CN111722028A - Load identification method based on high-frequency data - Google Patents
Load identification method based on high-frequency data Download PDFInfo
- Publication number
- CN111722028A CN111722028A CN201910211551.5A CN201910211551A CN111722028A CN 111722028 A CN111722028 A CN 111722028A CN 201910211551 A CN201910211551 A CN 201910211551A CN 111722028 A CN111722028 A CN 111722028A
- Authority
- CN
- China
- Prior art keywords
- current
- load
- steady
- state
- signal
- 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
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
Abstract
The invention belongs to the field of power demand side management, and particularly relates to a load identification method based on high-frequency data. The method is characterized in that firstly, each electric device is operated independently to obtain a prior current template library; then, detecting the switching event of the electric equipment; when a switching event is detected, extracting and storing the load steady-state current before the transient state occurs, and performing one-dimensional operation on the load steady-state current before the transient state occurs next time to obtain the steady-state current waveform of the newly switched load; establishing an objective function and constraining according to the operation habits of the residential electric equipment; and finally, optimizing the objective function by using a genetic algorithm to finish the load online identification work. The invention can effectively monitor the switching condition of the resident load under the condition of non-invasive power consumption data acquisition, can separate the switching signal on line and finish the load type identification work from the mixed data when a plurality of loads operate simultaneously through time delay feedback, solves the problems of separation and identification of the power consumption load, and is an effective realization method of the non-invasive load monitoring technology.
Description
Technical Field
The invention belongs to the field of demand side management, and particularly relates to a load identification method based on high-frequency data.
Background
With the rapid increase of the grid-connected capacity of new energy sources such as wind energy and solar energy, the intermittence, randomness and unpredictability of the new energy sources cause the stability of a power system to be challenged. The Demand Side Management (DSM) technology can guide power users to optimize the power utilization mode by taking effective measures, improve the terminal power utilization efficiency, optimize the resource allocation and stabilize the fluctuation of clean energy power generation. Load monitoring is an important content of DSM, and conventional intrusive load monitoring requires installation of information acquisition equipment at each electric device, and although the metering method is accurate, the metering method is high in cost and poor in operability. The Non-intrusive Load Monitoring (NILM) technology recognizes the power consumption of each appliance and participates in demand-side response by installing a set of current and voltage sensors at the electric service entrance of a house and analyzing a measured comprehensive signal based on signal analysis and processing theory.
Load decomposition and identification are the core content of the non-invasive load monitoring technology and are also the key points and difficulties of research. The high-frequency acquisition data has rich information content, relatively complete retention of the characteristics of the load equipment and high identification accuracy. However, due to the restriction of the power grid communication capacity and the limitation of hardware equipment, the high-frequency data acquisition needs to complete the load identification process on line. How to accurately separate the switching load current waveform and complete identification under the condition of high-frequency data acquisition is an important content of demand side management.
Disclosure of Invention
In order to solve the above problems, the present invention provides a load identification method based on high frequency data. The method is based on the principle that the load capacitance and inductance of the electric equipment are not changed, the steady-state periodic current is extracted under the same voltage background, and the load separation and identification can be completed by using the algorithm. The method comprises the following steps:
step 1: and independently operating each electric device and storing the steady-state current signal of one period in the circuit during independent operation of the electric devices to obtain a current template library of all electric appliances in the power utilization network.
Step 2: catch consumer switching signal, sample periodic current and periodic voltage waveform according to certain frequency promptly, real-time calculation electric current virtual value, when a periodic current virtual value contrasts last periodic current virtual value and takes place the sudden change, just think that load switching incident has taken place, specific process is as follows:
defining the current value of each period by setting the current signal i containing r continuous periods before transient occursEffective value of current iRMS:
N is the number of sampling points in a period, and i (N) is the current value of each sampling point. The current effective value variable quantity of two adjacent periods is as follows:
ΔiRMS=iRMS(j+1)-iRMS(j)
j refers to the jth cycle of the monitored current, j being 1, 2. If Δ iRMSAnd judging that the switching event of the electric load occurs at the moment, and judging a threshold value for the transient state.
