CN110969306A - Power distribution low-voltage distribution area load prediction method and device based on deep learning - Google Patents
Power distribution low-voltage distribution area load prediction method and device based on deep learning Download PDFInfo
- Publication number
- CN110969306A CN110969306A CN201911235433.4A CN201911235433A CN110969306A CN 110969306 A CN110969306 A CN 110969306A CN 201911235433 A CN201911235433 A CN 201911235433A CN 110969306 A CN110969306 A CN 110969306A
- Authority
- CN
- China
- Prior art keywords
- prediction model
- data
- distribution
- low
- voltage
- 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
- 238000009826 distribution Methods 0.000 title claims abstract description 155
- 238000013135 deep learning Methods 0.000 title claims abstract description 32
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 47
- 238000005259 measurement Methods 0.000 claims abstract description 34
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims abstract description 18
- 238000005457 optimization Methods 0.000 claims abstract description 14
- 238000013528 artificial neural network Methods 0.000 claims abstract description 9
- 238000012360 testing method Methods 0.000 claims description 21
- 230000002159 abnormal effect Effects 0.000 claims description 17
- 230000000694 effects Effects 0.000 claims description 16
- 238000012795 verification Methods 0.000 claims description 14
- 238000012937 correction Methods 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 9
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000007405 data analysis Methods 0.000 claims description 7
- 230000008030 elimination Effects 0.000 claims description 7
- 238000003379 elimination reaction Methods 0.000 claims description 7
- 238000001514 detection method Methods 0.000 claims description 4
- 238000005516 engineering process Methods 0.000 description 6
- 230000008901 benefit Effects 0.000 description 5
- 230000008859 change Effects 0.000 description 5
- 238000010801 machine learning Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000006467 substitution reaction 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
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Operations Research (AREA)
- Primary Health Care (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Entrepreneurship & Innovation (AREA)
- Quality & Reliability (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a method and a device for predicting load of a power distribution low-voltage distribution area based on deep learning, wherein the method comprises the following steps: collecting measurement data of a power distribution low-voltage transformer area; iteratively updating the weight of the neural network based on a back propagation algorithm through an Adam optimization algorithm according to the measured data of the distribution low-voltage transformer area, generating an input value of a prediction model, and training the prediction model; and after the prediction model meets the preset conditions, acquiring the current distribution low-voltage distribution area measurement data, and inputting the current distribution low-voltage distribution area measurement data into the trained prediction model to obtain a distribution low-voltage distribution area load prediction result. The method is based on an LSTM model, and a brand-new load prediction model is built by combining the information expression characteristics of a Wide module and a Deep module, so that the load prediction precision of a low-voltage distribution station area is improved.
Description
Technical Field
The invention relates to the technical field of power distribution network management of a power system, in particular to a method and a device for predicting load of a power distribution low-voltage transformer area based on deep learning.
Background
In recent years, the scale of a power system is continuously increased, and the characteristics of a power grid become more complex; with the massive access of new energy, the gradual popularization of new energy automobiles and the rise of intelligent household appliances, the load prediction of an electric power system, particularly a low-voltage distribution area, is increasingly difficult due to the complex multi-source factors.
The load prediction is the basis of the work of planning, dispatching and the like of the power system, and plays an important supporting role in ensuring the safe and economic operation of the power grid. Load prediction can be divided into long-term, medium-term, short-term and ultra-short term predictions from the time horizon of prediction. The short-term load prediction plays an important role in guiding and adjusting the daily operation of a power grid company, and the accurate prediction result can reasonably arrange the daily production plan of the power; meanwhile, the traditional machine learning algorithm cannot add expression to time sequence dimension information, so that the time sequence problem processing aspect is weak.
Disclosure of Invention
The present application is based on the recognition and discovery by the inventors of the following problems:
with the rise of a new generation of artificial intelligence technology, especially with the deep learning and the rapid development of big data technology, a good foundation is laid for further improving the accuracy of load prediction. Deep learning is a method based on data representation learning which is developed in recent years, is a new field in machine learning research, and aims to establish a neural network simulating human brain for analysis learning, mine the hidden intrinsic rules in data distribution and simulate the mechanism of the human brain to explain data; meanwhile, the model for deep learning is high in complexity, and mass data information can be well expressed.
