CN107274067B - Distribution transformer overload risk assessment method - Google Patents
Distribution transformer overload risk assessment method Download PDFInfo
- Publication number
- CN107274067B CN107274067B CN201710360958.5A CN201710360958A CN107274067B CN 107274067 B CN107274067 B CN 107274067B CN 201710360958 A CN201710360958 A CN 201710360958A CN 107274067 B CN107274067 B CN 107274067B
- Authority
- CN
- China
- Prior art keywords
- overload
- distribution transformer
- transformer
- value
- risk
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000012502 risk assessment Methods 0.000 title claims abstract description 15
- 238000013528 artificial neural network Methods 0.000 claims abstract description 30
- 238000012549 training Methods 0.000 claims abstract description 27
- 238000004088 simulation Methods 0.000 claims abstract description 5
- 238000005516 engineering process Methods 0.000 claims abstract description 4
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 210000002569 neuron Anatomy 0.000 claims description 23
- 239000011159 matrix material Substances 0.000 claims description 19
- 230000006870 function Effects 0.000 claims description 17
- 230000000875 corresponding effect Effects 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 9
- 238000012544 monitoring process Methods 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 230000002596 correlated effect Effects 0.000 claims description 3
- 238000010219 correlation analysis Methods 0.000 claims description 3
- 238000013178 mathematical model Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 2
- 230000009467 reduction Effects 0.000 claims description 2
- 230000001105 regulatory effect Effects 0.000 claims description 2
- 230000009286 beneficial effect Effects 0.000 abstract description 4
- 238000011156 evaluation Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 238000004804 winding 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- 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
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Educational Administration (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The distribution transformer overload risk assessment method comprises the steps of collecting distribution transformer operation data, and preprocessing original data according to overload attributes to obtain a distribution transformer overload attribute set; according to the method, a neural network technology is utilized to conduct network modeling and simulation training on an attribute set when the distribution transformer is overloaded, so that a distribution transformer overload risk network structure is obtained; according to the network structure, evaluating the outage risk degree of the distribution transformer in different operation states during overload, and simulating the overload return probability of the distribution transformer; the greater the probability, the greater the likelihood of transformer outage. The overload risk assessment of the distribution transformer can accurately determine the risk degree of overload outage or damage of the distribution transformer, can provide basis for finding out a distribution transformer with larger overload risk and adopting pertinence when the power load is unexpectedly changed, and further is beneficial to the controllability of overload operation of the transformer, and the accuracy reaches more than 87%.
Description
Technical Field
The invention relates to a distribution transformer overload risk assessment method, and belongs to the technical field of transformer operation.
Background
The overload distribution transformer occupies a certain proportion in the power distribution network, the overload proportion of the distribution transformer in a part of areas reaches 15%, the load of the distribution transformer in different time periods is changed, overload conditions of the distribution transformer easily occur during power utilization peaks, and the unbalance of power utilization loads requires the distribution transformer to have a certain overload capacity so as to meet the requirements of safe and reliable power utilization of users.
At present, less researches are needed for mining overload hidden modes and the like from massive operation data of a distribution transformer. The evaluation of the outage risk degree during overload of the distribution transformer is a premise of fully utilizing the overload capacity of the distribution transformer and ensuring safe and reliable electricity utilization, if the overload of the distribution transformer exits from running, stronger correlation exists between the running state and the running mode of the distribution transformer during overload, so that influence on the running mode of the distribution transformer during different running working conditions can be quantified by mining the correlation between the running state and the running mode of the distribution transformer during overload.
Disclosure of Invention
The invention aims to solve the problem that the overload power failure risk degree of a distribution transformer cannot be determined when an distribution transformer is overloaded.
The technical scheme of the invention is that the overload risk assessment method for the distribution transformer is characterized by collecting operation data of the distribution transformer, and processing the original data according to overload attributes to obtain an overload attribute set of the distribution transformer; modeling and training the attribute set when the distribution transformer is overloaded by utilizing a neural network technology to obtain an overload risk network structure of the distribution transformer; according to the network structure, evaluating the outage risk degree of the distribution transformer in different operation states during overload, and simulating the overload return probability of the distribution transformer; the greater the probability, the greater the likelihood of transformer outage.
