CN114598722A - Non-invasive load identification and energy consumption monitoring system of Internet of things and implementation method thereof - Google Patents

Non-invasive load identification and energy consumption monitoring system of Internet of things and implementation method thereof Download PDF

Info

Publication number
CN114598722A
CN114598722A CN202210254065.3A CN202210254065A CN114598722A CN 114598722 A CN114598722 A CN 114598722A CN 202210254065 A CN202210254065 A CN 202210254065A CN 114598722 A CN114598722 A CN 114598722A
Authority
CN
China
Prior art keywords
power
energy consumption
load identification
load
internet
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210254065.3A
Other languages
Chinese (zh)
Inventor
沐阿华
于晓丽
于芳
陈秀宁
于凌云
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Huizhi Boyi Technology Co ltd
Original Assignee
Beijing Huizhi Boyi Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Huizhi Boyi Technology Co ltd filed Critical Beijing Huizhi Boyi Technology Co ltd
Priority to CN202210254065.3A priority Critical patent/CN114598722A/en
Publication of CN114598722A publication Critical patent/CN114598722A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/02—Protocols based on web technology, e.g. hypertext transfer protocol [HTTP]
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S40/00—Systems for electrical power generation, transmission, distribution or end-user application management characterised by the use of communication or information technologies, or communication or information technology specific aspects supporting them
    • Y04S40/12—Systems for electrical power generation, transmission, distribution or end-user application management characterised by the use of communication or information technologies, or communication or information technology specific aspects supporting them characterised by data transport means between the monitoring, controlling or managing units and monitored, controlled or operated electrical equipment
    • Y04S40/128—Systems for electrical power generation, transmission, distribution or end-user application management characterised by the use of communication or information technologies, or communication or information technology specific aspects supporting them characterised by data transport means between the monitoring, controlling or managing units and monitored, controlled or operated electrical equipment involving the use of Internet protocol

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Remote Monitoring And Control Of Power-Distribution Networks (AREA)
  • Telephonic Communication Services (AREA)

Abstract

The invention discloses an Internet of things non-invasive load identification and energy consumption monitoring system and an implementation method thereof, wherein the system comprises non-invasive load monitoring equipment, a mobile APP and a software server, the output end of the non-invasive load monitoring equipment is electrically connected with the input end of the mobile APP, and the mobile APP is bidirectionally connected with the software server, and the system has the beneficial effects that: on the one hand, the system can be used for identifying electric loads, monitoring energy consumption, optimizing and controlling, and on the other hand, the system can also monitor whether self power generation equipment (such as photovoltaic equipment, cogeneration equipment and the like) generates enough electric power or not, so that self electric power energy is preferentially utilized to the maximum extent; compared with the traditional intrusive method, the method has the advantages that only one monitoring device is needed to be installed at the entrance of power supply, the load in the whole system can be monitored without modifying a circuit, the function of the Internet of things is achieved, and the problem that many field environments are inconvenient to wire is effectively solved.

