CN112083224A - Non-invasive real-time intelligent electric meter system based on characteristic quantity matching and identification method - Google Patents
Non-invasive real-time intelligent electric meter system based on characteristic quantity matching and identification method Download PDFInfo
- Publication number
- CN112083224A CN112083224A CN202010562333.9A CN202010562333A CN112083224A CN 112083224 A CN112083224 A CN 112083224A CN 202010562333 A CN202010562333 A CN 202010562333A CN 112083224 A CN112083224 A CN 112083224A
- Authority
- CN
- China
- Prior art keywords
- vector
- transient
- characteristic quantity
- representing
- appliance
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000005070 sampling Methods 0.000 claims abstract description 24
- 230000006854 communication Effects 0.000 claims abstract description 16
- 238000004891 communication Methods 0.000 claims abstract description 14
- 238000004364 calculation method Methods 0.000 claims abstract description 13
- 238000001514 detection method Methods 0.000 claims abstract description 10
- 230000003993 interaction Effects 0.000 claims abstract description 5
- 239000013598 vector Substances 0.000 claims description 103
- 230000001052 transient effect Effects 0.000 claims description 56
- 239000011159 matrix material Substances 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 3
- 230000009286 beneficial effect Effects 0.000 abstract description 4
- 230000006870 function Effects 0.000 description 15
- 230000008569 process Effects 0.000 description 9
- 230000005540 biological transmission Effects 0.000 description 8
- 238000013461 design Methods 0.000 description 8
- 238000005457 optimization Methods 0.000 description 7
- 238000011160 research Methods 0.000 description 4
- 230000005611 electricity Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 102100034458 Hepatitis A virus cellular receptor 2 Human genes 0.000 description 2
- 101001068133 Homo sapiens Hepatitis A virus cellular receptor 2 Proteins 0.000 description 2
- 101000669511 Homo sapiens T-cell immunoglobulin and mucin domain-containing protein 4 Proteins 0.000 description 2
- 102100039367 T-cell immunoglobulin and mucin domain-containing protein 4 Human genes 0.000 description 2
- 238000012850 discrimination method Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 101150025671 tim8 gene Proteins 0.000 description 2
- 230000001960 triggered effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R22/00—Arrangements for measuring time integral of electric power or current, e.g. electricity meters
- G01R22/06—Arrangements for measuring time integral of electric power or current, e.g. electricity meters by electronic methods
- G01R22/10—Arrangements for measuring time integral of electric power or current, e.g. electricity meters by electronic methods using digital techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/22—Indexing; Data structures therefor; Storage structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Educational Administration (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Analysis (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Mathematical Optimization (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Computing Systems (AREA)
- Algebra (AREA)
- Game Theory and Decision Science (AREA)
- Remote Monitoring And Control Of Power-Distribution Networks (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Power Engineering (AREA)
Abstract
The invention discloses a non-invasive real-time intelligent electric meter system and an identification method based on characteristic quantity matching, wherein the system comprises a sampling circuit, an electric energy metering module, a communication module, a learning machine and an HMI (human machine interface) serial port screen, the sampling circuit detects voltage and current signals of each electric appliance and sends the detection signals to the electric energy metering module, the electric energy metering module analyzes the voltage and current signals to generate steady-state characteristic quantity and transient-state characteristic quantity in each electric appliance and sends the characteristic quantities to the learning machine, the learning machine records characteristic quantity data as a database and calculates the newly detected characteristic quantity so as to judge the load state of each electric appliance, and the learning machine carries out data interaction with the HMI serial port screen through the communication module. The invention adopts multi-characteristic quantity to calculate, improves the universality of the electric meter, adopts a non-intrusive load identification algorithm, improves the calculation rate and the universality of the electric meter, and is beneficial to identifying the electric appliance with smaller household power.
Description
Technical Field
The invention relates to the field of electric energy metering, in particular to a non-invasive real-time intelligent electric meter system and an identification method based on characteristic quantity matching.
Background
The existing research of non-invasive load identification is mainly in a laboratory stage, but the research gradually shows the trend of commercialization marketing for better electric energy metering and management and the corresponding national smart grid call.
At present, the design uses a transient power characteristic vector at the time of starting as a characteristic quantity of identification to obtain a transient power curve with the time as long as 2s from the report of electrical technology on household appliance load identification based on the characteristic of starting transient load. The processing method is to introduce an opening coefficient and a delay coefficient, multiply the opening coefficient by the waveform vector of the power of the electric appliance in the database, and superpose after displacement according to the delay coefficient, and the superposed signal is used as a fitting signal to be compared with the acquired characteristic quantity. The similarity degree of the fitting signals and the collected characteristic quantities is specified to be the Pearson coefficients of two groups of vector waveforms, high-dimensional optimization is carried out through a particle swarm algorithm, namely a modern optimization algorithm, and an optimized objective function, namely the Pearson coefficients of two groups of vector waveforms are minimum. And finally, determining whether the electrical appliances are started or not according to the trained starting coefficients, and determining the accurate starting time of the electrical appliances according to the delay coefficients. But in practical use, the following disadvantages are very significant: firstly, the selection of characteristic quantity is less, so that the method is difficult to deal with the electric appliance with high waveform similarity when the electric appliance is put into use, is not beneficial to the increase of the capacity of a database, and has poor robustness to different electric appliances; and the processing time of the second high-dimensional optimization algorithm is long, so that the real-time performance of the electric meter is influenced. The algorithm of modern optimization algorithms solves the NP-hard problem, with its complexity rising rapidly as dimensionality increases. When there are N electrical appliance data in the database, the dimension of the optimization algorithm is 2 × N (each electrical appliance contains an on coefficient and a delay coefficient). From this point of view, the increase of the database capacity is also not facilitated; third, it presents disadvantages when addressing the general situation. There are generally fewer opportunities for both appliances to be switched on or off simultaneously (i.e., the time difference between the two appliances being switched on or off is less than 2 seconds). Under the condition that only one electric appliance is put into use at the same time, the algorithm has no advantages, the complexity of the algorithm is not reduced, and meanwhile, in the aspect of hardware, due to the fact that the electric meter is lack of control flexibility due to the fact that the electric meter is controlled by the keys, the function expansion of the electric meter is not utilized, the software and hardware design is suitable for being used as an experimental platform in a laboratory for research purposes, but is not suitable for being applied to actual load detection occasions and is not suitable for being popularized in the market.
