CN113609768A - Bidirectional LSTM network-based distribution room line loss rate calculation method - Google Patents
Bidirectional LSTM network-based distribution room line loss rate calculation method Download PDFInfo
- Publication number
- CN113609768A CN113609768A CN202110878731.6A CN202110878731A CN113609768A CN 113609768 A CN113609768 A CN 113609768A CN 202110878731 A CN202110878731 A CN 202110878731A CN 113609768 A CN113609768 A CN 113609768A
- Authority
- CN
- China
- Prior art keywords
- line loss
- loss rate
- area
- layer
- distribution
- 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
- 238000004364 calculation method Methods 0.000 title claims abstract description 40
- 230000002457 bidirectional effect Effects 0.000 title claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 22
- 239000010410 layer Substances 0.000 claims description 40
- 238000012549 training Methods 0.000 claims description 14
- 230000005611 electricity Effects 0.000 claims description 4
- 238000010248 power generation Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 239000002356 single layer Substances 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 210000005036 nerve Anatomy 0.000 claims description 2
- 238000004422 calculation algorithm Methods 0.000 abstract description 7
- 238000002790 cross-validation Methods 0.000 abstract description 2
- 238000012417 linear regression Methods 0.000 abstract description 2
- 238000012706 support-vector machine Methods 0.000 abstract description 2
- 210000004027 cell Anatomy 0.000 description 14
- 230000015654 memory Effects 0.000 description 14
- 230000006870 function Effects 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention belongs to the technical field of power distribution and utilization, and solves the problems of large calculation amount and low efficiency of the traditional algorithm by combining an intelligent algorithm and theoretical line loss based on the line loss rate calculation of a distribution area of a bidirectional LSTM network. And comparing and analyzing the prediction results of the linear regression model, the support vector machine regression model, the regression tree model and the bidirectional LSTM network, and obtaining a conclusion that the calculation result of the method is optimal by adopting cross validation of ten folds. And the bidirectional LSTM network used by the algorithm makes up the problem that the LSTM network lacks the following semantic information. Therefore, the method and the device can improve the accuracy of the theoretical line loss of the distribution network area and have great significance for the fine management of the line loss of the distribution network area.
Description
Technical Field
The invention belongs to the technical field of power distribution and utilization, and particularly relates to a method for calculating a line loss rate of a transformer area based on a bidirectional LSTM network.
Background
Nowadays, power grid companies achieve fine management on power distribution networks by taking quality improvement, efficiency enhancement, energy conservation and loss reduction as targets, and theoretical line loss is an important part for reflecting the loss condition of the power distribution networks.
The method for calculating the theoretical line loss is not uncommon, and examples thereof include a common root mean square current method, an average current method, a power flow method, a forward-backward substitution method, and the like. However, as the informatization level of a power grid is continuously improved, the power distribution network line is gradually complicated, the traditional calculation method is insufficient in mining strength of existing data such as a production management system, a marketing service system and a power utilization information acquisition system, so that a large amount of data resources are idle and are not efficiently utilized, and meanwhile, the traditional calculation method starts to have the problems of large calculation amount, low calculation precision and the like in the aspect of theoretical line loss calculation of the power distribution network.
Disclosure of Invention
The invention adopts the combination of an intelligent algorithm and theoretical line loss, designs the line loss rate calculation method of the distribution room based on the bidirectional LSTM network, improves the calculation precision of the theoretical line loss of the distribution room, reduces the calculation time and enhances the stability.
The technical scheme adopted by the invention is that,
the method for calculating the line loss rate of the distribution room based on the bidirectional LSTM network comprises the following steps:
step 1: screening out a platform area training sample set with stable line loss rate;
step 2: collecting and arranging basic information of distribution network distribution areas, and selecting operation state parameters of the distribution areas;
and step 3: establishing a platform area line loss rate calculation model based on the bidirectional LSTM network model respectively for each type of platform area;
and 4, step 4: training a platform area line loss rate calculation model;
and 5: and selecting test data to input into the transformer area line loss rate calculation model to obtain a transformer area theoretical line loss calculation result.
Further, the step 1 includes, when the sample data of the transformer area is selected, removing the transformer area samples which are not fully covered, have photovoltaic power generation users under the transformer area, adjust the user variation relationship, and have a negative line loss rate or exceed 10%.
Further, the operating state parameters of the distribution room comprise power selling amount, load rate, three-phase unbalance, power factor and ambient temperature.
