CN111242391A - Machine learning model training method and system for power load identification - Google Patents
Machine learning model training method and system for power load identification Download PDFInfo
- Publication number
- CN111242391A CN111242391A CN202010152521.4A CN202010152521A CN111242391A CN 111242391 A CN111242391 A CN 111242391A CN 202010152521 A CN202010152521 A CN 202010152521A CN 111242391 A CN111242391 A CN 111242391A
- Authority
- CN
- China
- Prior art keywords
- data
- training
- power
- model
- power load
- 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.)
- Granted
Links
- 238000012549 training Methods 0.000 title claims abstract description 114
- 238000000034 method Methods 0.000 title claims abstract description 70
- 238000010801 machine learning Methods 0.000 title claims abstract description 51
- 238000005070 sampling Methods 0.000 claims abstract description 61
- 238000012360 testing method Methods 0.000 claims abstract description 32
- 238000005457 optimization Methods 0.000 claims abstract description 20
- 230000000694 effects Effects 0.000 claims abstract description 8
- 238000004140 cleaning Methods 0.000 claims description 20
- 238000012795 verification Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 11
- 238000010606 normalization Methods 0.000 claims description 5
- 230000003190 augmentative effect Effects 0.000 claims description 3
- 239000012634 fragment Substances 0.000 claims description 3
- 238000011056 performance test Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 2
- 238000003062 neural network model Methods 0.000 abstract description 5
- 239000000284 extract Substances 0.000 abstract description 3
- 238000012544 monitoring process Methods 0.000 description 11
- 230000008569 process Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 5
- 238000010200 validation analysis Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 230000001052 transient effect Effects 0.000 description 3
- 238000005406 washing Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 2
- 238000004220 aggregation Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 244000068645 Carya illinoensis Species 0.000 description 1
- 235000009025 Carya illinoensis Nutrition 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 239000003990 capacitor Substances 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 239000004020 conductor Substances 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000011900 installation process Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 230000006798 recombination Effects 0.000 description 1
- 238000005215 recombination Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000010183 spectrum analysis Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Images
Classifications
-
- 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- 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
-
- 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
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The machine learning model training method and system for power load recognition provided by this embodiment specifically use the measured electrical parameter data as the basis, unify the format of the basic electrical parameter data, train and input the data into the neural network model for continuous optimization, continuously adjust the parameters of the model through verifying the data set to select the optimal model, and simultaneously evaluate the performance of the model by using the test data set to achieve the optimal effect, so that the model can be further used in the recognition system of the power load, the method in this application can train the model according to the input sampling data to realize the recognition of the electric equipment in use according to the waveform sampling data of specific voltage, current, active power, etc., therefore, the application does not need to manually adjust the parameters and extract the characteristics, the application can learn autonomously and automatically obtain the characteristic parameters required by recognizing the power load, therefore, the application range of the model is widened, and the accuracy of power load identification is improved.
Description
Technical Field
The application relates to the technical field of power load monitoring, in particular to a machine learning model training method and system for power load identification.
Background
The power load is characterized in that the active power and the reactive power which are drawn by the power load from a power supply of a power system change along with the voltage of a load endpoint and the frequency change of the system; the power load characteristics are important components of the power system; the electric equipment is identified through the power load characteristics, and the method plays an important role in the development of the smart grid technology.
The most common methods of electrical load identification are intrusive and non-intrusive identification methods. The intrusive identification method needs to establish a monitoring system to install the sensor to each load, and although the method can directly obtain the measurement data of the load, the installation cost is high, the installation process is complex, and the maintenance is relatively difficult; and the non-invasive identification method only needs to install monitoring equipment at the general inlet of the power supply to decompose, monitor and identify each load in the whole system. Specifically, the non-invasive identification method is based on the extraction and identification of the characteristics of the electric appliance load imprints; the electric appliance load imprint characteristics can reflect unique information of electric equipment in operation, such as voltage, active waveform, starting current and the like; these load signature features are repeated during operation of the device, and based on this, the consumer can be identified.
The design and extraction of the load imprint features are the main difficulties of the whole method; the characteristic design generally adopts the stable/transient characteristics of simpler current, voltage, active power and reactive power and the combination thereof. The more complicated method considers the time domain characteristics of the related signals and the frequency domain characteristics, extracts valuable signal characteristics through mathematical tools such as Fourier transform, wavelet transform, harmonic analysis and the like, establishes a matching database, matches load data acquired in real time according to the valuable signal characteristics, and detects the power consumption condition. However, the manually designed signal characteristics require manual parameter adjustment, and have the problems of low complexity and dimensionality, and the inherent characteristics of the signals cannot be fully reflected and the resolution between different signals cannot be improved; meanwhile, the matching accuracy of the traditional matching algorithm such as time domain waveform matching, feature point matching, spectrum analysis and other methods is not high, so that the power load identification accuracy is not high, and the actual application effect is not ideal.
Disclosure of Invention
The application provides a machine learning model training method and system for power load identification, and aims to solve the technical problem that in the existing method, the accuracy rate of power load identification is low due to the fact that parameters need to be manually adjusted manually by artificially designed signal characteristics.
In order to solve the technical problem, the embodiment of the application discloses the following technical scheme:
in a first aspect, the present application provides a machine learning model training method for power load identification, the method including:
acquiring historical electrical parameter data sets of all electrical appliances;
cleaning the historical electrical parameter data sets of each electrical appliance;
dividing a washed historical electrical parameter data set of a single electrical appliance into a training set, a verification set and a test set according to a proportion;
carrying out balancing treatment on the training set to obtain a new training set;
training parameters of a machine learning model of the target electrical appliance by using the balanced new training set, and obtaining an optimization model according to the training set and the test set;
the method comprises the steps of collecting current data of a user power load, inputting the current data into an optimization model of each target electrical appliance, separating the working state of the electrical appliance, and outputting a classification result of the power load.