And step 3: circuit steady state current is judged, namely the difference of current effective value is less than the threshold value of steady state current for a plurality of continuous cycles, and the concrete process is as follows:
will still be Δ iRMSThe difference between the effective values of the currents is used as a decision parameter. After a switching event is detected, if the circuit current signal is stable and unchanged in continuous periods, the circuit current signal is judged to be in a stable state, namely, the following conditions are met:
Δi<γ&T≥λ
and gamma is a steady-state current determination threshold value, T represents the number of continuous periods when delta i < gamma is satisfied, and lambda is the threshold value of T, and at the moment, the current signal is considered to stably run in continuous lambda periods, and the current in the lambda periods is determined to be the steady-state current.
And 4, step 4: extracting and storing a steady-state current waveform before transient generation under the same voltage background, and then obtaining a switching load current waveform by using current signal one-dimensional operation according to the current superposition characteristic to finish the online time-delay separation of the power load current signal, wherein the specific process comprises the following steps:
according to the principle that the load capacitance inductance is unchanged, the current at the voltage starting point position is taken to calculate the steady-state current before the transient state occurs, namely when the circuit voltage meets the following conditions:
U(X)>0&U(X-1)<0
the above equation detects a voltage zero crossing point of the terminal voltage U in λ cycles, X ═ X (1), X (2),.. multidot.. X (l) }TDenotes the x (l) th sampling point, l denotes the l thThe voltage crosses zero.
Extracting steady state current waveform before electric load k transient state occurrenceAnd storing, calculating the circuit steady-state current:
the above equation detects λ -1 voltage zero-crossings, l ═ 1, 2.
StoringAnd extracting the steady-state current waveform I of the circuit before the switching of the (k + 1) th electric load by the same method to finish the steady-state current waveform extraction work.
Then, from the current superposition characteristics:
subtracting the steady-state current waveform before the kth load switching from the steady-state current waveform I (t) before the kth load switchingObtaining the kth load steady-state current waveformAnd completing the online time-delay separation work of the electrical load.
And 5: establishing an objective function and a constraint condition, and carrying out comparison on the obtained steady-state current waveformAnd carrying out online identification. Namely, setting an objective function:
a1,a2,...aq, wherein aM is an electrical appliance turn-on coefficient of M electrical appliances, and when aq is 0, it indicates that the corresponding electrical appliance state is unchanged; when aq is 1, it indicates that the corresponding electric appliance state is changed, IqThe current signal of the q-th electric appliance is represented, and | L | · | |, represents the norm of L2.
And the operation habit of the resident electric equipment and the living habit of the user are restricted, namely the operation habit of the user determines that the possibility that more than two electric equipments are completely and synchronously started or closed is very low, so that the load can be started and closed as sequential execution. In addition, experimental results show that the similarity between the waveform obtained by separation and the current waveform of the corresponding electrical appliance in the template library should be more than 0.9. Namely, setting the constraint conditions:
0≤a1+a2+…+aM≤2
s ═ a1 · I1+ a2 · I2+. + aM · IM for the separation signalAn approximation signal of (a). r is the similarity coefficient between the separated signal and the signal S.Is a split signalAnd the covariance of the signal S is calculated,and muSIs the standard deviation of the two signals.
Step 6: and (5) optimizing and solving the objective function by using a genetic algorithm. Namely:
considering the operation habit of the user, the time difference exists between most load state changes, and the extracted currentWith a high probability of being usedThe current waveform of the electric equipment which operates independently, therefore, a unit array with the population size of M × M in a genetic algorithm is initialized according to M loads in an electricity utilization network, and then a group of a1, a2, when a, aqqWhen the load state is equal to 0, the corresponding load state is not changed; when a isqWhen 1, it means that the corresponding load is on. The aM determines a load state of the electric device in the circuit based on the obtained a1, a 2. Finally, the current I is updated toAnd monitoring the current in the circuit in real time, and repeating the steps 2 to 6.
Has the advantages that:
the method monitors the switching event of the power load in real time under the condition of non-invasive power consumption data acquisition, and when the circuit is detected to have the event, the current steady-state waveform of a new switching load is separated in an online time-delay manner according to the current superposition characteristic under the same voltage background, and the load type is identified through the current template library. The invention adopts a time delay feedback method to separate the current waveform, thereby ensuring the stability of the extracted current waveform and improving the accuracy of the separation of the steady-state current waveform. The invention can separate and identify switching loads on line, is an effective implementation method of a non-invasive load monitoring technology, and is also an important basis of a demand side management technology. The method is simple and feasible, convenient to implement, small in calculation amount and high in operation efficiency.
