A method and system for energy meter sampling and communication-based collaborative energy consumption reduction control

Abstract

This invention relates to the field of communication control technology for electricity meters, and particularly to a method and system for collaborative power consumption reduction control of electricity meter sampling communication. The method quantifies data timeliness by constructing an information value decay function, and obtains the collaborative wake-up urgency based on a weighted fusion of event levels output by multi-level event detection and a comprehensive information residual value coefficient. This urgency is compared with a wake-up threshold to determine communication wake-up, solving the problem of the disconnect between the value of sampled data and the timeliness of communication reporting. While ensuring timely transmission of key data, it significantly reduces the number of invalid wake-ups and substantially reduces the power consumption of the communication module. Through a bidirectional closed-loop adjustment mechanism of weight coefficients based on the proportion of effective data feedback, the weights of event level, information value, and cache occupancy rate can be adaptively optimized according to historical reporting results. When there are too many invalid reports, the weight of the event level is increased; when there is a possibility of missed reports, the weight is adjusted in the opposite direction, achieving continuous self-learning of the wake-up decision strategy.

Description

Technical Field

[0001] This invention relates to the field of communication control technology for electricity meters, and in particular to a method and system for energy meter sampling communication-coordinated energy consumption reduction control. Background Technology

[0002] Smart meters acquire analog voltage and current signals through sampling circuits, which are then converted by a metering chip to complete energy metering and data acquisition. The data is then periodically reported to a remote master station via its internal communication module. To ensure the integrity and accuracy of energy metering data, existing smart meters generally employ a periodic wake-up mechanism, where both the sampling module and communication module are periodically woken up at a fixed frequency to collect and report data. For example, patent CN118604442B discloses an energy metering method, device, and smart meter that can flexibly adjust the sampling frequency according to the power load, reducing the sampling frequency to save energy under low fluctuation conditions. However, this method mainly focuses on adjusting the wake-up frequency of the sampling module, while the wake-up reporting cycle of the communication module still operates independently according to a pre-set fixed cycle.

[0003] However, the data transmission power consumption of smart meters largely depends on the number of times the communication module is woken up and the amount of data transmitted. Each time it is woken up and performs a wireless transmission operation, its power consumption can reach tens or even hundreds of milliamps, far exceeding the power consumption of MCUs and ADCs in operating mode. Furthermore, the value of sampled data is not constant. When the power load tends to stabilize, data from multiple consecutive samples is relatively redundant, with minimal information increment. However, when the load undergoes rapid changes, such as the data from the start-up and shutdown of high-power equipment, it carries extremely critical transient change information. Its value decays exponentially over time. If it is not reported in time, the fault or abnormal information reflected in that frame of data will be overwhelmed by the normal data from subsequent sampling points within the next few seconds, thus losing its unique diagnostic value.

[0004] Therefore, the communication wake-up cycle setting of the existing periodic wake-up mechanism must strike a balance between two objectives: "frequent enough to avoid loss of critical data" and "as sparse as possible to reduce power consumption." If the communication wake-up cycle is set too sparsely, although power consumption can be reduced, the high-value transient data collected by the sampling module will lose its timeliness due to untimely reporting, and the sampling buffer data may overflow due to the long waiting period. If the communication wake-up cycle is set too frequently, the communication module will generate a large number of unnecessary wake-up operations, transmitting a large amount of redundant data during stable load periods, resulting in energy waste. Therefore, it is difficult to simultaneously achieve the two objectives of low power consumption and high timeliness within the existing periodic wake-up framework. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a sampling and communication-coordinated power consumption reduction control method for electricity meters, which solves the technical problem that it is difficult to simultaneously achieve both low power consumption and high timeliness within the periodic wake-up framework of existing electricity meters.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for energy meter sampling and communication-coordinated energy consumption reduction control, the method comprising the following steps: S1. Continuously collect the instantaneous voltage and current values ​​of the electricity meter at fixed time intervals to construct a sampling data sequence; S2. The sampling data sequence of multiple consecutive sampling points is packaged into a sampling data packet according to the preset packaging rules and stored in the data buffer to be sent. Each sampling data packet has a unique generation time. S3. Using a multi-level event detection model, the sampled data sequence is detected and the highest event level is output. S4. Construct an information value decay function and calculate the comprehensive information residual value coefficient of the data buffer to be sent based on the time elapsed since the generation time of each sampled data packet in the data buffer to be sent. S5. Construct a collaborative wake-up urgency function, which weights and integrates the current event level with the comprehensive information residual value coefficient, and outputs the collaborative wake-up urgency value. S6. Compare the collaborative wake-up urgency value with the preset wake-up threshold, and decide whether to wake up the communication module to report data based on the comparison result.