The present invention is directed to solving, at least to some extent, one of the technical problems in the related art.
Therefore, the invention aims to provide a power distribution low-voltage distribution area load prediction method based on Deep learning, which is based on an LSTM model and combines the information expression characteristics of a Wide module and a Deep module to build a brand-new load prediction model, so that the load prediction precision of a low-voltage distribution area is improved.
The invention further aims to provide a power distribution low-voltage distribution area load prediction device based on deep learning.
In order to achieve the above object, an embodiment of the present invention provides a method for predicting a load of a distribution low-voltage distribution area based on deep learning, including the following steps: collecting measurement data of a power distribution low-voltage transformer area; iteratively updating the weight of the neural network based on a back propagation algorithm through an Adam optimization algorithm according to the measured data of the distribution low-voltage transformer area, generating an input value of a prediction model, and training the prediction model; and after the prediction model meets the preset condition, acquiring the current distribution low-voltage distribution area measurement data, and inputting the current distribution low-voltage distribution area measurement data into the trained prediction model to obtain a distribution low-voltage distribution area load prediction result.
The power distribution low-voltage distribution area load prediction method based on deep learning fully exerts the precious value of mass historical data of a power grid company and the popularization and application of a big data technology in the field of load prediction, and can effectively improve the expression effect of a model on the data distribution rule along with the continuous increase of the number of training samples and the diversified change of data characteristics; the prediction can be carried out only by data normally acquired by the power grid without depending on excessive external data, so that additional information acquisition equipment is not needed, the prediction can be economically realized, and meanwhile, a more accurate prediction result can be obtained; the comprehensive effect of the model is good, and the power load data information of the multi-feature dimension and the time sequence feature can be well expressed.
In addition, the method for predicting the load of the distribution low-voltage transformer area based on deep learning according to the above embodiment of the invention may further have the following additional technical features:
further, in an embodiment of the present invention, the collecting power distribution low voltage station area measurement data includes: and carrying out abnormal data analysis, data distribution analysis, default value filling, abnormal data elimination, numerical data, category data and characteristic engineering processing on the measured data of the distribution low-voltage distribution area.
Further, in one embodiment of the present invention, training the predictive model includes: dividing the processed measurement data of the distribution low-voltage transformer area into a training set, a verification set and a test set; training the predictive model and adjusting model parameters through the training set; optimizing and fine-tuning the prediction model through the verification set to obtain an optimal prediction model; and testing the effect of the optimal prediction model through the test set to generate a final prediction model.
Further, in an embodiment of the present invention, after training the prediction model, the method further includes: detecting a load prediction error of the final prediction model; if the load prediction error is smaller than or equal to a preset value, judging that the preset condition is met; and if the load prediction error is larger than the preset value, judging that the preset condition is not met.
Further, in an embodiment of the present invention, the formula of the Adam optimization algorithm is:
wherein ,mt、vtRespectively a first moment estimate and a second moment estimate of the gradient,
is a corrected value, β1 and β2For correction factor, α is the learning rate, gtIn order to be a gradient of the magnetic field,
in order to modify the coefficients after the iteration,
for the post-iteration correction factor, mt-1For the first moment estimation result of the previous iteration, vt-1And estimating the result of the second moment of the previous iteration.
In order to achieve the above object, an embodiment of another aspect of the present invention provides an apparatus for predicting load of a distribution low-voltage distribution area based on deep learning, including: the acquisition module is used for acquiring measurement data of the distribution low-voltage transformer area; the iterative updating module is used for iteratively updating the weight of the neural network based on a back propagation algorithm through an Adam optimization algorithm according to the measured data of the distribution low-voltage distribution area, generating an input value of a prediction model and training the prediction model; and the prediction module is used for acquiring the current power distribution low-voltage distribution area measurement data after the prediction model meets the preset condition, and inputting the current power distribution low-voltage distribution area measurement data into the trained prediction model to obtain a power distribution low-voltage distribution area load prediction result.
The power distribution low-voltage distribution area load prediction device based on deep learning fully exerts the precious value of mass historical data of a power grid company and the popularization and application of a big data technology in the field of load prediction, and can effectively improve the expression effect of a model on a data distribution rule along with the continuous increase of the number of training samples and the diversified change of data characteristics; the prediction can be carried out only by data normally acquired by the power grid without depending on excessive external data, so that additional information acquisition equipment is not needed, the prediction can be economically realized, and meanwhile, a more accurate prediction result can be obtained; the comprehensive effect of the model is good, and the power load data information of the multi-feature dimension and the time sequence feature can be well expressed.