Drawings
FIG. 1 is a flow chart of a delayed feedback load identification method based on high frequency data V-I characteristics;
FIG. 2 is a current waveform template for some appliances in the UK-DALE data set, including current templates for when the dishwasher and treadmill are operating alone;
FIGS. 3(a) - (b) illustrate the process of isolating and identifying the steady state current waveform of a dishwasher.
FIGS. 4(a) - (b) illustrate the separation and identification process of the steady state current waveform of the treadmill.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings. The invention provides a non-invasive load identification method, which selects a dishwasher and a treadmill with UK-DALE data set as a typical load to simulate, and further explains the invention by combining the attached drawings:
fig. 1 is a flowchart of the load recognition algorithm of the present invention, which specifically includes:
step 1: and (3) operating each electric equipment independently and storing the steady-state current waveform of each electric equipment in one period to obtain a current template library of all electric appliances in the electric network.
Step 2: sampling the periodic current and the periodic voltage waveform, and calculating the current effective value i in real timeRMS:
The current effective value variable quantity of two adjacent periods is as follows: Δ iRMS=iRMS(j+1)-iRMS(j) If Δ iRMSAnd if yes, the switching event of the electric load occurs at the moment, and the threshold value is judged for the transient state.
And step 3: and (3) circuit steady-state current judgment, namely judging that the circuit is in a steady state if the difference delta i of the current effective values is less than gamma in lambda periods, wherein gamma is a steady-state current judgment threshold value.
And 4, step 4: according to the principle of constant load capacitance and inductance, the current at the voltage starting point position is taken to carry out steady-state current before the transient state of the kth electric equipment occursThe calculation of (2):
the above equation detects λ -1 voltage zero-crossings, l ═ 1, 2.
StoringAnd extracting the circuit steady-state current waveform I before the switching of the (k + 1) th electric load by the same method.
And 5:according to the superposition characteristic of the current in the circuit, subtracting the steady-state current waveform I (t) before the kth load switching from the steady-state current waveform I (t) before the kth load switchingObtaining the kth load steady-state current waveformNamely, the steady-state current waveform of the kth load can be obtained through time-delay separation:
step 6: setting a target function and a constraint condition, and obtaining a steady-state current waveformAnd comparing the current waveform with the current waveform in the template library to judge which electric appliance the separation signal belongs to. Setting an objective function:
a1, a 2.,. aq.,. aM, appliance on factor for M appliances, IqThe current signal of the q-th electric appliance is represented, and | L | · | |, represents the norm of L2.
According to the operation habit of a user and the similarity between the separated current waveform and the current waveform in the template library, which is more than 90%, setting the constraint conditions as follows:
0≤a1+a2+…+aM≤2
s ═ a1 · I1+ a2 · I2+. + aM · IM for the separation signalAn approximation signal of (a). r is the similarity coefficient between the separated signal and the signal S.Is a split signalAnd the covariance of the signal S is calculated,μSis the standard deviation of the two signals.
And 7: considering a group of optimal solutions a1, a2, a, aq, a, M solved by the algorithm only have 0 and 1, a genetic algorithm is adopted for optimization solution, a unit array with the population size of M × M is set, and then the load state of the electric equipment in the circuit is judged according to the obtained a1, a2, a, aq, a.
And decomposing and identifying signals when the dishwasher and the treadmill are switched by using the electricity utilization data in the UK-DALE data set.