[0007] Preferably, the multi-level event detection model includes: The L1 level event detection model is used to calculate the absolute value of the difference between the current voltage sample value and the current sample value and the previous sample value, respectively, and to determine that an L1 level event is triggered when at least one absolute value of the difference is greater than the corresponding difference threshold. The L2-level event detection model is used to calculate the mean of the effective voltage value and the mean of the effective current value within a sliding window with the current time as the endpoint, and to determine that an L2-level event is triggered when the rate of change of at least one mean relative to the corresponding preset benchmark reaches the corresponding rate of change threshold. The L3 event detection model is used to record the cumulative number of L2 events occurring per unit time, and determines that an L3 event is triggered when the number of L2 events exceeds a preset cumulative number threshold. The comprehensive judgment model is used to output the event level at the current moment based on the event levels that have been triggered. If no event level has been triggered, it outputs no event level. The event levels of L1, L2 and L3 events increase sequentially.

[0008] Preferably, in the L1-level event detection model, the difference threshold includes a voltage difference threshold that is compared with the absolute value of the difference between the voltage sample value and the previous sample value, and a current difference threshold that is compared with the absolute value of the difference between the current sample value and the previous sample value. The current difference threshold is positively correlated with the standard deviation of the current sample values ​​at multiple times prior to the current time.

[0009] Preferably, in the L2-level event detection model, the rate of change is a relative rate of change; The relative rate of change of the mean effective voltage value is determined based on the ratio of the absolute value of the difference between the mean effective voltage value and the voltage reference value to the voltage reference value; the calculation formula is as follows: ; In the above formula, express The relative rate of change of the mean effective value of voltage at time step express The average effective voltage value at time t. express The voltage reference value at that moment; The relative rate of change of the mean effective value of the current is determined based on the ratio of the absolute value of the difference between the mean effective value of the current and the reference value of the current to the reference value of the current: ; In the above formula, express The relative rate of change of the mean effective value of the current at any given time. express The mean effective value of the current at time t. express The current reference value at that moment; The voltage and current reference values ​​are determined based on the average of the effective voltage and the average of the effective current values ​​at all sampling times between the two most recent L2 level event trigger times, respectively, and their expressions are as follows: ; In the above formula, This indicates the number of sampling times between the two most recent L2 level event trigger times. This represents the time interval between two adjacent sampling times.

[0010] Preferably, the residual value coefficient of the comprehensive information is determined based on the minimum value of the residual value coefficient of each sampled data packet in the data buffer to be sent; the residual value coefficient of each sampled data packet is exponentially negatively correlated with the time elapsed since its generation.

[0011] Preferably, the exponential term of the formula for calculating the residual value coefficient of information also includes a decay time constant that is positively correlated with it. This decay time constant is negatively correlated with the triggering frequency of L2 and L3 events per unit time between preset high-frequency and low-frequency thresholds, and its calculation formula is as follows: ; In the above formula, Represents the decay time constant. and These represent the maximum and minimum values ​​of the decay time constant, respectively. and Preset high-frequency threshold and low-frequency threshold, This indicates the triggering frequency of L2 and L3 events per unit of time. Greater than or equal to the high frequency threshold hour, Pick ,when Less than or equal to the low frequency threshold hour, Pick .

[0012] Preferably, the collaborative wake-up urgency function is at least a weighted sum of three terms: the event level mapping function term, the difference between 1 and the comprehensive information residual value coefficient, and the ratio of the data length of the data buffer to the maximum capacity of the buffer at the current moment. Let the weight coefficients of the three terms be respectively... , and It dynamically adjusts based on the feedback results after the communication module performed data reporting during the last wake-up, and the specific steps are as follows: Obtain the proportion of valid data in the feedback results after the last wake-up communication module executes data reporting; If the proportion of valid data is less than or equal to the preset lower limit of the valid data proportion threshold, then the proportion will increase by the preset step size. and reduce and ; If the proportion of valid data is greater than or equal to the preset upper limit of the valid data proportion threshold, then the proportion will decrease by a preset step size. and increase and ; If the proportion of valid data is between the preset lower and upper limits of the valid data proportion threshold, then maintain... , and constant.