The distribution transformer overload risk network structure is an artificial neural network, the artificial neuron is a mathematical model for simulating a biological neuron, and the output quantity of the neuron is a function of the weighted input quantity, and the following formula is as follows:
wherein y is the neuron output; x is x i Is the input of neurons; w (w) i The corresponding weight is input to the neuron;
the BP neural network is a multi-layer perceptron structure and comprises an input layer, an output layer and a plurality of hidden layers, wherein the multi-layer perceptron structure is divided into a forward propagation stage and a backward propagation stage, and in the forward propagation stage, information is transmitted to the output layer from the input layer through the hidden layers in a step-by-step transformation way, and the following formula is shown:
O i ＝f n (f n-1 ((...f 1 ([X i ][W 1 ])+[B 1 ]...)[W n-1 ])+[B n-1 ])
X i represents the ith sample input, W 1 …W n-1 Weight matrix representing hidden layer, B 1 …B n-1 A bias matrix representing hidden layers, f 1 …f n-1 Weight function representing hidden layer, f n Representing the output function, 0 i Is the expected value;
in the backward propagation stage, the weight and deviation of the network are repeatedly regulated and trained, the training process needs to provide an input vector X and a target value Y, the network training process is a process of minimizing the mean square error, and the error precision E of the ith sample is set i The following formula is shown:
the error E for the entire m sample sets is shown as follows:
the simulation training sets a network training error, such as 10 -2 Determining the number of neurons of an input layer, an implicit layer and an output layer through multiple tests, wherein the input layer is an overload attribute vector, the output layer is whether the corresponding allocation is out of operation or not, if so, the number is set to be 1, and if not, the number is set to be 0; the training algorithm adopts BP algorithm, and activates the function; the training samples randomly extract 120 overload shutdown samples and 200 overload non-shutdown samples, continuously change weights and deviations along the direction of the fastest reduction of the error function, and stop until the training error is smaller than a specified value, thereby obtaining a matrix containing trained weights and a bias matrixAnd further, a network structure which can be used for predicting the overload risk of the distribution transformer is obtained.
The collection of the distribution transformer operation data comprises information contained in a distribution transformer monitoring related system, wherein the information is recorded in a platform region operation record: capacity, voltage value, current value, acquisition time, active power, reactive power; and weather conditions when the distribution transformer is overloaded are acquired through a meteorological monitoring system, so that distribution transformer operation original data are formed.
Preprocessing original data according to overload attributes, wherein discrete operation data of a distribution transformer cannot be directly input as a neural network, the operation data needs to be processed according to the overload attributes of the distribution transformer, and noise data and empty data in the original data need to be removed;
1) An overload shutdown criterion of the distribution transformer, wherein if two continuous load coefficients of the distribution transformer are smaller than 0.01 in overload, the distribution transformer is considered to be shutdown;
2) The factor K1 before overload, the average value of n load factor K acquisition points before overload is taken as the factor K1 value before overload;
3) The overload time coefficient K2 is obtained by taking the average value of n load coefficient K acquisition points in overload as the overload time coefficient K2 value;
4) The overload coefficient K3 is obtained, and the average value of n load coefficient K acquisition points after overload is taken as the overload coefficient K3 value;
5) The overload time length T is the time interval from the first acquisition point to the last acquisition point when overload occurs;
6) The current unbalance rate beta 1 during overload is taken as the unbalance rate beta 1 value during overload, wherein the average value of the three-phase unbalance rates during overload is taken as the unbalance rate beta 1 value during overload;
by determination of the overload attribute, all the overload attribute data are represented in the form of a matrix, as follows:
wherein D is an overload attribute matrix; Λ type m Is a certain attribute vector; a, a nm Representing an overload property unit.
The selection of the samples can directly influence the overload risk assessment prediction result of the distribution transformer, some singular samples in the samples should be removed, including the capacity of 5kVA, the time immediately after the operation and the special transformer, and the sample selection principle is as follows:
1) Selecting an S9 model and an upper transformer;
2) Selecting a transformer with a capacity of more than or equal to 30kVA and less than or equal to 800 kVA;
3) Only a conventional transformer is considered, and a high overload distribution transformer and an on-load capacity-regulating transformer are not considered;
4) Distribution transformer operation data for two years or more;
5) Samples are selected according to the ratio of the number of the overload distribution transformer with different capacities.