The other publication is from graduation design of 'research on residential electrical load identification and design of intelligent meters', the design mainly adopts a steady-state active power average value and each current harmonic as characteristic quantities, and matches characteristic library data by using a mode of combining a KNN algorithm and an SVM support vector machine algorithm, and the design is actually comprehensive application of a linear discrimination method based on Euclidean distance and combining a nonlinear discrimination method of an SVM. And after the switching algorithm is added into the filtering algorithm, performing variable point detection, and when the switching algorithm detects that an electric appliance is switched in or a cutting event occurs, performing pre-classification according to the load type and the power threshold value and entering different classifier options. And respectively training each class of classifier, and comprehensively utilizing an improved KNN method and an SVM method for judging.
The algorithm design is used for verifying the feasibility of the KNN algorithm and the SVM algorithm in judging the load type, and the algorithm is simple and has high operation speed. But the practical application has very large limitation and the following disadvantages: firstly, parameters of each classifier need to be trained independently (in a thesis, a grid method is adopted for optimization, namely a traversal method), and the robustness for different electrical appliances is poor; second, before untraining, it is unpredictable whether a classifier utilizes the KNN algorithm or the SVM algorithm, which leads to further deterioration in the utility of such algorithms. The electric meter can only be suitable for a classifier with high recognition rate by carrying out prior training under the condition of known electric appliance types, only has the capability of non-intrusive load recognition of the electric meter and basic peripherals such as storage, communication, alarm and the like, and needs to be combined with external functional components for use when in actual use. The electric meter has no learning function, can only identify the trained electric appliances, and has no capacity of expanding a database.
In summary, the feature quantity of the current electric meter system is less, the verified data are only several electric appliances which are trained, and the universality which can fully meet the practical requirements is difficult to achieve.
The capability of applying complex occasions is pursued, the calculation speed and the algorithm universality are sacrificed, and the real-time tracking capability is poor.
In order to pursue higher accuracy, the accuracy of the electric meter is excessively sacrificed, and the electric meter is not recognizable for household appliances with lower power (less than 100w active power), such as electric fans, fluorescent lamps, computers and the like. In fact, the availability of the electricity meter is reduced.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a non-invasive real-time intelligent electric meter system and an identification method based on characteristic quantity matching.
In order to solve the technical problems, the invention provides the following technical scheme:
the invention relates to a non-invasive real-time intelligent electric meter system based on characteristic quantity matching, which comprises a sampling circuit, an electric energy metering module, a communication module, a learning machine and an HMI (human machine interface) serial port screen, wherein the sampling circuit detects voltage and current signals of each electric appliance and sends the detection signals to the electric energy metering module, the electric energy metering module analyzes the voltage and current signals to generate steady-state characteristic quantity and transient-state characteristic quantity in each electric appliance and sends the characteristic quantities to the learning machine, the learning machine records characteristic quantity data as a database and calculates the newly detected characteristic quantity so as to judge the load state of each electric appliance, and the learning machine carries out data interaction with the HMI serial port screen through the communication module.
As a preferred technical scheme of the invention, the learning machine sends the load state data of each electric appliance to the HMI serial port screen, the HMI serial port screen displays real-time parameters of each electric appliance, the user sends a control instruction to the HMI serial port screen, the HMI serial port screen sends the control instruction to the learning machine through the communication module, and the learning machine executes the control of the electric meters according to the control instruction.
As a preferred technical solution of the present invention, the sampling circuit includes a voltage sampling and a current sampling, the voltage sampling is bridged between a zero line and a live line by a resistance voltage division, the current sampling uses a current transformer, and a transformation ratio is 1000: 1.
as a preferred technical scheme of the invention, the electric energy meter further comprises a power supply circuit, wherein the power supply circuit adopts a 220V-5V switching power supply as a main power supply to ensure that the system supplies power to the system under the condition of no commercial power, and the learning machine and the electric energy metering module adopt a microprocessor to store, calculate and identify parameters of each electric appliance.
The invention relates to a non-invasive load identification method based on characteristic quantity matching, which comprises the following steps:
1. detecting values of steady-state characteristic quantities and transient-state characteristic quantities in each electrical appliance;
2. respectively calculating the score value of the steady-state characteristic quantity and the score value of the transient-state characteristic quantity in each electric appliance by adopting a load identification method, and obtaining the total score value of each electric appliance;
3. and determining the load state of each electric appliance according to the total score value of each electric appliance.