According to the method of the invention:
1) as shown in fig. 2, first, the current waveforms of a part of the electric loads individually operated by the electric loads are stored as a current template, including the current waveforms of the dishwasher and the treadmill;
2) before the dishwasher is started, the effective value of the current in the circuit is monitored and calculated in real time:opening the dishwasher, the difference Δ i between the effective values of the currentsRMSWhere λ ═ 5, 1/20 for the last cycle current intensity;
3) then starting the treadmill, and calculating the effective value of the current by using the method before starting the treadmill, and marking the effective value as I;
4) according to the current superposition characteristicsSeparating to obtain the steady-state current waveform of the dish-washing machine
5) According to an objective function:
and the constraint condition is as follows:
0≤a1+a2+…+aM≤2
as shown in fig. 3(a) and 3(b), when the result of the separation is compared with the template current in fig. 2, the objective function value is 2.024, the correlation coefficient is 0.9938, and is greater than the correlation coefficient threshold value of 0.9, the waveform obtained by the separation is judged to beIs a dishwasher;
6) as shown in fig. 4(a) and 4(b), the current waveform separation and identification of the treadmill can be accomplished in the same manner. The treadmill was successfully identified with an objective function value of 4.163, a correlation coefficient of 0.9835, greater than a correlation coefficient threshold of 0.9.
Claims (1)
1. A load identification method based on high frequency data, the method comprising:
step 1: and independently operating each electric device and storing the steady-state current signal of one period in the circuit during independent operation of the electric devices to obtain a current template library of all electric appliances in the power utilization network.
Step 2: catch consumer switching signal, sample periodic current and periodic voltage waveform according to certain frequency promptly, real-time calculation electric current virtual value, when a periodic current virtual value contrasts last periodic current virtual value and takes place the sudden change, just think that load switching incident has taken place, specific process is as follows:
setting the current value of r continuous periods in the current signal i before transient occurrence, defining the effective value i of the current in each periodRMS:
N is the number of sampling points in a period, and i (N) is the current value of each sampling point. The current effective value variable quantity of two adjacent periods is as follows:
ΔiRMS=iRMS(j+1)-iRMS(j)
j refers to the jth cycle of the monitored current, j being 1, 2. If Δ iRMSAnd judging that the switching event of the electric load occurs at the moment, and judging a threshold value for the transient state.
And step 3: circuit steady state current is judged, namely the difference of current effective value is less than the threshold value of steady state current for a plurality of continuous cycles, and the concrete process is as follows:
will still be Δ iRMSThe difference between the effective values of the currents is used as a decision parameter. After a switching event is detected, if the circuit current signal is stable and unchanged in continuous periods, the circuit current signal is judged to be in a stable state, namely, the following conditions are met:
Δi<γ&T≥λ
and gamma is a steady-state current determination threshold value, T represents the number of continuous periods when delta i < gamma is satisfied, and lambda is the threshold value of T, and at the moment, the current signal is considered to stably run in continuous lambda periods, and the current in the lambda periods is determined to be the steady-state current.
And 4, step 4: extracting and storing a steady-state current waveform before transient generation under the same voltage background, and then obtaining a switching load current waveform by using current signal one-dimensional operation according to the current superposition characteristic to finish the online time-delay separation of the power load current signal, wherein the specific process comprises the following steps:
according to the principle that the load capacitance inductance is unchanged, the current at the voltage starting point position is taken to calculate the steady-state current before the transient state occurs, namely when the circuit voltage meets the following conditions:
U(X)>0&U(X-1)<0
the above equation detects a voltage zero crossing point of the terminal voltage U in λ cycles, X ═ X (1), X (2),.. multidot.. X (l) }TDenotes the x (l) th sampling point, l represents the l th voltage zero crossing point.
Extracting steady state current waveform before electric load k transient state occurrenceAnd storing, calculating the circuit steady-state current:
the above equation detects λ -1 voltage zero-crossings, l ═ 1, 2.
StoringAnd extracting the steady-state current waveform I of the circuit before the switching of the (k + 1) th electric load by the same method to finish the steady-state current waveform extraction work.
Then, from the current superposition characteristics:
subtracting the steady-state current waveform before the kth load switching from the steady-state current waveform I (t) before the kth load switchingObtaining the kth load steady-state current waveformAnd completing the online time-delay separation work of the electrical load.
And 5: establishing an objective function and a constraint condition, and carrying out comparison on the obtained steady-state current waveformCarry out on-line recognitionOtherwise. Namely, setting an objective function:
a1, a2, aq, aM is the electrical appliance turn-on coefficient of the M electrical appliances, and when aq is 0, it indicates that the corresponding electrical appliance state is unchanged; when aq is 1, it indicates that the corresponding electric appliance state is changed, IqThe current signal of the q-th electric appliance is represented, and | L | · | |, represents the norm of L2.