The overload attribute is determined by carrying out correlation calculation, and whether the input requirement of the neural network is met is verified; when two or more independent variables injected into the neural network are highly correlated, the learning ability of the neural network will be negatively impacted, removing the redundant variables will result in faster training time, and an adaptive neural network can be used to streamline redundant connections and neurons; the correlation analysis of the association degree of 2 overload attributes is as follows:
correlation coefficient matrix:
wherein r is ij Representing overload attribute vector X i And X is j The calculation formula of the correlation coefficient is as follows:
the degree of correlation is defined as shown in table 1.
TABLE 1 definition of degree of correlation
| Correlation coefficient | Degree of correlation |
| 0.00～±0.3 | Microcorrelation |
| ±0.3～±0.5 | Low degree of correlation |
| ±0.5～±0.8 | Moderate correlation |
| ±0.8～±1.0 | Significant correlation |
According to the overload attribute vector r ij And carrying out correlation calculation, and verifying whether the input requirements of the neural network are met according to the table 1.
The invention has the beneficial effects that the overload risk assessment of the distribution transformer can accurately determine the risk degree of overload outage or damage of the distribution transformer, can provide basis for finding out the distribution transformer with larger overload risk and taking pertinence when the power load is unexpectedly changed, and further is beneficial to the controllability of the overload operation of the transformer, and the accuracy reaches more than 87%.
Compared with a winding hot spot temperature thermal model, the distribution transformer overload risk assessment method is characterized in that a general hidden model of a transformer group is obtained through a large amount of overload operation data mining, rather than a thermal model formed by overload temperature characteristics of the transformer, and the distribution transformer overload risk assessment method has strong practicability.
When the neural network is used for evaluating the overload outage risk of the distribution transformer, the internal parameters of the distribution transformer do not need to be considered; the traditional method does not consider that the distribution transformer has a certain overload capacity, corresponding treatment measures are adopted once the distribution transformer is overloaded, and the evaluation method achieves the aim of managing and controlling the overload operation risk of the transformer under the condition of ensuring the overload operation of the transformer.
According to the distribution transformer overload risk assessment method, historical operation discrete data are extracted from the operation angle of the distribution transformer, correlation among all attributes in overload is analyzed, the overload operation data are analyzed and mined by utilizing a neural network technology, the influence degree of different operation states on an operation mode when the distribution transformer is overloaded is assessed, the assessment result is beneficial to top-level decision and actual production, and effective management and control on overload operation of the distribution transformer are realized when the overload risk degree of the distribution transformer is accurately assessed.
Drawings
FIG. 1 is a flow chart of an overload risk assessment model according to the present invention.
Detailed Description
Detailed description of the invention as shown in fig. 1, the specific steps of the invention are as follows:
(1) Data acquisition
In the distribution transformer monitoring related system, the operation record of the transformer area contains rich information such as capacity, voltage value, current value, acquisition time, active power, reactive power and the like, and weather conditions when the distribution transformer is overloaded are obtained through the meteorological monitoring system, so that the distribution transformer operation original data is formed.
(2) Data preprocessing
The discrete operation data of the distribution transformer cannot be directly input as a neural network, the operation data needs to be processed according to overload attributes of the distribution transformer, and noise data and empty data in the original data need to be removed.
1) And (5) an overload shutdown criterion of the distribution transformer. A distribution transformer is considered to be out of service if there are two consecutive load factors of less than 0.01 in overload.
2) The pre-overload coefficient K1. Taking the average value of n load factor K acquisition points before overload as the value of the coefficient K1 before overload.
3) And a factor K2 at overload. And taking the average value of n load factor K acquisition points in overload as the value of the overload factor K2.
4) And an overload post-coefficient K3. And taking the average value of n load factor K acquisition points after overload as the overload time factor K3 value.
5) Overload duration T. The time interval from the first acquisition point at overload to the last acquisition point at overload.
6) Current imbalance ratio β1 at overload. And taking the average value of the three-phase unbalance rate at the overload as the value of the unbalance rate beta 1 at the overload.
By determining overload attributes, all overload attribute data can be represented in matrix form, e.g. common
Formula 1:
(3) Sample selection
The selection of samples can directly influence the overload risk assessment prediction result of the distribution transformer, and some singular samples in the samples should be removed, such as the capacity of 5kVA, the time of operation, a special transformer and the like, and the sample selection principle is as follows:
1) And selecting the S9 model and the upper transformer.
2) A transformer having a capacity greater than or equal to 30kVA and less than or equal to 800kVA is selected.
3) Only the conventional transformer is considered, and the high overload distribution transformer and the on-load capacity-regulating transformer are not considered.
4) Distribution transformers operate for two years or more.
5) Samples are selected according to the ratio of the number of the overload distribution transformer with different capacities.
(4) Correlation verification
And carrying out correlation calculation on the overload attribute vector, and verifying whether the input requirement of the neural network is met.
By correlation analysis of the degree of correlation of the 2 overload attributes, when two or more independent variables injected into the neural network are highly correlated, the learning ability of the neural network will be negatively affected, removing the redundant variables will result in faster training time, and an adaptive neural network can be used to streamline redundant connections and neurons.
1) Correlation coefficient matrix
Wherein r is ij Representing overload attribute vector X i And X is j The calculation formula of the correlation coefficient is as follows:
2) Calculation formula
3) The degree of correlation is defined as shown in table 1.
TABLE 1 definition of degree of correlation
And carrying out correlation calculation on the overload attribute vector, and verifying whether the overload attribute vector meets the input requirement of the neural network according to a table 1.
(5) Network modeling
The overload risk network structure of the distribution transformer is an artificial neural network, the artificial neuron is a mathematical model for simulating biological neurons, the distribution transformer is a multi-input single-output nonlinear element, and each input quantity x of the neurons i All have a corresponding weight w i . The processing unit quantizes the weighted inputs, adds the weighted sums, sums the weighted sums with the deviation to form the input of the neuron transfer function,
the output of neurons is a function of weighted input as shown in equation 4:
the BP neural network is a multi-layer perceptron structure and comprises an input layer, an output layer and a plurality of hidden layers, and is mainly divided into a forward propagation stage and a backward propagation stage, wherein in the forward propagation stage, information is transmitted to the output layer from the input layer through hidden layer step-by-step transformation, as shown in a formula (5):
O i ＝f n (f n-1 ((...f 1 ([X i ][W 1 ])+[B 1 ]...)[W n-1 ])+[B n-1 ]) (5)
X i represents the ith sample input, W 1 …W n-1 Weight matrix representing hidden layer, B 1 …B n-1 A bias matrix representing hidden layers, f 1 …f n-1 Weight function representing hidden layer, f n Representing the output function, O i Is the expected value.
In the backward propagation stage, the weight and deviation of the network are repeatedly adjusted and trained, the training process needs to provide an input vector X and a target value Y, the network training process is a process of minimizing the mean square error, and the error precision of the ith sample is set as shown in the formula (6):
the error of the entire m sample sets is shown as equation (7):
(6) Simulation training
Setting network training errors, e.g. 10 -2 The number of neurons of an input layer, an implied layer and an output layer is determined through multiple tests, wherein the input layer is an overload attribute vector, the output layer is whether the corresponding allocation is stopped or not, if the corresponding allocation is stopped, the corresponding allocation is set to be 1, if the corresponding allocation is not stopped, the corresponding allocation is set to be 0, the corresponding allocation is trained by adopting a BP algorithm, and an activation function is used by adopting a tan sig function. Training samples randomly draw 120 overloaded shutdown samples and 200 overloaded non-shutdown samples, along the direction in which the error function decreases fastestAnd continuously changing the weight and the deviation until the training error is smaller than a specified value, and stopping to obtain a trained weight matrix and a trained bias matrix, thereby obtaining a network structure capable of being used for predicting the overload risk of the distribution transformer.