As a preferred technical solution of the present invention, the calculation formula of the total score value of each electrical appliance is:
Pgeneral assembly＝ω5Pt3+(1-ω5)Ps
Pt3Score value, P, representing transient characteristic quantity of each electric appliancesScore value, ω, representing steady-state characteristic quantity of each electric appliance5Weight, 1-omega, representing a transient characteristic quantity5And representing the weight of the steady-state characteristic quantity.
As a preferred technical solution of the present invention, the calculation formula of the score of the transient characteristic amount is:
Pt3＝[ω2Pt11+(1-ω2)Pt21]·ω3+[ω2Pt12+(1-ω2)Pt22]·(1-ω3)
Pt3＝ω4Pt13+(1-ω4)Pt23
wherein, Pt13Representing the value of the transient vector waveform score when each electrical appliance is turned off; pt23Representing the first order difference vector score value when each electrical appliance is turned off; omega4Weight, 1-omega, representing the score of the transient vector waveform at turn-off4A weight representing a first order difference vector score at turn-off; pt11The temporary vector waveform score value of the reactive vector when each electric appliance is started is represented; pt21First order difference vector score, P, representing reactive vectors at turn-on of each appliancet12The method comprises the steps of representing a transient vector waveform score value of an active vector when each electric appliance is started; pt22First order difference vector score value, omega, representing the active vector at the time of start of each appliance3Weight, 1-omega, representing active vector at start-up3Representing the weight of the reactive vector when the switch is switched on; omega2Weight, 1-omega, representing the score of the transient vector waveform at turn-on2A weight representing a first order difference vector score at startup;
the calculation formula of the score value of the steady-state characteristic quantity is as follows:
Ps＝d2(x，Gi)＝x1∑-1x-2x1∑-1ui+ui’∑-1ui
wherein x is the collected steady-state characteristic value vector, Gi is the barycentric coordinate of the characteristic value of the ith type of electrical appliance in the database, Σ is the covariance matrix of six kinds of steady-state quantities corresponding to the type of electrical appliance, and μ i is the mean value of the steady-state characteristic quantity vectors of the ith type of electrical appliance in the database.
As a preferred technical solution of the present invention, a calculation formula of the transient vector waveform score values of the electrical appliances is:
Pt1＝ω1D(x’，y’)+(1-ω1)·Na，b
where D (x, y) describes the closeness of the waveform trend between the transient vector and the vector in the database,
n represents vector dimension, xi and yi respectively represent ith value in database and sampled vector, and x 'i and y' i are corresponding values of normalized vector elements and meet the requirement
N (a, b) represents the closeness of the vector magnitude;
a. b is the maximum of the elements in vector x and vector y, ω1Weight, 1-omega, representing the trend of the waveform at turn-on1Representing the weight of the waveform amplitude;
the calculation formula of the transient vector first-order difference vector waveform score value of each electric appliance is as follows:
Pt2＝ω1D(Δx’，Δy’)+(1-ω1)·Nm，n
where D (Δ x, Δ y) represents the closeness of the waveform trend between the transient vector and the vector in the database when the transient vector waveform vector length is less than 1, and N (m, N) represents the closeness of the vector magnitude when the transient vector waveform vector length is less than 1.
As a preferred embodiment of the present invention, the step of determining the load type of each electrical appliance includes: calculating steady state scores ten times before and after the transient score is obtained, comprehensively obtaining total scores ten times, respectively setting an upper limit threshold value DH1 and DH2 for starting and stopping, when the database electrical appliance has the score lower than the upper limit threshold value, taking the electrical appliance with the lowest score, counting the position of the corresponding electrical appliance by one, and when the ten times are finished, considering the electrical appliance with the value greater than 5 times as the currently-switched-in or switched-off electrical appliance, and obtaining a judgment result.
Compared with the prior art, the invention has the following beneficial effects:
the invention adopts multi-characteristic quantity to calculate, improves the universality of the electric meter, adopts a non-intrusive load identification algorithm, improves the calculation rate and the universality of the electric meter, is beneficial to identifying electric appliances with lower household power, and simultaneously utilizes the HMI serial port screen to enable the learning machine to interact with users, thereby increasing the flexibility of the electric meter and improving the experience of the users.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
FIG. 1 is a flow chart of the operation of an electric meter system;
FIG. 2 is a workflow diagram of a learning machine;
fig. 3 is a flow chart of the structure of the electricity meter system.
Detailed Description
The preferred embodiments of the present invention will be described in conjunction with the accompanying drawings, and it will be understood that they are described herein for the purpose of illustration and explanation and not limitation.
In addition, if a detailed description of the known art is not necessary to show the features of the present invention, it is omitted.
Example 1
The invention provides a non-invasive load identification method based on characteristic quantity matching, which uses characteristic quantities including steady-state characteristic quantities: the active power average value, and the amplitude of each current harmonic (including current fundamental wave, third harmonic wave, fifth harmonic wave, seventh harmonic wave and ninth harmonic wave); transient characteristic quantity: the average vector of active power with the opening time of 700ms (empirical value), the average vector of reactive power and the average vector of reactive power with the closing time of 400ms (empirical value).