[0013] The present invention also provides a power meter sampling and communication collaborative energy saving control system, including a processor and a memory, wherein the memory is used to store a computer program, and the computer program is executed by the processor to implement the power meter sampling and communication collaborative energy saving control method.

[0014] By employing the above technical solution, the present invention provides a method for energy meter sampling and communication-coordinated energy consumption reduction control, which has at least the following beneficial effects: 1. This invention quantifies data timeliness by constructing an information value decay function, and obtains the collaborative wake-up urgency by weighted fusion of the event level output by multi-level event detection and the comprehensive information residual value coefficient. After comparing with the wake-up threshold, a communication wake-up decision is made. This solves the problem of the disconnect between the value of sampled data and the timeliness of communication reporting, and realizes communication reporting driven by the actual value of data. While ensuring the timely transmission of key data, it significantly reduces the number of invalid wake-ups and significantly reduces the power consumption of the communication module.

[0015] 2. This invention adopts a multi-level event detection model and sets a threshold that is positively correlated with the historical standard deviation to capture load changes at different time scales, thus avoiding missed or false detections caused by a single fixed threshold and improving the accuracy and robustness of event detection.

[0016] 3. This invention uses a two-way closed-loop adjustment mechanism for weight coefficients based on the proportion of effective data feedback to enable the three weights of event level, information value and cache occupancy rate to be adaptively optimized according to the historical reporting effect. When there are too many invalid reports, the weight of event level is increased, and when there is a possibility of missed reports, the weight is adjusted in the opposite direction. This realizes the continuous self-learning of the wake-up decision strategy and maintains the optimal balance between power consumption and real-time performance during long-term operation. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart of the energy meter sampling communication collaborative energy consumption reduction control method of the present invention. Detailed Implementation

[0018] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. This will allow for a full understanding of how the present application uses technical means to solve technical problems and achieve technical effects, and to facilitate its implementation.

[0019] To address the technical challenge of simultaneously achieving low power consumption and high timeliness within the existing periodic wake-up framework of electricity meters, this embodiment provides a collaborative power consumption reduction control method for electricity meter sampling and communication. By establishing a multi-level event judgment model and an information value quantification evaluation system, the degree of change and timeliness decay law of the sampled data can drive the communication wake-up decision in real time, realizing the linkage optimization control between the sampling end and the communication end to achieve both low power consumption and high timeliness. This embodiment takes a single-phase smart electricity meter of a residential user as an example. This electricity meter is powered by a battery or a low-power power supply, and requires collaborative power consumption reduction control for the two high-power modules of sampling and communication. The specific steps include the following: S1. Construct the sampled data sequence.

[0020] The core function of an electricity meter is to measure electrical energy, which requires continuously acquiring instantaneous values ​​of voltage and current. In this embodiment, the electricity meter's MCU controls the ADC module to continuously acquire instantaneous voltage values ​​at fixed time intervals. and instantaneous current value , such as Data is acquired at microsecond intervals, instantaneous voltage value. and instantaneous current value These represent the instantaneous voltage and current values ​​at the k-th sampling time, respectively. The collected raw data constitutes the sampling data sequence, which is the basic data source for all subsequent event detection and decision-making.

[0021] S2. Since the original sampling data is continuous streaming data, it cannot be directly sent to the data buffer to be sent for management. Therefore, it is necessary to encapsulate the sampling data sequence of multiple consecutive sampling points into independent sampling data packets according to a preset packaging rule and store them in the data buffer to be sent. Each sampling data packet has a unique generation time. In this embodiment, a fixed time window packaging rule can be adopted: every 200 milliseconds (i.e., 10 power frequency cycles) of sampling data is encapsulated into a sampling data packet. Each data packet contains all voltage and current sampling points within that time period and is assigned a unique generation time, which can be the sampling time of the first sampling point in the data packet.

