CN112598173B - Self-organizing learning modeling method for time sequence data of energy storage system under cloud platform - Google Patents
Self-organizing learning modeling method for time sequence data of energy storage system under cloud platform Download PDFInfo
- Publication number
- CN112598173B CN112598173B CN202011524755.3A CN202011524755A CN112598173B CN 112598173 B CN112598173 B CN 112598173B CN 202011524755 A CN202011524755 A CN 202011524755A CN 112598173 B CN112598173 B CN 112598173B
- Authority
- CN
- China
- Prior art keywords
- energy storage
- data
- storage system
- cloud platform
- model
- 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
- 238000004146 energy storage Methods 0.000 title claims abstract description 61
- 238000000034 method Methods 0.000 title claims abstract description 19
- 230000002159 abnormal effect Effects 0.000 claims abstract description 4
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 238000005516 engineering process Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000012938 design process Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06N20/00—Machine learning
Abstract
The invention provides a self-organizing learning modeling method of time sequence data of energy storage systems under a cloud platform, which is characterized in that operation data generated by each energy storage system is stored in the cloud platform and used as training data; preprocessing training data, removing abnormal data in the training data, normalizing the data, and outputting the data as time sequence data; judging whether a history learning model exists or not, and if so, directly predicting by using the learning model; if the history learning model does not exist, entering the next step; according to the length of time series data generated by different energy storage systems under the cloud platform, corresponding self-organizing learning modeling is carried out, and a model is obtained and used for load prediction and peak value prediction of the energy storage systems; the method can solve the problems of occupation of a large amount of resources and unreasonable allocation caused by time series data learning modeling of the energy storage equipment under the cloud platform.
Description
Technical Field
The invention relates to a self-organizing learning modeling method of time sequence data of an energy storage system under a cloud platform.
Background
The time series data shows the state that one variable value or a plurality of variable values continuously change along with the change of another variable, wherein the time series data mainly changes along with the change of time, and has two forms for the data, namely, a unit time series refers to a state sequence of single variable values which change along with the change of time, and a multi-element time series refers to a state sequence of multi-variables which change along with the change of time. In general, time series data can be abstracted into the form of a binary group (t, d), wherein t represents a time variable, d represents a data variable, the variable reflects a variable value and an actual meaning of a data unit, and the variable can represent the operation condition of an energy storage system. The energy storage systems are connected with each other in a distributed mode, and operation data of each energy storage system are uploaded to the cloud platform.
Because of its low cost and openness, cloud platform technology is being widely used in various fields such as commerce, military, and academia, which makes the reliability of the overall resources of the cloud platform particularly important. The effective data learning method and strategy can efficiently use calculation and network resources, reduce unnecessary cost waste and provide a reliable environment with high availability for users.
The existing time series data learning modeling method under the cloud platform has the problems of no consideration of resource occupation and unreasonable allocation, and the existing technology aims at increasing the investment of the cloud platform or dynamically increasing the memory capacity of the cloud platform, thereby increasing the cost and being a great burden for enterprises with insufficient funds.
The above-mentioned problem is a problem that should be considered and solved in the design process of the self-organizing learning modeling method of time-series data of the energy storage system under the cloud platform.
Disclosure of Invention
The invention aims to provide a self-organizing learning modeling method for time series data of an energy storage system under a cloud platform, which solves the problems of large resource occupation and unreasonable allocation caused by time series data learning modeling of energy storage equipment under the cloud platform in the prior art.
The technical scheme of the invention is as follows:
a self-organizing learning modeling method of time sequence data of an energy storage system under a cloud platform, which comprises the following steps,
s1, storing operation data generated by each energy storage system in a cloud platform to serve as training data;
s2, preprocessing training data, removing abnormal data in the training data, normalizing the data, and outputting the normalized data as time sequence data;
s3, judging whether a history learning model exists, and if so, directly predicting by using the learning model; if the history learning model does not exist, entering the next step;
and S4, performing corresponding self-organizing learning modeling according to the length of time sequence data generated by different energy storage systems under the cloud platform to obtain a model for load prediction and peak prediction of the energy storage systems.
Further, the self-organizing learning modeling method of the time sequence data of the energy storage system under the cloud platform is characterized by comprising the following steps of: in step S4, corresponding self-organizing learning modeling is performed, specifically,
s41, judging the length of the time series data, if the length of the time series data is smaller than a set value, performing online learning on the data, and directly using a model obtained by learning for load prediction and peak prediction of an energy storage system; if the length of the time series data is not less than the set value, entering the next step;
s42, if the length of the time series data is not smaller than a set value, performing offline learning on the time series data, querying the energy storage system under the cloud platform again, querying whether a history learning model exists, and if the history learning model does not exist, storing the learned model parameters, and using the model parameters for load prediction and peak prediction; when the same energy storage system performs more than two times of prediction, the previous model parameters are directly used for prediction, the time series data of the energy storage system are periodically learned to obtain a corresponding model, the corresponding model is updated and reserved, and the learned model is used for load prediction and peak value prediction of the energy storage system.
The beneficial effects of the invention are as follows: according to the self-organizing learning modeling method for the time series data of the energy storage system under the cloud platform, the corresponding self-organizing learning modeling is selected according to the sizes of the time series data generated by different energy storage systems under the cloud platform, so that the problems that a large amount of resources are occupied and the distribution is unreasonable during the time series data learning modeling of the energy storage device under the cloud platform can be solved.
Drawings
Fig. 1 is a frame diagram of an energy storage system based on a cloud platform in an embodiment.