In addition, the distribution low-voltage platform load prediction device based on deep learning according to the above embodiment of the present invention may further have the following additional technical features:
further, in an embodiment of the present invention, the acquisition module is further configured to perform abnormal data analysis, data distribution analysis, default value padding, abnormal data elimination, numerical data, category data, and feature engineering processing on the measured data of the distribution low-voltage distribution area.
Further, in one embodiment of the present invention, training the predictive model includes: dividing the processed measurement data of the distribution low-voltage transformer area into a training set, a verification set and a test set; training the predictive model and adjusting model parameters through the training set; optimizing and fine-tuning the prediction model through the verification set to obtain an optimal prediction model; and testing the effect of the optimal prediction model through the test set to generate a final prediction model.
Further, in an embodiment of the present invention, the method further includes: a detection module for detecting a load prediction error of the final prediction model after training the prediction model; the judging module is used for judging that the preset condition is met when the load prediction error is smaller than or equal to a preset value; and when the load prediction error is larger than the preset value, judging that the preset condition is not met.
Further, in an embodiment of the present invention, the formula of the Adam optimization algorithm is:
wherein ,mt、vtRespectively a first moment estimate and a second moment estimate of the gradient,
is a corrected value, β1 and β2For correction factor, α is learning rate, gtIn order to be a gradient of the magnetic field,
in order to modify the coefficients after the iteration,
for the post-iteration correction factor, mt-1For the first moment estimation result of the previous iteration, vt-1And estimating the result of the second moment of the previous iteration.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
fig. 1 is a flowchart of a deep learning based load prediction method for a distribution low voltage distribution area according to an embodiment of the present invention;
FIG. 2 is a flow chart of a deep learning based distribution low voltage distribution area load prediction method according to an embodiment of the invention;
FIG. 3 is an illustrative diagram of model input and output layer data constructs in accordance with an embodiment of the invention;
fig. 4 is a schematic structural diagram of a distribution low-voltage distribution area load prediction device based on deep learning according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
Aiming at the technical problem of the background art, the embodiment of the invention provides a load prediction method and a load prediction device based on the load prediction requirement of an actual low-voltage distribution station area, and a brand-new load prediction model is built based on an LSTM model and in combination with the information expression characteristics of a Wide module and a Deep module, so that the load prediction precision of the low-voltage distribution station area is improved.
The method and the device for predicting the load of the low-voltage distribution area based on deep learning provided by the embodiment of the invention are described below with reference to the accompanying drawings, and firstly, the method for predicting the load of the low-voltage distribution area based on deep learning provided by the embodiment of the invention is described with reference to the accompanying drawings.
Fig. 1 is a flowchart of a method for predicting load of a distribution low-voltage distribution area based on deep learning according to an embodiment of the present invention.
As shown in fig. 1, the method for predicting the load of the distribution low-voltage distribution area based on deep learning includes the following steps:
in step S101, power distribution low voltage distribution area measurement data is collected.
In an embodiment of the present invention, collecting measurement data of a distribution low voltage distribution area includes: and carrying out abnormal data analysis, data distribution analysis, default value filling, abnormal data elimination, numerical data, category data and characteristic engineering processing on the measured data of the distribution low-voltage distribution area.
Specifically, as shown in fig. 2, "data determines an upper limit of a machine learning model, and an algorithm approaches the upper limit as much as possible", so that in order to train an effective model, data of the training model needs to be processed, which specifically includes:
1) and (5) abnormal data analysis. According to the outlier data judgment principle of the boxed graph, defining the abnormal criterion of the data w:
in the formula Q1 and Q3A first quartile and a third quartile, respectively; i isQRIs a quarter-bit pitch.
2) And (5) analyzing data distribution. To determine the time distribution of the active load, the time nodes of the peak load period are roughly determined by plotting the active power of the collected samples.