[0022] S3. The current periodic wake-up mechanism has the drawback of being unable to detect the drastic degree of load changes. Therefore, a multi-level event detection model is used to detect the sampled data sequence and output the highest event level, thereby achieving the goal of detecting the drastic degree of load changes. In this embodiment, a three-level event detection model is designed, specifically including an L1-level event detection model, an L2-level event detection model, an L3-level event detection model, and a comprehensive judgment model, as detailed below: For the L1-level event detection model, considering that voltage and current are independent but highly correlated physical quantities, voltage fluctuations typically reflect grid-side events such as voltage sags, while current fluctuations reflect load-side events such as equipment startup. Detecting only a single physical quantity may miss crucial information carried by the other. Therefore, this embodiment calculates voltage and current separately, specifically calculating the absolute value of the difference between the current voltage sample value and the current sample value relative to the previous sample value. An L1-level event is triggered when at least one absolute value of the difference is greater than the corresponding difference threshold. Furthermore, the difference thresholds include a voltage difference threshold compared to the absolute value of the difference between the voltage sample value and the previous sample value, and a current difference threshold compared to the absolute value of the difference between the current sample value and the previous sample value. Specifically, the MCU calculates the instantaneous voltage value of the currently sampled value in real time. The instantaneous voltage value sampled at the previous sampling time absolute value of the difference and the currently sampled instantaneous current value The instantaneous current value sampled at the previous sampling time absolute value of the difference Its calculation formula can be expressed as: ; Let the absolute value of the difference The corresponding difference threshold is absolute value of difference The corresponding difference threshold is ,but or When this occurs, an L1 level event is determined to be triggered. Since the grid voltage is relatively stable, fluctuations are typically within ±10%. To ensure the requirements are met by a fixed threshold, it is generally set to 0.5% to 2% of the rated voltage of the electricity meter. In this embodiment, the rated voltage is 220V, so 1% is 2.2V. However, since the dynamic range of current is extremely large under both light and heavy load scenarios, typically ranging from tens of milliamps to tens of amperes, a fixed threshold cannot simultaneously accommodate both scenarios. When the load background current is small, such as only 50mA in standby mode, the standard deviation of the current sample value is small, therefore the current difference threshold should also be correspondingly small to capture minute changes of tens of milliamps. When the load background current is large, such as 15A when the air conditioner is running, the standard deviation of the current sample value is large, therefore the current difference threshold should also be correspondingly large to avoid normal fluctuations being misinterpreted as events. Therefore, the current difference threshold in this embodiment is positively correlated with the standard deviation of the current sample values ​​from multiple previous moments, and its expression can be expressed as: ; In the above formula, Indicates the current time The current difference threshold, This represents the standard deviation of the current sample values ​​over multiple time intervals prior to the current time. In this embodiment, the standard deviation of the current sample values ​​within 0.5 seconds prior to the current time is used. This represents a preset dimensionless proportionality coefficient, which is set to 3 in this embodiment. Thus, the L1-level event detection model can capture instantaneous changes in voltage and current between adjacent sampling points, corresponding to scenarios where voltage and current change rapidly, such as load switching operations or equipment start-up and shutdown.

[0023] For the L2-level event detection model, the L1-level event detection model only focuses on the instantaneous difference between adjacent sampling points. For slowly changing loads, such as a motor gradually starting over dozens of cycles, the absolute value of the current difference in each sampling interval may always be lower than the corresponding current difference threshold, but the current value gradually changes. Therefore, the L2-level event detection model is used to capture statistical feature changes over a longer time scale, corresponding to scenarios such as slow load increases and decreases and ramp-like changes. It is used to calculate the mean of the effective voltage value and the mean of the effective current value within a sliding window ending at the current time, and to determine the triggering of an L2-level event when the rate of change of at least one mean relative to the corresponding preset benchmark reaches the corresponding rate of change threshold. Specifically, in this embodiment, a sliding window of length W is constructed with the current time as the endpoint. The length of W can be 160, which corresponds to a duration of 50 milliseconds, or 2.5 power frequency cycles. Then, the mean of the effective voltage value and the mean of the effective current value within the sliding window are calculated respectively. Alternatively, a relative rate of change can be used. For the relative rate of change of the mean effective voltage value, it can be determined by the ratio of the absolute value of the difference between the mean effective voltage value and the voltage reference value to the voltage reference value. The calculation formula is as follows: ; In the above formula, express The relative rate of change of the mean effective value of voltage at time step express The average effective voltage value at time t. express The voltage reference value at that moment; The relative rate of change of the mean effective value of the current can be determined by the ratio of the absolute value of the difference between the mean effective value of the current and the reference value of the current to the reference value of the current. The calculation formula is as follows: ; In the above formula, express The relative rate of change of the mean effective value of the current at any given time. express The mean effective value of the current at time t. express The current reference value at any given time.

[0024] when or When an L2 level event is triggered, in this embodiment, the voltage change rate threshold is used. A value of 5% can be used as the threshold for the rate of change of current. An 8% is acceptable.