Further, the step 3 comprises the steps of,
step 301: dividing a hidden layer nerve unit into a positive time direction and a negative time direction which are independent;
step 302: feeding forward two independent hidden layers to the same output layer, including both past and future sequence information;
step 303: the layer 1 LSTM calculates the current time point sequence information, and the layer 2 LSTM reads the same sequence in the reverse direction;
step 304: adding the reverse order information, each layer of LSTM has different parameters.
Further, the input attributes of the line loss rate calculation model comprise the operation age of the transformer area, the occupancy ratio of the residential capacity, the average capacity of the residential households, the number of the transformer area users, the daily electricity sales, the daily load rate, the daily average three-phase unbalance degree, the daily average power factor and the daily average air temperature.
Further, the step 4 includes the steps of,
step 401: searching the optimal hidden unit number of the single-layer model and fixing the optimal hidden unit number;
step 402: adding a hidden layer, and searching the optimal number of hidden units of the layer on the basis of the step 401;
step 403: counting the mean square error of each training sample input to the platform area line loss rate calculation model;
step 404: and setting the model parameter with the minimum root mean square error as the final parameter of the line loss rate calculation model of the transformer area.
The working principle and the beneficial effects of the invention are as follows:
the method is based on the line loss rate calculation of the two-way LSTM network, and the problem that the traditional algorithm is large in calculation amount and low in efficiency is solved by combining an intelligent algorithm and theoretical line loss. And comparing and analyzing the prediction results of the linear regression model, the support vector machine regression model, the regression tree model and the bidirectional LSTM network, and obtaining a conclusion that the calculation result of the method is optimal by adopting cross validation of ten folds. And the bidirectional LSTM network used by the algorithm makes up the problem that the LSTM network lacks the following semantic information. Therefore, the method and the device can improve the accuracy of the theoretical line loss of the distribution network area and have great significance for the fine management of the line loss of the distribution network area.
The present invention will be described in detail with reference to the accompanying drawings.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The technical solutions of the present invention are further described in detail below with reference to specific examples and drawings, but the scope and implementation of the present invention are not limited thereto.
In the specific embodiment, as shown in fig. 1, the present invention is a method for calculating the line loss rate of a distribution area based on a bidirectional LSTM network,
(1) selecting sample data of the transformer area, and removing the following unqualified transformer areas in a key way: collecting incomplete coverage; special users are arranged under the platform area, such as photovoltaic power generation; service change occurs, such as user change relationship adjustment; the line loss rate is negative or exceeds 10%. Screening out a platform area training sample set with stable line loss rate;
(2) after a platform area training sample set is determined, selecting platform area running state parameters which mainly comprise electricity sales amount, load rate, three-phase unbalance, power factor and environment temperature;
(3) and (3) respectively constructing a platform area line loss rate calculation model for each type of platform area based on a bidirectional LSTM network model, wherein the input attributes of the model comprise the operation age of the platform area, the occupation capacity ratio, the average capacity of residents, the number of platform area users, the daily electricity sales, the daily load rate, the daily average three-phase unbalance degree, the daily average power factor and the daily average air temperature.
The step of building a bidirectional LSTM network computational model is to first divide the hidden layer neurons into 2 parts in the positive and negative temporal directions, with 2 independent hidden layers, and then feed forward to the same output layer, including both past and future sequence information. The layer 1 LSTM calculates the current time point sequence information, the layer 2 LSTM reads the same sequence in reverse direction, adds the reverse sequence information, and each layer of LSTM has different parameters.
In a bidirectional LSTM network, by implying layer states forward
And backward hidden layer states
To obtain an output pt. In the direction of T-1 to T,
is the hidden layer of the current neuron's input x to this moment
The weight of (a) is determined,
the weight of the state quantity to the current state quantity at the last moment,
to imply the layer state output value at the previous time,
as an offset term, it can be expressed as:
the method is consistent with the positive sequence calculation method in the direction of T ═ T to 1:
the bidirectional LSTM iterates in two directions simultaneously, and the state of the hidden layer is weighted and calculated to obtain a predicted value ptExpressed as:
the hidden layer of the bi-directional LSTM eventually computes and saves two weights, including the forward output vector
And a reverse output vector
The calculation process is similar to the LSTM, and the specific calculation process is as follows:
the bidirectional LSTM is an extension of the traditional LSTM and can improve the model performance of the sequence classification problem. In the problem that all time steps of the input sequence are available, the bi-directional LSTM trains two instead of one LSTM on the input sequence. The first of the input sequence is the original and the second is the inverted copy of the input sequence. This may provide additional context for the network to learn the problem faster and more fully. Based on the prediction of the time series, both past and future series information near the current point in time can be used to estimate the current time instant, and does not rely on predefined parameters. However, the conventional LSTM neural network can only predict the result using input information before a certain time, and certain information is ignored. Bi-directional LSTM predicts the output based on the entire time series.