Description

Non-invasive load identification and energy consumption monitoring system of Internet of things and implementation method thereof
Technical Field
The invention relates to the technical field of energy consumption monitoring, in particular to a non-intrusive load identification and energy consumption monitoring system of an Internet of things and an implementation method thereof.
Background
The traditional intrusive energy consumption monitoring adopts a subentry energy consumption metering method, namely, a power supply circuit is subjected to subentry metering transformation, and an electric energy meter with a communication function is installed on each type of equipment or even each equipment as required, so that the data acquisition, monitoring and analysis of energy consumption can be realized, and the installation and maintenance are time-consuming and labor-consuming;
a Non-intrusive Load Monitoring (NILM) system, which is to install Monitoring equipment at an electric power inlet, and analyze and obtain the type and operation condition of a single Load in a Load cluster by Monitoring signals such as voltage, current and the like at the position;
however, the existing non-intrusive load identification and energy consumption monitoring system is inconvenient for identifying and monitoring the electrical load and energy consumption, and cannot perform optimization and control processing after monitoring, so that the utilization of electrical energy is wasted.
Disclosure of Invention
The invention aims to provide an Internet of things non-intrusive load identification and energy consumption monitoring system and an implementation method thereof, and aims to solve the problems that the existing non-intrusive load identification and energy consumption monitoring system proposed in the background art is inconvenient for identification and energy consumption monitoring processing of electric loads, cannot be optimized and controlled after monitoring, and causes waste in utilization of electric energy.
In order to achieve the purpose, the invention provides the following technical scheme: the non-invasive load identification and energy consumption monitoring system comprises non-invasive load monitoring equipment, a mobile APP and a software server, wherein the output end of the non-invasive load monitoring equipment is electrically connected with the input end of the mobile APP, the mobile APP is in bidirectional connection with the software server, the non-invasive load monitoring equipment comprises a sensor interface unit, a control and processing module, an electric energy metering module and a power management module, and the sensor interface unit and the control and processing module are electrically connected with the electric energy metering module and the power management module;
the sensor interface unit is used for installing the current sensor and simultaneously collecting parameters of current and voltage of the current sensor;
the control and processing module is used for accessing load identification and energy consumption conditions through network browsing and is also used for analyzing and taking charge of load identification and energy consumption analysis;
the electric energy metering module is used for collecting voltage and current data, calculating to obtain parameters such as active power, reactive power, apparent power, power factors and the like, and sending the power utilization data to the control and processing module;
the power management module is used for providing a DC5V working power supply for the non-invasive load monitoring equipment and also used for providing a DC3.3V working power supply for the electric energy metering module and the control and processing module;
the mobile APP is used for browsing energy consumption conditions on the mobile application program by the consumer at any time and any place;
the software server is used for storing the electricity utilization data and deploying the load decomposition, identification and energy consumption analysis software.
As a preferred embodiment of the present invention: the sensor interface unit comprises four groups of eight current sockets and four paths of voltage sockets, the sensor interface supports contact type and non-contact type sensors, the sampling rate can reach 2 kHz-8 kHz, the current sockets are used for monitoring parameters such as current and flow, the current sockets comprise four groups of eight ports of L1 k, L1L, L2 k, L2L, L3 k, L3L, L N k and N L and are used for connecting the contact type and non-contact type current sensors and simultaneously used for collecting current parameters of L1, L2, L3 and N, and the voltage sockets comprise four ports of L1, L2, L3 and N and are used for connecting 220/400VAC three-phase four-wire system alternating current and simultaneously used for collecting voltage parameters of L1, L2, L3 and N.
As a preferred embodiment of the present invention: the control and processing module comprises an independent operation mode and a server operation mode, the independent operation mode utilizes an edge calculation and an embedded load identification model to complete the identification of the power load, a web service and a browser page are arranged in a software firmware, other software outside the browser does not need to be installed during use, the load identification and the energy consumption condition are browsed and accessed through a mobile phone, a tablet computer or a computer network, the embedded load identification model is trained through an LSTM-RNN model established by the server application, and after the training is completed, TensorFlow Lite is used for compression processing and is led into equipment through OTA;
the server operation mode adopts an MQTT Internet of things protocol to send power utilization data to a designated analysis server, and analysis software in the server is responsible for load identification and energy consumption analysis and sends an analysis result to a subscriber.
As a preferable scheme of the invention: the control and processing module is provided with an integrated 2.4GHz wireless communication unit, and the integrated 2.4GHz wireless communication unit is provided with Wi-Fi and Bluetooth connection functions.
As a preferred embodiment of the present invention: the characteristic quantities of the electricity utilization data comprise steady state, transient state and operation mode, the steady state and the transient state are determined by the characteristics of components inside the equipment, and the operation mode is determined by the operation control strategy of the equipment.