And the operation habit of the resident electric equipment and the living habit of the user are restricted, namely the operation habit of the user determines that the possibility that more than two electric equipments are completely and synchronously started or closed is very low, so that the load can be started and closed as sequential execution. In addition, experimental results show that the similarity between the waveform obtained by separation and the current waveform of the corresponding electrical appliance in the template library should be more than 0.9. Namely, setting the constraint conditions:
0≤a1+a2+…+aM≤2
s ═ a1 · I1+ a2 · I2+. + aM · IM for the separation signalAn approximation signal of (a). r is the similarity coefficient between the separated signal and the signal S.Is a split signalAnd the covariance of the signal S is calculated,and muSIs the standard deviation of the two signals.
Step 6: and (5) optimizing and solving the objective function by using a genetic algorithm. Namely:
considering the operation habit of the user, the time difference exists between most load state changes, and the extracted currentThe current waveform of a certain electric device which operates independently is high in probability, so that a unit array with the population size of M × M in a genetic algorithm is initialized according to M loads in an electric network, and then a group of a1, a2, aqqWhen the load state is equal to 0, the corresponding load state is not changed; when a isqWhen 1, it means that the corresponding load is on. The aM determines a load state of the electric device in the circuit based on the obtained a1, a 2. Finally, the current I is updated toAnd monitoring the current in the circuit in real time, and repeating the steps 2 to 6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910211551.5A CN111722028A (en) | 2019-03-19 | 2019-03-19 | Load identification method based on high-frequency data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910211551.5A CN111722028A (en) | 2019-03-19 | 2019-03-19 | Load identification method based on high-frequency data |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111722028A true CN111722028A (en) | 2020-09-29 |
Family
ID=72562860
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910211551.5A Pending CN111722028A (en) | 2019-03-19 | 2019-03-19 | Load identification method based on high-frequency data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111722028A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112701683A (en) * | 2020-12-23 | 2021-04-23 | 国网黑龙江省电力有限公司双鸭山供电公司 | Non-invasive load detection and decomposition method, system and storage medium |
CN113036754A (en) * | 2021-03-05 | 2021-06-25 | 重庆邮电大学 | Non-invasive load identification method based on improved DAG-SVMS |
CN113723590A (en) * | 2021-07-27 | 2021-11-30 | 清华大学 | Load identification method and device for edge calculation, electronic equipment and storage medium |
CN113899944A (en) * | 2021-09-30 | 2022-01-07 | 广东电网有限责任公司 | Detection method and device for power load switching point |
CN114184870A (en) * | 2021-12-14 | 2022-03-15 | 河北科技大学 | Non-invasive load identification method and equipment |
CN115407157A (en) * | 2022-10-31 | 2022-11-29 | 南昌工程学院 | Complex multi-state electrical appliance load event distinguishing method and device |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150309092A1 (en) * | 2012-11-16 | 2015-10-29 | Tianjin University | Current Pattern Matching Method for Non-Intrusive Power Load Monitoring and Disaggregation |
CN105404784A (en) * | 2015-12-07 | 2016-03-16 | 河南许继仪表有限公司 | Non-invasive power load decomposition method |
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 |
CN107025365A (en) * | 2017-05-16 | 2017-08-08 | 电子科技大学 | A kind of non-intruding load discrimination method for user side |
US20170351288A1 (en) * | 2016-06-07 | 2017-12-07 | Shenzhen Zhidian Energy Technology LTD | Non-invasive online real-time electric load identification method and identification system |
CN108197817A (en) * | 2017-09-19 | 2018-06-22 | 济南大学 | A kind of method of the non-intrusion type load transient state monitoring based on big data |
CN108429254A (en) * | 2018-03-07 | 2018-08-21 | 国网四川省电力公司电力科学研究院 | A kind of non-intrusion type power load recognition methods |
-
2019
- 2019-03-19 CN CN201910211551.5A patent/CN111722028A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150309092A1 (en) * | 2012-11-16 | 2015-10-29 | Tianjin University | Current Pattern Matching Method for Non-Intrusive Power Load Monitoring and Disaggregation |