In order to organically combine these characteristic quantities together, the following definitions are made:
steady state score for each appliance:
Ps＝d2(x，Gi)＝(x-ui)T∑-1(x-ui)
after the deployment:
Ps＝d2(x，Gi)＝x1∑-1x-2x1∑-1ui+ui’∑-1ui
the meaning of the above variables is explained. x is the vector of the collected steady-state characteristic value, and Gi is the barycentric coordinate of the characteristic value of the ith type of electrical appliance in the database. And the sigma is a covariance matrix of six steady-state quantities corresponding to the type of the electric appliance, and the mean value of the steady-state characteristic quantity vectors of the type i electric appliance in the mu i database.
Transient vector waveform score for each appliance:
Pt1＝ω1D(x，y)+(1-ω1)·Na，b
the meaning of the above variables is explained. D (x ', y') describes the closeness of the waveform trend between the transient vector and the vector in the database. n represents the vector dimension, xi and yi respectively represent ith values in the database and the vectors obtained by sampling, x 'i and y' i are corresponding values of vector elements after normalization, and a and b are respectively the maximum values of the elements in the vector x and the vector y. The normalization satisfies the following relationship:
n (a, b) represents the closeness of the vector magnitude. The method is used for compensating the amplitude relation which cannot be reflected in the waveform trend. w1 and (1-w1) represent the weight relationship between the waveform trend and the amplitude.
Transient vector first order difference vector waveform score for each appliance:
Pt2＝ω1D(Δx’，Δy’)+(1-ω1)·Nm，n
the meaning of each variable in the above equation is similar to the definition of the transient vector waveform score. But the length of the first order difference vector is 1 less than the transient vector waveform score.
Transient score when each appliance was turned on:
Pt3＝[ω2Pt11+(1-ω2)Pt21]·ω3+[ω2Pt12+(1-ω2)Pt22]·(1-ω3)
w2, (1-w2) represents the weighted relation between the transient vector waveform score and the first order difference vector score when turned on.
w3, and (1-w3) represent the weight relation between reactive vectors and active vector scores when the switch is switched on.
Transient score when each appliance was off:
Pt3＝ω4Pt12+(1-ω4)Pt22
w4, (1-w4) represents the weighted relation of the transient vector waveform score and the first order difference vector score when turned off.
Total score of steady state transient state of each electric appliance:
Pgeneral assembly＝ω5Pt3+(1-ω5)Ps
w5, (1-w5) represents the weighted relationship of steady-state score and transient score.
Finally, the decision is made according to the total score of each electric appliance, and the lower the score is, the better the score is. And in decision making, the thought of majority voting in KNN is used for reference, ten times of steady-state scores are calculated before and after the transient score is solved, and ten times of total scores are obtained comprehensively. An upper threshold value DH1, DH2 is set for turning on and off respectively, when the database electrical appliances have scores lower than the upper threshold value, the electrical appliance with the lowest score is taken, and the position count of the corresponding electrical appliance is increased by one. And after the ten times, considering the electric appliance with the value more than 5 times as the electric appliance which is currently put into or cut off, and obtaining a judgment result.
In the whole process of calculating the decision score, there are 5 weights, which are w1, w2, w3, w4 and w 5. The optimal value can be obtained through modern optimization algorithm training, and a wider range can be obtained through experience debugging.
The switching algorithm is that a power threshold value is set through a variable point detection mode of an active average value, and when the power threshold value is continuously larger than a set value and meets the input condition, the fact that an electric appliance is input is indicated; and when the power threshold is smaller than the set value and the cutting-off condition is met, indicating that the electric appliance is cut off. And judging the electric appliance once when the electric appliance is switched on or switched off.
Example 2
As shown in fig. 3, the non-intrusive real-time smart meter system based on feature quantity matching mainly comprises a smart meter and a matched learning machine component, wherein the smart meter is installed at a household power bus inlet during working, a learning machine is placed at any position in a house and is matched with a plug and a socket for use during execution of a learning function, and the smart meter comprises a power supply circuit, a sampling circuit, an electric energy metering module, a communication module and a microprocessor. Besides the above, the learning machine also has an HMI serial port screen for interacting with users.
The power supply circuit adopts a 220V-5V switching power supply as a main power supply, and the AMS1117 continuously reduces the voltage to 3.3V for power supply. This is to ensure that the meter can supply power directly to the device via the mains supply without an external dc power supply. When the learning machine does not execute the learning function, the learning machine does not need to be connected with the mains supply, so that a USB interface is reserved on the learning machine to provide 5V power.
The sampling circuit is divided into voltage sampling and current sampling. The voltage sampling is divided by a resistor, and the voltage is bridged between a zero line and a live line, wherein the voltage division ratio is 526: 1. the current sampling utilizes a current transformer, and the transformation ratio is 1000: 1. The collected current signal and voltage signal are respectively sent to VP1, VN1, VP3 and VN3 of the electric energy metering chip.
The ATT7053AU is selected as the electric energy metering chip, and the electric energy metering chip is provided with one voltage channel and two current channels. When in use, only the voltage channels VP3 and VN3 and the first current channels VP1 and VN1 are used. The effective characteristic quantity finally output by the chip comprises a current waveform numerical value, an active power waveform numerical value, a reactive power waveform numerical value and an active power average value. The chip can be configured with interrupt and output through an IRQ pin, and only voltage zero-crossing interrupt is configured in the invention.
The model of the HMI serial port screen is TJC4827K043, the HMI serial port screen has rich functional controls and instruction sets, and the HMI serial port screen is communicated with the single chip microcomputer through a serial port. The mode of controlling the HMI display by the single chip microcomputer is a character string mode of directly inputting instructions in a serial port, and 30 xff are added as end characters at last.
The communication module adopts an NRF24L01 wireless communication module.
The microprocessor uses stm32f407VET6, which lists the internal resources and applications used by the chip.
Interrupt configuration and priority settings (ammeter)
| Interrupt type | Preemption priority | Response priority |
| TIM3 timed interrupts | 0x00 | 0x01 |
| TIM7 timed interrupts | 0x01 | 0x00 |
| TIM5 timed interrupts | 0x01 | 0x01 |
| PE8 external interrupts | 0x01 | 0x02 |
| PA3 external interrupt | 0x01 | 0x03 |
| USART1 serial port interrupt | 0x02 | 0x01 |
| TIM4 timed interrupts | 0x02 | 0x02 |
| TIM8 timed interrupts | 0x03 | 0x03 |
(learning machine)
System design system defined important global variable
In addition, there are many global variables defined for implementing some local functions, for caching data, counting, etc. These variables are not listed again because they are not mentioned below in the flow of specific functional implementations.
In addition to the global variables defined, there are many macro definitions, which are described herein.
A set of macro definitions whose tag data is stored at a specific location in the database. To facilitate storing and transferring database data, the database is stored in a contiguous space. And defines a series of starting addresses and data lengths for convenient recall of a set of data of a particular meaning.
2. Defining the initial address of the 7 th sector of the database stored in the FLASH inside the database, and defining the address as FLASH _ SAVE _ ADDR which is a 16-system number 0x 08060000.
Serial port instruction set sent by HMI (human machine interface) during user operation
Communication instruction set of learning machine in interaction
Communication instruction set of electricity meter in interaction
As shown in fig. 1-2, the working process of the main functions is explained in detail as follows:
the process of switching detection and load identification control is performed by the currsta.
In normal operation, currsta is 0. And entering a PA3 external interrupt, calculating average power and carrying out switching detection. When no event occurs, the currsta value cannot be changed; and executing a transient active and reactive power acquiring process when a switching event occurs until the transient process is ended, wherein a flag cursta bit0 is 1. Transient information is stored in pvector and qvector when the switch is switched on, the lower two bits of a flag cursta are 2 when the switch is switched off, and the transient information is stored in qvector and.
Since the end of the interruption is not just zero crossing of the voltage, an external interruption is waited for again. When the currsta bit0 is detected to be 1, switching detection is not executed, and only the currsta bit2 is detected to be 1. When bit2 is detected to be 1 in the main function, TIM5 is enabled and an FFT operation is performed, and when the FFT operation is completed, currsta bit3 is 1. When the master function detects that currsta bit3 is 1, TIM4 is disabled, the active average is read again, and the steady state information is saved in fft _ output putbuf and apr. When the lower two bits are detected to be 2, then
And then, carrying out a distinguishing process, and updating the SWI _ STA zone bit after the distinguishing is finished. currsta ═ 0, initializes all function variables simultaneously.
And (5) controlling the switching detection and characteristic quantity learning process by the currsta.
When entering the learning mode, the HMI prompts a signal to switch on the electrical appliance. When an electric appliance is switched on, a transient active and reactive power acquisition process is collected in an external interrupt function of the PA3 until the transient process is finished, and a flag cursta bit0 is 1. When the external interrupt is entered again, currsta bit2 is made 1, indicating that the voltage has just passed zero. TIM3 is turned off, TIM5 is enabled, samples are taken and FFT is performed ten times, and steady state data is stored in record. After each FFT operation, currsta bit3 is 1. And after the learning function is detected, reserving two lower bits of the currsta and clearing other bits of the currsta. After ten operations are completed, currsta is 0.
At the moment, the HMI sends a cut-off signal again, the two lower bits of the cursta are changed into 2 when the cut-off of the electric appliance is detected, and the cursta is 0 after the transient state signal at the cut-off is stored.
The input and the cut-off are repeated for ten times, and the transient power vector is directly averaged and directly stored in the Data _ Base. 100 groups of data are stored in each characteristic quantity of the steady-state data, firstly, sorting is carried out, after sorting, the middle 80 groups are selected, the middle 50 groups are reserved, the mean vector and the covariance matrix of the characteristic values are obtained, and a constant term is obtained. The parameters as a function of the steady state criterion are stored in the Data Base.
And controlling the communication process by flag and Pro in the electric meter.
When an external interrupt is triggered by PE8, the value of the STATUS register is read, and the event of entering the interrupt is determined based on the value. When the event is that the maximum retransmission number is reached, the flag is 0x 10.
When the event is a successful reception, there are several cases: when rxbuf [0] is 0x30, the database of a certain electrical appliance is synchronized, rxbuf [1] is the corresponding position of the electrical appliance, a variable EPLOC is given, flag is set to be 0x40, and the next time the database content is marked to enter; when rxbuf [0] ═ 0x31, the information indicating that a certain electrical appliance is deleted, at this time, the database information of the electrical appliance is changed to 0, and flag is changed to 0; when rxbuf [0] is equal to 0x33 and indicates that power information is requested, the timer 7 is started, and flag is equal to 0x41, which marks that the next entry is successful in sending a power instruction; when rxbuf [0] ═ 0x34, timer 7 starts, and flag ═ 0x42, indicating that the next entry was successful in assigning transmission energy. When flag is 0x40, it is decided whether data transmission ends and the position to which data transmission is currently carried out according to Pro. When Pro 8+8> NUMERREAD is satisfied, it indicates the end of transmission, Pro is set to 0, and flag is set to 0.
When the event is a successful transmission, there are several cases: when flag is 0x41, it shows that the power command has been successfully sent, then the timer 7 is started, and the power information is sent after a timing of 10 ms. The power information is 8 floating point data stored in the run, is converted into 32-byte unsigned character type, is stored in the txbuf and is sent out, and flag is 0x 43; when the flag is 0x42, it indicates that the energy command has been successfully sent, then the timer 7 is started to send the energy information, and the flag is 0x 44. When the flag is 0x43 or 0x44, indicating that the power or energy information has been transmitted, the NRF24L01 is switched to the reception mode.
And controlling the communication process by flag in the learning machine.
As above, when an external interrupt is triggered by PE8, the STATUS register is read. When the event is the maximum retransmission times, the flag is 0x 10; when the transmission is successful, 0x20 is marked; when the reception is successful, the following cases are classified:
when rxbuf [0] is received, the rxbuf [0] is 0x33, the rxbuf [0] indicates that the frame data is a power transmission command, and the rxbuf [0] is 0x 41; when rxbuf [0] ═ 0x34, flag ═ 0x42 indicates that this frame data is an energy transmission instruction. When the flag is 0x41, the frame data is power information, and the flag is 0x44, and when the flag is detected to be 0x44 in the power query function, the power is modified, and the current power is sent to the HMI; when the flag is 0x42, the frame data is indicated as energy information, and the flag is 0x48, and when the flag is detected to be 0x48 in the energy query function, the energy is modified, and the current energy is sent to the HMI.
And (4) processing modes of electric meter multi-electric appliance operation.
And when multiple electric appliances are used, the characteristic quantity is calculated by adopting an superposition method. When the current characteristic value is calculated, the amplitude and the phase of each harmonic of the current are calculated by taking the voltage zero crossing as a reference. The sampling time reference before and after the electric appliance is put into the electric appliance is the same, so the amplitude phase can be directly added or subtracted by complex numbers. Thus obtaining the amplitude phase of each harmonic of the current newly input into the electric appliance. When the electric appliance is cut off, the electric appliance is cut off only by reducing the electric appliance before cutting off. The active power has no phase relation, and the added and cut electrical appliance power characteristic quantity can be obtained by directly carrying out addition and subtraction operation.
The transient characteristic quantity has no problems, and the transient quantity is gradually reduced to 0 after the previous electric appliance is put into operation. Therefore, the input or cut-off transient state quantity of the new electric appliance can be directly utilized.
And (4) processing the power and energy calculation by the electric meter.
When multiple electrical appliances are in use, the power is divided into total power according to the rated power of the electrical appliances. The energy is superimposed in terms of the power of the last second, multiplied by the corresponding time, i.e. 1 s.
In operation, the timer is controlled by TIM8 for 1s, energy is first calculated, and power is calculated after energy calculation is completed. After the electric meter is electrified again, the accumulated use electric quantity of each electric appliance is reset and cannot be stored.
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that changes may be made in the embodiments and/or equivalents thereof without departing from the spirit and scope of the invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (9)
1. The non-invasive real-time intelligent electric meter system based on characteristic quantity matching comprises a sampling circuit, an electric energy metering module, a communication module, a learning machine and an HMI (human machine interface) serial port screen and is characterized in that the sampling circuit detects voltage and current signals of each electric appliance and sends detection signals to the electric energy metering module, the electric energy metering module analyzes the voltage and current signals to generate steady-state characteristic quantity and transient-state characteristic quantity in each electric appliance and sends the characteristic quantities to the learning machine, the learning machine records characteristic quantity data as a database and calculates the newly detected characteristic quantity so as to judge the load state of each electric appliance, and the learning machine carries out data interaction with the HMI serial port screen through the communication module.
2. The non-invasive real-time smart meter system based on feature quantity matching according to claim 1, wherein the learning machine sends load state data of each electrical appliance to an HMI serial port screen, the HMI serial port screen displays real-time parameters of each electrical appliance, the user sends a control instruction to the HMI serial port screen, the HMI serial port screen sends the control instruction to the learning machine through the communication module, and the learning machine executes control over the electric meter according to the control instruction.
3. The non-intrusive real-time smart meter system based on feature quantity matching as claimed in claim 1, wherein the sampling circuit comprises a voltage sampling and a current sampling, the voltage sampling is connected across the zero line and the live line through resistance voltage division, the current sampling utilizes a current transformer, and the transformation ratio is 1000: 1.
4. the non-invasive real-time smart meter system based on feature quantity matching according to claim 1, further comprising a power supply circuit, wherein the power supply circuit adopts a 220V-5V switching power supply as a main power supply to ensure that the system supplies power to the system without mains supply, and the learning machine and the electric energy metering module both adopt a microprocessor to store, calculate and identify parameters of each electric appliance.