As a preferred embodiment of the present invention: the load identification comprises a load identification model whose power supply entries p represent the sum of the active power of all the individual devices at time t, the total power can then be expressed as
Figure BDA0003548124030000041
Wherein y isi(T) represents the power consumption of the ith device of the I available devices at time T, knowing the total power over time duration
Figure BDA0003548124030000042
The objective is to obtain the power consumption of the ith device
Figure BDA0003548124030000043
As a preferable scheme of the invention: the load recognition model has P and dP input time series vectors: p is the total power P recorded by the non-invasive load monitoring equipment, and a training data set PT，iFor setting up the ith deviceT training samples of which
Figure BDA0003548124030000044
Figure BDA0003548124030000045
Is derived from the aggregate signal P and
Figure BDA0003548124030000046
represents the response device energy consumption of the nth sample;
the dP represents the change in power. The power change is introduced, so that transient characteristics are provided for load identification on one hand, and the influence of noise on learning model training is reduced on the other hand.
As a preferred embodiment of the present invention: the data of the transient characteristics are time series power data, a power data window is determined by a start time stamp and an end time stamp, and the transient characteristics are defined by selecting the following three characteristics:
firstly, the power change delta P of the equipment from starting to reaching a steady state;
Figure BDA0003548124030000047
Figure BDA0003548124030000048
is the median value of the post-transient window,
Figure BDA0003548124030000049
is the median of the previous transient window.
Second, maximum peak power variation Pmax；
Figure BDA00035481240300000410
max(x1) Is the maximum value of the transient window and,
Figure BDA0003548124030000051
is the median of the previous transient window.
Third is the minimum peak power variation Pmin。
Figure BDA0003548124030000052
min(x1) Is the minimum value of the transient window and,
Figure BDA0003548124030000053
is the median of the previous transient window.
As a preferred embodiment of the present invention: the load identification belongs to binary classification, and the accuracy of the identification is evaluated by adopting F scoring:
Figure 1
where precision is a positive predictive value, recall is a true positive predictive value ratio, tp is true positive, fp is false positive, fp is the prediction device is on but off, fn is false negative, fn is the device is on but expected to be off, and tp, fp, and fn can be determined by time data street.
The implementation method of the non-intrusive load identification and energy consumption monitoring system of the Internet of things is characterized by comprising the following steps:
s1, collecting: recording the measurement data;
s2, analysis: analyzing the measurement data and determining any user configuration;
s3, identifying: using AI and machine learning method to identify and create user consumption pattern, and sending to analysis service software;
s4, prediction: the analysis system detects any deviation of the user mode, carries out prediction maintenance, finds any fault in the early stage, predicts a future event according to historical data and realizes prediction alarm;
s5, operation control: and controlling the equipment by using an optimal power consumption strategy through the technology of the Internet of things.
Compared with the prior art, the invention has the beneficial effects that: the invention can be used for identifying, monitoring, optimizing and controlling the power load on one hand, and monitoring whether the self power generation equipment (such as photovoltaic, cogeneration and the like) generates enough power on the other hand, thereby ensuring that the self power energy is preferentially utilized to the maximum extent;
compared with the traditional intrusive method, the method has the advantages that only one monitoring device is needed to be installed at an inlet of power supply, the load in the whole system can be monitored without modifying a circuit, and meanwhile, the method has the function of the Internet of things, so that the problem that many field environments are inconvenient to wire is effectively solved;
the automatic clustering identification of dozens of household appliances, common industrial electromechanical equipment and unknown electric equipment is completed by configuring strong analysis software;
according to the total load power requirement, 1 monitoring device can be optionally arranged in one room, floor, building or power transmission station, so that the cost is saved, and the downtime is reduced.
Drawings
FIG. 1 is a schematic diagram of the operation of the apparatus of the present invention;
FIG. 2 is a flow chart of the main operation of the apparatus of the present invention;
FIG. 3 is a schematic diagram of an application scenario of the present invention;
FIG. 4 is a layout diagram of the apparatus of the present invention;
FIG. 5 is a schematic diagram of the LSTM-RNN learning model for NILM according to the present invention;
FIG. 6 is a schematic diagram of a neuron node structure according to the present invention;
FIG. 7 is a flow chart of a transient peak and boundary detection method of the present invention;
FIG. 8 is a schematic diagram of the operation mechanism of the present invention;
fig. 9 is a schematic view of the installation of the apparatus of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1-9, the present invention provides a technical solution: the non-invasive load identification and energy consumption monitoring system comprises non-invasive load monitoring equipment, a mobile APP and a software server, wherein the output end of the non-invasive load monitoring equipment is electrically connected with the input end of the mobile APP, the mobile APP is in bidirectional connection with the software server, the non-invasive load monitoring equipment comprises a sensor interface unit, a control and processing module, an electric energy metering module and a power management module, and the sensor interface unit and the control and processing module are electrically connected with the electric energy metering module and the power management module;