Optionally, the balancing the training set to obtain a new training set includes:
dividing a training set, a verification set and a test set from the cleaned user data fragment set according to a certain proportion;
dividing the data segments into two types according to the standard of whether the data segments contain the opening state of the electric appliance;
presetting a target proportion p% of data segments containing an opening state, performing oversampling processing on all the segments, and randomly and repeatedly copying the segments to increase the proportion to p%;
and taking the augmented and recombined data set as a new training set.
Optionally, the acquiring historical electrical parameter data sets of each electrical appliance includes:
unifying the data formats of the original public data sets to form a public data set, and integrating and summarizing the current public data set to obtain a first data set;
installing an electric energy meter at a main incoming line end of a user to obtain electrical parameters of a main electric load and a single electric load in one or more spaces to obtain a second data set;
and summarizing according to the first data set and the second data set to obtain historical electrical parameter data sets of the electrical appliances.
Optionally, unifying the data formats of the original common data sets to form a common data set, where the unifying includes:
extracting information in each original public data set, wherein the information comprises a timestamp, voltage, current, active power and reactive power;
with the timestamps as markers, generating a line of data with each new timestamp, the data comprising voltage, current, active power and reactive power;
and describing the circuit and the power load represented by each original common data set by using a metadata description file.
Optionally, the cleaning the historical electrical parameter data set of each electrical appliance includes: unifying and downsampling the data format to the specified frequency and normalizing the voltage.
Optionally, the unifying of the data formats includes:
and S [ i ] represents a sampling value, namely instantaneous active power, C is a power load type, and sa is sample data.
Optionally, the down-sampling to a specific frequency includes:
if the sampling rate is lower than 1Hz, recording according to the original sampling rate;
if the sampling rate is higher than 1Hz, the sampling rate is reduced to 1 Hz;
wherein the down-sampling the sampling rate to 1Hz comprises:
discarding all other sampling values within 1 second by using the values of the sampling points at intervals of every 1 second;
calculating the average value of the original sampling points in adjacent 1 second as a 1 second boundary data value;
and calculating the median value of the original sampling points in the adjacent 1 second as the 1 second boundary data value.
Optionally, the voltage normalization includes:
Powernormalisedrepresenting normalized Power values, PowerobservedRepresenting the measured power value, VoltagenominalRepresenting the nominal Voltage value VoltageobservedRepresenting the measured voltage value.
Optionally, the training of parameters of the machine learning model of the target electrical appliance using the balanced new training set and the obtaining of the optimization model according to the training set and the test set include:
training parameters of a machine learning model using a new training set of the target appliance;
verifying and testing different models obtained in different training stages on a verification set until the effect is best as a corresponding model of the target electrical appliance;
and performing performance test on the corresponding model of the target electrical appliance by using the test set until the performance is optimal to obtain an optimized model of the target electrical appliance.
In a second aspect, based on the above-mentioned machine learning model training method for power load identification, the present application further provides a machine learning model training system for power load identification, the system includes:
the data acquisition module is used for acquiring historical electrical parameter data sets of all electrical appliances;
the data cleaning module is used for cleaning the historical electrical parameter data sets of the electrical appliances;
the data dividing module is used for dividing the washed historical electrical parameter data set of a single electrical appliance into a training set, a verification set and a test set according to a proportion;
the data balancing module is used for carrying out balancing processing on the training set to obtain a new training set;
the model optimization module is used for training parameters of a machine learning model of the target electrical appliance by using the balanced new training set and obtaining an optimization model according to the training set and the test set;
and the power load type output module is used for acquiring the current data of the power load of the user, inputting the current data into the optimization model of each target electrical appliance, separating the working state of the electrical appliance and outputting the type result of the power load.
Compared with the prior art, the beneficial effect of this application is:
it can be seen from the above technical solutions that, in the machine learning model training method and system for power load identification provided in this embodiment, based on the measured electrical parameter data including current, voltage, active power, reactive power, etc., the basic electrical parameter data is formatted, trained, and input into the machine model for continuous optimization, the parameters of the model are continuously adjusted in a verification set to select an optimal model, and the performance of the model is evaluated by using a test set to achieve an optimal effect, so that the model is further used in the identification system for power loads, the method in this application can identify the power consumption device in use according to the input sampling data, which includes waveforms of voltage, current, active power, etc., and the specific power consumption device has specific waveforms of voltage, current, active power, etc., so that the model can be trained to identify the power consumption device in use with the specific waveforms of voltage, current, active power, etc., the machine learning training method and the system for recognizing the power load do not need manual parameter adjustment, compared with the traditional method, the matching accuracy rate is high, the machine learning training method and the system for recognizing the power load can automatically learn and automatically obtain the characteristic parameters needed by recognizing the power load, so that the application range of the model is widened, and the accuracy rate of recognizing the power load is improved.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application.
Drawings
In order to more clearly explain the technical solution of the present application, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious to those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flowchart of a machine learning model training method for power load identification according to an embodiment of the present disclosure;
FIG. 2 is a schematic structural diagram of a machine learning model training system for power load identification according to an embodiment of the present disclosure;
fig. 3 is a schematic structural diagram of a regional monitoring total electric meter and a sub-electric meter according to an embodiment of the present application.
Detailed Description
In order to make those skilled in the art better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all 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 application.
The power load mark features can reflect unique information reflecting power utilization states, such as voltage, active power waveform, current and the like, of one power utilization device in the running process, and the load marks can repeatedly appear in the running process of the device, so that each power utilization device can be identified.
The application provides a machine learning model training method for power load recognition, and specifically refers to fig. 1, where fig. 1 is a schematic flow diagram of the machine learning model training method for load recognition provided in the embodiment of the application, and the method includes:
s110: historical electrical parameter data sets for each appliance are obtained.