(7) Prediction result
The test samples are taken as 30 overload shutdown samples and 50 overload non-shutdown samples, corresponding neural network weights and deviations are obtained through training, the test samples are tested by using a trained network, an output result is more than 0.4 and is a shutdown area, otherwise, the test samples are non-shutdown areas, the trained neural network structure is used for the actual distribution transformer operation, and the test results are shown in the table 2:
table 2 test results
Claims (4)
1. The method is characterized in that the method utilizes a neural network technology to perform network modeling and simulation training on the distribution transformer overload attribute set to obtain a distribution transformer overload risk network structure; according to the network structure, evaluating the outage risk degree of the distribution transformer in different operation states during overload, and simulating the overload return probability of the distribution transformer; the larger the probability is, the greater the possibility that the transformer is out of operation is;
the operation data comprises information contained in the platform region operation record in the distribution transformer monitoring related system: capacity, voltage value, current value, acquisition time, active power, reactive power; acquiring weather conditions when the distribution transformer is overloaded through a weather monitoring system;
the distribution transformer overload risk network structure is an artificial neural network, the artificial neuron is a mathematical model for simulating a biological neuron, and the output quantity of the neuron is a function of the weighted input quantity, and the following formula is as follows:
wherein y is the neuron output; x is x i Is the input of neurons; w (w) i The corresponding weight is input to the neuron;
the BP neural network is a multi-layer perceptron structure and comprises an input layer, an output layer and a plurality of hidden layers, wherein the multi-layer perceptron structure is divided into a forward propagation stage and a backward propagation stage, and in the forward propagation stage, information is transmitted to the output layer from the input layer through the hidden layers in a step-by-step transformation way, and the following formula is shown:
O i ＝f n (f n-1 ((...f 1 ([X i ][W 1 ])+[B 1 ]...)[W n-1 ])+[B n-1 ])
X i represents the ith sample input, W 1 …W n-1 Weight matrix representing hidden layer, B 1 …B n-1 A bias matrix representing hidden layers, f 1 …f n-1 Weight function representing hidden layer, f n Representing the output function, 0 i Is the expected value;
in the backward propagation stage, the weight and deviation of the network are repeatedly regulated and trained, the training process needs to provide an input vector X and a target value Y, the network training process is a process of minimizing the mean square error, and the error precision E of the ith sample is set i The following formula is shown:
the error E for the entire m sample sets is shown as follows:
preprocessing original data according to overload attributes, wherein discrete operation data of a distribution transformer cannot be directly input as a neural network, the operation data needs to be processed according to the overload attributes of the distribution transformer, and noise data and empty data in the original data need to be removed;
1) An overload shutdown criterion of the distribution transformer, wherein if two continuous load coefficients of the distribution transformer are smaller than 0.01 in overload, the distribution transformer is considered to be shutdown;
2) The factor K1 before overload, the average value of n load factor K acquisition points before overload is taken as the factor K1 value before overload;
3) The overload time coefficient K2 is obtained by taking the average value of n load coefficient K acquisition points in overload as the overload time coefficient K2 value;
4) The overload coefficient K3 is obtained, and the average value of n load coefficient K acquisition points after overload is taken as the overload coefficient K3 value;
5) The overload time length T is the time interval from the first acquisition point to the last acquisition point when overload occurs;
6) The current unbalance rate beta 1 during overload is taken as the unbalance rate beta 1 value during overload, wherein the average value of the three-phase unbalance rates during overload is taken as the unbalance rate beta 1 value during overload;
by determination of the overload attribute, all the overload attribute data are represented in the form of a matrix, as follows:
wherein D is an overload attribute matrix; Λ type m Is a certain attribute vector; a, a nm An overload attribute unit;
the overload attribute is determined by carrying out correlation calculation, and whether the input requirement of the neural network is met is verified; when two or more independent variables injected into the neural network are highly correlated, the learning ability of the neural network will be negatively impacted, removing the redundant variables will result in faster training time, and an adaptive neural network is used to reduce redundant connections and neurons; the correlation analysis of the association degree of 2 overload attributes is as follows:
correlation coefficient matrix:
wherein r is ij Representing overload attribute vector X i And X is j The calculation formula of the correlation coefficient is as follows:
2. the method for evaluating overload risk of distribution transformer according to claim 1, wherein the simulation training is performed, a network training error is set, and the number of neurons of an input layer, an hidden layer and an output layer is determined through multiple tests, wherein the input layer is an overload attribute vector, the output layer is whether the corresponding distribution transformer is out of operation or not, if the distribution transformer is out of operation, the distribution transformer is set to 1, and if the distribution transformer is not out of operation, the distribution transformer is set to 0; the training algorithm adopts BP algorithm, and activates the function; the training samples randomly extract 120 overload shutdown samples and 200 overload non-shutdown samples, continuously change weights and deviations along the direction of the fastest reduction of the error function, and stop until the training error is smaller than a specified value, so that a trained weight matrix and a trained bias matrix are obtained, and a network structure for predicting overload risk of the distribution transformer is further obtained.
3. The method for evaluating overload risk of distribution transformer according to claim 1, wherein the collecting the operation data of the distribution transformer includes information contained in the operation record of the transformer area in the distribution transformer monitoring related system: capacity, voltage value, current value, acquisition time, active power, reactive power; and weather conditions when the distribution transformer is overloaded are acquired through a meteorological monitoring system, so that distribution transformer operation original data are formed.
4. A distribution transformer overload risk assessment method according to claim 1, wherein the selection of the samples directly affects the prediction result of the distribution transformer overload risk assessment, and some singular samples in the samples should be removed, including a capacity of 5kVA, a time after operation, and a special transformer, and the sample selection principle is as follows:
1) Selecting an S9 model and an upper transformer;
2) Selecting a transformer with a capacity of greater than or equal to 30kVA and less than or equal to 800 kVA;
3) Only a conventional transformer is considered, and a high overload distribution transformer and an on-load capacity-regulating transformer are not considered;
4) Distribution transformer operation data for two years or more;
5) Samples are selected according to the ratio of the number of the overload distribution transformer with different capacities.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710360958.5A CN107274067B (en) | 2017-05-22 | 2017-05-22 | Distribution transformer overload risk assessment method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710360958.5A CN107274067B (en) | 2017-05-22 | 2017-05-22 | Distribution transformer overload risk assessment method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107274067A CN107274067A (en) | 2017-10-20 |
| CN107274067B true CN107274067B (en) | 2023-05-23 |
Family
ID=60064089
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710360958.5A Active CN107274067B (en) | 2017-05-22 | 2017-05-22 | Distribution transformer overload risk assessment method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107274067B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110097123B (en) * | 2019-05-05 | 2021-05-14 | 北京印刷学院 | Express mail logistics process state detection multi-classification system |
| CN110457770B (en) * | 2019-07-18 | 2022-07-01 | 中国电力科学研究院有限公司 | Time scale-oriented overload judgment method for distribution transformer |
| CN110261749A (en) * | 2019-07-24 | 2019-09-20 | 广东电网有限责任公司 | A kind of GIS partial discharge fault identification model building method, device and fault recognition method |
| CN112434839B (en) * | 2019-08-26 | 2023-05-30 | 电力规划总院有限公司 | Distribution transformer heavy overload risk prediction method and electronic equipment |
| CN111984638A (en) * | 2020-07-10 | 2020-11-24 | 广州汇通国信科技有限公司 | Distribution transformer heavy overload data management method based on BP neural network algorithm |
| CN112434929B (en) * | 2020-11-19 | 2023-04-18 | 广西电网有限责任公司电力科学研究院 | Method for assessing risk resistance of distribution transformer |
| CN113567783A (en) * | 2021-07-20 | 2021-10-29 | 云南电网有限责任公司 | Real-time simulation system for judging element overload |
| CN115588993B (en) * | 2022-09-19 | 2023-07-25 | 广东电网有限责任公司佛山供电局 | On-load real-time online capacity adjustment method and device for distribution transformer |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103914791A (en) * | 2014-03-17 | 2014-07-09 | 中国南方电网有限责任公司超高压输电公司检修试验中心 | Electrical equipment state maintenance system |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20140046358A (en) * | 2012-10-10 | 2014-04-18 | 이명진 | Landslide hazard analysis system considering climate change and method thereof |
| CN103150635B (en) * | 2013-03-29 | 2016-01-06 | 广东电网公司 | Power equipment O&M method |
| CA2827701A1 (en) * | 2013-09-23 | 2015-03-23 | Sureshchandra B. Patel | Methods of patel decoupled loadlow computation for electrical power system |
| JP5980195B2 (en) * | 2013-12-20 | 2016-08-31 | 三菱電機株式会社 | Distribution system load prediction device and distribution system load prediction method |
| CN104217104A (en) * | 2014-08-19 | 2014-12-17 | 上海交通大学 | Power transformer service life analysis method and system based on risk evaluation |
| CN104599189B (en) * | 2014-12-25 | 2019-02-22 | 中国电力科学研究院 | A kind of power network planning scheme methods of risk assessment being included in power system operation mode |
| KR102107709B1 (en) * | 2015-06-05 | 2020-05-07 | 구글 엘엘씨 | Spatial transformer modules |
| CN105404780B (en) * | 2015-11-25 | 2019-03-26 | 国网山东省电力公司 | A kind of transformer overload capability assessment method of multiparameter exponential family |
| CN105488740A (en) * | 2015-12-17 | 2016-04-13 | 深圳供电局有限公司 | Risk level evaluation method and system for urban power distribution network operation mode |
| CN106156485A (en) * | 2016-06-16 | 2016-11-23 | 广州供电局有限公司 | Method for diagnosing fault of power transformer and device |
| CN106483947A (en) * | 2016-09-21 | 2017-03-08 | 国网江苏省电力公司南通供电公司 | Distribution Running State assessment based on big data and method for early warning |
-
2017
- 2017-05-22 CN CN201710360958.5A patent/CN107274067B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103914791A (en) * | 2014-03-17 | 2014-07-09 | 中国南方电网有限责任公司超高压输电公司检修试验中心 | Electrical equipment state maintenance system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107274067A (en) | 2017-10-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107274067B (en) | Distribution transformer overload risk assessment method | |
| CN105117602B (en) | A kind of metering device running status method for early warning | |
| CN104809658B (en) | A kind of rapid analysis method of low-voltage distribution network taiwan area line loss | |
| CN111525587B (en) | Reactive load situation-based power grid reactive voltage control method and system | |
| CN105719048A (en) | Intermediate-voltage distribution operation state fuzzy integrated evaluation method based on principle component analysis method and entropy weight method | |
| CN103105246A (en) | Greenhouse environment forecasting feedback method of back propagation (BP) neural network based on improvement of genetic algorithm | |
| CN105956779A (en) | Power transformer running state evaluation method and power transformer running state evaluation device | |
| CN110222897A (en) | A kind of distribution network reliability analysis method | |
| CN112818604A (en) | Wind turbine generator risk degree assessment method based on wind power prediction | |
| CN112308427A (en) | New energy consumption restriction factor evaluation method and system based on combined empowerment-grey correlation | |
| CN112990627B (en) | Power quality evaluation method | |
| CN111797365B (en) | Converter transformer temperature abnormity judgment method and system | |
| CN110705859A (en) | PCA-self-organizing neural network-based method for evaluating running state of medium and low voltage distribution network | |
| CN108876114A (en) | Consider the power grid efficiency benefit appraisal procedure of new energy access | |
| CN107169655A (en) | A kind of method and device of preferred power distribution network project compatibility | |
| CN115689114A (en) | Submarine cable running state prediction method based on combined neural network | |
| CN110837915A (en) | Low-voltage load point prediction and probability prediction method for power system based on hybrid integrated deep learning | |
| CN111932081B (en) | Method and system for evaluating running state of power information system | |
| CN106845711A (en) | Processing method and processing device for power supply reliability data | |
| CN113379116A (en) | Cluster and convolutional neural network-based line loss prediction method for transformer area | |
| CN110045237A (en) | Transformer state parametric data prediction technique and system based on drosophila algorithm optimization | |
| CN115358437A (en) | Power supply load prediction method based on convolutional neural network | |
| CN113065715A (en) | Multi-load ultra-short-term prediction method for comprehensive energy system | |
| CN116415732A (en) | User side power load data processing method based on improved ARNN | |
| CN108694479A (en) | Consider the distribution network reliability prediction technique that weather influences time between overhaul |
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 |