the sensor interface unit is used for installing the current sensor and simultaneously collecting parameters of current and voltage of the current sensor;
the control and processing module is used for accessing load identification and energy consumption conditions through network browsing and is also used for analyzing and taking charge of load identification and energy consumption analysis;
the electric energy metering module is used for collecting voltage and current data and calculating to obtain parameters such as active power, reactive power, apparent power, power factors and the like, and sending the power consumption data to the control and processing module;
the power management module is used for providing a DC5V working power supply for the non-invasive load monitoring equipment and providing a DC3.3V working power supply for the electric energy metering module and the control and processing module, the power management module comprises an AC/DC conversion unit and a DC/DC isolation conversion unit, the AC/DC conversion unit provides a DC5V working power supply for the equipment, and the input end of the AC/DC conversion unit is electrically connected with the voltage socket. The DC/DC isolation conversion unit provides DC3.3V working power supply for the electric energy metering module and the control and processing module, and the input end is the output end of the AC/DC conversion unit;
the mobile APP is used for browsing energy consumption conditions on a mobile application program by a consumer at any time and any place, and also supports a load recognition training learning function;
the software server is used for storing power utilization data and deploying load decomposition, identification and energy consumption analysis software, and machine learning is adopted for load identification, consumption curve optimization and relay coordination control, so that optimal overall consumption control is ensured.
Wherein, the sensor interface unit comprises four groups of eight current sockets and four paths of voltage sockets, the sensor interface supports contact type and non-contact type sensors, the installation form of the sensors is open-close type, can be quickly installed and disassembled without power cut and power cut, has simple and convenient operation, the sampling rate can reach 2 kHz-8 kHz, the current sockets are used for monitoring parameters such as current, flow and the like, the current sockets comprise four groups of eight ports of L1 k, L1L, L2 k, L2L, L3 k, L3L, N k and N L, for connecting contact and non-contact current sensors and simultaneously acquiring current parameters of L1, L2, L3 and N, the voltage socket comprises four ports of L1, L2, L3 and N, the device is used for connecting 220/400VAC (50/60Hz) three-phase four-wire system alternating current and is also used for acquiring voltage parameters of L1, L2, L3 and N.
The control and processing module comprises an independent operation mode and a server operation mode, the independent operation mode utilizes an edge calculation and an embedded load identification model to complete the identification of the power load, a web service and a page of a browser are arranged in a software firmware, other software outside the browser is not required to be installed during use, the load identification and the energy consumption condition are browsed and accessed through a mobile phone, a tablet personal computer or a computer network, the embedded load identification model is trained through an LSTM-RNN model established by the server application, after the training is completed, TensorFlow Lite is used for compression processing and is led into equipment through OTA, the memory use and the calculation time of the edge equipment can be reduced, and the load identification model capable of running on the edge equipment with high efficiency is formed;
the server operation mode adopts an MQTT Internet of things protocol to send power utilization data to a designated analysis server, and analysis software in the server is responsible for load identification and energy consumption analysis and sends an analysis result to a subscriber.
The control and processing module is provided with an integrated 2.4GHz wireless communication unit, and the integrated 2.4GHz wireless communication unit is provided with Wi-Fi and Bluetooth connection functions.
The characteristic quantities of the electricity utilization data comprise a steady state, a transient state and an operation mode, the steady state and the transient state are determined by the characteristics of components inside the equipment, and the operation mode is determined by an operation control strategy of the equipment.
Wherein the load identification comprises a load identification model whose power supply entries p represent the sum of the active power of all individual devices at time t, the total power can then be expressed as
Figure BDA0003548124030000091
Wherein y isi(T) represents the power consumption of the ith device of the I available devices at time T, knowing the total power over time duration
Figure BDA0003548124030000092
The objective is to obtain the power consumption of the ith device
Figure BDA0003548124030000093
In the current research and application, load increase, multi-state of the load and similar power consumption of the load are three major problems faced by NILM, and a long-short term memory recurrent neural network (LSTM-RNN) model is proposed for load decomposition and identification, so that the difficulty of various feature reproducibility of the load can be effectively solved.
Wherein, the load identification model has P and dP input time sequence vectors: p is the total power P recorded by the non-invasive load monitoring equipment, and a training data set PT，iT training samples for establishing the ith device, wherein
Figure BDA0003548124030000101
Wherein
Figure BDA0003548124030000102
Is derived from the aggregate signal P and
Figure BDA0003548124030000103
represents the response device energy consumption of the nth sample;
the dP represents the change in power. The power change is introduced, so that transient characteristics are provided for load identification on one hand, and the influence of noise on learning model training is reduced on the other hand.