[0025] In addition, voltage reference value and current reference value The average effective voltage and current values ​​are determined based on the average of all sampled moments between the two most recent L2 event trigger times. For example, if the two most recent L2 event trigger times are 10 seconds and 20 seconds from the current time, then the average effective voltage and current values ​​are calculated for all moments between 10 seconds and 20 seconds from the current time. The calculation formula can be expressed as: ; In the above formula, This indicates the number of sampling times between the two most recent L2 level event trigger times. This represents the time interval between two adjacent sampling moments. This scheme freezes the steady-state data between events as a baseline when an L2-level event is triggered, ensuring that the baseline value more accurately reflects the normal values ​​between events and avoiding mutual interference between baseline updates and event detection.

[0026] For the L3 event detection model, since a single L2 event reflects a significant load change, while multiple L2 events within a unit of time reflect a high-frequency load fluctuation, in this state, the data packet generation rate is faster than the normal consumption rate. If the communication module is not woken up in time, the buffer may overflow, leading to the loss of critical data. Therefore, the L3 event detection model is used to determine whether the circuit system is in a frequent fluctuation state, corresponding to scenarios such as multiple load starts and stops within a short period and the cumulative effect of harmonic distortion. It records the cumulative number of L2 events occurring within a unit of time, and determines that an L3 event is triggered when the number exceeds a preset cumulative number threshold. In this embodiment, the unit of time can be set to 30 seconds, that is, counting the cumulative number of L2 events occurring within 30 seconds. If this number exceeds the preset cumulative number threshold, such as 5, an L3 event is determined to be triggered.

[0027] For the comprehensive judgment model, since the three-level event detection models run in parallel and independently, each will output its own judgment result. Therefore, the comprehensive judgment model needs to output the event level at the current moment based on the event levels that have been triggered. If no event level has been triggered, it outputs "no event level." Furthermore, the event levels of L1, L2, and L3 events increase sequentially. For example, if L1, L2, and L3 events are triggered simultaneously, the comprehensive judgment model will output the current event as an L3 event. The event level directly reflects the severity and urgency of load changes; the higher the level, the more immediately the communication module needs to be woken up. If none of the three-level event detection models output the current event level, it can output "no event level," indicating that no anomaly has occurred, and L0 can be output. The three-level event detection models characterize the load changes at different time scales. The L1 event detection model captures instantaneous changes, the L2 event detection model captures gradual drift, and the L3 event detection model captures frequent fluctuations. Their outputs provide the data basis for subsequent wake-up of the communication module.

[0028] S4. Construct an information value decay function. Based on the time elapsed since the generation of each sampled data packet in the data buffer to be sent, calculate the comprehensive remaining information value coefficient of the data buffer to be sent. In S2, sampled data is encapsulated into independent data packets and stored in the data buffer to be sent. Multiple data packets may exist simultaneously in the buffer, generated at different times, and therefore possessing different time-sensitive values. Simply taking the average or weighted sum of the values ​​of all data packets may mask the value decay of the worst-performing data packet in the buffer. For example, if the buffer contains a low-value data packet generated a long time ago and multiple newly generated high-value data packets, using the average to calculate the comprehensive value will smooth out the impact of the low-value data packet by the high-value data packet. Since the overall average value of the buffer is still relatively high, the system determines that the overall value of the buffer is acceptable and therefore does not wake up the communication module. As time progresses, due to the time-sensitive nature of the data, the value of the low-value data packet will continue to decrease until it is lost, losing its subsequent analytical value. It may even be discarded directly after the buffer is full, thus completely losing the value of the data acquisition. Therefore, this embodiment adopts a minimum value strategy, that is, the comprehensive information residual value coefficient is taken as the minimum value of the information residual value coefficient of each data packet in the buffer. The communication module is then activated based on the data packet with the lowest value, thus ensuring low power consumption without any data packet losing its value for subsequent analysis. Furthermore, the information residual value coefficient of each data packet is exponentially negatively correlated with the time elapsed since its generation. Additionally, the exponential term in the information residual value coefficient calculation formula also includes a decay time constant that is positively correlated with it. The decay time constant is negatively correlated with the triggering frequency of L2 and L3 events per unit time between preset high-frequency and low-frequency thresholds. The expression for the decay time constant can be: ; In the above formula, Represents the decay time constant. and These represent the maximum and minimum values ​​of the decay time constant, for example, set to 2 seconds and 120 seconds respectively. and The preset high-frequency threshold and low-frequency threshold can be set to 0.2 times / second and 0.02 times / second, respectively. This indicates the triggering frequency of L2 and L3 events per unit time. Greater than the high frequency threshold hour, Pick ,when Less than the low frequency threshold hour, Pick .