The bidirectional LSTM network makes up the defect that the traditional LSTM lacks the following semantic information, can utilize the past information and the future information simultaneously, can better learn the sequence information of the source code, and has smaller prediction error compared with the traditional LSTM structure. It is very important to observe the whole period when training the neural network using the time series, and the bidirectional structure has more advantages than the conventional structure.
The specific steps for constructing the LSTM network model are as follows:
input gate, forgetting gate and output gate
The input of the gate with long and short term memory is the current time step input XtHidden state H from previous time stept-1The output by value range is [0,1 ]]And calculating the full connection layer of the sigmoid activation function. Specifically, assuming the number of hidden units is h, a small batch of inputs X at a given time step t is enteredt∈Rn×d(n samples, d inputs) and hidden state H of last time stept-1∈Rn×h. Input gate I at time step tt∈Rn×hForgetting door Ft∈Rn×hAnd an output gate Ot∈Rn×hThe following are calculated respectively:
It＝σ(XtWxi+Ht-1Whi+bi)
Ft＝σ(XtWxf+Ht-1Whf+bf)
Ft＝σ(XtWxo+Ht-1Who+bo)
wherein, Wxi，Wxf，Wxo∈Rd×h，Whi，Whf，Who∈Rh×hIs a weight parameter, bi，bf，boIs a deviation parameter.
② candidate memory cell, memory cell
Candidate memory cells at time step t
Is calculated as:
wherein, Wxc∈Rd×hAnd Whc∈Rh×hIs a weight parameter, bc∈Rl×hIs a bias weight parameter, and the tanh function is a range of values [ -1,1 []The activation function of (2). The calculation of the memory cell of the current time step combines the information of the memory cell of the previous time step and the candidate memory cell of the current time step, and controls the flow of the information through a forgetting gate and an input gate:
the forgetting gate controls whether information in memory cells of a previous time step can flow into a current time step, and the input gate controls how input of the current time step flows into memory cells of the current time step through the candidate memory cells. If the forgetting gate is always approximately 1 and the input gate is always approximately 0, the past memory cells will always be saved by time and passed to the current time step.
State of hiding
On the basis of the memory cell, the output gate can control the memory cell to the hidden state Ht∈Rn×hThe information flow of (2):
Ht＝Ot⊙tanh(Ct)
wherein the tanh function ensures that the hidden state element is between-1 and 1. It should be noted that when the output gate is approximately 1, the memory cell information will be passed to the hidden state for use by the output layer; when the output gate is approximately 0, the memory cell information is self-retained.
Based on the prediction of the time series, both past and future series information near the current point in time can be used to estimate the current time instant, and does not rely on predefined parameters. However, the LSTM neural network can only predict the outcome using input information up to a certain time. Bidirectional LSTM predicts the output based on the entire time series, so a bidirectional LSTM model needs to be built.
(4) Training a platform area line loss rate calculation model;
in the training stage of the line loss rate calculation model of the transformer area, the most important parameters are the number of layers of the bidirectional LSTM network and the number of hidden units in each layer. The more the number of layers and the more the number of hidden units, the stronger the nonlinear fitting capability of the model is, but the complexity of the model is also greatly increased. The specific steps for determining the layer number of the LSTM model and the number of the hidden units are as follows:
firstly, searching the optimal hidden unit quantity of the single-layer model and fixing the optimal hidden unit quantity;
secondly, adding a layer of hidden layer, and searching the optimal number of hidden units of the layer on the basis of the step 401;
by analogy, the training data is substituted into the double-layer LSTM model to carry out model training, and the number of layers and the number of hidden units corresponding to the minimum mean square error obtained by carrying out model training are the final parameters of the LSTM model.
And analyzing the mean square errors of the models with different learning parameters to determine that the model has the best effect when the number of layers of the models is a certain value, and taking the bidirectional LSTM model under the parameters as a final station area line loss rate calculation model.