The data of the transient characteristics are time series power data, a power data window is determined through a start timestamp and an end timestamp, and the power data can be divided into three areas: in practical application, a first derivative of a Gaussian window function and power data are subjected to convolution calculation and then are derived, a peak value is detected and the left and right boundaries of the transient window are determined through a zero crossing method, and the transient characteristics are defined by selecting the following three characteristics:
firstly, the power change delta P of the equipment from starting to reaching a steady state;
Figure BDA0003548124030000104
Figure BDA0003548124030000105
is the median value of the post-transient window,
Figure BDA0003548124030000106
is the median of the previous transient window.
Second, maximum peak power variation Pmax；
Figure BDA0003548124030000107
max(x1) Is the maximum value of the transient window and,
Figure BDA0003548124030000108
is the median of the previous transient window.
Third is the minimum peak power variation Pmin。
Figure BDA0003548124030000109
min(x1) Is the minimum value of the transient window and,
Figure BDA0003548124030000111
is the median of the preceding transient window, and the learning method is
The server is matched with the mobile APP, and online and historical data training learning can be performed.
A typical supervised learning process is as follows:
a1, training start: disconnecting all the devices;
a2, opening the equipment to be identified;
a3, waiting for 5 seconds;
a4, selecting the equipment;
a5, closing;
a6, waiting for 5 seconds;
a7, opening;
a8, waiting for 5 seconds;
a9, selecting the equipment;
a10, closing;
a11, waiting for 5 seconds;
a12, opening;
a13, waiting for 5 seconds;
a14, selecting the equipment;
and A15, finishing.
Wherein, load identification belongs to binary classification, and F scoring is adopted to evaluate identification accuracy:
Figure 1
where precision is a positive predictive value, call is a true positive predictive value ratio, tp is true positive, fp is false positive, fp is the predictive device is on but off, fn is false negative, fn is the device is on but expected to be off, and tp, fp and fn can be determined by a time data block.
The implementation method of the non-intrusive load identification and energy consumption monitoring system of the Internet of things is characterized by comprising the following steps:
s1, collecting: recording the measurement data;
s2, analysis: analyzing the measurement data and determining any user configuration;
s3, identifying: recognizing and creating a user consumption pattern by using an AI (artificial intelligence) and machine learning methods, and sending the user consumption pattern to analysis service software;
s4, prediction: the analysis system detects any deviation of the user mode, carries out prediction maintenance, finds any fault in the early stage, predicts a future event according to historical data and realizes prediction alarm;
s5, operation control: and controlling the equipment by using an optimal power consumption strategy through the technology of the Internet of things.
Specifically, through a sensor interface unit of the non-invasive load monitoring equipment, the sensor interface unit comprises four groups of eight current sockets and four paths of voltage sockets, the sensor interface supports a contact type sensor and a non-contact type sensor, the current sensor is installed and processed, parameters of current and voltage of the current sensor are collected simultaneously, a control and processing module of the non-invasive load monitoring equipment completes control, processing and communication functions, the non-invasive load monitoring equipment is provided with a dual-core processor, a high-capacity onboard flash memory and a wireless module, and has two modes of independent operation and server operation, when in the independent operation mode, the identification of an electric load is completed by utilizing an edge calculation and an embedded load identification model, a web service and a browser page are built in a software firmware, when in use, other software outside a browser is not required to be installed, and the load identification and energy consumption condition are browsed and accessed through a mobile phone, a tablet computer or a computer network, in the operation mode of the server, the software firmware adopts an MQTT internet of things protocol to send power consumption data to a designated analysis server, the analysis software in the server is responsible for load identification and energy consumption analysis and sends an analysis result to a subscriber, the control and processing module is provided with an integrated 2.4GHz wireless communication unit, the integrated 2.4GHz wireless communication unit has a Wi-Fi and Bluetooth connection function, the electric energy metering module is responsible for collecting voltage and current data and calculating to obtain parameters such as active power, reactive power, apparent power, power factors and the like, and sends the power consumption data to the control and processing module, according to the operation mode, the independent operation and server operation mode is adopted, according to the characteristic quantity for identifying the type of the electric equipment, the load type identification and the energy consumption analysis are completed, the characteristic quantity of the power consumption data is used for providing a DC5V working power supply for the non-invasive load monitoring equipment, the power supply monitoring system is also used for providing DC3.3V working power supplies for the electric energy metering module and the control and processing module, and comprises an AC/DC conversion unit and a DC/DC isolation conversion unit, wherein the AC/DC conversion unit provides DC5V working power supplies for equipment, the input end of the AC/DC conversion unit is electrically connected with a voltage socket, the DC/DC isolation conversion unit provides DC3.3V working power supplies for the electric energy metering module and the control and processing module, the input end of the DC/DC conversion unit is the output end of the AC/DC conversion unit, the mobile APP is used for a consumer to browse the energy consumption situation on the mobile application program anytime and anywhere, and the software server is used for storing the electricity consumption data and deploying load decomposition, identification and energy consumption analysis software.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (10)