[0029] Therefore, the formula for calculating the information residual value coefficient in this embodiment can be expressed as: ; Where t represents the current time, This represents the residual value coefficient of the i-th data packet in the buffer. This indicates the generation time of the i-th sampled data packet, which is the sampling time of the first sample point in that data packet.

[0030] Therefore, a data packet has the highest value and the highest residual value coefficient when it is first generated, but as time goes on... Increase the information surplus value coefficient In this embodiment, for residential users, the typical load change is relatively slow, and the attenuation time constant is generally around 30 seconds. For industrial and commercial users, the load change is fast, and the attenuation time constant is generally around 10 seconds.

[0031] For example, if the current buffer contains three data packets with a decay time constant of 30, generated 10 seconds ago, 5 seconds ago, and 1 second ago respectively, their corresponding residual value coefficients are approximately 0.716, 0.846, and 0.967. Following the previous example, if we simply take the average value of all data packets, the average is 0.843. In the subsequent wake-up urgency function... However, according to the method of this embodiment, the comprehensive information remaining value coefficient is taken as the minimum value of the remaining value coefficients of each data packet in the buffer, that is, when the minimum value of 0.716 is taken, then in the subsequent wake-up urgency function... Since 0.284 > 0.157, when calculating according to the method of this embodiment, This component makes a significant contribution and is more likely to trigger wake-up, which forces the system to report the earliest data packet in a timely manner when its value decays to a certain level, preventing it from being discarded.

[0032] The above example uses a fixed decay time constant. However, the actual rate of change in power load varies greatly at different times of day: at night, the load is stable, and the data value decays more slowly, requiring a larger decay time constant to reduce unnecessary wake-ups; during the day, the load fluctuates frequently, meaning the event frequency is high, and the data value decays rapidly, requiring a smaller decay time constant to encourage more frequent system wake-ups and avoid cache backlog. Therefore, it needs to be dynamically adjusted according to the actual load fluctuations. That is, the decay time constant mentioned earlier is negatively correlated with the triggering frequency of L2 and L3 events per unit time. Of course, to prevent the decay time constant from being too large or too small, an upper limit is set. and lower limit value This mechanism allows for the effective saving of energy while preventing the omission of important information.

[0033] S5. Construct a collaborative wake-up urgency function, which weights and fuses the current event level with the comprehensive information residual value coefficient to output the collaborative wake-up urgency value. In this embodiment, the collaborative wake-up urgency function is at least a weighted sum of three terms: the event level mapping function term, the difference between 1 and the comprehensive information residual value coefficient, and the ratio of the data length of the data buffer to the maximum capacity of the buffer at the current moment. Let the weight coefficients of the three terms be as follows: , and Then its expression can be represented as: ; In the above formula, This indicates the urgency value of the coordinated wake-up call. The event level mapping function is a mathematical mapping that converts discrete event levels into numerical values. This refers to the event levels output by the multi-level event detection model, such as the event levels L0, L1, L2, and L3 mentioned earlier. The function can then be set as follows: This converts different event levels into specific numerical values ​​for calculation. In the co-awakening urgency function, this is the event level mapping function term. This is the difference between 1 and the comprehensive information residual value coefficient, also known as the information decay urgency term, and the comprehensive information residual value coefficient. The larger the value, the smaller this item becomes. That is, the length of the data buffer to be sent at the current moment. With the maximum capacity of the cache The ratio also represents the cache utilization rate, reflecting the pressure of data backlog. , and These are the coefficients for the corresponding items, and the sum of the three coefficients is 1. However, fixed weight coefficients are difficult to adapt to different power consumption scenarios. For example: 1. In scenarios with frequent fluctuations, the event level L best reflects the true urgency, and its corresponding weight should be increased. 1. Size; 2. Cache occupancy rate in scenarios where there are no events for a long time but the cache accumulates slowly. If it becomes the primary contradiction, its corresponding weight should be increased. 3. In scenarios where data is outdated (such as recovery after a communication failure); Most importantly, the weighting of this item should be increased. The size, therefore, this embodiment further sets the size of, therefore, to , and The steps for dynamic adjustment are as follows: Each time the communication module wakes up and reports data, the master station returns an acknowledgment message indicating which data is considered valid—that is, non-redundant and valuable data. Weighting can be adjusted based on the proportion of valid data. Therefore, the first step is to obtain the proportion of valid data from the feedback result after the last wake-up of the communication module and its data reporting—that is, the proportion of data deemed valid by the master station in the last report out of all previously reported data.