We propose to use sampling and cleaning of basic electrical parameter data (including current, voltage, active power and reactive power) (as shown in table one), and combine tuning of model parameters (parameters include but are not limited to weight, kernel function, dimension, learning rate, regularization, etc.), to implement training of machine learning model, and store the trained model in a specific and unified file format, so that the model can be further used in the identification system of power load. In the training process, the system does not need to artificially design the characteristic parameters of the power load, so that the application range of the model is improved, the accuracy of load identification is improved, and the utilization efficiency of basic electrical parameter data is improved.
Table-based electrical parameter data
The acquiring of the historical electrical parameter data set of each electrical appliance comprises:
unifying the data formats of the original public data sets to form a public data set, and integrating and summarizing the current public data set to obtain a first data set;
installing an electric energy meter at a main incoming line end of a user to obtain electrical parameters of a main electric load and a single electric load in one or more spaces to obtain a second data set;
and summarizing according to the first data set and the second data set to obtain historical electrical parameter data sets of the electrical appliances.
The original public data sets adopted in the embodiment of the invention comprise a reference energy decomposition data set, a building-level electronic aggregation data set of data from a single family, a Smart data set of total family power data and high-frequency and low-frequency data from three families, and a Pecan Street data set of summary and measured power data from ten families. Since these common data sets originate from different universities and institutions, the reading and identification of each data set results in the use of disparate data formats, and thus there is a need to unify the data formats of the original common data sets. The specific data format unification steps are as follows:
extracting the following information in each data set, including specific data such as time stamps, voltages, currents, apparent power, reactive power and the like;
taking the time stamp as a mark, generating a row of data by each new time stamp, wherein the data comprises voltage, current, apparent power and reactive power;
the time stamp interval is 1 second, and the data set with the sampling interval lower than 1 second in the original data set is subjected to down-sampling to obtain a time stamp sequence with the interval of 1 second;
and describing the circuit and the electric appliance represented by each data set and the general scores and the connection relations thereof by using a metadata description file (metadata).
We can specifically perform real-time monitoring tasks for an area in a certain space, where there is not only integral power load monitoring but also monitoring of a single electric device, such as a house, and all of televisions, kettles, microwave ovens, refrigerators, air conditioners, washing machines, and the like belong to a single electric device, specifically refer to fig. 3, and fig. 3 is a schematic structural diagram of a total electric meter and a branch electric meter for area monitoring provided by the embodiment of the present application.
Due to the transient and steady state signals of the total load obtained by data measurement, there are several cases of measurement errors: firstly, the inconsistency of the measuring devices is that different measuring devices have different measured values for the same electric equipment; secondly, the sensor can cause data loss in the process of compressing and transmitting the original data. It is because data acquisition and transmission can cause data deviation or loss, so it is necessary to process the data to improve the noise resistance of the load identification method; and thirdly, researching the influence of the data sampling period on load identification and discussing the balance point of the data sampling overhead and the system modeling complexity.
The data set collected by the user is a basic electrical parameter data record of energy consumption, multiple sets of monitoring instrument and equipment are used for monitoring various electric appliances in a room or a space within a certain time range, and two data collection modes of low frequency and high frequency are adopted. The low frequency acquisition mode has an acquisition frequency of 1Hz, while the high frequency acquisition can reach 10kHz to 100 kHz. The low-frequency signals are mainly extracted aiming at load steady-state characteristics, and the high-frequency signals can obtain load transient characteristics and high-frequency harmonic characteristics. Generally, the high-frequency signal can include more load electricity utilization characteristics, which is beneficial to the training of the model and the improvement of the accuracy, but also puts higher requirements on the data acquisition, transmission, compression and processing capabilities, and increases the complexity of the system. In the whole research process, the complexity and the accuracy of the system need to be chosen according to actual conditions.
In addition, need use single-phase multi-functional electric energy meter to monitor, this table adopts RS485 remote link control panel to go up all circuit breakers, and data acquisition inquires once per minute and in time links data real-time collection server, and we can adopt single-phase multi-functional electric energy meter alternately to compare current, voltage, power, and the advantage of this equipment is: the DIN35MM guide rail is installed, and has the characteristic of convenient assembly and disassembly; the communication speed can reach 9600bps, and the transmission speed is high; six paths of switching value input and output are adopted, and the requirements on the input and the output of the measured data are met; the DSP chip can be reconstructed and developed according to actual requirements, and the requirements of experimental environments are met.
In the system, the first step needs to extract and sample the power load characteristics, which is also the key point of the system, through long-term monitoring of the total voltage, current, power and partial voltage, current and power in a space area, a data set of real power load characteristics can be obtained, through detailed information provided in the data set, the use of actual power (P) and current (I) can be compared, and since the voltage loss comes from the voltage difference between the head end and the tail end of the line (which is related to the cross-sectional area and the size of the wire), when the fundamental voltage reaches an extreme value, the voltage is reduced, so that a method of correcting the power factor by using a capacitor can be used, and the power factor cos (theta) is the ratio between the actual power (P) and the apparent power (S) in the circuit. The power supply formula:
S=I·V;
P=S·COS(θ),
where θ is the angle between the voltage (V) and the current (I). The power factor is 1 when the voltage and current are in phase and zero when the current leads or lags the voltage.
We found that the actual power reading has a high fluctuation (up to 10 times) compared to the current. This is because the meter uses two sensor readings that can be independently fluctuated to measure actual power, current on the same conductor as the load, and voltage on the circuit breaker power panel. The problem of significant voltage drop occurs when measuring the voltage at the top and bottom of the breaker power strip. This means that if the voltage is measured at a single point, the further from the transformer, the less accurate the reading. This can result in less accurate power readings when calculating the associated power. Furthermore, the resistivity (R) of the load also presents problems with voltage drop due to factors such as wire size, but the current is not affected by these problems and does not drop due to these factors, from which we conclude that the use of current readings allows better determination of the load condition from historical data, and with high accuracy.
S120: and cleaning the historical electrical parameter data sets of the various electrical appliances.