1. The non-invasive load identification and energy consumption monitoring system of the Internet of things is characterized by comprising non-invasive load monitoring equipment, a mobile APP and a software server, wherein the output end of the non-invasive load monitoring equipment is electrically connected with the input end of the mobile APP, the mobile APP is in bidirectional connection with the software server, the non-invasive load monitoring equipment comprises a sensor interface unit, a control and processing module, an electric energy metering module and a power management module, and the sensor interface unit and the control and processing module are electrically connected with the electric energy metering module and the power management module;
the sensor interface unit is used for installing the current sensor and simultaneously collecting parameters of current and voltage of the current sensor;
the control and processing module is used for accessing load identification and energy consumption conditions through network browsing and is also used for analyzing and taking charge of load identification and energy consumption analysis;
the electric energy metering module is used for collecting voltage and current data, calculating to obtain parameters such as active power, reactive power, apparent power, power factors and the like, and sending the power utilization data to the control and processing module;
the power management module is used for providing a DC5V working power supply for the non-invasive load monitoring equipment and also used for providing a DC3.3V working power supply for the electric energy metering module and the control and processing module;
the mobile APP is used for browsing energy consumption conditions on the mobile application program by the consumer at any time and any place;
the software server is used for storing the electricity utilization data and deploying the load decomposition, identification and energy consumption analysis software.
2. The internet of things non-intrusive load identification and energy consumption monitoring system of claim 1, wherein: the sensor interface unit comprises four groups of eight current sockets and four paths of voltage sockets, the sensor interface supports contact type and non-contact type sensors, the sampling rate can reach 2 kHz-8 kHz, the current sockets are used for monitoring parameters such as current and flow, the current sockets comprise four groups of eight ports of L1 k, L1L, L2 k, L2L, L3 k, L3L, L N k and N L and are used for connecting the contact type and non-contact type current sensors and simultaneously used for collecting current parameters of L1, L2, L3 and N, and the voltage sockets comprise four ports of L1, L2, L3 and N and are used for connecting 220/400VAC three-phase four-wire system alternating current and simultaneously used for collecting voltage parameters of L1, L2, L3 and N.
3. The internet of things non-intrusive load identification and energy consumption monitoring system of claim 2, wherein: the control and processing module comprises an independent operation mode and a server operation mode, the independent operation mode utilizes an edge calculation and an embedded load identification model to complete the identification of the power load, a web service and a browser page are arranged in a software firmware, other software outside the browser does not need to be installed during use, the load identification and the energy consumption condition are browsed and accessed through a mobile phone, a tablet computer or a computer network, the embedded load identification model is trained through an LSTM-RNN model established by the server application, and after the training is completed, TensorFlow Lite is used for compression processing and is led into equipment through OTA;
the server operation mode adopts an MQTT internet of things protocol to send power utilization data to a designated analysis server, and analysis software in the server is responsible for load identification and energy consumption analysis and sends an analysis result to a subscriber.
4. The internet of things non-intrusive load identification and energy consumption monitoring system of claim 3, wherein: the control and processing module is provided with an integrated 2.4GHz wireless communication unit, and the integrated 2.4GHz wireless communication unit is provided with Wi-Fi and Bluetooth connection functions.
5. The internet of things non-intrusive load identification and energy consumption monitoring system of claim 4, wherein: the characteristic quantities of the electricity utilization data comprise steady state, transient state and operation mode, the steady state and the transient state are determined by the characteristics of components inside the equipment, and the operation mode is determined by the operation control strategy of the equipment.
