Power consumption quantity prediction system of intelligent electric meter
Through the smart meter electricity consumption prediction system, a deep neural network is used to generate an electricity consumption prediction model, which solves the problem that the total electricity consumption in the set area is difficult to accurately predict, and improves the utilization rate of reserved electricity.
Patent Information
- Application Number
- CN202510077098.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2024-10-11
- Filing Date
- 2025-01-17
- Publication Date
- 2025-04-11
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art is difficult to accurately predict the total electricity consumption of a set area in the future time interval, resulting in a low utilization rate of reserved electricity.
A smart meter electricity consumption prediction system is adopted. This system uses a smart electricity meter electricity consumption quantity prediction system. By obtaining information about the electricity consumption equipment in the set area and past electricity consumption data during the same period, and using deep neural networks to learn multiple times to generate electricity consumption prediction models to improve the reliability and stability of prediction.
By accurately predicting the total amount of electricity consumption, the system can allocate reserved power before the future time interval reaches to improve the utilization rate of reserved power.
Abstract
Description
Technical Field
[0001] The present invention relates to the field of smart electric meters, and more specifically, to a smart electric meter electricity consumption prediction system. Background Art
[0002] The main types of smart meters are IC card meters. Users take their IC cards to the power supply management department to pay for electricity. The power supply management department uses the electricity sales management machine to write the purchased electricity into the IC card. Users swipe the contactless IC card in the sensing area (simply swipe the card) to close the switch and supply power. After the power is supplied, the card is taken away. When the remaining power in the meter is equal to the alarm power, the power is cut off and the alarm (or buzzer alarm) is sounded. At this time, the user can restore the power supply by swiping the card in the sensing area; when the remaining power is zero, the power is automatically cut off, and the user must pay for electricity with the card again before the power can be restored.
[0003] For a set area with multiple smart meters, it is necessary to clarify its total electricity consumption in the future time period. In this way, a matching amount of reserve electricity can be allocated for the set area in the future time period based on the specific value of the total electricity consumption of the set area in the future time period, thereby improving the utilization rate of the relatively limited amount of reserve electricity. Obviously, the current obstacle lies in the difficulty in accurately and effectively predicting the total electricity consumption of the set area in the future time period. Summary of the invention
[0004] In order to solve the technical problems in the related fields, the present invention provides a smart meter electricity consumption prediction system. In order to improve the reliability and stability of the predicted value of the total electricity consumption of a set area in the future time interval of the day, a sufficient and comprehensive number of basic data are selected for the prediction, including the number, rated power, reference clock frequency and working current of various types of electrical equipment in the set area, and multiple total electricity consumption corresponding to the same time interval of the set area in the past multiple days. A layer-by-layer learning mechanism is introduced to perform multiple learning on a deep neural network to obtain a deep neural network after multiple learning and output it as a power consumption prediction model. The number of times the deep neural network is learned is proportional to the duration of the future time interval of the day. Proportional, thereby completing the targeted model design of the electricity consumption quantity prediction model, and inputting multiple electricity consumption information of the set area, multiple total electricity consumption corresponding to the same time interval of the set area in the past multiple days, and the duration of the future time interval of the day in parallel into the electricity consumption quantity prediction model, and running the electricity consumption quantity prediction model to obtain the predicted value of the total electricity consumption of the set area in the future time interval of the day output by the electricity consumption quantity prediction model, and before the future time interval of the day arrives, based on the predicted value of the total electricity consumption of the set area in the future time interval of the day, allocating a matching amount of reserve electric energy for the set area in the future time interval of the day, thereby improving the utilization rate of the relatively limited amount of reserve electric energy.
[0005] According to the present invention, a smart meter electricity consumption prediction system is provided, the system comprising: A first capture mechanism is used to obtain the number, rated power, reference clock frequency and operating current of various types of electrical equipment in a set area to output multiple electrical consumption information of the set area; The second capture mechanism is used to obtain multiple sets of total electricity consumption corresponding to the same time interval of the past multiple days in the set area, where the single set of total electricity consumption corresponding to the same time interval of a certain day in the past is the sum of the electricity consumption of each smart meter in the same time interval of the set area in the past day, and the same time interval of the past day and the future time interval of the current day have the same distribution position on the time axis of each day; A layer-by-layer learning mechanism is used to perform multiple learning on the deep neural network to obtain the deep neural network after the multiple learning and output it as the electricity consumption prediction model, and the number of times the deep neural network is learned is proportional to the duration of the future time interval of the day; an intelligent prediction device, connected to the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, respectively, for inputting multiple pieces of electricity consumption information of a set area, multiple copies of total electricity consumption corresponding to the same time interval of the set area in the past multiple days, and the duration of the future time interval of the day in parallel into the electricity consumption quantity prediction model, and running the electricity consumption quantity prediction model to obtain a predicted value of the total electricity consumption of the set area in the future time interval of the day output by the electricity consumption quantity prediction model; an electric power allocation device, connected to the intelligent prediction device, for allocating a matching amount of reserve electric energy for a set area in a future time interval of the day based on a predicted value of the total power consumption of the set area in the future time interval of the day before the future time interval of the day arrives; Wherein, before the future time interval of the day arrives, allocating a matching amount of reserve electric energy for the set area in the future time interval of the day based on the predicted value of the total power consumption of the set area in the future time interval of the day includes: before the future time interval of the day arrives, the amount of reserve electric energy allocated for the set area in the future time interval of the day is a multiple of the predicted value of the total power consumption of the set area in the future time interval of the day; Among them, the first capture mechanism is used to obtain the number, rated power, reference clock frequency and operating current of various types of electrical equipment in the set area as multiple power consumption information outputs of the set area; including: the set area is a single block, a single industrial park or a single residential building.
