Intelligent ammeter high-precision low-power consumption metering method and system

Abstract

The invention relates to the field of electric energy metering, and discloses a high-precision low-power consumption metering method for an intelligent electric meter, which comprises the following steps of: 1, predicting a power consumption mode in a future preset time period; step 2, dynamically calibrating a trigger sensitivity threshold value based on the power consumption mode; 3, monitoring a transient electrical event on the line in a low power consumption state according to the trigger sensitivity threshold; and step 4, in response to the monitored transient electrical event, executing a preset metering operation, and in the step 2, dynamically calibrating a trigger sensitivity threshold based on the power consumption mode, including: determining an expected load fluctuation from the power consumption mode. According to the method, the trigger threshold value can be actively increased in the time period when stable power utilization is predicted so as to reduce the sensitivity of the system to tiny fluctuation, so that unnecessary wakeup and calculation are avoided, and the duration time of a low-power-consumption monitoring state is prolonged to the maximum extent.

Description

Technical Field

[0001] The present invention relates to the technical field of electric energy metering, and in particular to a high-precision and low-power consumption metering method and system for smart electric meters. Background Art

[0002] With the development of smart grid technology, the capabilities of energy metering devices have expanded far beyond traditional energy accumulation. Modern smart meters are required to provide more refined electricity usage data to support advanced applications such as demand-side management, load identification, fault diagnosis, and energy efficiency analysis. These capabilities rely on the meter's ability to accurately monitor and capture transient electrical events on the line, such as the start and stop of various electrical appliances and switching between operating modes. The waveform characteristics of these events contain rich load information. To ensure that critical information is not missed and to obtain sufficient resolution for feature analysis, metering chips typically need to continuously sample voltage and current signals at frequencies of several kilohertz (kHz) or even higher.

[0003] However, as a public metering device deployed on a large scale, smart meters are subject to strict power consumption restrictions. Maintaining high-frequency sampling and high-performance computing for a long time will cause the meter's own power consumption to exceed the standard, failing to meet relevant industry standards and energy efficiency requirements, and increasing the operating costs of power grid companies. Therefore, how to minimize the average power consumption of the meter while ensuring high-precision and high-fidelity capture of key electrical events has become a technical problem that needs to be solved in this field. To meet this challenge, existing technologies generally adopt a low-power operating mode based on event triggering. In this mode, the system is in sleep or standby state most of the time. Only when it detects that the change in a certain electrical parameter (such as current or power) exceeds a preset fixed threshold, it is awakened and switched to high-frequency sampling state to perform precise metering.

[0004] While this approach achieves a certain balance between power consumption and performance, its inherent flaws are also significant. Load behavior in power systems is highly time-varying and complex, with normal load fluctuations varying significantly across different times of the day and between residential and industrial users. The fixed trigger thresholds used in existing technologies are unable to adapt to this dynamic power consumption environment. If the fixed threshold is set low to ensure sensitivity to minor events, normal, non-critical load fluctuations in the line during periods of high power consumption will frequently and unnecessarily wake the system, significantly increasing meter power consumption and defeating the purpose of low-power mode. Conversely, if the threshold is set high to avoid frequent interruptions during periods of high power consumption, many low-power but important electrical events (such as the startup of standby appliances) will be ignored during periods of stable power consumption because they fail to meet the trigger conditions, resulting in the permanent loss of critical data. Therefore, due to the static and non-adaptive nature of monitoring sensitivity, existing technologies have been unable to achieve an ideal balance between power consumption and metering integrity, and have significant limitations. Summary of the Invention

[0005] The purpose of the present invention is to provide a high-precision, low-power metering method and system for smart meters, which solves the problem that the existing technology uses a fixed trigger threshold and cannot strike a balance between power consumption and metering event capture integrity.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a high-precision and low-power consumption measurement method for a smart meter, comprising the following steps: Step 1: predicting the electricity consumption pattern within a preset time period in the future; Step 2: dynamically calibrating a trigger sensitivity threshold based on the power usage pattern; Step 3: monitoring transient electrical events on the circuit in a low power consumption state according to the trigger sensitivity threshold; Step 4: In response to detecting the transient electrical event, executing a preset metering operation.

[0007] Preferably, in step 2, dynamically calibrating a trigger sensitivity threshold based on the power usage pattern includes: determining expected load fluctuations from the electricity usage pattern; Furthermore, the trigger sensitivity threshold is set according to the load fluctuation, so that the trigger sensitivity threshold is inversely correlated with the load fluctuation.

[0008] Preferably, the trigger sensitivity threshold is determined by the following formula: Among them, T base is the basic trigger threshold, V pis the load fluctuation, and α is the preset sensitivity calibration coefficient.

[0009] Preferably, in step 4, in response to monitoring the transient electrical event, performing a preset metering operation includes: capturing waveform data containing the transient electrical event with high fidelity; The waveform data is analyzed using a preset feature recognition model to identify the load type.

[0010] Preferably, in the step 4, after identifying the load type, the step further includes: Based on the identified load type, an optimal metering strategy is selected and executed from a preset strategy library.

[0011] Preferably, the analyzing the waveform data using a preset feature recognition model to identify the load type includes: extracting a feature vector comprising at least one electrical parameter from the waveform data; The feature vector is input into the feature recognition model to obtain the load type.

[0012] Preferably, the characteristic vector includes the current crest factor: It is determined by the following formula: Among them, I peak is the peak current, I rms is the effective value of current.

[0013] Preferably, after capturing the waveform data containing the transient electrical event with high fidelity, the method further comprises: The waveform data is used to update and optimize the prediction model or the feature recognition model used in step one.

[0014] Preferably, in step 1, predicting the electricity usage pattern within a preset time period in the future is achieved by analyzing the stored historical electricity usage data of the smart meter.

[0015] The present invention also provides a high-precision and low-power consumption metering system for smart meters, comprising: The power consumption pattern prediction module is used to predict the power consumption pattern within a preset time period in the future; A dynamic calibration module, configured to dynamically calibrate a trigger sensitivity threshold based on the power usage pattern; An event monitoring module is used to monitor transient electrical events on the line in a low power consumption state according to the trigger sensitivity threshold; and a metering execution module is used to execute a preset metering operation in response to monitoring the transient electrical event.

[0016] In summary, the present invention includes at least one of the following beneficial technical effects: 1. This invention achieves proactive power management by first predicting future power usage patterns and dynamically calibrating the trigger sensitivity threshold for transient electrical events based on this prediction. Compared to technologies that reactively adjust based solely on current load levels, this invention proactively raises the trigger threshold during periods of predicted stable power usage to reduce the system's sensitivity to minor fluctuations. This avoids unnecessary wake-ups and calculations, maximizes the duration of low-power monitoring states, and significantly reduces the system's average power consumption.

[0017] 2. This invention ensures high-precision measurement of critical electrical events by employing an event-triggered mechanism based on dynamic thresholds. This method can respond instantly to changes in load conditions and capture the complete event waveform with high fidelity, rather than reacting only after detecting that power exceeds a fixed threshold. This ability to instantly capture transient processes effectively avoids oversight of short-duration, high-frequency energy, such as motor startup surges and switching power supply surges, caused by sampling frequency switching lags, thereby improving measurement accuracy in complex power usage scenarios.

[0018] 3. By introducing load type identification and model feedback optimization mechanisms, this invention endows the system with excellent adaptability and long-term stability. The system not only matches the optimal metering strategy based on the identified specific load type, achieving refined resource allocation, but also continuously updates its prediction and identification models using newly captured waveform data. This closed-loop self-optimization capability enables the meter to adapt to changes in user electricity usage habits or the addition of new appliances, maintaining high efficiency and accuracy throughout its lifecycle, enhancing the system's intelligence and environmental adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flow chart of the method of the present invention; Figure 2 It is a schematic diagram of the main framework of the present invention. DETAILED DESCRIPTION

[0020] The following is combined with Figure 1 -Attached Figure 2 , the present invention is described in further detail.

[0021] This invention provides a high-precision, low-power metering method for smart meters. This method implements proactive power management by first predicting future power usage patterns and then dynamically calibrating the trigger sensitivity threshold for transient electrical events based on this prediction. Compared to technologies that reactively adjust based solely on current load levels, this method proactively raises the trigger threshold during periods of predicted stable power usage to reduce the system's sensitivity to minor fluctuations. This avoids unnecessary wake-ups and calculations, maximizes the duration of the low-power monitoring state, and significantly reduces the system's average power consumption.

[0022] like Figure 1 As shown, the high-precision and low-power consumption measurement method of the smart meter may include the following steps: Step 1: predicting the electricity consumption pattern within a preset time period in the future; Step 2: Dynamically calibrate a trigger sensitivity threshold based on the power usage pattern; Step 3: Monitor transient electrical events on the line in a low power state according to a trigger sensitivity threshold; Step 4: In response to detecting a transient electrical event, a preset metering operation is performed.

[0023] The method can be implemented in a smart meter device. In addition to conventional hardware components such as voltage and current sampling circuits, an analog-to-digital converter (ADC), and a communication interface, the device also includes a processor and a connected memory. The memory stores computer program instructions, which the processor executes to implement the method provided by the present invention.

[0024] In this embodiment, the system specifically includes the following functional modules: a power consumption pattern prediction module, a dynamic calibration module, an event monitoring module, and a metering execution module. These modules are functional modules that can be implemented by the processor by executing specific program segments.

[0025] The power consumption pattern prediction module is configured to analyze historical power consumption data stored in the memory to predict the power consumption pattern within a preset time period in the future. The output of this module is power consumption pattern data that includes information such as expected load fluctuation.

[0026] The dynamic calibration module, whose input is connected to the output of the power usage pattern prediction module, is configured to receive power usage pattern data and, based on the data, calculate and calibrate a trigger sensitivity threshold for subsequent event monitoring.

[0027] The event monitoring module has an input connected to the output of the dynamic calibration module. This module is configured to receive a trigger sensitivity threshold and, based on this threshold, continuously monitor transient electrical parameters on the power line during low-power operation. When the monitored parameter exceeds the threshold, the module outputs an event trigger signal.

[0028] The metering execution module, whose input is connected to the output of the event monitoring module, is configured to switch from a low-power state to a full-function operation state upon receiving an event trigger signal, and perform subsequent operations such as high-fidelity waveform capture, load identification, and metering strategy selection.

[0029] When the system implements the high-precision and low-power consumption measurement method for smart meters, its overall workflow may include the following steps: S100, the power consumption pattern prediction module is started to predict the future power consumption pattern; S200, the dynamic calibration module sets the trigger sensitivity threshold according to the prediction result of S100; S300, the event monitoring module performs low power transient event monitoring according to the threshold value set in S200; S400 , the metering execution module responds to the trigger signal of S300 and executes a preset metering operation.

[0030] In specific operation, the output data of the power consumption pattern prediction module is transmitted to the dynamic calibration module. After the dynamic calibration module processes the data, it writes the generated trigger sensitivity threshold parameters into the specified register or memory address. When the event monitoring module is running, it reads the threshold parameters in the register or memory address as its judgment basis. Once the judgment conditions of the event monitoring module are met, it will transfer control to the metering execution module through an interrupt or status flag to complete the subsequent precise metering task. Through the orderly data transmission and functional collaboration between the above modules, a complete execution link of the method of the present invention is formed.

[0031] In a specific embodiment, the method begins with a prediction of future electricity usage behavior. The electricity usage pattern prediction module is configured to perform the prediction task periodically, for example, once every hour. The module retrieves historical electricity usage data from the memory, which is time series data containing parameters such as timestamps, active power, and effective current values. The module uses a preset prediction model to analyze the historical electricity usage data to output the electricity usage pattern for a preset time period in the future. The electricity usage pattern is a collection of one or more parameters, including at least one quantitative parameter for characterizing the severity of the expected load changes, namely load volatility.

[0032] After obtaining the power consumption pattern, the dynamic calibration module dynamically calibrates a trigger sensitivity threshold based on the power consumption pattern. Specifically, the module extracts the load fluctuation parameter (V p ), and according to a preset functional relationship, use the load fluctuation parameter to calculate and set a trigger sensitivity threshold (T d ). The functional relationship is configured so that T d With V p When the predicted load fluctuation is high, the system sets a lower trigger sensitivity threshold to improve the ability to detect small changes; when the predicted load fluctuation is low, the system sets a higher trigger sensitivity threshold to reduce the probability of false triggering due to non-power events such as circuit noise.

[0033] In one embodiment, the functional relationship can be expressed by the following formula: Among them, T d is the trigger sensitivity threshold finally generated; T base V is a basic trigger threshold preset by a system and used when there is no prediction information; p is the quantitative value of load fluctuation extracted from the power consumption pattern; α is a preset, non-negative sensitivity calibration coefficient used to adjust the weight of the load fluctuation on the trigger threshold. After the calculation is completed, the dynamic calibration module will obtain T d The value is written to a specified register or memory address for use in subsequent steps.

[0034] After setting the trigger sensitivity threshold, the system enters a low-power operating state. In this state, most of the processor's computational units, as well as the high-power digital signal processing units in the metering chip, are in a dormant or clock-gated state. The event monitoring module is activated, which continuously monitors one or more transient electrical parameters on the power line, such as the instantaneous rate of change (di / dt) of the current, using a low-power comparator circuit.

[0035] The event monitoring module continuously compares the absolute value of the instantaneous rate of change of the monitored current with the trigger sensitivity threshold (T d ) for comparison. In a discrete digital system, this comparison operation is specifically: Among them, i[k] is the instantaneous value of the current at the current sampling point, i[k-1] is the instantaneous value of the current at the previous sampling point, Δt s is the sampling time interval in the low power monitoring state. When the above conditions are met, the event monitoring module determines that a transient electrical event has occurred and immediately generates an event trigger signal.

[0036] In response to the event trigger signal, the metering execution module is activated and switches the processor and metering chip from a low-power state to a fully functional state. Subsequently, the metering execution module performs a series of pre-set metering operations, including but not limited to: capturing voltage and current waveform data at a high sampling frequency and high fidelity for a period of time before and after the event, analyzing the captured waveform data to identify the load type, and selecting and executing an optimal metering strategy based on the identification results. The specific implementation of these operations will be described in detail later.

[0037] In one specific embodiment, the dynamic calibration of trigger sensitivity thresholds essentially establishes a direct mapping from macro-level power usage predictions to micro-level event monitoring sensitivities. This process, performed by a dynamic calibration module, primarily involves determining a key quantitative parameter from the predicted power usage pattern: expected load volatility.

[0038] Load volatility is a numerical value that characterizes the degree of load fluctuation within a predetermined time period. This parameter can be calculated in a variety of ways. For example, it can be defined as the variance of the power signal within the forecast time window, or as the mathematical expectation of the square of the instantaneous rate of change of the current signal. By calculating this parameter, the system can transform an abstract electricity consumption trend into a precise, forward-looking quantitative indicator.

[0039] After determining the load fluctuation parameter, the system sets the trigger sensitivity threshold based on the parameter to achieve an inverse correlation between the two. This inverse correlation is an important technical feature of the present invention, and its technical effect is that when the system predicts that the load will fluctuate frequently and drastically in the future, the system will automatically set a lower trigger sensitivity threshold, making the event monitoring module more sensitive to small changes in the line, thereby ensuring that no important electrical events are missed; conversely, when it is predicted that future power consumption will be stable and load changes will be rare and gentle, the system will set a higher trigger sensitivity threshold to effectively filter out small disturbances caused by non-real power consumption behaviors such as inherent line noise, avoiding unnecessary system wake-up and computing resource consumption.

[0040] In a preferred embodiment of the present invention, the functional relationship used to determine the trigger sensitivity threshold includes several key parameters with clear technical definitions. For example, the basic trigger threshold is a system parameter preset based on the hardware characteristics of the meter. It represents the system's baseline monitoring sensitivity. Its value can be determined based on the circuit's inherent noise level or relevant metering specifications. It defines the upper limit or baseline of the dynamic threshold.

[0041] For example, the functional relationship also includes a sensitivity calibration coefficient. This coefficient is a configurable parameter used to adjust the weight of the load fluctuation on the final threshold (i.e., the steepness of the adjustment curve). By adjusting this coefficient, the method of the present invention can be adapted to different power grid environments or meet specific metering accuracy requirements. For example, in an industrial environment with high electrical noise, this coefficient can be appropriately reduced to reduce the severity of threshold fluctuations and enhance system stability.

[0042] Using an exponential decay function to construct this anti-correlation relationship can achieve a nonlinear and efficient regulation effect. When the load volatility parameter increases from zero, the trigger threshold will drop rapidly, allowing the system to quickly adapt to the transition from stable to active power consumption. When the load volatility parameter is very large, the rate of decrease of the threshold will slow down and approach a minimum value rather than zero. This ensures that the system still maintains the necessary monitoring capabilities in extremely active power consumption scenarios, avoids the situation where the threshold is set to an invalid value, and enhances the robustness of the method. After the dynamic calibration module sets and loads the trigger sensitivity threshold, the system enters a low-power operating state. In this state, to minimize energy consumption, the core units in the processor that perform the main computing functions and the circuit modules in the metering chip responsible for high-frequency sampling and complex digital signal processing (DSP) are placed in clock gating or deep sleep mode. At this time, only the minimum functional units required to maintain basic system activities remain running, including a low-power timer and event monitoring module.

[0043] In this low-power state, the event monitoring module is responsible for continuously monitoring one or more preset transient electrical parameters on the power line. In one specific embodiment, the module samples the line current using a low-power analog-to-digital converter (ADC) at a sampling rate much lower than the normal metering frequency (e.g., hundreds of hertz) to obtain a discrete sequence of instantaneous current values ​​(i[k]).

[0044] The core functionality of this module is to calculate the instantaneous rate of change of current in real time and compare it to the trigger sensitivity threshold stored in a designated hardware register. The instantaneous rate of change is calculated by dividing the difference between the current values ​​at two adjacent sampling points by the sampling interval. When the absolute value of the instantaneous rate of change exceeds the trigger sensitivity threshold, the trigger condition is met.

[0045] The satisfaction of this trigger condition indicates that a transient electrical event with sufficient energy or characteristics has occurred on the line and exceeds the normal fluctuation range under the current prediction model. This event indicates a change in the state of the power load, such as the startup or shutdown of a high-power appliance.

[0046] Once the event monitoring module detects that a trigger condition has been met, it immediately generates a hardware interrupt signal. This interrupt signal is transmitted to the system's processor. Upon receiving the interrupt signal, the processor immediately wakes up from its sleep state and simultaneously removes clock gating from the metering chip and other related peripherals, allowing the entire system to resume full functionality in a very short time, ready for subsequent high-fidelity waveform capture and precise metering operations. This approach allows the system to maintain extremely low power consumption levels most of the time, mobilizing all system resources only when electrical events of actual metering significance occur, thus achieving an effective balance between power consumption and accuracy.

[0047] When the event monitoring module outputs an event trigger signal, the metering execution module is activated, and the system switches from a low-power state to a fully functional operating state. Thereafter, the module performs a series of pre-set metering operations to accurately analyze and measure the transient electrical event.

[0048] First, the metering execution module performs the operation of capturing waveform data with high fidelity. To ensure that complete information about the transient electrical event is captured, the system uses a circular buffer to continuously store low-frequency sampling data from the recent period. After receiving the event trigger signal, the system immediately increases the sampling frequency of the analog-to-digital converter (ADC) to a preset high sampling frequency (for example, several thousand hertz or higher) and continues to collect data for a period of time. At the same time, the system splices the data stored in the circular buffer before the trigger point with the newly collected high-frequency sampling data after the trigger point. In this way, the system obtains a section of high-fidelity voltage and current waveform data that covers the entire process before, during, and after the event.

[0049] After obtaining high-fidelity waveform data, the metering execution module uses a pre-configured feature recognition model to analyze the waveform data to identify the load type that triggered the event. This recognition process specifically involves extracting a feature vector containing at least one electrical parameter from the high-fidelity waveform data. This feature vector is a multidimensional number that describes the electrical characteristics of the load.

[0050] In one embodiment, the feature vector may include a parameter of current crest factor (CF). The current crest factor is defined as the ratio of the current peak value to the current effective value, which is determined by the following formula: Among them, I peak is the peak current value extracted from the waveform data, I rms It is the effective value of current calculated from the same waveform data. This parameter plays a significant role in distinguishing different types of loads. For example, the current crest factor of a pure resistive load is close to Electronic devices containing switching power supplies usually have higher current crest factors.

[0051] After being extracted, the feature vector is sent as input to a pre-set feature recognition model (e.g., a pre-trained support vector machine or neural network model). The model processes the input feature vector and outputs an identification result, which is the type of load that caused the transient event, such as "inductive load," "capacitive load," or "non-linear load."

[0052] After identifying the load type, the metering execution module further performs an optimal metering strategy selection operation based on the identification result. A strategy library is preset within the system, which stores the mapping relationship between multiple metering strategies and different load types. The metering execution module queries and selects an optimal metering strategy corresponding to the identified load type from the strategy library and executes the strategy immediately. For example, if it is identified as a motor start-up, the optimal strategy may define maintaining high-frequency sampling for the next few seconds to accurately calculate its starting energy; if it is identified as a resistive load, the strategy may be defined as returning to a lower sampling frequency after completing an accurate power calculation.

[0053] The present invention also includes a feedback optimization mechanism. High-fidelity waveform data that has been successfully identified as a load type constitutes a high-quality, labeled data sample. This data sample can be stored and used to update and optimize the prediction model or feature recognition model used by the power consumption pattern prediction module. For example, by adding these new data samples to the training set for incremental training of the feature recognition model, its recognition accuracy can be continuously improved, enabling the system to adapt and evolve.

[0054] In another specific embodiment, a more detailed implementation method is provided for the technical solution of the present invention.

[0055] The event monitoring module can monitor transient electrical parameters, including the instantaneous rate of change of current and the rate of change of active power. By monitoring the rate of change of power, the system can directly respond to changes in load energy consumption, providing an alternative or supplementary triggering basis for certain application scenarios.

[0056] The feature recognition model can specifically be a support vector machine (SVM) model. This model is trained offline using a large number of labeled waveform data samples to establish a hyperplane decision boundary for distinguishing different load types. After training, the model parameters are fixed and stored in memory as firmware, allowing the metering execution module to directly call them during runtime.

[0057] The optimal metering strategy library can specifically be a lookup table. This lookup table exists as a static data structure in memory and establishes a direct mapping relationship between load type identifiers and one or more metering program entry addresses. When the metering execution module queries this lookup table based on the output of the feature recognition model (i.e., a specific load type identifier), it can directly obtain the starting address of the specific metering subroutine to be executed, thereby enabling rapid switching and execution of metering strategies.

[0058] The specific implementation process of the self-optimization mechanism includes the following steps: a counter is installed within the system to accumulate the number of newly acquired and successfully labeled waveform data samples. When the counter reaches a preset threshold (for example, 100 samples), the processor will start a model retraining program as a background task during an idle period when system resources are low. This program incorporates the newly accumulated sample data into the original training data set, iteratively updates the internal parameters of the feature recognition model, and overwrites the original model parameters in memory with the updated model parameters. This process does not require external intervention, allowing the device's recognition accuracy to gradually improve with increasing operating time.