6. The internet-of-things non-intrusive load identification and energy consumption monitoring system as defined in claim 5, wherein: the load identification comprises a load identification model whose power supply entries p represent the sum of the active power of all the individual devices at time t, the total power can then be expressed as
Figure FDA0003548124020000021
Wherein y isi(T) represents the power consumption of the ith device of the I available devices at time T, knowing the total power over time duration
Figure FDA0003548124020000025
The goal is to obtain the energy consumption of the ith device
Figure FDA0003548124020000026
7. The internet-of-things non-intrusive load identification and energy consumption monitoring system as defined in claim 6, wherein: the load recognition model has P and dP input time series vectors: p is the total power P recorded by the non-invasive load monitoring equipment, and a training data set PT，iT training samples for establishing the ith device, wherein
Figure FDA0003548124020000022
Wherein
Figure FDA0003548124020000023
Is derived from the aggregate signal P and
Figure FDA0003548124020000024
represents the response device energy consumption of the nth sample;
the dP represents the change in power. The power change is introduced, so that transient characteristics are provided for load identification on one hand, and the influence of noise on learning model training is reduced on the other hand.
8. The internet of things non-intrusive load identification and energy consumption monitoring system of claim 7, wherein: the data of the transient characteristics are time series power data, a power data window is determined through a starting timestamp and an ending timestamp, and the transient characteristics are defined by selecting the following three characteristics:
firstly, the power change delta P of the equipment from starting to reaching a steady state;
Figure FDA0003548124020000027
Figure FDA0003548124020000028
is the median value of the post-transient window,
Figure FDA0003548124020000029
is the median of the previous transient window.
Second, maximum peak power variation Pmax；
Figure FDA00035481240200000210
max(x1) Is the maximum value of the transient window and,
Figure FDA00035481240200000211
is the median of the previous transient window.
Third is the minimum peak power variation Pmin。
Figure FDA00035481240200000212
min(x1) Is the minimum value of the transient window and,
Figure FDA00035481240200000213
is the median of the previous transient window.
9. The internet of things non-intrusive load identification and energy consumption monitoring system of claim 8, wherein: the load identification belongs to binary classification, and the accuracy of the identification is evaluated by adopting F scoring:
Figure FDA0003548124020000031
where precision is a positive predictive value, recall is a true positive predictive value ratio, tp is true positive, fp is false positive, fp is the prediction device is on but off, fn is false negative, fn is the device is on but expected to be off, and tp, fp, and fn can be determined by time data street.
10. The method for implementing the non-intrusive load identification and energy consumption monitoring system of the internet of things as claimed in claims 1 to 9, is characterized by comprising the following steps:
s1, collecting: recording the measurement data;
s2, analysis: analyzing the measurement data and determining any user configuration;
s3, identifying: using AI and machine learning method to identify and create user consumption pattern, and sending to analysis service software;
s4, prediction: the analysis system detects any deviation of the user mode, carries out prediction maintenance, finds any fault in the early stage, predicts a future event according to historical data and realizes prediction alarm;
s5, operation control: and controlling the equipment by using an optimal power consumption strategy through the technology of the Internet of things.
CN202210254065.3A 2022-03-15 2022-03-15 Non-invasive load identification and energy consumption monitoring system of Internet of things and implementation method thereof Pending CN114598722A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210254065.3A CN114598722A (en) 2022-03-15 2022-03-15 Non-invasive load identification and energy consumption monitoring system of Internet of things and implementation method thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210254065.3A CN114598722A (en) 2022-03-15 2022-03-15 Non-invasive load identification and energy consumption monitoring system of Internet of things and implementation method thereof