[0006] Therefore, the present invention has at least the following beneficial technical effects: Technical effect 1: In order to improve the reliability and stability of the predicted value of the total electricity consumption of the set area in the future time interval of the day, a number of comprehensive basic data are selected for the prediction, including the number, rated power, reference clock frequency and working current of various types of electrical equipment in the set area, and multiple sets of total electricity consumption corresponding to the same time interval of the set area in the past multiple days; Technical effect 2: A layer-by-layer learning mechanism is introduced to perform multiple learning on the deep neural network to obtain a deep neural network after multiple learning and output it as a power consumption prediction model. The number of times the deep neural network is learned is proportional to the duration of the future time interval of the day, thereby completing the targeted model design of the power consumption prediction model; Technical effect three: multiple electricity consumption information of the set area, multiple total electricity consumption corresponding to the same time interval of the set area in the past multiple days, and the duration of the future time interval of the day are input into the electricity quantity prediction model in parallel, and the electricity quantity prediction model is run to obtain the predicted value of the total electricity consumption of the set area in the future time interval of the day output by the electricity quantity prediction model, and before the future time interval of the day arrives, a matching amount of reserve electric energy is allocated for the set area in the future time interval of the day based on the predicted value of the total electricity consumption of the set area in the future time interval of the day, thereby improving the utilization rate of the relatively limited amount of reserve electric energy.
[0007] The intelligent meter power consumption prediction system of the present invention is intelligently controlled and reliably designed. Since the intelligent prediction scheme can be used to predict the predicted value of the total power consumption of a set area in the future time interval of the day, a matching amount of reserve power is allocated for the set area in the future time interval of the day, thereby improving the utilization rate of the relatively limited amount of reserve power. DETAILED DESCRIPTION
[0008] The implementation scheme of the smart meter electricity consumption prediction system of the present invention will be described in detail below.
[0009] The smart meter electricity consumption prediction system according to the first embodiment of the present invention includes: A first capture mechanism is used to obtain the number, rated power, reference clock frequency and operating current of various types of electrical equipment in a set area to output multiple electrical consumption information of the set area; Specifically, the first capture mechanism is used to obtain the number, rated power, reference clock frequency and working current of various types of electrical equipment in the set area as multiple power consumption information outputs of the set area, including: the first capture mechanism includes multiple information capture components, which are used to respectively capture the number, rated power, reference clock frequency and working current of various types of electrical equipment in the set area; The second capture mechanism is used to obtain multiple sets of total electricity consumption corresponding to the same time interval of the past multiple days in the set area, where the single set of total electricity consumption corresponding to the same time interval of a certain day in the past is the sum of the electricity consumption of each smart meter in the same time interval of the set area in the past day, and the same time interval of the past day and the future time interval of the current day have the same distribution position on the time axis of each day; A layer-by-layer learning mechanism is used to perform multiple learning on the deep neural network to obtain the deep neural network after the multiple learning and output it as the electricity consumption prediction model, and the number of times the deep neural network is learned is proportional to the duration of the future time interval of the day; an intelligent prediction device, connected to the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, respectively, for inputting multiple pieces of electricity consumption information of a set area, multiple copies of total electricity consumption corresponding to the same time interval of the set area in the past multiple days, and the duration of the future time interval of the day in parallel into the electricity consumption quantity prediction model, and running the electricity consumption quantity prediction model to obtain a predicted value of the total electricity consumption of the set area in the future time interval of the day output by the electricity consumption quantity prediction model; an electric power allocation device, connected to the intelligent prediction device, for allocating a matching amount of reserve electric energy for a set area in a future time interval of the day based on a predicted value of the total power consumption of the set area in the future time interval of the day before the future time interval of the day arrives; Wherein, before the future time interval of the day arrives, allocating a matching amount of reserve electric energy for the set area in the future time interval of the day based on the predicted value of the total power consumption of the set area in the future time interval of the day includes: before the future time interval of the day arrives, the amount of reserve electric energy allocated for the set area in the future time interval of the day is a multiple of the predicted value of the total power consumption of the set area in the future time interval of the day; The first capture mechanism is used to obtain the number, rated power, reference clock frequency and working current of various types of electrical equipment in a set area as multiple power consumption information outputs of the set area; including: the set area is a single block, a single industrial park or a single residential building; Wherein, before the future time interval of the day arrives, the amount of reserve electric energy allocated to the set area in the future time interval of the day is a multiple of the predicted value of the total power consumption of the set area in the future time interval of the day, including: the specific value of the multiple is between 1.05-1.10; Among them, the layer-by-layer learning mechanism is used to perform multiple learning on the deep neural network to obtain the deep neural network after multiple learning and output it as the electricity consumption prediction model, and the number of times the deep neural network is learned is proportional to the duration of the future time interval of the day. It includes: using FPGA devices to implement the layer-by-layer learning mechanism, and using MATLAB toolbox to complete the simulation and testing of the processing process of performing multiple learning on the deep neural network to obtain the deep neural network after multiple learning and output it as the electricity consumption prediction model.
[0010] Compared with the first embodiment of the present invention, the smart meter electricity consumption quantity prediction system according to the second embodiment of the present invention may further include the following components: an electric quantity measuring mechanism, connected to the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, respectively, and used to measure the current electric quantity used by the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively; Wherein, the power measurement mechanism is respectively connected to the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, and is used to respectively measure the current power usage of the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, including: the power measurement mechanism includes a plurality of power measurement units, which are respectively connected to the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, so as to complete the respective measurement of the current power usage of the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism; Wherein, the power measurement mechanism includes a plurality of power measurement units, which are used to connect with the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively, so as to complete the measurement of the current power usage of the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively, including: the plurality of power measurement units are a plurality of power sensing circuits, which are used to connect with the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively, so as to complete the measurement of the current power usage of the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively; Wherein, the multiple power measurement units are multiple power sensing circuits, which are used to respectively connect with the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism to respectively measure the current power usage of the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, including: the multiple power sensing circuits have the same structure; And wherein, the multiple power measurement units are multiple power sensing circuits, which are used to respectively connect to the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism to complete the separate measurement of the current power usage of the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, and also include: the multiple power sensing circuits have the same power measurement upper limit value and power measurement lower limit value.
[0011] Compared with the first embodiment of the present invention, the smart meter electricity consumption quantity prediction system according to the third embodiment of the present invention may further include the following components: A fan heat dissipation mechanism is arranged near the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, and is used to provide the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism with the heat dissipation services required by each of them; Wherein, the fan heat dissipation mechanism is arranged near the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, and is used to provide the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism with the heat dissipation services required by each of them, including: the fan heat dissipation mechanism includes a plurality of fan heat dissipation units, which are respectively arranged near the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism to provide the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism with the heat dissipation services required by each of them; And wherein, the fan cooling mechanism includes multiple fan cooling units, which are respectively arranged near the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism to provide the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism with the required cooling services respectively, including: the rated output power of the multiple fan cooling units is equal.
[0012] In addition, in the smart meter electricity consumption quantity prediction system, multiple electricity consumption information of the set area, multiple total electricity consumption corresponding to the same time interval of the set area in the past multiple days, and the duration of the future time interval of the day are input into the electricity consumption quantity prediction model in parallel, and the electricity consumption quantity prediction model is run to obtain the predicted value of the total electricity consumption of the set area in the future time interval of the day output by the electricity consumption quantity prediction model, which also includes: using a parallel control component to complete the parallel input of multiple electricity consumption information of the set area, multiple total electricity consumption corresponding to the same time interval of the past multiple days, and the duration of the future time interval of the day into the electricity consumption quantity prediction model.
[0013] Because certain changes can be made without departing from the scope of the present invention, all contents contained in the above description should be interpreted as illustrative rather than just literal meanings. It should be appreciated by those skilled in the art that the sequential steps or structures described in the description can be changed without departing from the scope of the present invention, and the descriptions contained herein are only individual examples of various possible descriptions of the present invention.
Claims
1. A smart meter electricity consumption prediction system, characterized in that: The system comprises: A first capture mechanism is used to obtain the number, rated power, reference clock frequency and operating current of various types of electrical equipment in a set area to output multiple electrical consumption information of the set area; The second capture mechanism is used to obtain multiple sets of total electricity consumption corresponding to the same time interval of the past multiple days in the set area, where the single set of total electricity consumption corresponding to the same time interval of a certain day in the past is the sum of the electricity consumption of each smart meter in the same time interval of the set area in the past day, and the same time interval of the past day and the future time interval of the current day have the same distribution position on the time axis of each day; A layer-by-layer learning mechanism is used to perform multiple learning on the deep neural network to obtain the deep neural network after the multiple learning and output it as the electricity consumption prediction model, and the number of times the deep neural network is learned is proportional to the duration of the future time interval of the day; an intelligent prediction device, connected to the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, respectively, for inputting multiple pieces of electricity consumption information of a set area, multiple copies of total electricity consumption corresponding to the same time interval of the set area in the past multiple days, and the duration of the future time interval of the day in parallel into the electricity consumption quantity prediction model, and running the electricity consumption quantity prediction model to obtain a predicted value of the total electricity consumption of the set area in the future time interval of the day output by the electricity consumption quantity prediction model; an electric power allocation device, connected to the intelligent prediction device, for allocating a matching amount of reserve electric energy for a set area in a future time interval of the day based on a predicted value of the total power consumption of the set area in the future time interval of the day before the future time interval of the day arrives; Wherein, before the future time interval of the day arrives, allocating a matching amount of reserve electric energy for the set area in the future time interval of the day based on the predicted value of the total power consumption of the set area in the future time interval of the day includes: before the future time interval of the day arrives, the amount of reserve electric energy allocated for the set area in the future time interval of the day is a multiple of the predicted value of the total power consumption of the set area in the future time interval of the day; Among them, the first capture mechanism is used to obtain the number, rated power, reference clock frequency and operating current of various types of electrical equipment in the set area as multiple power consumption information outputs of the set area; including: the set area is a single block, a single industrial park or a single residential building.
2. The smart meter electricity consumption prediction system according to claim 1, characterized in that: Before the future time interval of the day arrives, the amount of reserve electric energy allocated to the set area in the future time interval of the day is a multiple of the predicted value of the total power consumption of the set area in the future time interval of the day, including: the specific value of the multiple is between 1.05-1.10; Among them, the layer-by-layer learning mechanism is used to perform multiple learning on the deep neural network to obtain the deep neural network after multiple learning and output it as the electricity consumption prediction model, and the number of times the deep neural network is learned is proportional to the duration of the future time interval of the day. It includes: using FPGA devices to implement the layer-by-layer learning mechanism, and using MATLAB toolbox to complete the simulation and testing of the processing process of performing multiple learning on the deep neural network to obtain the deep neural network after multiple learning and output it as the electricity consumption prediction model.
3. The smart meter electricity consumption prediction system according to claim 2, characterized in that: The system further comprises: an electric quantity measuring mechanism, connected to the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, respectively, and used to measure the current electric quantity used by the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively; Among them, the power measurement mechanism is respectively connected to the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, and is used to respectively measure the current power usage of the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism. The power measurement mechanism includes multiple power measurement units, which are respectively connected to the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism to complete the respective measurements of the current power usage of the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism.
4. The smart meter electricity consumption prediction system according to claim 3, characterized in that: The power measurement mechanism includes multiple power measurement units, which are used to connect with the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively to complete the separate measurement of the current power usage of the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively. The multiple power measurement units are multiple power sensing circuits, which are used to connect with the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively to complete the separate measurement of the current power usage of the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively.
5. The smart meter electricity consumption prediction system according to claim 4, characterized in that: The multiple power measurement units are multiple power sensing circuits, which are used to connect to the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively, so as to complete the respective measurements of the current power usage of the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, including: the structures of the multiple power sensing circuits are the same.
6. The smart meter electricity consumption prediction system according to claim 5, characterized in that: The multiple power measurement units are multiple power sensing circuits, which are used to connect to the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively to complete the separate measurement of the current power usage of the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively, and also include: the multiple power sensing circuits have the same power measurement upper limit value and power measurement lower limit value.
7. The smart meter electricity consumption prediction system according to claim 3, characterized in that: The system further comprises: A fan heat dissipation mechanism is arranged near the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, and is used to provide the required heat dissipation services for the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism respectively.
8. The smart meter electricity consumption prediction system according to claim 7, characterized in that: A fan heat dissipation mechanism is arranged near the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism, and is used to provide the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism with the heat dissipation services they need respectively. The fan heat dissipation mechanism includes multiple fan heat dissipation units, which are respectively arranged near the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism to provide the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism with the heat dissipation services they need respectively.
9. The smart meter electricity consumption prediction system according to claim 8, characterized in that: The fan cooling mechanism includes multiple fan cooling units, which are respectively arranged near the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism to provide the intelligent prediction device, the first capture mechanism, the second capture mechanism and the layer-by-layer learning mechanism with their respective required cooling services, including: the rated output powers of the multiple fan cooling units are equal.