After the information of the power load data is obtained, the power load data cleaning is the core technology of the system and the method, and mainly comprises automatic screening and abnormal data cleaning, noise identification and separation, down sampling, discarding rate, data normalization, missing data compensation and processing, Top-k, error data elimination and the like.
The system needs to create a data CSV file, and finishes the work of cleaning a data set by deleting incomplete data, such as the condition that part of instrument and equipment data has incomplete data or is lost due to different timestamps. After the data set is created and the CSV file import is generated, the data will reside in our in-memory data structure, which can be used throughout the training process. In order to solve the problem of non-uniformity of data formats of different data sets, several preprocessing operations are required.
The DSP single-phase multifunctional electric energy meter is used for collecting general basic electric parameter data (including voltage, current, active power, reactive power and the like), and sampling data in the same parameter are arranged according to a time sequence. Taking active power as an example, piRepresenting the voltage degree detected by the electric energy meter at the ith sampling time point, and originally collected data according to timeSequentially arranged as active power sequence Pactive={p1,p2,…pi…, forming an output of the power load characteristic data sampling module and acting as an input to the power load data cleansing module.
In the data cleaning module, the electrical time sequence data are down-sampled to a specified frequency f according to actual application requirements, and if the frequency f corresponds to an interval period of 5 sampling points in the original sampling data, the active power data sequence after down-sampling is P'active={p1,p6,p11,…p5i+1,…}={q1,q2,…qi… }. And then carrying out other preprocessing operations such as normalization and the like on the obtained relative low-frequency time sequence data to generate continuous time sequence data which can be directly used as training by a model. Because the power load characteristic identification model receives a plurality of electrical parameters at the same time, but the input dimension of a specific electrical parameter is limited and is fixed length n, and all continuous time sequence data of the electrical parameter cannot be input at one time, the power load characteristic identification model adopts a sliding window (sliding window) mode with the length n to continuously slide with the step length l, and intercepts a data subsequence with the fixed length n in the electrical parameter sequence after each sliding, such as an active power subsequence Pi={qin+1,qin+2,…qin+nAnd the input is used as the input of the power load characteristic identification module.
Different electrical appliances have independent corresponding deep neural network models in the load characteristic identification module, the data subsequence can simultaneously input the deep neural network models corresponding to the different electrical appliances, n sampling values of each electrical parameter in the same time period correspond to n corresponding nodes of a deep neural network input layer, and the mth target electrical appliance working state sequence S in the data subsequence is output through network forward propagation operationm,i={sm,in+1,sm,in+2,…sm,in+n}. The state sequence length is n, but not limited to n, determined by the model architecture. All the appliance types in use in the sequence can be judged based on the working state of each appliance.
The method specifically comprises the following steps:
(1) unified format
Because the formats of the original data sets are not uniform, the method needs to extract the characteristics of each data set for evaluation, in order to avoid generating great interference on judgment due to too large difference of consumed power of different electrical appliances, the data needs to be cleaned, and meanwhile, the data is normalized, namely, the data is converted into a value between [0 and 1], Si represents a sampling value, namely instantaneous active power, C is a power load type, sa is sample data, and the formula is as follows:
(2) down sampling
The sampling rate of the device monitor is between 0.008Hz and 16kHz in the data set, so the system down samples the data set to a specified frequency using aggregation functions such as mean, mode, and median.
And if the sampling rate is lower than 1Hz, recording according to the original sampling rate.
If the sampling rate is higher than 1Hz, the sampling rate is down-sampled to 1 Hz. The specific down-sampling method comprises the following steps:
1. discarding all other sampling values within 1 second by using the values of the sampling points at intervals of every 1 second;
2. calculating the average value of the original sampling points in adjacent 1 second as a 1 second boundary data value;
3. calculating the median of the original sampling points in the adjacent 1 second as a 1 second boundary data value;
(3) voltage normalization
Due to the fluctuation of the collected voltage, for example, the voltage in the same data set is changed from 180-250V, while the voltage in the other data set is changed within the range of 118-123V. The present system must take into account the effects of these voltage fluctuations as they can significantly affect power consumption.
Powernormalisedrepresenting normalized Power values, PowerobservedRepresenting the measured power value, VoltagenominalRepresenting the nominal Voltage value VoltageobservedRepresenting the measured voltage value.
(4)Top-k
Generally speaking, our identification system is directed to the high-energy consumption devices ranked k top (where k is an adjustable parameter) rather than all devices, because there are three points that first of all, the top k power consumption devices can already provide most of the reference information for the overall power consumption situation; secondly, these devices have the most significant characteristics, and the remaining devices can be considered to produce only noise; third, modeling and identification for larger-duty power consuming devices can greatly improve the reliability of the data.
During data cleansing, the present system will also address other common problems with data sets, such as: device sensors do not report readings, small data missing, remove abnormal values such as observed voltages exceeding twice the rated voltage, loss of mains power data, etc.
(5) Detecting a gap
Many algorithms today assume that the communication of each data acquisition device is continuous, however, in practical situations, disconnection or failure of the data acquisition device may occur, and if a parameter value is set, when the time of disconnection or failure is longer than the set parameter value, it can be considered that a gap exists in a continuous power data sample. For example, we calculate the difference between the timestamps of adjacent sampling points, and if it is greater than a certain parameter, such as 10 seconds, we consider that there is a gap in the data set. For the data sequence with gaps, the data sequence cannot be directly used for a system training set and a test set, and all the training data and test data sequences must select the data sequence without gaps in the middle.
(6) Uptime
Uptime is the total time recorded by the sensor. It is the last timestamp, minus the first timestamp, minus the duration of all gaps present.
S130: and proportionally dividing the cleaned historical electrical parameter data set of the single electrical appliance into a training set, a verification set and a test set.
In this application, we divide the sample into three separate parts: training set (train set), validation set (validation set), and test set (test set). The training set is used to estimate the model, the validation set is used to determine the network structure or parameters that control the complexity of the model, and the test set examines how well the model is performing to the final selection of the optimal model. We divide the training set, validation set, and test set into 50% of the total samples in the training set, and 25% in the others, all three randomly drawn from the power load samples we collected.
The training set is used for fitting the model, and the model is built by using data acquired and cleaned by actual power loads.
The data set obtained after power load collection and cleaning is input into a model for continuous learning, and the optimal parameters of the model are searched through fitting.
The effect of the model is only to represent the training data, but not necessarily to fit other data of the same class, so we split the data into two parts before modeling, one part being the training data and one part being the validation data (the ratio of the two parts of data is approximately 7:3 or 6: 4). Meanwhile, a plurality of models can be trained, and the verification data can be input into different models for comparison to obtain the optimal model.
The biggest difference between the test set and the training and verification set is that the data of the training and verification set are used in the model training process, and the test set is only used as the reference basis for the performance test after the model is finally trained.
In the patent, the emphasis is to train the model on a support vector machine, a decision tree, a principal component analysis and an LSTM long and short term memory network model, and meanwhile, the emphasis is also on the variation of the loss value of the model and the convergence condition of the model.
S140: and carrying out balancing treatment on the training set to obtain new training.
Due to the time imbalance of appliance usage (switch states), the skewness of the model is affected if the original imbalance distributed data is used directly to train the model, and therefore the model is more prone to make the same judgment as most categories or states in the data set. The method adopts an innovative training data set balancing method, remarkably improves the performance of a machine learning model for power load identification, constructs a complete and commercially-realized model training system and method for power load identification together with innovative data acquisition, cleaning and other methods, realizes training of the machine learning model for power load identification, automatically extracts power load characteristics through a large amount of training data by using a machine learning algorithm, and lays a foundation for further realizing the power load identification system.
The balancing processing on the training set to obtain a new training set comprises the following steps:
dividing a training set, a verification set and a test set from the cleaned user data fragment set according to a certain proportion;
dividing the data segments into two types according to the standard of whether the data segments contain the opening state of the electric appliance;
presetting a target proportion p% of data segments containing an opening state, performing oversampling processing on all the segments, and randomly and repeatedly copying the segments to increase the proportion to p%;
and taking the augmented and recombined data set as a new training set.
The application provides a data set balancing method, which has the core idea that a training set is subjected to amplification and recombination, and load data segments containing an electric appliance starting state are subjected to oversampling (i.e., all data segments containing the electric appliance starting state in the training set are copied for multiple times and merged into an original training set, so that the data segment containing the starting state accounts for a higher percentage than p%, the data segments are used as a new training set, data segments are continuously extracted from the new training set and input into a model, and model parameters are iteratively adjusted and optimized through an optimization algorithm. The problem of distribution unbalance of the power load data set has non-negligible influence on almost all non-inertia machine learning models, so the balancing method is suitable for various machine learning models and obviously improves the identification accuracy of the machine learning models.
S150: and training parameters of a machine learning model of the target electrical appliance by using the balanced new training set, and then obtaining an optimization model according to the training set and the test set.
S160: the method comprises the steps of collecting current data of a user power load, inputting the current data into an optimization model of each target electrical appliance, separating the working state of the electrical appliance, and outputting a classification result of the power load.
Different electrical appliances have independent corresponding deep neural network models in the load characteristic identification module, the data subsequence can simultaneously input the deep neural network models corresponding to the different electrical appliances, n sampling values of each electrical parameter in the same time period correspond to n corresponding nodes of a deep neural network input layer, and the mth target electrical appliance working state sequence S in the data subsequence is output through network forward propagation operationm,i={sm,in+1,sm,in+2,…sm,in+n}. The state sequence length is n, but not limited to n, determined by the model architecture. All the appliance types in use in the sequence can be judged based on the working state of each appliance.
The basic electrical parameter data (including current, voltage, active power and reactive power) processed by the aforementioned power load collection and cleaning module are input into the power load identification module as input data, and the aim is to deduce the equipment category of each electric equipment through the electrical data.
The selected data and identification method are as follows:
selecting data basic electrical parameter data (including current, voltage, active power and reactive power) of a total circuit;
selecting the data for 1 to 5 minutes, wherein 60 to 300 sampling points exist when the sampling rate is 1 Hz;
processing the data into a plurality of sequences with 16 data points as one sequence, wherein the sequences with gaps are removed;
loading a machine learning model program for power load identification which is trained in advance, and can be realized by any existing machine learning or neural network programming framework, such as tensor flow, keras and the like);
the data sequence is input into a machine learning model, and the identification module outputs current equipment category data (expressed by integers, such as 1 ═ air conditioner, 2 ═ kettle, 3 ═ washing machine, 4 ═ electric lamp, and the like) corresponding to sampling points at each moment.
The basic electrical parameter data processed by the aforementioned power load collection and cleaning module, including current, voltage, active power and reactive power, are input into the power load identification module, and the purpose is to deduce the equipment category of each electric equipment through the electrical data.
The selected data and identification method are as follows:
selecting basic data of electrical parameters of the total circuit, such as current, voltage, active power and reactive power;
selecting the data for 1 to 5 minutes, wherein 60 to 300 sampling points exist when the sampling rate is 1 Hz;
processing the data into a plurality of sequences with 16 data points as one sequence, wherein the sequences with gaps are removed;
loading a machine learning model program which is trained in advance and used for power load identification can be realized by any existing machine learning or neural network programming framework, such as tensor flow, keras and the like;
the data sequence is input into a machine learning model, and the identification module outputs current equipment category data (expressed by integers, such as 1 ═ air conditioner, 2 ═ kettle, 3 ═ washing machine, 4 ═ electric lamp, and the like) corresponding to sampling points at each moment.
Based on the inventive concept of the above method, the present embodiment further provides a machine learning model training system for power load identification, as shown in fig. 2, fig. 2 is a schematic structural diagram of the machine learning model training system for power load identification provided in the embodiment of the present application; the system comprises:
the data acquisition module is used for acquiring historical electrical parameter data sets of all electrical appliances;
the data cleaning module is used for cleaning the historical electrical parameter data sets of the electrical appliances;
the data dividing module is used for dividing the washed historical electrical parameter data set of a single electrical appliance into a training set, a verification set and a test set according to a proportion;
the data balancing module is used for carrying out balancing processing on the training set to obtain a new training set;
the model optimization module is used for training parameters of a machine learning model of the target electrical appliance by using the balanced new training set and obtaining an optimization model according to the training set and the test set;
and the power load type output module is used for acquiring the current data of the power load of the user, inputting the current data into the optimization model of each target electrical appliance, separating the working state of the electrical appliance and outputting the type result of the power load.
It can be seen from the above technical solutions that, in the machine learning model training method and system for power load identification provided in this embodiment, based on the measured electrical parameter data including current, voltage, active power, reactive power, etc., the basic electrical parameter data is formatted, trained, and input into the machine model for continuous optimization, the parameters of the model are continuously adjusted in a verification set to select an optimal model, and the performance of the model is evaluated by using a test set to achieve an optimal effect, so that the model is further used in the identification system for power loads, the method in this application can identify the power consumption device in use according to the input sampling data, which includes waveforms of voltage, current, active power, etc., and the specific power consumption device has specific waveforms of voltage, current, active power, etc., so that the model can be trained to identify the power consumption device in use with the specific waveforms of voltage, current, active power, etc., the machine learning training method and the system for recognizing the power load do not need manual parameter adjustment, compared with the traditional method, the matching accuracy rate is high, the machine learning training method and the system for recognizing the power load can automatically learn and automatically obtain the characteristic parameters needed by recognizing the power load, so that the application range of the model is widened, and the accuracy rate of recognizing the power load is improved.
Since the above embodiments are all described by referring to and combining with other embodiments, the same portions are provided between different embodiments, and the same and similar portions between the various embodiments in this specification may be referred to each other. And will not be described in detail herein.
Other embodiments of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the application being indicated by the following claims.
The above-described embodiments of the present application do not limit the scope of the present application.
Claims (10)
1. A method of machine learning model training for power load recognition, the method comprising:
acquiring historical electrical parameter data sets of all electrical appliances;
cleaning the historical electrical parameter data sets of each electrical appliance;
dividing a washed historical electrical parameter data set of a single electrical appliance into a training set, a verification set and a test set according to a proportion;
carrying out balancing treatment on the training set to obtain a new training set;
training parameters of a machine learning model of the target electrical appliance by using the balanced new training set, and obtaining an optimization model according to the training set and the test set;
the method comprises the steps of collecting current data of a user power load, inputting the current data into an optimization model of each target electrical appliance, separating the working state of the electrical appliance, and outputting a classification result of the power load.
2. The method of claim 1, wherein the balancing the training set to obtain a new training set comprises:
dividing a training set, a verification set and a test set from the cleaned user data fragment set according to a certain proportion;
dividing the data segments into two types according to the standard of whether the data segments contain the opening state of the electric appliance;
presetting a target proportion p% of data segments containing an opening state, performing oversampling processing on all the segments, and randomly and repeatedly copying the segments to increase the proportion to p%;
and taking the augmented and recombined data set as a new training set.
3. The machine learning model training method for power load identification according to claim 1, wherein the obtaining historical electrical parameter data sets of individual appliances comprises:
unifying the data formats of the original public data sets to form a public data set, and integrating and summarizing the current public data set to obtain a first data set;
installing an electric energy meter at a main incoming line end of a user to obtain electrical parameters of a main electric load and a single electric load in one or more spaces to obtain a second data set;
and summarizing according to the first data set and the second data set to obtain historical electrical parameter data sets of the electrical appliances.
4. The method for training the machine learning model for power load identification according to claim 3, wherein unifying the data formats of the original common data sets to form the common data set comprises:
extracting information in each original public data set, wherein the information comprises a timestamp, voltage, current, active power and reactive power;
with the timestamps as markers, generating a line of data with each new timestamp, the data comprising voltage, current, active power and reactive power;
and describing the circuit and the power load represented by each original common data set by using a metadata description file.
5. The machine learning model training method for power load recognition of claim 1, wherein the cleaning the historical electrical parameter data sets of individual appliances comprises: unifying and downsampling the data format to the specified frequency and normalizing the voltage.
6. The method of machine learning model training for power load recognition of claim 5, wherein the unification of the data formats comprises:
and S [ i ] represents a sampling value, namely instantaneous active power, C is a power load type, and sa is sample data.
7. The machine learning model training method for power load identification of claim 5, wherein the down-sampling to a specified frequency comprises:
if the sampling rate is lower than 1Hz, recording according to the original sampling rate;
if the sampling rate is higher than 1Hz, the sampling rate is reduced to 1 Hz;
wherein the down-sampling the sampling rate to 1Hz comprises:
discarding all other sampling values within 1 second by using the values of the sampling points at intervals of every 1 second;
calculating the average value of the original sampling points in adjacent 1 second as a 1 second boundary data value;
and calculating the median value of the original sampling points in the adjacent 1 second as the 1 second boundary data value.
8. The machine learning model training method for power load identification according to claim 5, wherein the voltage normalization comprises:
Powernormalisedrepresenting normalized Power values, PowerobservedRepresenting the measured power value, VoltagenominalRepresenting the nominal Voltage value VoltageobservedRepresenting the measured voltage value.
9. The method of claim 1, wherein the training the parameters of the machine learning model of the target electrical appliance using the balanced new training set and obtaining the optimized model according to the training set and the testing set comprises:
training parameters of a machine learning model using a new training set of the target appliance;
verifying and testing different models obtained in different training stages on a verification set until the effect is best as a corresponding model of the target electrical appliance;
and performing performance test on the corresponding model of the target electrical appliance by using the test set until the performance is optimal to obtain an optimized model of the target electrical appliance.