Publications (1)

Publication Number Publication Date
CN114598722A true CN114598722A (en) 2022-06-07

Family

ID=81809173

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210254065.3A Pending CN114598722A (en) 2022-03-15 2022-03-15 Non-invasive load identification and energy consumption monitoring system of Internet of things and implementation method thereof

Country Status (1)

Country Link
CN (1) CN114598722A (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106655160A (en) * 2016-10-27 2017-05-10 国家电网公司 Non-intrusion electric power load decomposition identification decision method and system
CN108132379A (en) * 2017-12-11 2018-06-08 武汉大学 Non-intrusion type load monitor system and recognition methods based on cloud platform
CN211856726U (en) * 2020-03-24 2020-11-03 陈焕秋 Non-invasive load monitoring system based on optical coupling isolation circuit
CN112434799A (en) * 2020-12-18 2021-03-02 宁波迦南智能电气股份有限公司 Non-invasive load identification method based on full convolution neural network
CN112909923A (en) * 2021-01-21 2021-06-04 北京理工大学 Non-invasive household load behavior recognition device based on DTW algorithm
CN113206547A (en) * 2021-04-28 2021-08-03 湖北工业大学 Transformer substation power distribution system based on non-invasive power load monitoring

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106655160A (en) * 2016-10-27 2017-05-10 国家电网公司 Non-intrusion electric power load decomposition identification decision method and system
CN108132379A (en) * 2017-12-11 2018-06-08 武汉大学 Non-intrusion type load monitor system and recognition methods based on cloud platform
CN211856726U (en) * 2020-03-24 2020-11-03 陈焕秋 Non-invasive load monitoring system based on optical coupling isolation circuit
CN112434799A (en) * 2020-12-18 2021-03-02 宁波迦南智能电气股份有限公司 Non-invasive load identification method based on full convolution neural network
CN112909923A (en) * 2021-01-21 2021-06-04 北京理工大学 Non-invasive household load behavior recognition device based on DTW algorithm
CN113206547A (en) * 2021-04-28 2021-08-03 湖北工业大学 Transformer substation power distribution system based on non-invasive power load monitoring

Similar Documents

Publication Publication Date Title
CN112180193B (en) Non-invasive load identification system and method based on track image identification
US11605036B2 (en) System and methods for power system forecasting using deep neural networks
Schoofs et al. Annot: Automated electricity data annotation using wireless sensor networks
US20140005853A1 (en) Method and system for monitoring electrical load of electric devices
CN106464988B (en) Energy management system
CN112434799B (en) Non-invasive load identification method based on full convolution neural network
CN111580449A (en) Energy management control system based on narrow-band Internet of things technology
CN108964270A (en) A kind of intelligent appliance load detecting and control system and its method
CN112947127A (en) Intelligent electricity consumption control management system
CN110309984B (en) Non-invasive load identification and short-term user behavior prediction method
US11016129B1 (en) Voltage event tracking and classification
CN105320097A (en) Method, cloud platform and system for realization of intelligent control of household appliances
KR20190062741A (en) Electric Power Consumption Analyzing Method Using Deep Learning in Remote Power Control System
CN106225932A (en) Wireless low-power consumption infrared sensor and control method thereof and application
CN110545290A (en) Electrical equipment intelligent lock management device based on internet of things
CN102116810B (en) Power supply smart register system and method
CN103018611B (en) Non-invasive load monitoring method and system based on current decomposition
CN113328527A (en) Intelligent power distribution automatic terminal device and intelligent power distribution fault diagnosis method
CN116125204A (en) Fault prediction system based on power grid digitization
CN203338106U (en) Wireless intelligent housing system based on wifi
CN213243562U (en) Smart home life management system
CN113890024A (en) Non-invasive load intelligent decomposition and optimization control method
CN114598722A (en) Non-invasive load identification and energy consumption monitoring system of Internet of things and implementation method thereof
CN102156207A (en) Distributed intelligent electric energy recording system and detecting method thereof
CN111324073A (en) Machine tool detection and analysis method and analysis platform thereof

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination