctional smart water meter within a target time period, the monitoring parameters including: date information of the target time period, actual water consumption measurement value, and a time series of temperature and pressure of water flowing through the multifunctional smart water meter;

[0009] Determining an initial water consumption based on a time series of the temperature and pressure of the water flowing through the multifunctional smart water meter, wherein determining the initial water consumption comprises: performing time series fine-grained integral encoding and decoding prediction on the time series of the temperature and pressure of the water flowing through the multifunctional smart water meter to obtain the initial water consumption;

[0010] Based on the historical monthly water consumption data of the month information corresponding to the target time period, the initial water consumption is corrected to obtain a second water consumption;

[0011] The actual water consumption measurement value of the target time period is used as the first water consumption, and based on the difference between the first water consumption and the second water consumption, the multifunctional smart water meter is error-checked.

[0012] According to another aspect of the present application, there is provided an error checking system for a smart water meter, comprising:

[0013] A monitoring parameter acquisition module is used to obtain monitoring parameters of the multifunctional smart water meter within a target time period, wherein the monitoring parameters include: date information of the target time period, actual water consumption measurement value, and time series of temperature and pressure of the water body flowing through the multifunctional smart water meter;

[0014] an initial water consumption determination module, configured to determine the initial water consumption based on the time series of the temperature and pressure of the water body flowing through the multifunctional smart water meter, wherein determining the initial water consumption comprises: performing time series fine-grained integral encoding and decoding prediction on the time series of the temperature and pressure of the water body flowing through the multifunctional smart water meter to obtain the initial water consumption;

[0015] an initial water consumption correction module, configured to correct the initial water consumption based on historical monthly water consumption data of the month information corresponding to the target time period to obtain a second water consumption;

[0016] An error checking module is used to use the actual water consumption measurement value in the target time period as the first water consumption, and perform error checking on the multifunctional smart water meter based on the difference between the first water consumption and the second water consumption.

[0017] Compared with the prior art, the error checking method and system for smart water meters provided by this application use a neural network model based on deep learning to extract fine-grained time series features from the time series data of the temperature and pressure of the water body flowing through the multifunctional smart water meter during the target time period, so as to capture the time series variation pattern of the water body temperature and pressure in each local time domain, and to understand the time series cumulative effect of the water body temperature and pressure on the water consumption in the global time domain by performing time series integral encoding on the time series variation characteristics of the water body temperature and pressure in each local time domain, so as to realize intelligent prediction of water consumption, and thus perform error correction on the smart water meter based on the difference between the prediction result and the water meter measurement value. In this way, the measurement error of the smart water meter can be reflected more accurately, thereby improving the accuracy of error correction and the intelligent level of water management. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above and other purposes, features, and advantages of the present application will become more apparent through a more detailed description of the embodiments of the present application in conjunction with the accompanying drawings. The accompanying drawings are intended to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0019] Figure 1 Flowchart of an error checking method for a smart water meter according to an embodiment of the present application.

[0020] Figure 2 This is a flowchart of sub-step S2 of the error checking method for a smart water meter according to an embodiment of the present application.

[0021] Figure 3 Schematic diagram of data flow of sub-step S2 of the error checking method for a smart water meter according to an embodiment of the present application.

[0022] Figure 4 This is a flowchart of sub-step S23 of the error checking method for a smart water meter according to an embodiment of the present application.

[0023] Figure 5 This is a flowchart of sub-step S232 of the error checking method for a smart water meter according to an embodiment of the present application.

[0024] Figure 6 This is a flowchart of sub-step S3 of the error checking method for a smart water meter according to an embodiment of the present application.

[0025] Figure 7 4 is a block diagram of an error checking system for a smart water meter according to an embodiment of the present application. DETAILED DESCRIPTION

[0026] As used in this application and the claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not intended to refer to the singular but may include the plural. Generally speaking, the terms "comprises" and "include" only indicate the inclusion of the steps and elements specifically identified, and these steps and elements do not constitute an exclusive list. A method or apparatus may also include other steps or elements.

[0027] Although the present application makes various references to certain modules in the system according to embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are illustrative only, and different aspects of the system and method can use different modules.

[0028] Flowcharts are used in this application to illustrate the operations performed by the systems according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the various steps may be processed in reverse order or simultaneously, as needed. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0029] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0030] It is worth noting that in this application, all actions to obtain data are carried out in compliance with the relevant data protection laws and policies of the country where they are located and with the authorization given by the owner of the corresponding device.

[0031] As mentioned in the above background technology, invention patent application CN118776645A proposes an error calibration method for a multifunctional smart water meter, which uses a pre-established multivariate linear regression model to predict the second water consumption within the target time period based on the date information of the target time period, the temperature and pressure of the water flowing through the multifunctional smart water meter, and combines the first water consumption within the target time period collected by the multifunctional smart water meter to realize error calibration of the multifunctional smart water meter, thereby ensuring the accurate measurement of water consumption by the multifunctional smart water meter.

[0032] In existing technologies, the multivariate linear regression model has some obvious flaws when predicting water consumption. First, the model assumes a linear relationship between variables, but in reality, the relationship between water temperature, pressure and water consumption is often more complex and may exist in a nonlinear relationship. For example, under extreme weather conditions, such as extreme heat or cold, changes in water temperature and pressure can have a significant nonlinear effect on water consumption, and this association pattern is difficult to accurately describe with a simple linear relationship. In addition, the multivariate linear regression model is highly dependent on data. If the historical data is noisy or incomplete, it will significantly affect the model's prediction accuracy. In response to the above technical problems, this application proposes an optimized error correction method for smart water meters, which uses a deep learning-based neural network model to perform fine-grained time series feature extraction on the time series data of the temperature and pressure of the water body flowing through the multifunctional smart water meter during the target time period, so as to capture the time series change pattern of the water body temperature and pressure in each local time domain, and to understand the time series cumulative effect of the water body temperature and pressure on the water consumption in the global time domain by performing time series integral encoding on the time series change characteristics of the water body temperature and pressure in each local time domain, so as to realize intelligent prediction of water consumption, and thus perform error correction on the smart water meter based on the difference between the prediction result and the water meter measurement value. In this way, the measurement error of the smart water meter can be reflected more accurately, thereby improving the accuracy of error correction and the level of intelligence of water management.