CN105404784A (en) * | 2015-12-07 | 2016-03-16 | 河南许继仪表有限公司 | Non-invasive power load decomposition method |
US20170351288A1 (en) * | 2016-06-07 | 2017-12-07 | Shenzhen Zhidian Energy Technology LTD | Non-invasive online real-time electric load identification method and identification system |
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 |
CN107025365A (en) * | 2017-05-16 | 2017-08-08 | 电子科技大学 | A kind of non-intruding load discrimination method for user side |
CN108197817A (en) * | 2017-09-19 | 2018-06-22 | 济南大学 | A kind of method of the non-intrusion type load transient state monitoring based on big data |
CN108429254A (en) * | 2018-03-07 | 2018-08-21 | 国网四川省电力公司电力科学研究院 | A kind of non-intrusion type power load recognition methods |
Non-Patent Citations (1)
Title |
---|
曹敏 等: "基于暂态过程的非侵入式负荷监测研究", 《水电能源科学》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112701683A (en) * | 2020-12-23 | 2021-04-23 | 国网黑龙江省电力有限公司双鸭山供电公司 | Non-invasive load detection and decomposition method, system and storage medium |
CN113036754A (en) * | 2021-03-05 | 2021-06-25 | 重庆邮电大学 | Non-invasive load identification method based on improved DAG-SVMS |
CN113723590A (en) * | 2021-07-27 | 2021-11-30 | 清华大学 | Load identification method and device for edge calculation, electronic equipment and storage medium |
CN113899944A (en) * | 2021-09-30 | 2022-01-07 | 广东电网有限责任公司 | Detection method and device for power load switching point |
CN113899944B (en) * | 2021-09-30 | 2023-11-10 | 广东电网有限责任公司 | Detection method and device for power load switching point |
CN114184870A (en) * | 2021-12-14 | 2022-03-15 | 河北科技大学 | Non-invasive load identification method and equipment |
CN115407157A (en) * | 2022-10-31 | 2022-11-29 | 南昌工程学院 | Complex multi-state electrical appliance load event distinguishing method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111722028A (en) | Load identification method based on high-frequency data | |
CN110956220B (en) | Non-invasive household appliance load identification method | |
CN105972761B (en) | A kind of non-intrusion type air conditioner load monitoring method | |
CN109799474B (en) | Multi-epitope meter box electric energy meter phase identification method and system based on data correlation | |
CN109345409B (en) | Comprehensive energy efficiency management method for residential users based on broadband carrier | |
CN106093565B (en) | A kind of electricity subentry measurement method and device based on steady state characteristic Waveform Matching | |
CN105823948A (en) | Non-invasive resident load identification method | |
Ghosh et al. | Artificial bee colony optimization based non-intrusive appliances load monitoring technique in a smart home | |
CN111242391B (en) | Machine learning model training method and system for power load identification | |
Yu et al. | Nonintrusive appliance load monitoring for smart homes: Recent advances and future issues | |
CN111209976B (en) | Non-invasive load monitoring method and system | |
CN106872824A (en) | A kind of network load appliance type identification and the method and its device of different electrical equipment electricity statistics | |
CN109470957B (en) | Non-invasive electrical appliance identification method based on Frey's distance | |
CN112101110A (en) | Non-invasive load identification method for user side of power system | |
CN111242161A (en) | Non-invasive non-resident user load identification method based on intelligent learning | |
CN109377052B (en) | Power load matching method, device, equipment and readable storage medium | |
Wang et al. | An online load identification algorithm for non-intrusive load monitoring in homes | |
CN107767037B (en) | User electricity utilization composition analysis method | |
CN113193654A (en) | Event-driven non-intrusive power load monitoring method based on transient and steady state combination characteristics | |
Bucci et al. | A non-intrusive load identification system based on frequency response analysis | |
CN113010985A (en) | Non-invasive load identification method based on parallel AANN | |
CN112949563B (en) | Non-invasive load identification method based on variable point detection and improved KNN algorithm | |
CN106228244A (en) | A kind of energy based on self adaptation association rule mining depolymerizes method | |
CN115828091A (en) | Non-invasive load identification method and system based on end-cloud cooperation | |
CN110146758B (en) | Non-invasive electrical appliance identification method based on cross entropy |
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 | ||
AD01 | Patent right deemed abandoned | ||
AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20221230 |