[0034] Then, two thresholds are set: a lower limit for the effective data ratio and an upper limit for the effective data ratio, which can be set to 0.6 and 0.8 respectively. Then, the following judgment is performed: If the proportion of valid data is less than or equal to the preset lower limit of the valid data proportion threshold, it indicates that most of the data reported last time was redundant (i.e., invalid). This suggests that the wake-up decision is too frequent or the threshold is too low, and the threshold will be increased by the preset step size. and reduce and The specific adjustments can be calculated using the following expression: ; In the above formula, , and They represent the results after the (n+1)th adjustment. , and The value, , and They represent the nth adjustment. , and The value, Indicates weighting coefficient The adjustment step size is generally between 0.01 and 0.05; in this embodiment, it is set to 0.02. and They represent The upper and lower limits can take values ​​ranging from [0.6, 0.8] to [0.2, 0.4], respectively. In this embodiment... The upper and lower limits are set to 0.7 and 0.3, respectively. and Weighting coefficients The upper and lower limits of the value range from [0.4, 0.5] to [0.1, 0.2], respectively. In this embodiment, the values ​​are 0.4 and 0.15, respectively. Similarly, the weighting coefficient... The upper and lower limits are set to 0.4 and 0.15 respectively. Initially, , and The values ​​can be set to 0.4, 0.3, and 0.3 respectively. Of course, to prevent the values ​​from changing after each adjustment... , and If the sum of the values ​​is not 1, normalization can be performed to ensure that the sum of the three is 1.

[0035] If the proportion of valid data is greater than or equal to the preset upper limit of the valid data proportion threshold, it indicates that almost all the data reported last time was valid, but there may be omissions, i.e., important data that should have been reported but was not. In this case, the proportion of valid data will be reduced by the preset step size. and increase and The specific adjustments can be calculated using the following expression: ; If the proportion of valid data is between the preset lower and upper limits of the valid data proportion threshold, then maintain... , and constant.

[0036] S6. Compare the collaborative wake-up urgency value with a preset wake-up threshold, and decide whether to wake up the communication module to report data based on the comparison result. The collaborative wake-up urgency value U output in step S5 is a dimensionless number between 0 and 1. In this embodiment, the wake-up threshold can be set to 0.6. If the collaborative wake-up urgency value is greater than or equal to the wake-up threshold, a wake-up command can be generated to wake up the communication module and package all data packets in the data buffer to be sent to be reported to the master station. Otherwise, the communication module remains in a dormant state and does not perform any communication operations. In addition, it should be noted that the wake-up threshold can be adjusted according to actual needs. For example, if it is desired to further reduce power consumption, i.e., reduce the number of wake-ups, the wake-up threshold can be increased. If it is desired to improve data real-time performance, i.e., increase the number of wake-ups, the wake-up threshold can be decreased. Once woken up and successfully reported data, the buffer is cleared. After receiving the data, the master station will perform validity analysis and feed back the proportion of valid data to the energy meter through the downlink channel. The energy meter uses this feedback value to dynamically adjust the weight coefficient in the collaborative wake-up urgency function to prepare for the next wake-up decision.

[0037] This embodiment also provides a power meter sampling and communication collaborative energy saving control system, including a processor and a memory. The memory is used to store a computer program, and when the computer program is executed by the processor, it implements the power meter sampling and communication collaborative energy saving control method.

[0038] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0039] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Since the above embodiments are substantially similar to the method embodiments, their descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0040] The above embodiments provide a detailed description of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for energy meter sampling and communication-coordinated energy consumption reduction control, characterized in that, The method includes the following steps: S1. Continuously collect the instantaneous voltage and current values ​​of the electricity meter at fixed time intervals to construct a sampling data sequence; S2. The sampling data sequence of multiple consecutive sampling points is packaged into a sampling data packet according to the preset packaging rules and stored in the data buffer to be sent. Each sampling data packet has a unique generation time. S3. Using a multi-level event detection model, the sampled data sequence is detected and the highest event level is output. S4. Construct an information value decay function and calculate the comprehensive information residual value coefficient of the data buffer to be sent based on the time elapsed since the generation time of each sampled data packet in the data buffer to be sent. S5. Construct a collaborative wake-up urgency function, which weights and integrates the current event level with the comprehensive information residual value coefficient, and outputs the collaborative wake-up urgency value. S6. Compare the collaborative wake-up urgency value with the preset wake-up threshold, and decide whether to wake up the communication module to report data based on the comparison result.

2. The energy meter sampling and communication collaborative energy consumption reduction control method according to claim 1, characterized in that, The multi-level event detection model includes: The L1 level event detection model is used to calculate the absolute value of the difference between the current voltage sample value and the current sample value and the previous sample value, respectively, and to determine that an L1 level event is triggered when at least one absolute value of the difference is greater than the corresponding difference threshold. The L2-level event detection model is used to calculate the mean of the effective voltage value and the mean of the effective current value within a sliding window with the current time as the endpoint, and to determine that an L2-level event is triggered when the rate of change of at least one mean relative to the corresponding preset benchmark reaches the corresponding rate of change threshold. The L3 event detection model is used to record the cumulative number of L2 events occurring per unit time, and determines that an L3 event is triggered when the number of L2 events exceeds a preset cumulative number threshold. The comprehensive judgment model is used to output the event level at the current moment based on the event levels that have been triggered. If no event level has been triggered, it outputs no event level. The event levels of L1, L2 and L3 events increase sequentially.

3. The energy meter sampling and communication collaborative energy consumption reduction control method according to claim 2, characterized in that, In the L1-level event detection model, the difference threshold includes a voltage difference threshold that is compared with the absolute value of the difference between the voltage sample value and the previous sample value, and a current difference threshold that is compared with the absolute value of the difference between the current sample value and the previous sample value. The current difference threshold is positively correlated with the standard deviation of the current sample values ​​at multiple times prior to the current time.

4. The energy meter sampling and communication collaborative energy consumption reduction control method according to claim 2, characterized in that, In the L2 level event detection model, the rate of change is a relative rate of change; The relative rate of change of the mean effective voltage value is determined based on the ratio of the absolute value of the difference between the mean effective voltage value and the voltage reference value to the voltage reference value. The relative rate of change of the mean effective value of the current is determined based on the ratio of the absolute value of the difference between the mean effective value of the current and the reference value of the current to the reference value of the current. The voltage reference value and current reference value are determined based on the average effective voltage value and the average effective current value of all sampling times between the two most recent L2 level event trigger times, respectively.

5. The energy meter sampling and communication collaborative energy consumption reduction control method according to claim 1, characterized in that, The comprehensive information residual value coefficient is determined based on the minimum value of the residual value coefficient of each sampled data packet in the data buffer to be sent; the residual value coefficient of each sampled data packet is exponentially negatively correlated with the time elapsed since its generation.

6. The energy meter sampling and communication collaborative energy consumption reduction control method according to claim 5, characterized in that, The exponential term of the formula for calculating the residual value coefficient of information also includes a decay time constant that is positively correlated with it. The decay time constant is negatively correlated with the triggering frequency of L2 and L3 events per unit time between preset high-frequency thresholds and low-frequency thresholds.

7. The energy meter sampling communication collaborative energy consumption reduction control method according to claim 1, characterized in that, The collaborative wake-up urgency function is at least a weighted sum of three terms: the event level mapping function term, the difference between 1 and the comprehensive information residual value coefficient, and the ratio of the data length of the data buffer to the maximum capacity of the buffer at the current moment. Let the weight coefficients of these three terms be respectively... , and It dynamically adjusts based on the feedback results after the communication module performed data reporting during the last wake-up, and the specific steps are as follows: Obtain the proportion of valid data in the feedback results after the last wake-up communication module executes data reporting; If the proportion of valid data is less than or equal to the preset lower limit of the valid data proportion threshold, then the proportion will increase by the preset step size. and reduce and ; If the proportion of valid data is greater than or equal to the preset upper limit of the valid data proportion threshold, then the proportion will decrease by a preset step size. and increase and ; If the proportion of valid data is between the preset lower and upper limits of the valid data proportion threshold, then maintain... , and constant.

8. A system for implementing the energy meter sampling communication collaborative energy consumption reduction control method according to any one of claims 1-7, characterized in that, It includes a processor and a memory, the memory being used to store a computer program, which, when executed by the processor, implements the energy meter sampling communication collaborative energy saving control method as described in any one of claims 1-7.

Images