3) And filling default values. The default value processing methods are numerous, and different processing methods are selected according to different practical problems, and the filling average value, the mode, the median and the like are common. Because the power measurement parameters have time continuity and the numerical jump in a short time often does not exist at the non-fault time, the default value of the t time is approximately filled by the numerical value of the t-1 time.
4) And (5) abnormal data elimination. Screening abnormal data through a formula (1), removing abnormal outlier data, and filling according to a default filling method in the step 3).
5) Numerical data. The value ranges of the input numerical characteristic data are different, and an obvious size relationship exists on the numerical value, so that the input numerical characteristic data needs to be normalized according to each characteristic:
in the formula ：
is the nth sample of the ith sampleNumerical characteristic data; min (x)(n)) And max (x)(n)) The minimum value and the maximum value of the nth numerical characteristic data are respectively set;
the characteristic data is the nth characteristic data of the ith sample after normalization processing.
6) And (4) type data. The class type can reflect a certain characteristic attribute specific to a certain sample, and in order to enable a computer to "read" the specificity of the class, one-hot-encoding (one-hot-encoding) needs to be performed on the class type data.
7) And (5) characteristic engineering. The feature engineering refers to a process of converting original data into training data of a model, and aims to obtain better training data features so that a machine learning model approaches to the upper limit of the data. The characteristic engineering plays a very important role in machine learning, and the performance of the model can be improved. Namely, the hidden information of the original feature data is extracted through feature extraction, so that the important feature information contained in the original feature data is mined. The time characteristics are subjected to characteristic extraction in consideration of the power utilization characteristics of the power load, and the new characteristic data are month attributes, week attributes, hour attributes, working day attributes and peak load time attributes and are encoded by using a unique heat vector encoding rule.
In step S102, according to the measurement data of the distribution low-voltage distribution area, iteratively updating the neural network weights through an Adam optimization algorithm based on a back propagation algorithm, generating input values of a prediction model, and training the prediction model.
The adam (adaptive motion) optimization algorithm is an extended form of a random gradient descent algorithm, is widely applied to deep learning in recent years, is a first-order optimization algorithm capable of replacing the traditional random gradient descent process, and can iteratively update the weight of a neural network based on training data.
The Adam algorithm differs from the traditional random gradient descent. The stochastic gradient descent keeps a single learning rate (i.e., alpha) updating all weights, and the learning rate does not change during the training process. And Adam dynamically adjusts the learning rate for each parameter by using the first moment estimate and the second moment estimate of the gradient. The Adam algorithm achieves the advantages of both the AdaGrad and RMSProp algorithms. Adam not only calculates the adaptive parameter learning rate based on the first moment mean as in the RMSProp algorithm, it also takes full advantage of the second moment mean of the gradient (i.e., biased variance/unconcentrated variance).
Adam has the advantages that after offset correction, the learning rate of each iteration has a certain range, so that the parameters are relatively stable. The formula is as follows:
wherein mt、vtRespectively a first moment estimate and a second moment estimate of the gradient,
is a corrected value, and the learning rate α is 0.006 and the correction coefficient β is adopted herein1 and β20.9 and 0.999 respectively, and epsilon is 10-8。
It can be understood that the collected data of the low-voltage distribution station area often includes information records of active power, reactive power, voltage, current and collection time. Therefore, the processed data is used as a model input through data processing and characteristic engineering. The specific flow chart is shown in figure 3.
Further, in one embodiment of the present invention, training a predictive model includes: dividing the processed measurement data of the distribution low-voltage transformer area into a training set, a verification set and a test set; training a prediction model through a training set and adjusting model parameters; optimizing and fine-tuning the prediction model through a verification set to obtain an optimal prediction model; and testing the effect of the optimal prediction model through the test set to generate a final prediction model.
It can be understood that high-quality data after data processing is generally divided into a training set, a verification set and a test set according to the proportion of 50%, 30% and 20%. The model adjustment model parameters are trained through a training set. And optimizing and fine-tuning the model through the verification set to obtain the optimized model. And finally, testing the optimized model effect through a test set.
In step S103, after the prediction model meets the preset condition, the current distribution low-voltage distribution area measurement data is collected, and the current distribution low-voltage distribution area measurement data is input into the trained prediction model, so as to obtain a distribution low-voltage distribution area load prediction result.
In one embodiment of the present invention, after training the prediction model, the method further includes: detecting a load prediction error of the final prediction model; if the load prediction error is smaller than or equal to a preset value, judging that a preset condition is met; and if the load prediction error is larger than the preset value, judging that the preset condition is not met.
Specifically, the error evaluation indexes of the load prediction are many, and include an Absolute Error (AE), a Relative Error (RE), a maximum/minimum relative error (MaxRE/MinRE), a Mean Absolute Percentage Error (MAPE), a Root Mean Square Error (RMSE), a failure rate of a prediction result, and the like. In order to ensure the sensitivity of indexes and predict the actual application requirements by combining the load of the distribution low-voltage distribution area, the root mean square percentage error and the average absolute percentage error are selected as the basis for judging the prediction effect.
Evaluation index 1-root mean square error RMSPE, formula:
evaluation index 2-percent absolute error MAPE, formula:
in summary, the power distribution low-voltage distribution area load prediction method based on deep learning provided by the embodiment of the invention fully exerts the precious value of massive historical data of a power grid company and the popularization and application of a big data technology in the field of load prediction, and can effectively improve the expression effect of a model on the data distribution rule along with the continuous increase of the number of training samples and the diversified change of data characteristics; the prediction can be carried out only by data normally acquired by the power grid without depending on excessive external data, so that additional information acquisition equipment is not needed, the prediction can be economically realized, and meanwhile, a more accurate prediction result can be obtained; the comprehensive effect of the model is good, and the power load data information of the multi-feature dimension and the time sequence feature can be well expressed.
The following describes a deep learning-based distribution low-voltage transformer area load prediction device according to an embodiment of the present invention with reference to the drawings.
Fig. 4 is a schematic structural diagram of a distribution low-voltage distribution area load prediction device based on deep learning according to an embodiment of the present invention.
As shown in fig. 4, the deep learning based distribution low voltage distribution area load prediction apparatus 10 includes: an acquisition module 100, an iterative update module 200, and a prediction module 300.
The acquisition module 100 is used for acquiring measurement data of the distribution low-voltage distribution area; the iterative updating module 200 is used for iteratively updating the neural network weight through an Adam optimization algorithm based on a back propagation algorithm according to the measured data of the distribution low-voltage distribution area, generating an input value of a prediction model, and training the prediction model; the prediction module 300 is configured to acquire current power distribution low-voltage distribution area measurement data after the prediction model meets a preset condition, and input the current power distribution low-voltage distribution area measurement data into the trained prediction model to obtain a power distribution low-voltage distribution area load prediction result. The device 10 provided by the embodiment of the invention is based on an LSTM model, and a brand-new load prediction model is built by combining the information expression characteristics of the Wide module and the Deep module, so that the load prediction precision of a low-voltage distribution station area is improved.
Further, in an embodiment of the present invention, the acquisition module 100 is further configured to perform abnormal data analysis, data distribution analysis, default value padding, abnormal data elimination, numerical data, category data, and feature engineering processing on the measurement data of the distribution low-voltage distribution area.
Further, in one embodiment of the present invention, training a predictive model includes: dividing the processed measurement data of the distribution low-voltage transformer area into a training set, a verification set and a test set; training a prediction model through a training set and adjusting model parameters; optimizing and fine-tuning the prediction model through a verification set to obtain an optimal prediction model; and testing the effect of the optimal prediction model through the test set to generate a final prediction model.
Further, in one embodiment of the present invention, the apparatus 10 of the embodiment of the present invention further comprises: the device comprises a detection module and a judgment module. The detection module is used for detecting the load prediction error of the final prediction model after the prediction model is trained; the judging module is used for judging that a preset condition is met when the load prediction error is smaller than or equal to a preset value; and when the load prediction error is larger than a preset value, judging that the preset condition is not met.
Further, in one embodiment of the present invention, the formula of the Adam optimization algorithm is:
wherein ,mt、vtRespectively a first moment estimate and a second moment estimate of the gradient,
is a corrected value, β1 and β2For correction factor, α is the learning rate, gtIn order to be a gradient of the magnetic field,
in order to modify the coefficients after the iteration,
for the post-iteration correction factor, mt-1For the first moment estimation result of the previous iteration, vt-1And estimating the result of the second moment of the previous iteration.
It should be noted that the foregoing explanation of the embodiment of the deep learning-based power distribution low-voltage distribution area load prediction method is also applicable to the deep learning-based power distribution low-voltage distribution area load prediction apparatus of the embodiment, and details are not described here.
According to the power distribution low-voltage distribution area load prediction device based on deep learning, the precious value of mass historical data of a power grid company and the popularization and application of a big data technology in the field of load prediction are fully exerted, and the expression effect of a model on the data distribution rule can be effectively improved along with the continuous increase of the number of training samples and the diversified change of data characteristics; the prediction can be carried out only by data normally acquired by the power grid without depending on excessive external data, so that additional information acquisition equipment is not needed, the prediction can be economically realized, and meanwhile, a more accurate prediction result can be obtained; the comprehensive effect of the model is good, and the power load data information of the multi-feature dimension and the time sequence feature can be well expressed.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
In the present invention, unless otherwise expressly stated or limited, the first feature "on" or "under" the second feature may be directly contacting the first and second features or indirectly contacting the first and second features through an intermediate. Also, a first feature "on," "over," and "above" a second feature may be directly or diagonally above the second feature, or may simply indicate that the first feature is at a higher level than the second feature. A first feature being "under," "below," and "beneath" a second feature may be directly under or obliquely under the first feature, or may simply mean that the first feature is at a lesser elevation than the second feature.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (10)
1. A power distribution low-voltage distribution area load prediction method based on deep learning is characterized by comprising the following steps:
collecting measurement data of a power distribution low-voltage transformer area;
iteratively updating the weight of the neural network based on a back propagation algorithm through an Adam optimization algorithm according to the measured data of the distribution low-voltage transformer area, generating an input value of a prediction model, and training the prediction model; and
and after the prediction model meets the preset condition, acquiring the current distribution low-voltage distribution area measurement data, and inputting the current distribution low-voltage distribution area measurement data into the trained prediction model to obtain a distribution low-voltage distribution area load prediction result.
2. The method of claim 1, wherein collecting distribution low voltage bay measurement data comprises:
and carrying out abnormal data analysis, data distribution analysis, default value filling, abnormal data elimination, numerical data, category data and characteristic engineering processing on the measured data of the distribution low-voltage distribution area.
3. The method of claim 2, wherein training the predictive model comprises:
dividing the processed measurement data of the distribution low-voltage transformer area into a training set, a verification set and a test set;
training the predictive model and adjusting model parameters through the training set;
optimizing and fine-tuning the prediction model through the verification set to obtain an optimal prediction model;
and testing the effect of the optimal prediction model through the test set to generate a final prediction model.
4. The method of claim 3, further comprising, after training the predictive model:
detecting a load prediction error of the final prediction model;
if the load prediction error is smaller than or equal to a preset value, judging that the preset condition is met;
and if the load prediction error is larger than the preset value, judging that the preset condition is not met.
5. The method according to any of claims 1-4, wherein the formula of the Adam optimization algorithm is:
wherein ,mt、vtRespectively a first moment estimate and a second moment estimate of the gradient,
is a corrected value, β1 and β2For correction factor, α is learning rate, gtIn order to be a gradient of the magnetic field,
in order to modify the coefficients after the iteration,
for the post-iteration correction factor, mt-1For the first moment estimation result of the previous iteration, vt-1And estimating the result of the second moment of the previous iteration.
6. A distribution low-voltage transformer district load prediction device based on deep learning is characterized by comprising:
the acquisition module is used for acquiring measurement data of the distribution low-voltage transformer area;
the iterative updating module is used for iteratively updating the weight of the neural network based on a back propagation algorithm through an Adam optimization algorithm according to the measured data of the distribution low-voltage distribution area, generating an input value of a prediction model and training the prediction model; and
and the prediction module is used for acquiring the current power distribution low-voltage distribution area measurement data after the prediction model meets the preset condition, and inputting the current power distribution low-voltage distribution area measurement data into the trained prediction model to obtain a power distribution low-voltage distribution area load prediction result.
7. The device of claim 6, wherein the collection module is further configured to perform abnormal data analysis, data distribution analysis, default padding, abnormal data elimination, numerical data, classification data, and feature engineering on the distribution low voltage block measurement data.
8. The apparatus of claim 7, wherein training the predictive model comprises:
dividing the processed measurement data of the distribution low-voltage transformer area into a training set, a verification set and a test set;
training the predictive model and adjusting model parameters through the training set;
optimizing and fine-tuning the prediction model through the verification set to obtain an optimal prediction model;
and testing the effect of the optimal prediction model through the test set to generate a final prediction model.
9. The apparatus of claim 8, further comprising:
a detection module for detecting a load prediction error of the final prediction model after training the prediction model;
the judging module is used for judging that the preset condition is met when the load prediction error is smaller than or equal to a preset value; and when the load prediction error is larger than the preset value, judging that the preset condition is not met.
10. The apparatus according to any of claims 6-9, wherein the Adam optimization algorithm has the formula:
wherein ,mt、vtRespectively a first moment estimate and a second moment estimate of the gradient,
is a corrected value, β1 and β2For correction factor, α is learning rate, gtIn order to be a gradient of the magnetic field,
in order to modify the coefficients after the iteration,
for the post-iteration correction factor, mt-1For the first moment estimation result of the previous iteration, vt-1And estimating the result of the second moment of the previous iteration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911235433.4A CN110969306B (en) | 2019-12-05 | 2019-12-05 | Power distribution low-voltage transformer area load prediction method and device based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911235433.4A CN110969306B (en) | 2019-12-05 | 2019-12-05 | Power distribution low-voltage transformer area load prediction method and device based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110969306A true CN110969306A (en) | 2020-04-07 |
| CN110969306B CN110969306B (en) | 2023-09-19 |
Family
ID=70033041
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911235433.4A Active CN110969306B (en) | 2019-12-05 | 2019-12-05 | Power distribution low-voltage transformer area load prediction method and device based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110969306B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112050397A (en) * | 2020-08-27 | 2020-12-08 | 浙江省邮电工程建设有限公司 | Method and system for regulating and controlling temperature of machine room |
| CN112232714A (en) * | 2020-11-18 | 2021-01-15 | 中国科学院电工研究所 | Power distribution network risk assessment method under incomplete structural parameters based on deep learning |
| CN112508267A (en) * | 2020-12-04 | 2021-03-16 | 国网湖南省电力有限公司 | Distribution transformer area load prediction method and system based on artificial intelligence |
| CN114118862A (en) * | 2021-12-07 | 2022-03-01 | 双良节能系统股份有限公司 | Dynamic heat load prediction and regulation method and system for heat exchange station |
| CN114186739A (en) * | 2021-12-02 | 2022-03-15 | 国网辽宁省电力有限公司沈阳供电公司 | Platform area short-term load prediction method based on deep learning |
| CN114997477A (en) * | 2022-05-20 | 2022-09-02 | 国网江苏省电力有限公司苏州供电分公司 | Low-voltage residential district load prediction method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107247737A (en) * | 2017-05-10 | 2017-10-13 | 国家电网公司 | Platform area promise breaking electrical energy consumption analysis and method for digging based on power consumption |
| CN109002942A (en) * | 2018-09-28 | 2018-12-14 | 河南理工大学 | A kind of short-term load forecasting method based on stochastic neural net |
| CN109242180A (en) * | 2018-08-31 | 2019-01-18 | 合肥工业大学 | Long-medium term power load forecasting method and system |
| US20190050727A1 (en) * | 2018-01-12 | 2019-02-14 | Timothy Anderson | Neural network training |
| CN109766950A (en) * | 2019-01-18 | 2019-05-17 | 东北大学 | A kind of industrial user's short-term load forecasting method based on form cluster and LightGBM |
| CN110084424A (en) * | 2019-04-25 | 2019-08-02 | 国网浙江省电力有限公司 | A kind of Methods of electric load forecasting based on LSTM and LGBM |
-
2019
- 2019-12-05 CN CN201911235433.4A patent/CN110969306B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107247737A (en) * | 2017-05-10 | 2017-10-13 | 国家电网公司 | Platform area promise breaking electrical energy consumption analysis and method for digging based on power consumption |
| US20190050727A1 (en) * | 2018-01-12 | 2019-02-14 | Timothy Anderson | Neural network training |
| CN109242180A (en) * | 2018-08-31 | 2019-01-18 | 合肥工业大学 | Long-medium term power load forecasting method and system |
| CN109002942A (en) * | 2018-09-28 | 2018-12-14 | 河南理工大学 | A kind of short-term load forecasting method based on stochastic neural net |
| CN109766950A (en) * | 2019-01-18 | 2019-05-17 | 东北大学 | A kind of industrial user's short-term load forecasting method based on form cluster and LightGBM |
| CN110084424A (en) * | 2019-04-25 | 2019-08-02 | 国网浙江省电力有限公司 | A kind of Methods of electric load forecasting based on LSTM and LGBM |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112050397A (en) * | 2020-08-27 | 2020-12-08 | 浙江省邮电工程建设有限公司 | Method and system for regulating and controlling temperature of machine room |
| CN112232714A (en) * | 2020-11-18 | 2021-01-15 | 中国科学院电工研究所 | Power distribution network risk assessment method under incomplete structural parameters based on deep learning |
| CN112232714B (en) * | 2020-11-18 | 2023-06-20 | 中国科学院电工研究所 | Deep learning-based risk assessment method for distribution network under incomplete structural parameters |
| CN112508267A (en) * | 2020-12-04 | 2021-03-16 | 国网湖南省电力有限公司 | Distribution transformer area load prediction method and system based on artificial intelligence |
| CN114186739A (en) * | 2021-12-02 | 2022-03-15 | 国网辽宁省电力有限公司沈阳供电公司 | Platform area short-term load prediction method based on deep learning |
| CN114118862A (en) * | 2021-12-07 | 2022-03-01 | 双良节能系统股份有限公司 | Dynamic heat load prediction and regulation method and system for heat exchange station |
| CN114997477A (en) * | 2022-05-20 | 2022-09-02 | 国网江苏省电力有限公司苏州供电分公司 | Low-voltage residential district load prediction method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110969306B (en) | 2023-09-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110969306B (en) | Power distribution low-voltage transformer area load prediction method and device based on deep learning | |
| Lin et al. | An improved moth-flame optimization algorithm for support vector machine prediction of photovoltaic power generation | |
| CN110474339B (en) | Power grid reactive power control method based on deep power generation load prediction | |
| CN103023065B (en) | Wind power short-term power prediction method based on relative error entropy evaluation method | |
| CN111525587B (en) | Reactive load situation-based power grid reactive voltage control method and system | |
| CN110380444B (en) | Capacity planning method for distributed wind power orderly access to power grid under multiple scenes based on variable structure Copula | |
| CN105701562B (en) | Training method, applicable method for predicting generated power and respective system | |
| CN107015900B (en) | A kind of service performance prediction technique of video website | |
| CN107103184A (en) | A kind of high-voltage cable joint temperature predicting method | |
| CN112990500A (en) | Transformer area line loss analysis method and system based on improved weighted gray correlation analysis | |
| KR20200128232A (en) | Apparatus for predicting electricity demand and method thereof | |
| JP5923454B2 (en) | Power system management apparatus and method | |
| CN114021848A (en) | Generating capacity demand prediction method based on LSTM deep learning | |
| CN108647838B (en) | Power grid situation sensing method based on random matrix theory and online sequence extreme learning machine | |
| CN107194507A (en) | A kind of short-term wind speed forecasting method of wind farm based on combination SVMs | |
| CN117878979B (en) | Power balance and dynamic compensation system based on electric energy storage | |
| CN108256693A (en) | A kind of photovoltaic power generation power prediction method, apparatus and system | |
| CN109165770A (en) | Novel photovoltaic power forecasting method based on AFSA-Elman | |
| CN111612227A (en) | Load prediction method based on K-means clustering and bat optimization neural network | |
| CN111311001A (en) | Bi-LSTM network short-term load prediction method based on DBSCAN algorithm and feature selection | |
| CN107563641B (en) | Disaster-resistant multi-scene differentiation planning method for power distribution network considering disaster preference | |
| Sun et al. | Support vector machines with PSO algorithm for short-term load forecasting | |
| CN109034275A (en) | Prediction technique, system, medium and the equipment of polycrystalline reduction process energy consumption value | |
| CN114759579A (en) | Power grid active power optimization control system, method and medium based on data driving | |
| Xiong et al. | Cost prediction on fabricated substation considering support vector machine via optimized quantum particle swarm optimization |
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 |