[0059] The high-precision and low-power consumption metering system for smart electricity meters described below and the high-precision and low-power consumption metering method for smart electricity meters described above may refer to each other.

[0060] Please see the attached Figure 2 The present invention also provides a high-precision and low-power consumption metering system for smart meters, comprising: The power consumption pattern prediction module is used to predict the power consumption pattern within a preset time period in the future; A dynamic calibration module, used to dynamically calibrate a trigger sensitivity threshold based on the power consumption pattern; An event monitoring module is used to monitor transient electrical events on the line in a low power consumption state according to a trigger sensitivity threshold; and a metering execution module is used to execute a preset metering operation in response to detecting a transient electrical event.

[0061] The device of this embodiment can be used to execute the above method embodiment, and its principles and technical effects are similar, so they will not be repeated here.

[0062] While 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 these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A high-precision and low-power consumption measurement method for smart meters, characterized in that: The following steps are involved: Step 1: predicting the electricity consumption pattern within a preset time period in the future; Step 2: dynamically calibrating a trigger sensitivity threshold based on the power usage pattern; Step 3: monitoring transient electrical events on the circuit in a low power consumption state according to the trigger sensitivity threshold; Step 4: In response to detecting the transient electrical event, executing a preset metering operation.

2. The high-precision and low-power consumption measurement method for smart meters according to claim 1, characterized in that: In the second step, dynamically calibrating a trigger sensitivity threshold based on the power usage pattern includes: determining expected load fluctuations from the electricity usage pattern; Furthermore, the trigger sensitivity threshold is set according to the load fluctuation, so that the trigger sensitivity threshold is inversely correlated with the load fluctuation.

3. The high-precision and low-power consumption measurement method for smart meters according to claim 2, characterized in that: The trigger sensitivity threshold is determined by the following formula: Among them, T base is the basic trigger threshold, V p is the load fluctuation, and α is the preset sensitivity calibration coefficient.

4. The high-precision and low-power consumption measurement method for smart meters according to claim 1, characterized in that: In step 4, in response to detecting the transient electrical event, performing a preset metering operation includes: capturing waveform data containing the transient electrical event with high fidelity; The waveform data is analyzed using a preset feature recognition model to identify the load type.

5. The high-precision and low-power consumption measurement method for smart meters according to claim 1, characterized in that: In the fourth step, after the load type is identified, the following steps are further included: Based on the identified load type, an optimal metering strategy is selected and executed from a preset strategy library.

6. The high-precision and low-power consumption measurement method for smart meters according to claim 4, characterized in that: The method of analyzing the waveform data using a preset feature recognition model to identify the load type includes: extracting a feature vector comprising at least one electrical parameter from the waveform data; The feature vector is input into the feature recognition model to obtain the load type.

7. The high-precision and low-power consumption measurement method for smart meters according to claim 1, characterized in that: The characteristic vector includes a current crest factor; It is determined by the following formula: Among them, I peak is the peak current, I rms is the effective value of current.

8. The high-precision and low-power consumption measurement method for smart meters according to claim 4, characterized in that: After the waveform data containing the transient electrical event is captured with high fidelity, the method further includes: The waveform data is used to update and optimize the prediction model or the feature recognition model used in step one.

9. The high-precision and low-power consumption measurement method for smart meters according to claim 1, characterized in that: In the step 1, the power consumption pattern within a preset time period in the future is predicted by analyzing the stored historical power consumption data of the smart meter.

10. A high-precision, low-power consumption metering system for smart meters, configured to execute the high-precision, low-power consumption metering method for smart meters according to any one of claims 1 to 9, characterized in that: include: The power consumption pattern prediction module is used to predict the power consumption pattern within a preset time period in the future; A dynamic calibration module, configured to dynamically calibrate a trigger sensitivity threshold based on the power usage pattern; An event monitoring module, configured to monitor transient electrical events on the circuit in a low power consumption state according to the trigger sensitivity threshold; And, a metering execution module is used to execute a preset metering operation in response to detecting the transient electrical event.

Images