10. A machine learning model training system for electrical load recognition, the system comprising:
the data acquisition module is used for acquiring historical electrical parameter data sets of all electrical appliances;
the data cleaning module is used for cleaning the historical electrical parameter data sets of the electrical appliances;
the data dividing module is used for dividing the washed historical electrical parameter data set of a single electrical appliance into a training set, a verification set and a test set according to a proportion;
the data balancing module is used for carrying out balancing processing on the training set to obtain a new training set;
the model optimization module is used for training parameters of a machine learning model of the target electrical appliance by using the balanced new training set and obtaining an optimization model according to the training set and the test set;
and the power load type output module is used for acquiring the current data of the power load of the user, inputting the current data into the optimization model of each target electrical appliance, separating the working state of the electrical appliance and outputting the type result of the power load.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010152521.4A CN111242391B (en) | 2020-03-06 | 2020-03-06 | Machine learning model training method and system for power load identification |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010152521.4A CN111242391B (en) | 2020-03-06 | 2020-03-06 | Machine learning model training method and system for power load identification |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111242391A true CN111242391A (en) | 2020-06-05 |
CN111242391B CN111242391B (en) | 2023-10-31 |
Family
ID=70873388
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010152521.4A Active CN111242391B (en) | 2020-03-06 | 2020-03-06 | Machine learning model training method and system for power load identification |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111242391B (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111830350A (en) * | 2020-07-23 | 2020-10-27 | 珠海格力电器股份有限公司 | Energy consumption metering method and device and electric appliance |
CN112649644A (en) * | 2020-12-22 | 2021-04-13 | 常州常工电子科技股份有限公司 | Student apartment electricity utilization safety load learning method |
CN112801115A (en) * | 2021-01-26 | 2021-05-14 | 广西电网有限责任公司电力科学研究院 | Power supply fluctuation control method based on microgrid source-load comprehensive characteristic image and application |
CN113095425A (en) * | 2021-04-21 | 2021-07-09 | 广东电网有限责任公司 | Method, device, equipment and medium for automatically judging and deciding on site start-up conditions |
CN113704780A (en) * | 2021-07-16 | 2021-11-26 | 国网浙江省电力有限公司杭州供电公司 | Model-driven-based power distribution network user side information adaptive encryption method |
CN114977176A (en) * | 2022-07-19 | 2022-08-30 | 深圳江行联加智能科技有限公司 | Power load decomposition method, device, equipment and storage medium |
CN115018011A (en) * | 2022-07-19 | 2022-09-06 | 深圳江行联加智能科技有限公司 | Power load type identification method, device, equipment and storage medium |
CN115085274A (en) * | 2022-07-27 | 2022-09-20 | 北京智芯微电子科技有限公司 | Automatic identification method and device for new energy equipment access, electronic equipment and medium |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130138669A1 (en) * | 2011-11-28 | 2013-05-30 | Bin Lu | System and method employing a hierarchical load feature database to identify electric load types of different electric loads |
US20140067299A1 (en) * | 2012-08-29 | 2014-03-06 | Bin Lu | System and method for electric load identification and classification employing support vector machine |
US20160055419A1 (en) * | 2014-08-22 | 2016-02-25 | Opower, Inc. | Identifying electric vehicle owners |
CN106845553A (en) * | 2017-02-04 | 2017-06-13 | 武汉阿帕科技有限公司 | Load characteristic identification mechanism, method and auto-power-off device |
CN107944495A (en) * | 2017-12-04 | 2018-04-20 | 贵州电网有限责任公司电力科学研究院 | A kind of household electricity load classification recognition methods based on deep layer forest algorithm |
CN108899892A (en) * | 2018-06-08 | 2018-11-27 | 上海电力学院 | One kind being based on CNN non-intrusion type power load decomposition method |
CN109145949A (en) * | 2018-07-19 | 2019-01-04 | 山东师范大学 | Non-intrusive electrical load monitoring and decomposition method and system based on integrated study |
CN109359665A (en) * | 2018-08-28 | 2019-02-19 | 中国农业大学 | A kind of family's electric load recognition methods and device based on support vector machines |
CN109387712A (en) * | 2018-10-09 | 2019-02-26 | 厦门理工学院 | Non-intrusion type cutting load testing and decomposition method based on state matrix decision tree |
CN110188826A (en) * | 2019-05-31 | 2019-08-30 | 湖南科技大学 | Household electrical appliance operating status non-invasive inspection methods based on intelligent electric meter data |
CN110376457A (en) * | 2019-06-28 | 2019-10-25 | 同济大学 | Non-intrusion type load monitoring method and device based on semi-supervised learning algorithm |
CN110514889A (en) * | 2019-07-19 | 2019-11-29 | 浙江万胜智能科技股份有限公司 | A kind of method and system of non-intrusion type household electricity remained capacity |
-
2020
- 2020-03-06 CN CN202010152521.4A patent/CN111242391B/en active Active
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130138669A1 (en) * | 2011-11-28 | 2013-05-30 | Bin Lu | System and method employing a hierarchical load feature database to identify electric load types of different electric loads |
US20140067299A1 (en) * | 2012-08-29 | 2014-03-06 | Bin Lu | System and method for electric load identification and classification employing support vector machine |
US20160055419A1 (en) * | 2014-08-22 | 2016-02-25 | Opower, Inc. | Identifying electric vehicle owners |
CN106845553A (en) * | 2017-02-04 | 2017-06-13 | 武汉阿帕科技有限公司 | Load characteristic identification mechanism, method and auto-power-off device |
CN107944495A (en) * | 2017-12-04 | 2018-04-20 | 贵州电网有限责任公司电力科学研究院 | A kind of household electricity load classification recognition methods based on deep layer forest algorithm |
CN108899892A (en) * | 2018-06-08 | 2018-11-27 | 上海电力学院 | One kind being based on CNN non-intrusion type power load decomposition method |
CN109145949A (en) * | 2018-07-19 | 2019-01-04 | 山东师范大学 | Non-intrusive electrical load monitoring and decomposition method and system based on integrated study |
CN109359665A (en) * | 2018-08-28 | 2019-02-19 | 中国农业大学 | A kind of family's electric load recognition methods and device based on support vector machines |
CN109387712A (en) * | 2018-10-09 | 2019-02-26 | 厦门理工学院 | Non-intrusion type cutting load testing and decomposition method based on state matrix decision tree |
CN110188826A (en) * | 2019-05-31 | 2019-08-30 | 湖南科技大学 | Household electrical appliance operating status non-invasive inspection methods based on intelligent electric meter data |
CN110376457A (en) * | 2019-06-28 | 2019-10-25 | 同济大学 | Non-intrusion type load monitoring method and device based on semi-supervised learning algorithm |
CN110514889A (en) * | 2019-07-19 | 2019-11-29 | 浙江万胜智能科技股份有限公司 | A kind of method and system of non-intrusion type household electricity remained capacity |
Non-Patent Citations (2)
Title |
---|
唐璐等: "基于卷积神经网络的非侵入式负荷识别方法", 《云南电力技术》 * |
赵银菊: "人工神经网络在电力系统负荷预测中的应用", 《宁夏电力》 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111830350A (en) * | 2020-07-23 | 2020-10-27 | 珠海格力电器股份有限公司 | Energy consumption metering method and device and electric appliance |
CN112649644A (en) * | 2020-12-22 | 2021-04-13 | 常州常工电子科技股份有限公司 | Student apartment electricity utilization safety load learning method |
CN112649644B (en) * | 2020-12-22 | 2021-06-29 | 常州常工电子科技股份有限公司 | Student apartment electricity utilization safety load learning method |
CN112801115A (en) * | 2021-01-26 | 2021-05-14 | 广西电网有限责任公司电力科学研究院 | Power supply fluctuation control method based on microgrid source-load comprehensive characteristic image and application |
CN112801115B (en) * | 2021-01-26 | 2022-09-06 | 广西电网有限责任公司电力科学研究院 | Power supply fluctuation control method based on microgrid source-load comprehensive characteristic image and application |
CN113095425A (en) * | 2021-04-21 | 2021-07-09 | 广东电网有限责任公司 | Method, device, equipment and medium for automatically judging and deciding on site start-up conditions |
CN113704780A (en) * | 2021-07-16 | 2021-11-26 | 国网浙江省电力有限公司杭州供电公司 | Model-driven-based power distribution network user side information adaptive encryption method |
CN113704780B (en) * | 2021-07-16 | 2024-02-06 | 国网浙江省电力有限公司杭州供电公司 | Power distribution network user side information self-adaptive encryption method based on model driving |
CN114977176A (en) * | 2022-07-19 | 2022-08-30 | 深圳江行联加智能科技有限公司 | Power load decomposition method, device, equipment and storage medium |
CN115018011A (en) * | 2022-07-19 | 2022-09-06 | 深圳江行联加智能科技有限公司 | Power load type identification method, device, equipment and storage medium |
CN115085274A (en) * | 2022-07-27 | 2022-09-20 | 北京智芯微电子科技有限公司 | Automatic identification method and device for new energy equipment access, electronic equipment and medium |
CN115085274B (en) * | 2022-07-27 | 2023-01-20 | 北京智芯微电子科技有限公司 | Automatic identification method and device for new energy equipment access, electronic equipment and medium |
Also Published As
Publication number | Publication date |
---|---|
CN111242391B (en) | 2023-10-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111242391A (en) | Machine learning model training method and system for power load identification | |
CN111382789B (en) | Power load identification method and system based on machine learning | |
CN111026927B (en) | Low-voltage transformer area running state intelligent monitoring system | |
CN108021736B (en) | Load switching action monitoring method based on sliding window residual error model | |
US9658261B2 (en) | Current pattern matching method for non-intrusive power load monitoring and disaggregation | |
CN107025365B (en) | A kind of non-intruding load discrimination method for user side | |
Yu et al. | Nonintrusive appliance load monitoring for smart homes: Recent advances and future issues | |
CN106483370A (en) | Non-intrusion type household loads real-time identification method based on multi-feature fusion and device | |
CN102279326A (en) | Electrical appliance recognition method and household user carbon emission metering system applying same | |
CN106093565B (en) | A kind of electricity subentry measurement method and device based on steady state characteristic Waveform Matching | |
EP2217886A1 (en) | Improvements in hierarchy determination for power monitoring systems | |
CN107741577B (en) | Gateway meter accuracy online monitoring and analyzing method and system | |
CN109298379A (en) | A kind of recognition methods of the intelligent electric meter site error exception based on data monitoring | |
CN111027408A (en) | Load identification method based on support vector machine and V-I curve characteristics | |
CN115759708A (en) | Line loss analysis method and system considering power space-time distribution characteristics | |
CN105974220A (en) | Residential community power load identification system | |
CN113466548A (en) | Intelligent electric meter region identification method based on phasor measurement technology | |
CN105305437A (en) | Tri-reliability matching and identification method of electric load | |
CN111638401A (en) | Capacitor online monitoring system and method | |
Bier et al. | Smart meter systems measurements for the verification of the detection & classification algorithms | |
CN115133659A (en) | On-line monitoring method and system for low-voltage line loss of low-voltage distribution network | |
CN113422432B (en) | Electrical fire prevention and control system based on non-invasive load monitoring | |
Li | Abnormal energy consumption analysis based on big data mining technology | |
CN113985098A (en) | Method for improving electricity larceny prevention analysis accuracy based on real-time metering | |
CN114325107A (en) | Method for judging accuracy of harmonic impedance calculation result |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |