CN113327174A - Distribution transformer load prediction method - Google Patents
Distribution transformer load prediction method Download PDFInfo
- Publication number
- CN113327174A CN113327174A CN202110420221.4A CN202110420221A CN113327174A CN 113327174 A CN113327174 A CN 113327174A CN 202110420221 A CN202110420221 A CN 202110420221A CN 113327174 A CN113327174 A CN 113327174A
- Authority
- CN
- China
- Prior art keywords
- distribution transformer
- data
- transformer load
- load prediction
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000004458 analytical method Methods 0.000 claims abstract description 69
- 238000012423 maintenance Methods 0.000 claims abstract description 30
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 24
- 238000012549 training Methods 0.000 claims abstract description 24
- 238000004140 cleaning Methods 0.000 claims description 12
- 230000003068 static effect Effects 0.000 claims description 11
- 230000009466 transformation Effects 0.000 claims description 11
- 238000007726 management method Methods 0.000 claims description 9
- 230000000694 effects Effects 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 7
- 238000012800 visualization Methods 0.000 claims description 6
- 238000013499 data model Methods 0.000 claims description 5
- 238000013524 data verification Methods 0.000 claims description 5
- 238000005259 measurement Methods 0.000 claims description 5
- 238000012544 monitoring process Methods 0.000 claims description 5
- 238000012216 screening Methods 0.000 claims description 5
- 238000012937 correction Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 2
- 230000008676 import Effects 0.000 claims description 2
- 238000013277 forecasting method Methods 0.000 claims 2
- 238000011161 development Methods 0.000 description 10
- 238000007689 inspection Methods 0.000 description 9
- 230000008859 change Effects 0.000 description 6
- 230000003631 expected effect Effects 0.000 description 6
- 238000010219 correlation analysis Methods 0.000 description 4
- 238000007637 random forest analysis Methods 0.000 description 4
- 238000005457 optimization Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 2
- 238000005553 drilling Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000002123 temporal effect Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification 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
- 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
-
- 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"
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention belongs to the technical field of distribution transformer load prediction, and discloses a distribution transformer load prediction method, which comprises the following steps: training a distribution transformer load prediction model according to the load characteristic analysis result, the data wide table and the algorithm model; and predicting the distribution transformer load through the trained distribution transformer load prediction model to obtain a predicted value of the distribution transformer load. And judging whether the predicted value of the distribution transformer load exceeds a preset threshold value, and alarming when the predicted value of the distribution transformer load exceeds the preset threshold value. According to the distribution transformer operation analysis result, a corresponding threshold is set, when the predicted value is close to the threshold, early warning is carried out, when the predicted value exceeds the threshold, warning is carried out, short-term and ultra-short-term early warning and warning information is correspondingly output, the analysis result is pushed to operation and maintenance personnel through app to be checked, and the analysis result is pushed to management personnel to make a decision. And the line and distribution transformer containing sensitive users and important users are subjected to key early warning, so that the stable operation of the line is promoted, and the customer complaints are reduced.
Description
Technical Field
The invention belongs to the technical field of distribution transformer load prediction, and particularly relates to a distribution transformer load prediction method.
Background
In an electric power system, a distribution grid is an electric power grid that receives electric power from a transmission grid and distributes the electric power to individual electric power consumers. The distribution network is directly connected with power supply users and is an important component of the power system. The load prediction is an important component of power distribution network planning and also is the basis of the power distribution network planning, is closely related to construction investment, and is also an important condition for maintaining high-quality and economic operation of a power system. In the power distribution network planning work, in order to verify the reliability and economy of the power distribution network planning scheme, planning-state electrical simulation calculation and operation simulation need to be performed on the planning scheme.
The existing distribution transformer load prediction method lacks an early warning function and is easy to cause customer dissatisfaction and complaints.
Disclosure of Invention
The invention aims to provide a distribution transformer load prediction method to solve the existing problems.
In order to achieve the purpose, the invention provides the following technical scheme: a distribution transformer load prediction method comprises the following steps:
training a distribution transformer load prediction model according to the load characteristic analysis result, the data wide table and the algorithm model;
and predicting the distribution transformer load through the trained distribution transformer load prediction model to obtain a predicted value of the distribution transformer load.
And judging whether the predicted value of the distribution transformer load exceeds a preset threshold value, and alarming when the predicted value of the distribution transformer load exceeds the preset threshold value.
Preferably, the method for predicting a distribution transformer load of the present invention includes, before the step of determining whether the predicted value of the distribution transformer load exceeds a preset threshold, and when the predicted value of the distribution transformer load exceeds the preset threshold, performing an alarm, the steps of:
and setting the preset threshold value aiming at the distribution transformer load prediction result.
Preferably, the method for predicting the distribution transformer load of the present invention includes, before training the distribution transformer load prediction model according to the load characteristic analysis result, the data wide table and the algorithm model, the steps of:
collecting operation data, collecting and verifying statistical initial data in the historical load data to form first statistical data, and accessing meteorological data.
As a distribution transformer load prediction method of the present invention, preferably, before training the distribution transformer load prediction model according to the load characteristic analysis result, the data wide table and the algorithm model, the method further comprises the steps of:
and importing external static data.
As a distribution load prediction method of the present invention, it is preferable that the method includes, after the step of importing external static data in S20:
establishing a cleaning rule, and cleaning the operation data, the meteorological data and the external static data according to the cleaning rule to obtain accurate source data;
integrating the source data distributed in different tables to generate a uniform data model;
performing source tracing data verification and recalculation aiming at the first statistical data to obtain second statistical data;
and fusing multi-source data comprising the operation data, the meteorological data and the second statistical data, and processing to form a data wide table.
Preferably, the method for predicting distribution transformer load of the present invention includes the following steps after fusing the multi-source data including the operation data, the meteorological data and the second statistical data and processing to form the data wide table:
and carrying out load characteristic analysis according to the data wide table to obtain a load characteristic analysis result.
Preferably, as a distribution transformer load prediction method of the present invention, after training the distribution transformer load prediction model according to the load characteristic analysis result, the data wide table and the algorithm model, the method includes the steps of:
inputting the data of the previous N-1 days into the distribution transformer load prediction model, and predicting the distribution transformer load value of the Nth day;
acquiring a real distribution transformer load value of the Nth day;
and comparing the predicted distribution and transformation load value of the Nth day with the real distribution and transformation load value of the Nth day for deviation correction.
Respectively adopting an algorithm model and a mixing method to train and optimize a distribution transformer load prediction model;
and continuously monitoring the prediction accuracy, quantifying the prediction effect by adopting a measurement goodness of fit R square and a variance, and realizing the visualization of the estimation after prediction.
Preferably, the method for predicting the distribution transformer load of the present invention includes the steps of continuously monitoring the prediction accuracy, quantifying the prediction effect by using the metric goodness of fit R and the variance, and visualizing the post-prediction evaluation:
detecting whether the distribution transformer load prediction model has the condition that the accuracy rate is not expected or reduced;
when the situation that the accuracy rate of the distribution transformer load prediction model is not expected or reduced is detected, feedback is carried out and the distribution transformer load prediction model is adaptively optimized again;
and accessing the newly added historical operation data of the distribution transformer power, current and voltage and weather forecast data in preset time.
As a distribution transformer load prediction method of the present invention, it is preferable that the distribution transformer load prediction method includes, after the distribution transformer load prediction is performed by the trained distribution transformer load prediction model in S200 and a distribution transformer load prediction value is obtained, the steps of:
and displaying the prediction result, selecting the line and the distribution transformer for display, displaying the distribution transformer load prediction result of the current day by default, primarily displaying the first 10 distribution transformers with the highest overload probability under the selected line, and displaying 96-point load prediction values in a line graph form.
And clicking a display interface to drill and view detailed predicted values and historical predicted information, and screening through unit, line and distribution transformer names.
And displaying the factors which have the largest influence on the predicted value, and providing auxiliary data for operation and maintenance personnel.
Preferably, the method for predicting a distribution transformer load of the present invention, after the step of determining whether the predicted value of the distribution transformer load exceeds a preset threshold and when the predicted value of the distribution transformer load exceeds the preset threshold, alarming, comprises the steps of:
outputting short-term and ultra-short-term early warning and alarming information, pushing an analysis result to operation and maintenance personnel through app for checking, and pushing the analysis result to management personnel for decision making;
and performing key early warning on lines and distribution transformers containing sensitive users and important users.
Compared with the prior art, the invention has the following beneficial effects:
(1) the method has the function of showing the comparison curve of the distribution variation predicted load today and the actual load today and the comparison curve of the yesterday predicted load and the actual load to field base layer operation and maintenance personnel in the form of a broken line graph, is convenient to know the operation characteristics of equipment, and is further convenient for capacity increase decision.
(2) The method comprises the steps of performing carousel display on predicted heavily-overloaded equipment, checking the detail of the predicted heavily-overloaded equipment, performing key early warning on lines containing sensitive users and important users and distribution transformers, giving an early warning short message to an equipment owner, sending an active routing inspection work order to the distribution transformer equipment which has early warning for 5 times in a month, sending early warning on more than half of the distribution transformers on one line, sending a whole-line routing inspection work order, displaying in a character carousel mode, displaying the detail of the heavily-overloaded equipment in a list mode, and automatically sending the routing inspection work order by a background; the possibility of faults is reduced, stable operation of the line is promoted, and customer complaints are reduced.
(3) The invention counts the ratio of the distribution transformer quantity of the whole line aiming at the distribution transformer quantity of the specific line weight/overload to the operation and maintenance personnel of the basic level in a list form, and displays the ratio from large to small; the operation and maintenance personnel can predict risks in advance, auxiliary data are provided for the operation and maintenance personnel, and active operation and maintenance are realized.
(4) The method is based on a plurality of dimensions such as week, climate, holiday, economic development index and the like, deeply analyzes the change trend and influence factors of load data of model training, mainly comprises load periodicity analysis (such as week: Monday to Sunday), load correlation analysis (such as working day and non-working day), holiday load analysis (Yang/Yin calendar holiday), climate load analysis (such as temperature and humidity, wind speed, rainfall and the like), economic development index load analysis and the like, and displays the correlation of each dimension characteristic and load to site base-level operation and maintenance personnel in a form of combining a line graph and a bar graph; the influence factors of the load change trend can be clearer, the operation, maintenance and other auxiliary decisions are convenient to carry out, and support is provided for subsequent prediction modeling.
(5) The invention shows the load prediction accuracy rate of distribution change for about 10 days to the distribution places of province and city in a broken line graph mode. So as to know the accuracy of prediction and facilitate the optimization of the prediction model at any time.
(6) The invention analyzes the error of the predicted value and the actual value of the model training device during the training period, and displays the result to the province and city distribution places in a broken line graph mode, thereby being clear for the accuracy of the model training.
(7) According to the distribution transformer operation analysis result, a corresponding threshold is set, when the predicted value is close to the threshold, early warning is carried out, when the predicted value exceeds the threshold, warning is carried out, short-term and ultra-short-term early warning and warning information is correspondingly output, the analysis result is pushed to operation and maintenance personnel through app to be checked, and the analysis result is pushed to management personnel to make a decision. And the line and distribution transformer containing sensitive users and important users are subjected to key early warning, so that the stable operation of the line is promoted, and the customer complaints are reduced.
Drawings
FIG. 1 is a flow chart of a distribution transformer load prediction method according to the present invention;
FIG. 2 is a flowchart illustrating a distribution load prediction method according to a first embodiment of the present invention;
FIG. 3 is a flowchart illustrating a distribution load prediction method according to a second embodiment of the present invention;
FIG. 4 is a flowchart illustrating a distribution load prediction method according to a third embodiment of the present invention;
FIG. 5 is a flowchart illustrating a distribution load prediction method according to a fourth embodiment of the present invention;
FIG. 6 is a flowchart of a distribution load prediction method according to a fifth embodiment of the present invention;
fig. 7 is a flowchart of a distribution load prediction method according to a sixth embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, the present invention provides the following technical solutions: a distribution transformer load prediction method comprises the following steps:
s100, training a distribution transformer load prediction model according to a load characteristic analysis result, a data wide table and an algorithm model;
s200, carrying out distribution transformation load prediction through the trained distribution transformation load prediction model to obtain a distribution transformation load prediction value.
S300, judging whether the predicted value of the distribution transformer load exceeds a preset threshold value or not, and giving an alarm when the predicted value of the distribution transformer load exceeds the preset threshold value.
It is worth to be noted that the method adopts independent deployment of sub-project load prediction, and the power supply service command system load prediction module is connected in a menu mode in a function mode.
In this embodiment, carousel display is performed on the predicted heavy overload device, and details of the predicted heavy overload device can be checked. And the key early warning is carried out on the lines and distribution transformers containing sensitive users and important users. And giving an early warning short message to the equipment owner. And sending an active inspection work order to the distribution transformer equipment which has early warning for 5 times in one month. Over half of distribution transformers on one line are early warned, and a whole line inspection work order is sent.
Displaying in a character carousel mode, displaying the detailed heavy overload equipment in a list mode, and automatically sending a polling work order by a background; the possibility of faults is reduced, stable operation of the line is promoted, and customer complaints are reduced.
Referring to fig. 2, the present invention provides the following technical solutions:
s210, setting the preset threshold value aiming at the distribution transformer load prediction result.
In the embodiment, a corresponding threshold value is set for a distribution transformer operation analysis result, when a predicted value is close to the threshold value, early warning is performed, when the predicted value exceeds the threshold value, warning is performed, short-term and ultra-short-term early warning and warning information are correspondingly output, the analysis result is pushed to operation and maintenance personnel through app to be checked, and the analysis result is pushed to management personnel to make a decision. And the line and distribution transformer containing sensitive users and important users are subjected to key early warning, so that the stable operation of the line is promoted, and the customer complaints are reduced. It should be noted that the APP includes an operation and maintenance management APP with the publication number "CN 107832856A" and the patent name "a method for operation and maintenance management APP".
Referring to fig. 3, the present invention provides the following technical solutions:
s10, collecting operation data, collecting and verifying statistical initial data in the historical load data to form first statistical data, and accessing meteorological data;
s20 importing external static data;
s30, establishing a cleaning rule, and cleaning the operation data, the meteorological data and the external static data according to the cleaning rule to obtain accurate source data;
s40, integrating the source data distributed in different tables to generate a uniform data model;
s50, performing source tracing data verification and recalculation on the first statistical data to obtain second statistical data;
s60, fusing multi-source data including operation data, meteorological data and second statistical data, and processing to form a data wide table;
and S70, carrying out load characteristic analysis according to the data wide table to obtain a load characteristic analysis result.
In this embodiment, the power supply service command system is based on running data acquisition such as voltage, current, power data of the line/distribution transformer, and historical load data statistics data is acquired and verified at the same time, so as to ensure the accuracy of the data, and meteorological data including temperature, humidity, wind speed, weather, and the like are accessed at the same time.
And importing external static data such as holidays, economic development indexes and the like.
Many null values and abnormal data always exist in the original data, and in order to ensure the accuracy of prediction analysis, a set of proper cleaning rules needs to be established, so that an accurate data source is provided for load prediction and line loss analysis. In addition, the source data are distributed in different tables, and the data are required to be integrated to generate a uniform data model, which contains load values and other data characteristics (such as weather, etc.) for model establishment and load prediction.
Meanwhile, for various types of initial statistical data, source tracing data verification and recalculation are required, and accuracy of the statistical data is ensured.
And fusing multi-source data, and forming a wide table through a series of processing, wherein the wide table comprises the serial number, name, power, weather, temperature and humidity, holidays and the like of the distribution transformer and is used for constructing a prediction model.
Referring to fig. 4, the present invention provides the following technical solutions: s101, inputting data of the previous N-1 days into the distribution transformer load prediction model, and predicting a distribution transformer load value of the Nth day;
s102, acquiring a real distribution load value of the Nth day;
s103, comparing the predicted distribution and transformation load value of the Nth day with the real distribution and transformation load value of the Nth day for deviation correction.
S104, training and optimizing a distribution transformer load prediction model by adopting an algorithm model and a mixing method respectively;
s105, continuously monitoring the prediction accuracy, quantifying the prediction effect by adopting a measurement goodness of fit R square and a variance, and realizing the visualization of the estimation after prediction.
S106, detecting whether the distribution transformer load prediction model has the condition that the accuracy rate is not expected or reduced;
s107, when the situation that the accuracy rate of the distribution transformer load prediction model is not expected or reduced is detected, feedback is carried out and the distribution transformer load prediction model is adaptively optimized again;
and S108, accessing the newly added historical operation data of the distribution transformer power, current and voltage and weather forecast data in preset time.
In this embodiment, load characteristics are analyzed, data modeling is supported, and a prediction algorithm is selected based on a plurality of dimensions such as week, climate, holiday, economic development index, and the like. The load characteristic analysis mainly comprises load periodicity analysis, load correlation analysis, holiday load analysis, climate load analysis, economic development index load analysis and the like.
Based on the load characteristic analysis result, the short-term (daily) and ultra-short-term (temporal) load prediction analysis can be regarded as a regression problem, and algorithms such as dragline regression, ridge regression, tree models (decision trees, random forests, xgboost), multi-model fusion (such as random forests and xgboost) and the like are mainly adopted to carry out analysis.
And selecting proper data characteristics for modeling according to the conclusion of the load characteristic analysis and the algorithm type selection. And integrating the equipment operation data, the load data, the climate data, the holiday data, the economic index data and other data to form a multi-factor training set.
And (3) carrying out prediction model training by using historical data, inputting data of the previous N-1 days by adopting different algorithm models, predicting the distribution transformer load value of the Nth day, and then comparing the predicted distribution transformer load value with the real distribution transformer load value to correct the deviation.
After modeling and optimization are respectively carried out by adopting each algorithm and a mixing method, the prediction accuracy is quantized by adopting the R square and the variance, and the optimal model is adaptively selected for prediction.
In the operation process, the prediction accuracy is continuously monitored, the prediction effect is quantized by adopting a measurement goodness of fit R square and a variance, the visualization of estimation after prediction is realized, whether the accuracy rate of the model is not expected or reduced is detected, feedback is carried out, and the prediction model is adaptively optimized again.
And historical operation data of power, current and voltage of the distribution transformer newly added yesterday and weather forecast data are accessed every morning, and 24-hour distribution transformer load prediction is carried out through a trained prediction model.
Referring to fig. 5, the present invention provides the following technical solutions:
s201, displaying the prediction result, selecting the line and the distribution transformer to display, displaying the distribution transformer load prediction result of the current day by default, primarily displaying the first 10 distribution transformers with the highest overload probability under the selected line, and displaying the 96-point load prediction value in a line graph form.
S202, clicking a display interface to drill and view detailed predicted values and historical predicted information, and screening through unit, line and distribution transformer names.
S203, displaying the factors which have the largest influence on the predicted value, and providing auxiliary data for operation and maintenance personnel.
In this embodiment, the prediction result is displayed through the sail soft, the line and the distribution transformer can be selected for displaying, the distribution transformer load prediction result of the current day is displayed by default, the first 10 distribution transformers with the highest overload probability under the selected line are displayed firstly, and the 96-point load prediction value is displayed in a line graph form.
After clicking, drilling can be carried out to check detailed predicted values, screening can be carried out through unit, line and distribution and transformation names, and historical predicted information can also be checked.
In addition, factors with the largest influence on the predicted value are displayed, and auxiliary data are provided for operation and maintenance personnel.
Referring to fig. 6, the present invention provides the following technical solutions:
s301, outputting short-term and ultra-short-term early warning and alarming information, pushing an analysis result to operation and maintenance personnel through app for checking, and pushing the analysis result to management personnel for decision making;
s302, performing key early warning on the lines and distribution transformers containing sensitive users and important users.
In the embodiment, a corresponding threshold value is set for a distribution transformer operation analysis result, when a predicted value is close to the threshold value, early warning is performed, when the predicted value exceeds the threshold value, warning is performed, short-term and ultra-short-term early warning and warning information are correspondingly output, the analysis result is pushed to operation and maintenance personnel through app to be checked, and the analysis result is pushed to management personnel to make a decision. And the line and distribution transformer containing sensitive users and important users are subjected to key early warning, so that the stable operation of the line is promoted, and the customer complaints are reduced.
Referring to fig. 7, the present invention provides the following technical solutions:
the present/yesterday load forecast and actual value show.
Description of the function: and showing a comparison curve of the predicted load of the distribution transformer today and the actual load of the today. And displaying a yesterday predicted load and actual load comparison curve, and one moment every 15 minutes.
The display form is as follows: shown in the form of a line drawing.
The user: and (5) field basic operation and maintenance personnel.
The expected effect is as follows: and the running characteristics of the equipment are known, so that capacity increasing decision is facilitated.
And (4) early warning heavy/overload and sending an active inspection work order.
Description of the function: and performing carousel display on the predicted heavy overload equipment, and checking the detail of the predicted heavy overload equipment. And the key early warning is carried out on the lines and distribution transformers containing sensitive users and important users. And giving an early warning short message to the equipment owner. And sending an active inspection work order to the distribution transformer equipment which has early warning for 5 times in one month. Over half of distribution transformers on one line are early warned, and a whole line inspection work order is sent.
The display form is as follows: and displaying in a character carousel mode, displaying the details of the heavy overload equipment in a list mode, and automatically sending the inspection work order by a background.
The user: and (5) field basic operation and maintenance personnel.
The expected effect is as follows: the possibility of faults is reduced, stable operation of the line is promoted, and customer complaints are reduced.
And (4) displaying the weight/overload number ratio in a sorting mode.
Description of the function: and counting the ratio of the distribution transformer quantity of the whole line aiming at the distribution transformer quantity of the specific line heavy/overload, and displaying the ratio in a descending order.
The display form is as follows: shown in tabular form.
The user: and (5) field basic operation and maintenance personnel.
The expected effect is as follows: the operation and maintenance personnel can predict risks in advance, auxiliary data are provided for the operation and maintenance personnel, and active operation and maintenance are realized.
And (5) analyzing load characteristics.
Description of the function: based on a plurality of dimensions such as weeks, climates, holidays and economic development indexes, the change trend and the influence factors of the load data of the model training are deeply analyzed. The method mainly comprises load periodicity analysis (such as Monday to Sunday), load correlation analysis (such as working day and non-working day), holiday load analysis (Yang/Yin calendar holidays), climate load analysis (such as temperature, humidity, wind speed, rainfall and the like), economic development index load analysis and the like.
The display form is as follows: the correlation between each dimension characteristic and the load is shown in a form of a combination of a line graph and a bar graph.
The user: and (5) field basic operation and maintenance personnel.
The expected effect is as follows: the influence factors of the load change trend are clearer, the operation, maintenance and other auxiliary decisions are convenient to carry out, and support is provided for subsequent prediction modeling.
And displaying the load prediction accuracy.
Description of the function: and (5) displaying the load prediction accuracy rate of the distribution change for nearly 10 days.
The display form is as follows: shown in the form of a line drawing.
The user: province, city and power distribution.
The expected effect is as follows: the accuracy of prediction is known, and the prediction model is convenient to optimize at any time.
And displaying the error of the prediction model.
Description of the function: errors of the predicted values and the actual values of the model training apparatus during training are analyzed.
The display form is as follows: shown in the form of a line drawing.
The user: province, city and power distribution.
The expected effect is as follows: the accuracy of the model training is clear at a glance.
And (4) a function realization principle.
And fusing multi-source data.
And (5) acquiring basic data.
The power supply service command system is used for collecting operation data such as voltage, current and power data of lines/distribution transformers, collecting and verifying historical load data statistics data, ensuring the accuracy of the data, and accessing meteorological data including temperature, humidity, wind speed, weather and the like.
External static data import.
And importing external static data such as holidays, economic development indexes and the like.
And (5) cleaning and processing data.
Many null values and abnormal data always exist in the original data, and in order to ensure the accuracy of prediction analysis, a set of proper cleaning rules needs to be established, so that an accurate data source is provided for load prediction and line loss analysis. In addition, the source data are distributed in different tables, and the data are required to be integrated to generate a uniform data model, which contains load values and other data characteristics (such as weather, etc.) for model establishment and load prediction.
Meanwhile, for various types of initial statistical data, source tracing data verification and recalculation are required, and accuracy of the statistical data is ensured.
And fusing multi-source data, and forming a wide table through a series of processing, wherein the wide table comprises the serial number, name, power, weather, temperature and humidity, holidays and the like of the distribution transformer and is used for constructing a prediction model.
And (5) establishing a load prediction model.
And (5) analyzing load characteristics.
The load characteristics are analyzed, data modeling is supported, and a prediction algorithm is selected based on a plurality of dimensions such as week, climate, holiday, economic development index and the like. The load characteristic analysis mainly comprises load periodicity analysis, load correlation analysis, holiday load analysis, climate load analysis, economic development index load analysis and the like.
And (4) a load prediction algorithm.
Based on the load characteristic analysis result, the short-term (daily) and ultra-short-term (temporal) load prediction analysis can be regarded as a regression problem, and algorithms such as dragline regression, ridge regression, tree models (decision trees, random forests, xgboost), multi-model fusion (such as random forests and xgboost) and the like are mainly adopted to carry out analysis.
And establishing a load prediction model.
And selecting proper data characteristics for modeling according to the conclusion of the load characteristic analysis and the algorithm type selection. And integrating the equipment operation data, the load data, the climate data, the holiday data, the economic index data and other data to form a multi-factor training set.
And (3) carrying out prediction model training by using historical data, inputting data of the previous N-1 days by adopting different algorithm models, predicting the distribution transformer load value of the Nth day, and then comparing the predicted distribution transformer load value with the real distribution transformer load value to correct the deviation.
After modeling and optimization are respectively carried out by adopting each algorithm and a mixing method, the prediction accuracy is quantized by adopting the R square and the variance, and the optimal model is adaptively selected for prediction.
And (4) evaluating after load prediction.
In the operation process, the prediction accuracy is continuously monitored, the prediction effect is quantized by adopting a measurement goodness of fit R square and a variance, the visualization of estimation after prediction is realized, whether the accuracy rate of the model is not expected or reduced is detected, feedback is carried out, and the prediction model is adaptively optimized again.
And predicting the distribution load.
And historical operation data of power, current and voltage of the distribution transformer newly added yesterday and weather forecast data are accessed every morning, and 24-hour distribution transformer load prediction is carried out through a trained prediction model.
Line/distribution transformer overload prediction analysis.
Line/distribution operation analysis.
And displaying the prediction result through sail softwares, selecting lines and distribution transformers for displaying, displaying the distribution transformer load prediction result of the current day by default, primarily displaying the first 10 distribution transformers with the highest overload probability under the selected line, and displaying the 96-point load prediction value in a line graph form.
After clicking, drilling can be carried out to check detailed predicted values, screening can be carried out through unit, line and distribution and transformation names, and historical predicted information can also be checked.
In addition, factors with the largest influence on the predicted value are displayed, and auxiliary data are provided for operation and maintenance personnel.
And (5) early warning of operation of the line/distribution transformer.
And setting a corresponding threshold value aiming at the distribution transformer operation analysis result, performing early warning when the predicted value is close to the threshold value, performing alarm when the predicted value exceeds the threshold value, correspondingly outputting short-term and ultra-short-term early warning and alarm information, pushing the analysis result to operation and maintenance personnel through app for checking, and pushing the analysis result to management personnel for decision making. And the line and distribution transformer containing sensitive users and important users are subjected to key early warning, so that the stable operation of the line is promoted, and the customer complaints are reduced.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (10)
1. A distribution transformer load prediction method is characterized by comprising the following steps:
s100, training a distribution transformer load prediction model according to a load characteristic analysis result, a data wide table and an algorithm model;
s200, carrying out distribution transformer load prediction through the trained distribution transformer load prediction model to obtain a distribution transformer load prediction value;
s300, judging whether the predicted value of the distribution transformer load exceeds a preset threshold value or not, and giving an alarm when the predicted value of the distribution transformer load exceeds the preset threshold value.
2. The distribution transformer load prediction method according to claim 1, wherein the step of judging whether the predicted value of the distribution transformer load exceeds a preset threshold at S300, and before alarming when the predicted value of the distribution transformer load exceeds the preset threshold, comprises the steps of:
s210, setting the preset threshold value aiming at the distribution transformer load prediction result.