[0033] Figure 1 FIG. 1 is a flow chart of an error checking method for a smart water meter according to an embodiment of the present application. Figure 1As shown, the error verification method of the smart water meter includes the following steps: S1, obtaining monitoring parameters of the multifunctional smart water meter within a target time period, wherein the monitoring parameters include: date information of the target time period, actual water consumption measurement value, and time series of temperature and pressure of the water body flowing through the multifunctional smart water meter; S2, determining the initial water consumption based on the time series of temperature and pressure of the water body flowing through the multifunctional smart water meter, wherein determining the initial water consumption includes: performing time series fine-grained integral encoding and decoding prediction on the time series of temperature and pressure of the water body flowing through the multifunctional smart water meter to obtain the initial water consumption; S3, correcting the initial water consumption based on historical monthly water consumption data of the month information corresponding to the target time period to obtain a second water consumption; S4, taking the actual water consumption measurement value of the target time period as the first water consumption, and performing error verification on the multifunctional smart water meter based on the difference between the first water consumption and the second water consumption.

[0034] In the aforementioned smart water meter error correction method, step S1 obtains monitoring parameters of the multifunctional smart water meter during a target time period. These monitoring parameters include: date information for the target time period, actual water consumption measurements, and a time series of the temperature and pressure of the water flowing through the multifunctional smart water meter. It should be understood that date information can reflect differences in user water usage habits during different time periods, such as seasonal variations. Actual water consumption measurements are the direct result of current meter readings and serve as the core basis for error determination. Changes in water temperature and pressure can reflect the potential impact of the external environment on water consumption. For example, when water temperatures are low, such as in winter, users bathe less frequently, potentially reducing water consumption. However, during high temperatures in summer, users' water needs for bathing and laundry increase, leading to increased water consumption. Regarding pressure, low water pressure reduces water flow, affecting user water consumption. Normal or high water pressure results in smoother water flow, potentially increasing water consumption. Based on this, this application comprehensively considers multiple factors such as the date information, water temperature and pressure of the target time period to more comprehensively understand and predict the user's actual water consumption, and then performs error correction on the actual water consumption measurement values collected by the multi-functional smart water meter.

[0035] Specifically, obtaining date information requires accurately recording the specific dates of the target time period. This seemingly simple step is crucial, as date information directly impacts subsequent data analysis and model predictions. Specifically, smart water meters incorporate a high-precision clock module that records the current time in real time and associates these timestamps with corresponding water consumption data. This allows for precise tracking of daily and even hourly water usage, providing a solid foundation for subsequent analysis. Furthermore, considering the potential impact of different seasons and months on water consumption, such as increased water demand due to high temperatures in summer or frozen pipes due to low temperatures in winter, accurate date information helps better understand the changing patterns of water use patterns.

[0036] To obtain actual water consumption measurements, in addition to relying on built-in high-sensitivity sensors, attention must also be paid to the sensor's long-term stability and calibration mechanisms. Over time, sensor performance may drift, affecting the accuracy of measurement results. To address this, a self-calibration function can be introduced during the design phase to regularly self-check and calibrate the sensor. Furthermore, to address unexpected situations such as abnormal readings caused by pipe blockage or damage, smart water meters should be equipped with fault detection and alarm systems. Once an abnormal flow pattern is detected, the system can immediately issue an alarm and attempt to maintain basic functionality through backup channels or algorithmic compensation. Furthermore, by integrating big data analysis technology, normal water usage patterns can be learned from historical data, helping to identify potential problem areas and provide early warning of possible failures, thereby improving system reliability and user satisfaction.

[0037] Acquiring a time series of the temperature and pressure of water flowing through a multifunctional smart water meter requires high accuracy and real-time response. To achieve this, the selected sensor must not only offer high resolution and fast response speed but also withstand harsh operating conditions, including extreme temperatures, humidity, and corrosive media. To address these issues, the sensor housing is typically constructed of corrosion-resistant materials, and a waterproof seal is employed to protect the internal electronic components. Furthermore, during signal transmission, shielded cables or anti-interference coding schemes within wireless transmission technologies can be used to mitigate the effects of external electromagnetic interference on measurement results. Furthermore, considering the potential energy consumption associated with long-term continuous operation, optimizing power management strategies is crucial. For example, dynamically adjusting the sensor's operating frequency or utilizing energy harvesting technologies (such as solar energy) can extend battery life and reduce maintenance costs.

[0038] Finally, regarding data security and privacy protection, in addition to the encrypted storage and secure transmission measures mentioned above, a comprehensive data access control mechanism must be established. This means that only authorized personnel can access data at a specific level, and all access activities should be recorded for audit purposes.

[0039] In the error checking method of the above-mentioned smart water meter, the step S2 determines the initial water consumption based on the time series of the temperature and pressure of the water body flowing through the multifunctional smart water meter, wherein determining the initial water consumption includes: performing time series fine-grained integral coding and decoding prediction on the time series of the temperature and pressure of the water body flowing through the multifunctional smart water meter to obtain the initial water consumption. It should be understood that the user's water use pattern may change with the change of water body temperature and pressure, thereby affecting the total water consumption generated within the target time period. Therefore, the present application further captures the time series change pattern of water body temperature and pressure by performing time series fine-grained integral coding on the time series of water body temperature and pressure, thereby realizing decoding prediction of water consumption on this basis. Among them, Figure 2 This is a flowchart of sub-step S2 of the error checking method for a smart water meter according to an embodiment of the present application. Figure 3 FIG. 1 is a data flow diagram of sub-step S2 of the error checking method for a smart water meter according to an embodiment of the present application. Figure 2 and Figure 3 As shown, the step S2 includes the steps of: S21, sorting the time series of the temperature and pressure of the water body flowing through the multifunctional intelligent water meter to obtain the time series of the water body temperature and the time series of the water body pressure; S22, extracting local time series features of the time series of the water body temperature and the time series of the water body pressure to obtain the time series of the local time series implicit correlation feature vector of the water body temperature and the time series of the local time series implicit correlation feature vector of the water body pressure; S23, inputting the time series of the local time series implicit correlation feature vector of the water body temperature and the time series of the local time series implicit correlation feature vector of the water body pressure into the time series integrator respectively to obtain the water body temperature time series integral implicit coding vector and the water body pressure time series integral implicit coding vector; S24, predicting water consumption based on the water body temperature time series integral implicit coding vector and the water body pressure time series integral implicit coding vector to obtain the initial water consumption.

[0040] Specifically, in step S21, the time series of the temperature and pressure of the water body flowing through the multifunctional smart water meter are sorted to obtain the time series of the water body temperature and the time series of the water body pressure. Specifically, the present application takes into account that the water body temperature and the water body pressure have different physical properties and ranges of variation, and the influence patterns of the two on water consumption are also different. Therefore, in order to improve the pertinence of data analysis, the present application further groups and arranges the time series of the temperature and pressure of the water body flowing through the multifunctional smart water meter according to the type and time sequence of the data to form the time series of the water body temperature and the time series of the water body pressure, so as to facilitate independent time series encoding of the two and improve the accuracy of data analysis.

[0041] Specifically, step S22 performs local time series feature extraction on the water temperature time series and the water pressure time series to obtain a time series of implicit correlation feature vectors for the water temperature local time series and a time series of implicit correlation feature vectors for the water pressure local time series. In other words, in order to more meticulously capture short-term time series fluctuations in water temperature and pressure, and thereby more accurately understand the immediate impact of these time series fluctuations on water consumption, the present application further employs a deep learning algorithm to perform local time series feature extraction on the water temperature time series and the water pressure time series. Specifically, the present application uses a preset window to perform sliding segmentation processing on the time series of the water body temperature and the time series of the water body pressure, divides the original global time domain data into multiple local time domain data segments, and then uses a one-dimensional convolutional neural network model to extract time series features of the data in each local time domain. The powerful feature learning ability of the one-dimensional convolutional neural network is used to capture the time series fluctuation patterns of water body temperature and water body pressure in each local time window, thereby obtaining the time series of the implicit correlation feature vector of the local time series of water body temperature and the time series of the implicit correlation feature vector of the local time series of water body pressure.

[0042] Specifically, in step S23, the time series of the local time series implicitly associated characteristic vector of the water body temperature and the time series of the local time series implicitly associated characteristic vector of the water body pressure are respectively input into the time series integrator to obtain the water body temperature time series integral implicit coding vector and the water body pressure time series integral implicit coding vector. It should be understood that since the influence of water body temperature and water body pressure on water consumption is a cumulative process, in order to capture the cumulative effect of water body temperature and water body pressure on water consumption within the target time period, the present application further uses a time series integrator to perform feature aggregation processing on the time series of the local time series implicitly associated characteristic vector of the water body temperature and the time series implicitly associated characteristic vector of the water body pressure, respectively. By accumulating the water body temperature time series features and water body pressure time series features in each local time domain in the global time domain, the water body temperature time series integral implicit coding vector and the water body pressure time series implicit integration coding vector are obtained, thereby revealing the cumulative impact of water body temperature and water body pressure on water consumption on a long time scale, providing a more comprehensive input for subsequent intelligent prediction. Among them, Figure 4 FIG. 1 is a flow chart of sub-step S23 of the error checking method of the smart water meter according to an embodiment of the present application. Figure 4 As shown, the step S23 includes the steps of: S231, performing information kernel coarse-grained aggregation on the time series of the local time series implicit correlation feature vectors of the water body temperature to obtain a coarse-grained aggregation coding vector of the local time series features of the water body temperature; S232, performing dynamic compensation encoding on the coarse-grained aggregation coding vector of the local time series features of the water body temperature based on the feature differences of each local time series implicit correlation feature vector of the water body temperature in the time series of the local time series implicit correlation feature vector of the water body temperature relative to the coarse-grained aggregation coding vector of the local time series features of the water body temperature to obtain the water body temperature time series integral implicit coding vector.

[0043] More specifically, step S231 is expressed as follows:

[0044]

[0045] in, represents the time series of the implicit correlation eigenvector of the local time series of the water body temperature, 、 、 and Respectively represent the first, second, and and vectors, is the number of vectors in the time series of the implicitly associated eigenvectors of the local time series of water body temperature, and Respectively represent the maximum and minimum values, express The median of the characteristic distribution boundary, represents the normalized exponential function, express The attention weight, Represents the coarse-grained aggregated encoding vector of the local temporal characteristics of water body temperature.

[0046] That is, this application introduces the concept of information kernel, and compressively models the global characteristics of the time series of the implicit correlation feature vectors of the local time series of water body temperature by means of coarse-grained aggregation of information kernels, thereby generating a coarse-grained aggregation encoding vector of the local time series features of water body temperature. It should be understood that the introduction of information kernels provides an effective nonlinear mapping mechanism for the feature space, so that the node relationship in the sequence can be modeled in an implicit high-dimensional space, thereby accurately capturing the complex similarity structure between the nodes. The information kernel coarse-grained aggregation network summarizes the overall characteristics of the nodes in the sequence by extracting the main patterns between the nodes.

[0047] Figure 5 FIG. 1 is a flow chart of sub-step S232 of the error checking method of the smart water meter according to an embodiment of the present application. Figure 5 As shown, the step S232 includes the following steps: S2321, calculating the kernel convergence compensation factor of each water body temperature local time series implicit correlation feature vector in the time series of the water body temperature local time series implicit correlation feature vector relative to the water body temperature local time series feature coarse-grained convergence coding vector to obtain the time series of the water body temperature local time series feature kernel convergence compensation factor; S2322, performing compensation explicit modeling based on the gating function on the time series of the water body temperature local time series feature kernel convergence compensation factor to obtain the time series of the water body temperature local time series feature kernel convergence compensation weight factor. sequence; S2323, input the time series of the water body temperature local time series feature kernel convergence compensation weight factor, the water body temperature local time series feature coarse-grained convergence coding vector and the water body temperature local time series implicit correlation feature vector into the fine-grained dynamic compensation convergence network to obtain the water body temperature local time series compensation feature fine-grained convergence coding vector; S2324, input the water body temperature local time series compensation feature fine-grained convergence coding vector and the water body temperature local time series feature coarse-grained convergence coding vector into the residual unit to obtain the water body temperature time series integral implicit coding vector.

[0048] In a specific example of the present application, the step S2321 includes: performing point convolution encoding based on the Sigmoid activation function on the water body temperature local time series implicit correlation feature vector and the water body temperature local time series feature coarse-grained aggregation coding vector to obtain a standardized water body temperature local time series implicit correlation feature vector and a standardized water body temperature local time series feature coarse-grained aggregation coding vector; calculating the position difference vector between the standardized water body temperature local time series implicit correlation feature vector and the standardized water body temperature local time series feature coarse-grained aggregation coding vector, and taking the absolute value of the position difference vector to obtain a water body temperature local time series feature kernel convergence difference compensation coding vector; inputting the water body temperature local time series feature kernel convergence difference compensation coding vector into a compensation feature importance scoring module based on a neural network layer to obtain the water body temperature local time series feature kernel convergence compensation factor.

[0049]

[0050] in, represents the compensation factor calculation network, express activation function, represents a 1×1 convolution operation, and They represent the weight parameter matrix of implicit correlation feature of local time series of water body temperature and the weight parameter matrix of local time series feature of water body temperature, respectively. represents the implicit correlation eigenvector of the local time series of standardized water body temperature, represents the coarse-grained aggregation encoding vector of the local time series characteristics of the standardized water body temperature, Indicates point subtraction by position. Represents the kernel convergence difference compensation coding vector of the local time series feature of water body temperature, Represents the weight parameter matrix of the kernel convergence difference feature of the local time series of water body temperature, represents the bias term, represents the water temperature difference compensation feature importance score conversion vector, The first factor in the set of kernel convergence compensation factors representing the local time series characteristics of water body temperature is The kernel convergence compensation factor of the local time series characteristics of water body temperature.

[0051] That is, since the compressive aggregation process of global features may dilute or even completely lose the personalized information of certain important nodes, this application further introduces the calculation of the kernel convergence compensation factor. By measuring the deviation between the implicit correlation feature vector of each water body temperature local time series and the coarse-grained convergence encoding vector of the water body temperature local time series feature, the kernel convergence compensation factor of the water body temperature local time series feature is dynamically generated, so that the node personalized information can be retained and provide strong supplementary support for subsequent fine-grained compensation modeling.

[0052] In particular, in a preferred example of the present application, the water body temperature local time series feature core convergence difference compensation coding vector is input into the compensation feature importance scoring module based on the neural network layer to obtain the water body temperature local time series feature core convergence compensation factor, including: calculating the ratio between the Euclidean norm of the water body temperature local time series implicit correlation feature vector and the Euclidean norm of the water body temperature local time series feature coarse-grained convergence coding vector; if the ratio is less than 1, then adding one to the ratio and calculating the logarithm with base 2 as the bias term of the neural network layer of the compensation feature importance scoring module; if the ratio is greater than or equal to 1, then using the ratio as the bias term of the neural network layer of the compensation feature importance scoring module, which is expressed by the formula:

[0053]

[0054] in, represents the calculation of the Euclidean norm of the vector, Represents the base-2 logarithm function.

[0055] Here, to compensate for the deviation between the implicit correlation feature vector of the local water temperature time series and the coarse-grained aggregation encoding vector of the local water temperature time series features, the performance deviation of the kernel convergence strategy as a scenario strategy can be measured by quantifying the regret metric based on the information kernel compression assumption during the kernel convergence decision process, namely, the game-based counterfactual regret value. Specifically, the vector norm representation is used to provide a normalized description of the decision point loss based on the policy action, namely, the vector norm representation of the implicit correlation feature vector of the local water temperature time series and the coarse-grained aggregation encoding vector of the local water temperature time series features. Then, based on the differences in possible vector distribution action game scenarios, the compensation rules for the personalized information of the local water temperature time series nodes are modified using the information distribution degree and relative distribution amplitude of the regret value. This way, the personalized information of the local water temperature time series is considered as an untaken action in the decision, and biased compensation is performed based on the potential benefits of the information kernel convergence assumption.

[0056] In a specific example of the present application, step S2322 is expressed as follows:

[0057]

[0058] in, represents the gating threshold, is a natural constant, represents the gating function, The first time series of the kernel convergence compensation weight factor representing the local time series characteristics of water body temperature The kernel convergence compensation weight factor of the local time series characteristics of water body temperature.

[0059] That is, after obtaining the local temporal feature kernel convergence compensation factor of the water body temperature, the present application further regulates the effect of the local temporal feature kernel convergence compensation factor of the water body temperature through explicit compensation modeling based on the gating function. It should be understood that the gating function can dynamically select information based on nonlinear constraints. By explicitly modeling the local temporal feature kernel convergence compensation factor of the water body temperature, the system can screen its effect intensity and generate a local temporal feature kernel convergence compensation weight factor of the water body temperature to ensure that only locally significant features are highlighted, while irrelevant or redundant information is appropriately suppressed. Through this explicit regulation, the model can more accurately capture the dynamic interaction between global and local information.

[0060] In a specific example of the present application, step S2323 is expressed as follows:

[0061]

[0062] in, Represents the fine-grained aggregation encoding vector of the local time series compensation features of the water body temperature.

[0063] That is, based on the obtained water body temperature local time series feature kernel convergence compensation weight factor, the difference information between the water body temperature local time series feature coarse-grained converged coding vector and the water body temperature local time series implicit correlation feature vector is dynamically adjusted, and the water body temperature local time series compensation feature fine-grained converged coding vector is generated through the dynamic compensation convergence network, so that the compensated water body temperature local time series compensation feature fine-grained converged coding vector can more accurately describe the local details of the local time series node, while being able to adapt to the dynamic changes of the global feature constraints, thereby realizing the coordinated expression of features at the global and local levels.

[0064] In a specific example of the present application, step S2324 is expressed as follows:

[0065]

[0066] in, and Represents different weight parameters, represents the implicit coding vector of the time series integral of the water body temperature.

[0067] Specifically, a residual unit (RMU) performs linear weighted aggregation of the fine-grained aggregated encoding vector of the local temporal compensation features of water body temperature and the coarse-grained aggregated encoding vector of the local temporal features of water body temperature to generate an implicit encoding vector for the temporal integral of water body temperature. RMUs ensure feature transfer while avoiding the vanishing gradient problem, making the model easier to optimize and further improving the saliency and robustness of the encoding results.

[0068] Specifically, the step S24 includes: first, inputting the water body temperature time series integral implicit coding vector and the water body pressure time series integral implicit coding vector into the decoder-based water consumption prediction module to obtain a first water consumption prediction value and a second water consumption prediction value. Specifically, the decoder model has been trained with a large amount of historical data and has effectively learned the complex mapping relationship between water body temperature, water body pressure and water consumption. Therefore, when the decoder receives the water body temperature time series integral implicit coding vector and the water body pressure time series implicit integration coding vector, it can accurately decode the user's water consumption prediction result within the target time period based on the time series change information of the water body temperature and water body pressure contained in the vector. Here, the present application independently decodes and predicts the water body temperature time series integral implicit coding vector and the water body pressure time series implicit integration coding vector to obtain water consumption prediction results based on water body temperature and water body pressure, i.e., a first water consumption prediction value and a second water consumption prediction value, respectively. Among them, the first water consumption prediction value reflects the user's water consumption within the target time period under the influence of the current water temperature conditions, while the second water consumption prediction value reveals the user's water consumption within the target time period under the influence of water pressure conditions.

[0069] Then, the average of the first water consumption prediction value and the second water consumption prediction value is calculated to obtain the initial water consumption. That is, in order to comprehensively consider the combined effects of water temperature and water pressure, this application calculates the average of the water consumption prediction results based on water temperature and water pressure to integrate the combined effects of the two, thereby obtaining an initial water consumption that is closer to the user's actual water consumption.

[0070] Specifically, during each sampling cycle, the system collects the water temperature and pressure values at the current moment, combines them with the data from the previous cycle, and calculates the cumulative effect within that time period using numerical integration methods (such as the trapezoidal rule). The implicit coding vectors generated in this step can capture subtle changes on short time scales, providing a rich information basis for subsequent analysis. These coding vectors not only contain the changing trends of the physical quantities themselves, but also reflect the potential impact of environmental factors on water use behavior. In actual operation, the decoder receives the implicit coding vectors obtained by fine-grained integral coding as input and attempts to extract patterns related to water consumption from them. To improve prediction accuracy, a multi-layer structure can be used to increase the expressive power of the model, while regularization methods such as dropout can be used to prevent overfitting.

[0071] Before calculating the mean, the quality and integrity of the input data must be verified. If any predicted value significantly deviates from the expected range or exhibits unusual fluctuations, further investigation should be conducted to avoid using unreliable data to influence the final results. To this end, a range of data cleaning techniques can be employed, including identifying and removing outliers and filling in missing values, to ensure that every piece of data used in the calculation is highly reliable.

[0072] During the calculation process, it is necessary to adopt an appropriate numerical representation to take into account possible decimal precision issues. Generally, floating-point numbers are sufficient for most application scenarios, but in some situations where high precision is required, it may be necessary to use a higher-precision data type, such as the Decimal class provided by the decimal module in Python. This approach allows users to specify the desired level of precision, thereby reducing deviations caused by rounding errors. In addition, when calculating the average, attention should be paid to the issue of weight distribution. Although in the simplest case, the first and second water consumption forecast values are given the same weight (i.e., 0.5), in some special situations, adjusting the weight distribution based on the specific circumstances may produce better results. For example, if one variable has historically performed better than another, the former can be given a greater weight when calculating the average.

[0073] In the error checking method for the smart water meter, in step S3, the initial water consumption is corrected based on the historical monthly water consumption data of the month information corresponding to the target time period to obtain the second water consumption. Figure 6 FIG. 1 is a flow chart of sub-step S3 of the error checking method of the smart water meter according to an embodiment of the present application. Figure 6As shown, the step S3 includes the following steps: S31, extracting the historical monthly water consumption of the monthly information corresponding to the target time period to obtain a reference set of historical monthly water consumption; S32, determining a water consumption correction coefficient based on the reference set of historical monthly water consumption; S33, multiplying the water consumption correction coefficient by the initial water consumption to obtain the second water consumption.