5. A non-intrusive load identification method based on characteristic quantity matching is characterized by comprising the following steps:
1. detecting values of steady-state characteristic quantities and transient-state characteristic quantities in each electrical appliance;
2. respectively calculating the score value of the steady-state characteristic quantity and the score value of the transient-state characteristic quantity in each electric appliance by adopting a load identification method, and obtaining the total score value of each electric appliance;
3. and determining the load state of each electric appliance according to the total score value of each electric appliance.
6. The method for non-intrusive load identification based on feature quantity matching as defined in claim 5, wherein the calculation formula of the total score value of each electrical appliance is as follows:
Pgeneral assembly＝ω5Pt3+(1-ω5)Ps
Pt3Score value, P, representing transient characteristic quantity of each electric appliancesScore value, ω, indicating steady-state feature quantity of each electric appliance5Weight, 1-omega, representing a transient characteristic quantity5And representing the weight of the steady-state characteristic quantity.
7. The method of claim 6, wherein the score value of the transient characteristic quantity is calculated by the formula:
Pt3＝[ω2Pt11+(1-ω2)Pt21]·ω3+[ω2Pt12+(1-ω2)Pt22]·(1-ω3)
Pt3＝ω4Pt13+(1-ω4)Pt23
wherein, Pt13Representing the value of the transient vector waveform score when each electrical appliance is turned off; pt23Representing the first order difference vector score value when each electrical appliance is turned off; omega4Weight, 1-omega, representing the score of the transient vector waveform at turn-off4A weight representing a first order difference vector score at turn-off; pt11Representing the transient vector waveform score value of the reactive vector when each electric appliance is started; pt21First order difference vector score value, P, representing reactive vector at turn-on of each appliancet12The method comprises the steps of representing a transient vector waveform score value of an active vector when each electric appliance is started; pt22First order difference vector score value, omega, representing the active vector at the time of switching on of each appliance3Weight, 1-omega, representing active vector at start-up3Representing the weight of the reactive vector when the switch is switched on; omega2Weight, 1-omega, representing the score of the transient vector waveform at turn-on2A weight representing a first order difference vector score at startup;
the calculation formula of the score value of the steady-state characteristic quantity is as follows:
Ps＝d2(x，Gi)＝x1∑-1x-2x1∑-1ui+ui’∑-1ui
wherein x is the collected steady-state characteristic value vector, Gi is the barycentric coordinate of the characteristic value of the ith type of electrical appliance in the database, Σ is the covariance matrix of six kinds of steady-state quantities corresponding to the type of electrical appliance, and μ i is the mean value of the steady-state characteristic quantity vectors of the ith type of electrical appliance in the database.
8. The method of claim 7, wherein the transient vector waveform score value of each electrical appliance is calculated by the following formula:
Pt1＝ω1D(x’，y’)+(l-ω1)·Na，b
where D (x, y) describes the closeness of the waveform trend between the transient vector and the vector in the database,
n represents vector dimension, xi and yi respectively represent ith value in database and sampled vector, and x 'i and y' i are corresponding values of normalized vector elements and meet the requirement
N (a, b) represents the closeness of the vector magnitude;
a. b is the maximum of the elements in vector x and vector y, ω1Weight, 1-omega, representing the trend of the waveform at turn-on1Representing the weight of the waveform amplitude;
the calculation formula of the transient vector first-order difference vector waveform score value of each electric appliance is as follows:
Pt2＝ω1D(Δx’，Δy’)+(l-ω1)·Nm，n
where D (Δ x, Δ y) represents the closeness of the trend of the waveform between the transient vector and the vector in the database when the transient vector waveform vector length is less than 1, and N (m, N) represents the closeness of the vector magnitude when the transient vector waveform vector length is less than 1.
9. The non-intrusive load identification method based on feature quantity matching as claimed in claim 8, wherein the step of determining the load category of each electrical appliance is as follows: calculating steady state scores ten times before and after the transient score is obtained, comprehensively obtaining total scores ten times, respectively setting an upper limit threshold value DH1 and DH2 for starting and stopping, when the database electrical appliance has the score lower than the upper limit threshold value, taking the electrical appliance with the lowest score, counting the position of the corresponding electrical appliance by one, and when the ten times are finished, considering the electrical appliance with the value more than 5 times as the electrical appliance which is currently put in or cut off, and obtaining a judgment result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010562333.9A CN112083224A (en) | 2020-06-18 | 2020-06-18 | Non-invasive real-time intelligent electric meter system based on characteristic quantity matching and identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010562333.9A CN112083224A (en) | 2020-06-18 | 2020-06-18 | Non-invasive real-time intelligent electric meter system based on characteristic quantity matching and identification method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112083224A true CN112083224A (en) | 2020-12-15 |
Family
ID=73734677
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010562333.9A Pending CN112083224A (en) | 2020-06-18 | 2020-06-18 | Non-invasive real-time intelligent electric meter system based on characteristic quantity matching and identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112083224A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113223287A (en) * | 2021-04-02 | 2021-08-06 | 国电南瑞南京控制系统有限公司 | SPI-based communication method and system for intelligent electric energy meter and load identification module |
| CN113363974A (en) * | 2021-06-16 | 2021-09-07 | 广东电网有限责任公司 | Resident load composition analysis method and device based on accumulated electric quantity low-frequency sampling |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103217603A (en) * | 2013-03-22 | 2013-07-24 | 重庆大学 | Recognition method of on-line monitoring of power consumption of non-intrusive household appliances |
| CN203909162U (en) * | 2014-04-28 | 2014-10-29 | 深圳市双合电气股份有限公司 | Electric energy quality collection apparatus |
| CN105186693A (en) * | 2015-09-28 | 2015-12-23 | 南方电网科学研究院有限责任公司 | Non-intrusive mode electrical load identification system and method |
| CN106124850A (en) * | 2016-07-20 | 2016-11-16 | 国网江西省电力公司南昌供电分公司 | A kind of load identification system for domestic intelligent electricity meter based on similarity algorithm and load recognition methods thereof |
| CN208705399U (en) * | 2018-07-06 | 2019-04-05 | 淮阴师范学院 | A kind of smart grid non-intrusion type resident load monitoring device |
| CN109813978A (en) * | 2018-12-25 | 2019-05-28 | 武汉中原电子信息有限公司 | A kind of non-intruding load-type recognition methods of variation characteristic between comprehensive transient characteristic and stable state |
| CN111160798A (en) * | 2019-12-31 | 2020-05-15 | 华南理工大学 | Non-invasive household appliance load identification method based on bee colony algorithm |
-
2020
- 2020-06-18 CN CN202010562333.9A patent/CN112083224A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103217603A (en) * | 2013-03-22 | 2013-07-24 | 重庆大学 | Recognition method of on-line monitoring of power consumption of non-intrusive household appliances |
| CN203909162U (en) * | 2014-04-28 | 2014-10-29 | 深圳市双合电气股份有限公司 | Electric energy quality collection apparatus |
| CN105186693A (en) * | 2015-09-28 | 2015-12-23 | 南方电网科学研究院有限责任公司 | Non-intrusive mode electrical load identification system and method |
| CN106124850A (en) * | 2016-07-20 | 2016-11-16 | 国网江西省电力公司南昌供电分公司 | A kind of load identification system for domestic intelligent electricity meter based on similarity algorithm and load recognition methods thereof |
| CN208705399U (en) * | 2018-07-06 | 2019-04-05 | 淮阴师范学院 | A kind of smart grid non-intrusion type resident load monitoring device |
| CN109813978A (en) * | 2018-12-25 | 2019-05-28 | 武汉中原电子信息有限公司 | A kind of non-intruding load-type recognition methods of variation characteristic between comprehensive transient characteristic and stable state |
| CN111160798A (en) * | 2019-12-31 | 2020-05-15 | 华南理工大学 | Non-invasive household appliance load identification method based on bee colony algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 曲朝阳等: "基于开启瞬时负荷特征的家电负荷识别", 《电工技术学报》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113223287A (en) * | 2021-04-02 | 2021-08-06 | 国电南瑞南京控制系统有限公司 | SPI-based communication method and system for intelligent electric energy meter and load identification module |
| CN113363974A (en) * | 2021-06-16 | 2021-09-07 | 广东电网有限责任公司 | Resident load composition analysis method and device based on accumulated electric quantity low-frequency sampling |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112180193B (en) | Non-invasive load identification system and method based on track image identification | |
| CN111830347B (en) | Two-stage non-invasive load monitoring method based on event | |
| CN105972761B (en) | A kind of non-intrusion type air conditioner load monitoring method | |
| CN112083224A (en) | Non-invasive real-time intelligent electric meter system based on characteristic quantity matching and identification method | |
| Ahmadi et al. | Load decomposition at smart meters level using eigenloads approach | |
| CN109813978A (en) | A kind of non-intruding load-type recognition methods of variation characteristic between comprehensive transient characteristic and stable state | |
| CN102279326A (en) | Electrical appliance recognition method and household user carbon emission metering system applying same | |
| CN113036759B (en) | Fine granularity identification method and identification system for power consumer load | |
| Du et al. | Support vector machine based methods for non-intrusive identification of miscellaneous electric loads | |
| CN111242808B (en) | Power consumer classification method, electronic equipment and storage medium | |
| CN112909923A (en) | Non-invasive household load behavior recognition device based on DTW algorithm | |
| CN109470957A (en) | A kind of non-intrusion type electrical appliance recognition based on Fr é chet distance | |
| CN109495355A (en) | Intelligent household energy router | |
| CN115456034A (en) | Automatic identification and monitoring method and system for electric bicycle charging | |
| WO2018006579A1 (en) | Power-variable motor and intelligent controller therefor | |
| CN114709926A (en) | Non-invasive load identification method fused with transient and steady state characteristics | |
| Wenninger et al. | Recurrence plot spacial pyramid pooling network for appliance identification in non-intrusive load monitoring | |
| Tao et al. | Multi-level non-intrusive load identification based on k-NN | |
| Cannas et al. | NILM techniques applied to a real-time monitoring system of the electricity consumption | |
| CN110244150B (en) | Non-invasive electrical appliance identification method based on root mean square and standard deviation | |
| CN116317012A (en) | Battery charging identification control method and system of self-adaptive charging source | |
| CN114239762A (en) | Non-invasive load identification method and system based on structured load characteristic spectrum | |
| Li et al. | The investigation of residential load identification technologies | |
| CN115146709A (en) | Non-invasive load identification method facing surface end | |
| CN112862163A (en) | Short-term power load prediction system based on DTW algorithm |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201215 |
|
| RJ01 | Rejection of invention patent application after publication |