3. The distribution transformer load prediction method according to claim 1, characterized in that before training the distribution transformer load prediction model according to the load characteristic analysis result, the data wide table and the algorithm model, the method comprises the steps of:
s10 collects operation data, collects and verifies the initial statistical data in the historical load data to form first statistical data, and accesses the meteorological data at the same time.
4. The distribution transformer load forecasting method according to claim 3, further comprising the steps of, before training the distribution transformer load forecasting model according to the load characteristic analysis result, the data wide table and the algorithm model:
s20 imports external static data.
5. The distribution transformer load prediction method according to claim 4, characterized by comprising, after said importing external static data, the steps of:
s30, establishing a cleaning rule, and cleaning the operation data, the meteorological data and the external static data according to the cleaning rule to obtain accurate source data;
s40, integrating the source data distributed in different tables to generate a uniform data model;
s50, performing source tracing data verification and recalculation on the first statistical data to obtain second statistical data;
and S60, fusing multi-source data including the operation data, the meteorological data and the second statistical data, and processing to form a data wide table.
6. The distribution transformer load forecasting method according to claim 5, characterized in that after the multi-source data including the operation data, the meteorological data and the second statistical data are fused and processed to form the data wide table, the method comprises the following steps:
and S70, carrying out load characteristic analysis according to the data wide table to obtain a load characteristic analysis result.
7. The distribution transformer load prediction method according to claim 1, wherein the step of training the distribution transformer load prediction model based on the load characteristic analysis result, the data wide table and the algorithm model in S100 comprises:
s101, inputting data of the previous N-1 days into the distribution transformer load prediction model, and predicting a distribution transformer load value of the Nth day;
s102, acquiring a real distribution load value of the Nth day;
s103, comparing the predicted distribution and transformation load value of the Nth day with the real distribution and transformation load value of the Nth day for deviation correction;
s104, training and optimizing a distribution transformer load prediction model by adopting an algorithm model and a mixing method respectively;
s105, continuously monitoring the prediction accuracy, quantifying the prediction effect by adopting a measurement goodness of fit R square and a variance, and realizing the visualization of the estimation after prediction.
8. The distribution transformer load prediction method according to claim 7, wherein after the step S105 of continuously monitoring the prediction accuracy, quantifying the prediction effect by using the metric goodness-of-fit R-square and variance, and implementing the visualization of the post-prediction evaluation, the method comprises the steps of:
s106, detecting whether the distribution transformer load prediction model has the condition that the accuracy rate is not expected or reduced;
s107, when the situation that the accuracy rate of the distribution transformer load prediction model is not expected or reduced is detected, feedback is carried out and the distribution transformer load prediction model is adaptively optimized again;
and S108, accessing the newly added historical operation data of the distribution transformer power, current and voltage and weather forecast data in preset time.
9. The distribution transformer load prediction method according to claim 1, wherein the distribution transformer load prediction method in S200 through the trained distribution transformer load prediction model obtains a distribution transformer load prediction value, and then includes the steps of:
s201, displaying the prediction result, selecting a line and distribution transformers for displaying, displaying the distribution transformer load prediction result of the current day by default, primarily displaying the first 10 distribution transformers with the highest overload probability under the selected line, and displaying a 96-point load prediction value in a line graph form;
s202, clicking a display interface to drill and view detailed predicted values and historical predicted information, and screening the predicted values and the historical predicted information through unit, line and distribution transformer names;
s203, displaying the factors which have the largest influence on the predicted value, and providing auxiliary data for operation and maintenance personnel.
10. The distribution transformer load prediction method according to claim 1, wherein the step S300 of determining whether the predicted value of the distribution transformer load exceeds a preset threshold value, and when the predicted value of the distribution transformer load exceeds the preset threshold value, after performing an alarm, includes the steps of:
s301, outputting short-term and ultra-short-term early warning and alarming information, pushing an analysis result to operation and maintenance personnel through app for checking, and pushing the analysis result to management personnel for decision making;
s302, performing key early warning on the lines and distribution transformers containing sensitive users and important users.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110420221.4A CN113327174A (en) | 2021-04-19 | 2021-04-19 | Distribution transformer load prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110420221.4A CN113327174A (en) | 2021-04-19 | 2021-04-19 | Distribution transformer load prediction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113327174A true CN113327174A (en) | 2021-08-31 |
Family
ID=77414889
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110420221.4A Pending CN113327174A (en) | 2021-04-19 | 2021-04-19 | Distribution transformer load prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113327174A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113776164A (en) * | 2021-09-09 | 2021-12-10 | 广州珠江新城能源有限公司 | Automatic control method and control system for centralized cooling system |
| CN114169568A (en) * | 2021-11-03 | 2022-03-11 | 国网浙江省电力有限公司瑞安市供电公司 | Prophet model-based power distribution line current prediction and heavy overload early warning and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170011297A1 (en) * | 2015-01-06 | 2017-01-12 | Ming Li | Power distribution transformer load prediction analysis system |
| CN107977737A (en) * | 2017-11-19 | 2018-05-01 | 国网浙江省电力公司信息通信分公司 | Distribution transformer load Forecasting Methodology based on mxnet frame depth neutral nets |
| CN108985570A (en) * | 2018-08-17 | 2018-12-11 | 深圳供电局有限公司 | A kind of load forecasting method and its system |
| CN109063975A (en) * | 2018-07-11 | 2018-12-21 | 国网黑龙江省电力有限公司电力科学研究院 | A kind of electric power microclimate disaster monitoring and prior-warning device |
| CN110009136A (en) * | 2019-03-12 | 2019-07-12 | 国网江西省电力有限公司电力科学研究院 | A kind of load forecasting method of distribution transformer and distribution line |
| CN110263995A (en) * | 2019-06-18 | 2019-09-20 | 广西电网有限责任公司电力科学研究院 | Consider the distribution transforming heavy-overload prediction technique of load growth rate and user power utilization characteristic |
-
2021
- 2021-04-19 CN CN202110420221.4A patent/CN113327174A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170011297A1 (en) * | 2015-01-06 | 2017-01-12 | Ming Li | Power distribution transformer load prediction analysis system |
| CN107977737A (en) * | 2017-11-19 | 2018-05-01 | 国网浙江省电力公司信息通信分公司 | Distribution transformer load Forecasting Methodology based on mxnet frame depth neutral nets |
| CN109063975A (en) * | 2018-07-11 | 2018-12-21 | 国网黑龙江省电力有限公司电力科学研究院 | A kind of electric power microclimate disaster monitoring and prior-warning device |
| CN108985570A (en) * | 2018-08-17 | 2018-12-11 | 深圳供电局有限公司 | A kind of load forecasting method and its system |
| CN110009136A (en) * | 2019-03-12 | 2019-07-12 | 国网江西省电力有限公司电力科学研究院 | A kind of load forecasting method of distribution transformer and distribution line |
| CN110263995A (en) * | 2019-06-18 | 2019-09-20 | 广西电网有限责任公司电力科学研究院 | Consider the distribution transforming heavy-overload prediction technique of load growth rate and user power utilization characteristic |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113776164A (en) * | 2021-09-09 | 2021-12-10 | 广州珠江新城能源有限公司 | Automatic control method and control system for centralized cooling system |
| CN113776164B (en) * | 2021-09-09 | 2022-08-30 | 广州珠江新城能源有限公司 | Automatic control method and control system for centralized cooling system |
| CN114169568A (en) * | 2021-11-03 | 2022-03-11 | 国网浙江省电力有限公司瑞安市供电公司 | Prophet model-based power distribution line current prediction and heavy overload early warning and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109828182B (en) | Power grid system fault analysis and early warning method based on fault classification processing | |
| CN105811402B (en) | A kind of Electric Load Prediction System and its Forecasting Methodology | |
| CN109165763B (en) | Method and device for evaluating potential complaints of power grid customer service work order | |
| CN111177101A (en) | Power distribution network multidimensional visualization platform based on big data architecture | |
| CN105868373B (en) | Method and device for processing key data of power business information system | |
| US20030101009A1 (en) | Apparatus and method for determining days of the week with similar utility consumption profiles | |
| JP2009089594A (en) | Temporal-spatial load analysis system of power facility utilizing inspection data and calculation method of load | |
| CN106570784A (en) | Integrated model for voltage monitoring | |
| CN111582700B (en) | Method for calculating fault rate of power distribution network equipment | |
| CN113327174A (en) | Distribution transformer load prediction method | |
| US11774124B2 (en) | Systems and methods for managing building signature intelligent electronic devices | |
| US20190087762A1 (en) | Systems and methods for improving resource utilization | |
| Wang et al. | A load modeling algorithm for distribution system state estimation | |
| CN112702194B (en) | Indoor cell fault positioning method and device and electronic equipment | |
| CN114596693A (en) | Method, system, medium, and program product for energy monitoring and management | |
| CN112883062A (en) | Self-defined rule checking method not based on rule | |
| CN116011827A (en) | Power failure monitoring analysis and early warning system and method for key cells | |
| CN117273337A (en) | Intelligent electric energy meter evaluation method | |
| CN117335411B (en) | Medium-and-long-term power generation capacity prediction method for photovoltaic power station group | |
| CN113993141B (en) | Network optimization method and device | |
| CN106709059B (en) | Terminal online rate index monitoring method and device based on metering automation system | |
| CN112232662A (en) | Service monitoring system and method | |
| CN112184072A (en) | Machine room equipment management method and device | |
| Suciu et al. | Cloud-Based platform for enhancing energy consumption awareness and substantiating the adoption of energy efficiency measures within SMEs | |
| CN115018434A (en) | Remote operation and maintenance management system for new energy power station |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210831 |