(5) And selecting test data to input into the transformer area line loss rate calculation model to obtain a transformer area theoretical line loss calculation result. And comparing the result with the real line loss rate, and providing reference for service personnel.
Claims (6)
1. The method for calculating the line loss rate of the distribution room based on the bidirectional LSTM network is characterized by comprising the following steps of:
step 1: screening out a platform area training sample set with stable line loss rate;
step 2: collecting and arranging basic information of distribution network distribution areas, and selecting operation state parameters of the distribution areas;
and step 3: establishing a platform area line loss rate calculation model based on the bidirectional LSTM network model respectively for each type of platform area;
and 4, step 4: training a platform area line loss rate calculation model;
and 5: and selecting test data to input into the transformer area line loss rate calculation model to obtain a transformer area theoretical line loss calculation result.
2. The method according to claim 1, wherein the step 1 includes, when selecting the sample data of the station area, removing the samples of the station area which are not fully covered, have photovoltaic power generation users under the station area, have a relationship between user variables adjusted, and have a negative line loss rate or exceed 10%.
3. The method of calculating the line loss rate of the distribution room based on the bidirectional LSTM network of claim 1, wherein the distribution room operation status parameters comprise power selling amount, load rate, three-phase unbalance degree, power factor and ambient temperature.
4. The method of calculating line loss rate of a cell based on a bidirectional LSTM network of claim 1, wherein the step 3 comprises,
step 301: dividing a hidden layer nerve unit into a positive time direction and a negative time direction which are independent;
step 302: feeding forward two independent hidden layers to the same output layer, including both past and future sequence information;
step 303: the layer 1 LSTM calculates the current time point sequence information, and the layer 2 LSTM reads the same sequence in the reverse direction;
step 304: adding the reverse order information, each layer of LSTM has different parameters.
5. The method according to claim 1, wherein the input attributes of the line loss calculation model include operation age of the distribution area, occupancy ratio of residential capacity, average capacity of residential users, number of users in the distribution area, daily electricity consumption, daily load rate, daily three-phase imbalance, daily power factor, and daily air temperature.
6. The method of calculating line loss rate of a cell based on a bidirectional LSTM network of claim 1, wherein the step 4 comprises,
step 401: searching the optimal hidden unit number of the single-layer model and fixing the optimal hidden unit number;
step 402: adding a hidden layer, and searching the optimal number of hidden units of the layer on the basis of the step 401;
step 403: counting the mean square error of each training sample input to the platform area line loss rate calculation model;
step 404: and setting the model parameter with the minimum root mean square error as the final parameter of the line loss rate calculation model of the transformer area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110878731.6A CN113609768A (en) | 2021-08-02 | 2021-08-02 | Bidirectional LSTM network-based distribution room line loss rate calculation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110878731.6A CN113609768A (en) | 2021-08-02 | 2021-08-02 | Bidirectional LSTM network-based distribution room line loss rate calculation method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113609768A true CN113609768A (en) | 2021-11-05 |
Family
ID=78306391
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110878731.6A Pending CN113609768A (en) | 2021-08-02 | 2021-08-02 | Bidirectional LSTM network-based distribution room line loss rate calculation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113609768A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116930667A (en) * | 2023-09-15 | 2023-10-24 | 广东电网有限责任公司 | Method, device, equipment and storage medium for testing edge of power grid of transformer area |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107730039A (en) * | 2017-10-10 | 2018-02-23 | 中国南方电网有限责任公司电网技术研究中心 | The method and system of distribution network load prediction |
| CN110598854A (en) * | 2019-09-20 | 2019-12-20 | 国网福建省电力有限公司 | GRU model-based transformer area line loss rate prediction method |
| CN110928997A (en) * | 2019-12-04 | 2020-03-27 | 北京文思海辉金信软件有限公司 | Intention recognition method and device, electronic equipment and readable storage medium |
| CN111950928A (en) * | 2020-08-24 | 2020-11-17 | 国网冀北电力有限公司 | Loss reduction method and device for power distribution network, storage medium and computing equipment |
| CN112288172A (en) * | 2020-10-30 | 2021-01-29 | 国家电网有限公司 | Prediction method and device for line loss rate of transformer area |
| CN112971769A (en) * | 2021-02-04 | 2021-06-18 | 杭州慧光健康科技有限公司 | Home personnel tumble detection system and method based on biological radar |
-
2021
- 2021-08-02 CN CN202110878731.6A patent/CN113609768A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107730039A (en) * | 2017-10-10 | 2018-02-23 | 中国南方电网有限责任公司电网技术研究中心 | The method and system of distribution network load prediction |
| CN110598854A (en) * | 2019-09-20 | 2019-12-20 | 国网福建省电力有限公司 | GRU model-based transformer area line loss rate prediction method |
| CN110928997A (en) * | 2019-12-04 | 2020-03-27 | 北京文思海辉金信软件有限公司 | Intention recognition method and device, electronic equipment and readable storage medium |
| CN111950928A (en) * | 2020-08-24 | 2020-11-17 | 国网冀北电力有限公司 | Loss reduction method and device for power distribution network, storage medium and computing equipment |
| CN112288172A (en) * | 2020-10-30 | 2021-01-29 | 国家电网有限公司 | Prediction method and device for line loss rate of transformer area |
| CN112971769A (en) * | 2021-02-04 | 2021-06-18 | 杭州慧光健康科技有限公司 | Home personnel tumble detection system and method based on biological radar |
Non-Patent Citations (3)
| Title |
|---|
| 熊明兰;王华伟;徐怡;付强;: "基于鸟击事故征候预测的通用航空安全研究", 系统工程与电子技术, no. 09 * |
| 郭志民;张永浩;周兴华;孙志杰;康田园;耿俊成;: "考虑三相不平衡的低压用户业扩报装供电方案精准编制", 电力系统保护与控制, no. 12 * |
| 金亮;王飞;杨庆新;汪冬梅;寇晓斐;: "永磁同步电机性能分析的典型深度学习模型与训练方法", 电工技术学报, no. 1 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116930667A (en) * | 2023-09-15 | 2023-10-24 | 广东电网有限责任公司 | Method, device, equipment and storage medium for testing edge of power grid of transformer area |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11070056B1 (en) | Short-term interval prediction method for photovoltaic power output | |
| Ke et al. | Short-term electrical load forecasting method based on stacked auto-encoding and GRU neural network | |
| Yuce et al. | ANN–GA smart appliance scheduling for optimised energy management in the domestic sector | |
| US20190265768A1 (en) | Method, system and storage medium for predicting power load probability density based on deep learning | |
| CN111191854A (en) | Photovoltaic power generation prediction model and method based on linear regression and neural network | |
| CN110796307A (en) | Distributed load prediction method and system for comprehensive energy system | |
| CN107453396A (en) | A kind of Multiobjective Optimal Operation method that distributed photovoltaic power is contributed | |
| Xu et al. | Potential analysis of the attention-based LSTM model in ultra-short-term forecasting of building HVAC energy consumption | |
| CN112990587A (en) | Method, system, equipment and medium for accurately predicting power consumption of transformer area | |
| CN115115125A (en) | Photovoltaic power interval probability prediction method based on deep learning fusion model | |
| Guo et al. | A short-term load forecasting model of LSTM neural network considering demand response | |
| Zhao et al. | Heating load prediction of residential district using hybrid model based on CNN | |
| CN115189416A (en) | Power generation system control method and system based on day-ahead electricity price grading prediction model | |
| Poczeta et al. | Application of fuzzy cognitive maps to multi-step ahead prediction of electricity consumption | |
| CN113609768A (en) | Bidirectional LSTM network-based distribution room line loss rate calculation method | |
| Wang et al. | Electric vehicle charging load clustering and load forecasting based on long short term memory neural network | |
| CN113988436A (en) | Power consumption prediction method based on LSTM neural network and hierarchical relation correction | |
| CN113033898A (en) | Electrical load prediction method and system based on K-means clustering and BI-LSTM neural network | |
| CN116307287B (en) | Prediction method, system and prediction terminal for effective period of photovoltaic power generation | |
| Shang et al. | Distributed V2G dispatching via LSTM network within cloud-edge collaboration framework | |
| CN113762591B (en) | Short-term electric quantity prediction method and system based on GRU and multi-core SVM countermeasure learning | |
| Mori et al. | Development of GRBFN with global structure for PV generation output forecasting | |
| Li et al. | A gray rbf model improved by genetic algorithm for electrical power forecasting | |
| He et al. | Forecasting model of residential load based on general regression neural network and PSO-Bayes least squares support vector machine | |
| CN116596279B (en) | Intelligent park energy consumption scheduling system |
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 |