[0074] Specifically, step S31 extracts the historical monthly water consumption of the month information corresponding to the target time period to obtain a reference set of historical monthly water consumption. It should be understood that, considering that the user's life patterns and water needs have a certain degree of repeatability and similarity, that is, due to the certain stability of factors such as climatic conditions and social activities in the same month of each year, the user's water consumption behavior in the same month of each year also has a certain regularity. Therefore, in order to improve the accuracy and reliability of water consumption prediction, this application further extracts the water consumption data of the month in the past few years from the historical database based on the month information corresponding to the target time period to form a reference set of historical monthly water consumption, thereby analyzing and mining the user's water consumption behavior characteristics in the month based on the historical data, and correcting the above-mentioned water consumption prediction results to improve the accuracy and applicability of the water consumption prediction results.

[0075] Specifically, given the potentially massive data volumes, choosing the right database management system is crucial. Relational databases like MySQL or PostgreSQL are common choices due to their stability and powerful query capabilities. Meanwhile, non-relational databases like MongoDB also offer significant advantages for large datasets, particularly when processing semi-structured data. Regardless of the database type, scalability is crucial to ensure flexible adjustments to storage strategies as data volumes grow.

[0076] After determining the appropriate data storage solution, the next step is to design the corresponding query logic to accurately extract the required historical monthly water consumption from the massive data. In this process, the month information corresponding to the target time period must be clearly identified, which usually involves operations such as date parsing and time interval matching. For example, if the target time period is July 2024, it is necessary to locate the water consumption data belonging to that month in all records. To this end, special SQL statements or other forms of query scripts can be written to filter qualified records based on key fields such as user ID and collection date. It is worth noting that before executing the query, the data should be preprocessed, including steps such as cleaning invalid or erroneous data and filling missing values to ensure that subsequent analysis is based on high-quality information.

[0077] When extracting historical monthly water consumption, it's also important to consider how to effectively organize and present this data for subsequent analysis. One possible approach is to create a structured reference set containing water consumption records for all relevant users during each target month. This set should include not only raw values but also additional metadata, such as geographic location and user type (residential or commercial), to provide a more comprehensive perspective. Furthermore, to facilitate comparison and trend analysis, the data for each month can be further broken down by week or day to form a more detailed time series. This helps capture short-term fluctuations in water consumption, thereby improving the accuracy of the forecasting model.

[0078] Another important step in constructing a reference set of historical monthly water consumption is to standardize and normalize the data. Since water use habits may vary significantly between different regions and different types of users, directly comparing raw water consumption values may lead to misleading conclusions. Therefore, introducing standardization methods, such as Z-score standardization or Min-Max normalization, can eliminate the bias caused by scale effects and enable data from different backgrounds to be compared on a unified scale. In addition, statistical indicators such as average water consumption and standard deviation within each month can be calculated as an important basis for evaluating water use behavior characteristics. These statistics can not only reveal long-term trends, but also help identify anomalies, such as sudden high water consumption events or sustained low water consumption periods, providing clues for further in-depth analysis.

[0079] At the same time, to enhance the practical value of the reference set, it is also necessary to explore how to incorporate the impact of external factors on historical monthly water consumption. For example, factors such as temperature fluctuations, holiday schedules, and even the level of socioeconomic activity may indirectly affect users' water demand. Therefore, when constructing the reference set, it is advisable to integrate relevant data on these external variables to form a multidimensional, comprehensive dataset. Specifically, historical weather records can be obtained from meteorological departments, public holiday information can be collected from government websites, and even social media analysis tools can be used to track local community dynamics. In this way, more sophisticated models of water consumption behavior can be established, better understanding the interaction mechanisms between various internal and external factors.

[0080] Specifically, in a specific example of the present application, step S32 includes: inputting the reference set of historical monthly water consumption into a feature encoding and decoding module based on an LSTM model to obtain the water consumption correction coefficient. Those skilled in the art will be aware that LSTM (Long Short-Term Memory) models are widely used in time series forecasting tasks because they can capture long-term dependencies in sequence data. In the present application, the LSTM model learns the long-term dependencies and periodic patterns in the historical monthly water consumption data by encoding the time series features of the reference set of historical monthly water consumption. The decoder then decodes the time series features of the historical monthly water consumption data to obtain the water consumption correction coefficient for the historical monthly water consumption data. This coefficient is used to characterize the fluctuations and trend characteristics of the historical monthly water consumption data, thereby further revising the current water consumption forecast results.

[0081] Specifically, in step S33, the water consumption correction coefficient is multiplied by the initial water consumption to obtain the second water consumption. Specifically, if the correction coefficient is greater than 1, it indicates that the water consumption during the target time period may be higher than the initial estimate, possibly due to the typical high water consumption in that month. If it is less than 1, it indicates that water demand in that month is relatively low, and the water consumption may be lower than the initial estimate. In this way, historical monthly water consumption data is effectively utilized, and the initial forecast value is reasonably adjusted, thereby providing a more accurate water consumption forecast value for subsequent error correction.

[0082] In the aforementioned error-checking method for a smart water meter, step S4 uses the actual water consumption measurement during the target time period as the first water consumption, and based on the difference between the first and second water consumption, performs error checking on the multifunctional smart water meter. Accordingly, if the smart water meter's measurements are accurate, the first and second water consumption should be relatively close. If the difference between the first and second water consumption is significant, this indicates a possible measurement error in the smart water meter. A set difference threshold can be used to determine whether the error exceeds an acceptable range. If the error exceeds the acceptable range, an alarm mechanism is triggered, prompting inspection and maintenance of the smart water meter.

[0083] In summary, the error checking method of the smart water meter based on the embodiment of the present application is explained, which uses a neural network model based on deep learning to perform fine-grained time series feature extraction on the time series data of the temperature and pressure of the water body flowing through the multifunctional smart water meter during the target time period, so as to capture the time series change pattern of the water body temperature and pressure in each local time domain, and to understand the time series cumulative effect of the water body temperature and pressure on the water consumption in the global time domain by performing time series integral encoding on the time series change characteristics of the water body temperature and pressure in each local time domain, so as to realize the intelligent prediction of the water consumption, and thus perform error correction on the smart water meter according to the difference between the prediction result and the water meter measurement value. In this way, the measurement error of the smart water meter can be reflected more accurately, thereby improving the accuracy of error correction and the intelligent level of water management.

[0084] Furthermore, an error checking system for a smart water meter is also provided.

[0085] Figure 7 FIG is a block diagram of an error checking system for a smart water meter according to an embodiment of the present application. Figure 7 As shown, the error checking system 100 of the smart water meter according to the embodiment of the present application includes: a monitoring parameter acquisition module 110, which is used to obtain the monitoring parameters of the multifunctional smart water meter within a target time period, and the monitoring parameters include: date information of the target time period, actual water consumption measurement value, and time series of temperature and pressure of the water body flowing through the multifunctional smart water meter; an initial water consumption determination module 120, which is used to determine the initial water consumption based on the time series of temperature and pressure of the water body flowing through the multifunctional smart water meter, wherein determining the initial water consumption includes: performing time series fine-grained integral encoding and decoding prediction on the time series of temperature and pressure of the water body flowing through the multifunctional smart water meter to obtain the initial water consumption; an initial water consumption correction module 130, which is used to correct the initial water consumption based on historical monthly water consumption data of the month information corresponding to the target time period to obtain a second water consumption; an error checking module 140, which is used to use the actual water consumption measurement value of the target time period as the first water consumption, and perform error checking on the multifunctional smart water meter based on the difference between the first water consumption and the second water consumption.

[0086] Here, those skilled in the art will appreciate that the specific operations of each module in the error checking system of the smart water meter have been described in detail above. Figures 1 to 6 The error checking method of the smart water meter has been described in detail, and therefore, its repeated description will be omitted.

[0087] The basic principles of the present invention have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, and effects mentioned in the present invention are merely illustrative and non-limiting, and should not be construed as necessarily possessed by each embodiment of the present invention. Furthermore, the specific details of the above embodiments are provided for illustrative purposes and to facilitate understanding, and are not intended to be limiting. These details do not necessarily limit the present invention to being implemented using these specific details.

[0088] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiment described above is only schematic. For example, the unit division is only a logical function division, and there may be other division methods in actual implementation. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0089] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be encompassed therein. Any reference to a figure in a claim should not be construed as limiting the claim to which it relates.

[0090] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units stated in the system claims can also be implemented by one unit through software or hardware.

[0091] Finally, it should be noted that the above description has been provided for purposes of illustration and description. Furthermore, the above embodiments are intended only to illustrate the technical solutions of the present invention and are not intended to be limiting. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art will appreciate that the technical solutions of the present invention may be modified or replaced with equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for error checking of a smart water meter, characterized in that: include: Acquire monitoring parameters of the multifunctional smart water meter within a target time period, the monitoring parameters including: date information of the target time period, actual water consumption measurement value, and a time series of temperature and pressure of water flowing through the multifunctional smart water meter; Determining an initial water consumption based on a time series of temperature and pressure of the water flowing through the multifunctional smart water meter; Based on the historical monthly water consumption data of the month information corresponding to the target time period, the initial water consumption is corrected to obtain a second water consumption; using the actual water consumption measurement value of the target time period as a first water consumption, and performing an error check on the multifunctional smart water meter based on a difference between the first water consumption and the second water consumption; The method of determining the initial water consumption based on the time series of the temperature and pressure of the water flowing through the multifunctional smart water meter includes: Arrange the time series of the temperature and pressure of the water flowing through the multifunctional smart water meter to obtain a time series of the water temperature and a time series of the water pressure; Performing local time series feature extraction on the time series of the water body temperature and the time series of the water body pressure to obtain a time series of implicit correlation feature vectors of the local time series of the water body temperature and a time series of implicit correlation feature vectors of the local time series of the water body pressure; Inputting the time series of the local time series implicit correlation feature vector of the water body temperature and the time series of the local time series implicit correlation feature vector of the water body pressure into a time series integrator respectively to obtain a water body temperature time series integral implicit coding vector and a water body pressure time series integral implicit coding vector; Water consumption is predicted based on the water body temperature time series integral implicit coding vector and the water body pressure time series integral implicit coding vector to obtain the initial water consumption.

2. The error checking method for a smart water meter according to claim 1, characterized in that: Inputting the time series of the implicit correlation feature vector of the local time series of the water body temperature into a time series integrator to obtain the implicit coding vector of the water body temperature time series integration includes: Performing information kernel coarse-grained aggregation on the time series of the implicit correlation feature vector of the local time series of the water body temperature to obtain a coarse-grained aggregation coding vector of the local time series feature of the water body temperature; Based on the characteristic differences of each water body temperature local time series implicit correlation feature vector in the time series of the water body temperature local time series implicit correlation feature vector relative to the water body temperature local time series feature coarse-grained aggregation coding vector, the water body temperature local time series feature coarse-grained aggregation coding vector is dynamically compensated and encoded to obtain the water body temperature time series integral implicit coding vector.

3. The error checking method for a smart water meter according to claim 2, characterized in that: Based on the feature difference between each water body temperature local time series implicit correlation feature vector in the time series of the water body temperature local time series implicit correlation feature vector and the water body temperature local time series feature coarse-grained aggregated code vector, the water body temperature local time series feature coarse-grained aggregated code vector is dynamically compensated and encoded to obtain the water body temperature time series integral implicit code vector, including: Calculating the kernel convergence compensation factor of each water body temperature local time series implicit correlation feature vector in the time series of the water body temperature local time series implicit correlation feature vector relative to the water body temperature local time series feature coarse-grained convergence coding vector to obtain a time series of the water body temperature local time series feature kernel convergence compensation factors; Performing compensation explicit modeling based on a gating function on the time series of the water body temperature local time series characteristic kernel convergence compensation factor to obtain a time series of the water body temperature local time series characteristic kernel convergence compensation weight factor; Inputting the time series of the water body temperature local time series feature kernel convergence compensation weight factor, the water body temperature local time series feature coarse-grained convergence encoding vector and the water body temperature local time series implicit correlation feature vector into a fine-grained dynamic compensation convergence network to obtain the water body temperature local time series compensation feature fine-grained convergence encoding vector; The fine-grained aggregated coding vector of the local temporal compensation feature of the water body temperature and the coarse-grained aggregated coding vector of the local temporal feature of the water body temperature are input into a residual unit to obtain the temporal integral implicit coding vector of the water body temperature.

4. The error checking method for a smart water meter according to claim 3, characterized in that: Calculating the kernel convergence compensation factor of each water body temperature local time series implicit correlation feature vector in the time series of the water body temperature local time series implicit correlation feature vector relative to the water body temperature local time series feature coarse-grained convergence coding vector to obtain a time series of the water body temperature local time series feature kernel convergence compensation factors, including: Performing point convolution encoding based on a Sigmoid activation function on the water body temperature local time series implicit correlation feature vector and the water body temperature local time series feature coarse-grained aggregation coding vector respectively to obtain a standardized water body temperature local time series implicit correlation feature vector and a standardized water body temperature local time series feature coarse-grained aggregation coding vector; Calculating a position difference vector between the standardized water body temperature local time series implicit correlation feature vector and the standardized water body temperature local time series feature coarse-grained convergence coding vector, and taking the absolute value of the position difference vector to obtain a water body temperature local time series feature kernel convergence difference compensation coding vector; The water body temperature local time series feature kernel convergence difference compensation coding vector is input into the compensation feature importance scoring module based on the neural network layer to obtain the water body temperature local time series feature kernel convergence compensation factor.

5. The error checking method for a smart water meter according to claim 4, characterized in that: Inputting the water body temperature local time series feature kernel convergence difference compensation encoding vector into the compensation feature importance scoring module based on the neural network layer to obtain the water body temperature local time series feature kernel convergence compensation factor, including: Calculating the ratio between the Euclidean norm of the implicit correlation feature vector of the local time series of the water body temperature and the Euclidean norm of the coarse-grained aggregated coding vector of the local time series feature of the water body temperature; if the ratio is less than 1, adding one to the ratio and calculating the base-2 logarithmic value as the bias term of the neural network layer of the compensation feature importance scoring module; If the ratio is greater than or equal to 1, the ratio is used as a bias term of the neural network layer of the compensatory feature importance scoring module.

6. The error checking method for a smart water meter according to claim 5, characterized in that: The method of performing water consumption prediction based on the water body temperature time series integral implicit coding vector and the water body pressure time series integral implicit coding vector to obtain the initial water consumption includes: Inputting the water body temperature time series integral implicit coding vector and the water body pressure time series integral implicit coding vector into a water consumption prediction module based on a decoder to obtain a first water consumption prediction value and a second water consumption prediction value respectively; The average of the first water consumption prediction value and the second water consumption prediction value is calculated to obtain the initial water consumption.

7. The error checking method for a smart water meter according to claim 6, characterized in that: Based on historical monthly water consumption data of the month corresponding to the target time period, the initial water consumption is corrected to obtain a second water consumption, including: Extracting the historical monthly water consumption of the month information corresponding to the target time period to obtain a reference set of historical monthly water consumption; Determining a water consumption correction factor based on the reference set of historical monthly water consumption; The water consumption correction coefficient is multiplied by the initial water consumption to obtain the second water consumption.

8. The error checking method for a smart water meter according to claim 7, characterized in that: Based on the reference set of historical monthly water consumption, a water consumption correction factor is determined, including: The reference set of historical monthly water consumption is input into a feature encoding-decoding module based on an LSTM model to obtain the water consumption correction coefficient.

9. An error checking system for a smart water meter, characterized in that: include: A monitoring parameter acquisition module is used to obtain monitoring parameters of the multifunctional smart water meter within a target time period, wherein the monitoring parameters include: date information of the target time period, actual water consumption measurement value, and time series of temperature and pressure of the water body flowing through the multifunctional smart water meter; an initial water consumption determination module, configured to determine the initial water consumption based on a time series of temperature and pressure of the water flowing through the multifunctional smart water meter; an initial water consumption correction module, configured to correct the initial water consumption based on historical monthly water consumption data of the month information corresponding to the target time period to obtain a second water consumption; an error checking module, configured to use the actual water consumption measurement value during the target time period as a first water consumption, and perform an error check on the multifunctional smart water meter based on a difference between the first water consumption and the second water consumption; Wherein, the initial water consumption determination module includes: Arrange the time series of the temperature and pressure of the water flowing through the multifunctional smart water meter to obtain a time series of the water temperature and a time series of the water pressure; Performing local time series feature extraction on the time series of the water body temperature and the time series of the water body pressure to obtain a time series of implicit correlation feature vectors of the local time series of the water body temperature and a time series of implicit correlation feature vectors of the local time series of the water body pressure; Inputting the time series of the local time series implicit correlation feature vector of the water body temperature and the time series of the local time series implicit correlation feature vector of the water body pressure into a time series integrator respectively to obtain a water body temperature time series integral implicit coding vector and a water body pressure time series integral implicit coding vector; Water consumption is predicted based on the water body temperature time series integral implicit coding vector and the water body pressure time series integral implicit coding vector to obtain the initial water consumption.

Images