Fig. 2 is a flow chart of a self-organizing learning modeling method of time series data of an energy storage system under a cloud platform according to an embodiment of the invention.
Detailed Description
Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
Examples
A self-organizing learning modeling method of time sequence data of an energy storage system under a cloud platform, which comprises the following steps,
s1, storing operation data generated by each energy storage system in a cloud platform to serve as training data;
s2, preprocessing training data, removing abnormal data in the training data, normalizing the data, and outputting the normalized data as time sequence data;
s3, judging whether a history learning model exists, and if so, directly predicting by using the learning model; if the history learning model does not exist, entering the next step;
and S4, performing corresponding self-organizing learning modeling according to the length of time sequence data generated by different energy storage systems under the cloud platform to obtain a model for load prediction and peak prediction of the energy storage systems.
S41, judging the length of the time series data, if the length of the time series data is smaller than a set value, performing online learning on the data, and directly using a model obtained by learning for load prediction and peak prediction of an energy storage system; if the length of the time series data is not less than the set value, entering the next step; in one embodiment, the set point is preferably 1 month.
S42, if the length of the time series data is not smaller than a set value, performing offline learning on the time series data, querying the energy storage system under the cloud platform again, querying whether a history learning model exists, and if the history learning model does not exist, storing the learned model parameters, and using the model parameters for load prediction and peak prediction; when the same energy storage system performs more than two times of prediction, the previous model parameters are directly used for prediction, the time series data of the energy storage system are periodically learned to obtain a corresponding model, the corresponding model is updated and reserved, and the learned model is used for load prediction and peak value prediction of the energy storage system.
According to the self-organizing learning modeling method of the time series data of the energy storage system under the cloud platform, different self-organizing learning modeling methods can be selected according to the sizes of the time series data generated by different energy storage systems under the cloud platform, if the time series data generated by the energy storage system is larger and the time scale span is longer, model parameters are obtained through off-line learning, the model parameters are stored in the cloud platform, when the load of the energy storage system is predicted, the model parameters are directly used, the data are periodically re-learned off-line, and the model parameters are updated; if the time series data generated by the energy storage system is smaller, the time series data is studied online when the scale is shorter, so that the problems that a large amount of resources are occupied and the distribution is unreasonable in the time series data study modeling of the energy storage equipment under the cloud platform are solved, and the system model parameters after the study are reasonably utilized to conduct load prediction, peak value prediction and the like on the energy storage system.
According to the self-organizing learning modeling method for the time series data of the energy storage system under the cloud platform, the corresponding self-organizing learning modeling is selected according to the sizes of the time series data generated by different energy storage systems under the cloud platform, so that the problems that a large amount of resources are occupied and the distribution is unreasonable during the time series data learning modeling of the energy storage device under the cloud platform can be solved. The self-organizing learning modeling method of the time sequence data of the energy storage system under the cloud platform can solve the problems that the running data of the energy storage system is continuously increased along with the increase of the time scale under the cloud platform, the running data is used as the time sequence data, the resources occupied by online learning are large, and the energy storage system is connected through distributed construction, and if multiple users carry out load prediction on the energy storage system, the resource blockage and uneven distribution are easy to cause.
Claims (1)
1. A self-organizing learning modeling method of time sequence data of an energy storage system under a cloud platform is characterized by comprising the following steps of: comprises the steps of,
s1, storing operation data generated by each energy storage system in a cloud platform to serve as training data;
s2, preprocessing training data, removing abnormal data in the training data, normalizing the data, and outputting the normalized data as time sequence data;
s3, judging whether a history learning model exists, and if so, directly predicting by using the learning model; if the history learning model does not exist, entering the next step;
s4, performing corresponding self-organizing learning modeling according to the length of time sequence data generated by different energy storage systems under the cloud platform to obtain a model for load prediction and peak prediction of the energy storage systems; the corresponding self-organizing learning modeling is carried out, specifically: s41, judging the length of the time series data, if the length of the time series data is smaller than a set value, performing online learning on the data, and directly using a model obtained by learning for load prediction and peak prediction of an energy storage system; if the length of the time series data is not less than the set value, entering the next step; s42, if the length of the time series data is not smaller than a set value, performing offline learning on the time series data, querying the energy storage system under the cloud platform again, querying whether a history learning model exists, and if the history learning model does not exist, storing the learned model parameters, and using the model parameters for load prediction and peak prediction; when the same energy storage system performs more than two times of prediction, the previous model parameters are directly used for prediction, the time series data of the energy storage system are periodically learned to obtain a corresponding model, the corresponding model is updated and reserved, and the learned model is used for load prediction and peak value prediction of the energy storage system.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011524755.3A CN112598173B (en) | 2020-12-22 | 2020-12-22 | Self-organizing learning modeling method for time sequence data of energy storage system under cloud platform |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011524755.3A CN112598173B (en) | 2020-12-22 | 2020-12-22 | Self-organizing learning modeling method for time sequence data of energy storage system under cloud platform |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112598173A CN112598173A (en) | 2021-04-02 |
CN112598173B true CN112598173B (en) | 2024-03-08 |
Family
ID=75199994
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011524755.3A Active CN112598173B (en) | 2020-12-22 | 2020-12-22 | Self-organizing learning modeling method for time sequence data of energy storage system under cloud platform |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112598173B (en) |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110225075A (en) * | 2019-03-25 | 2019-09-10 | 北京快电科技有限公司 | A kind of building energy internet wisdom operation cloud operating system |
CN111415027A (en) * | 2019-01-08 | 2020-07-14 | 顺丰科技有限公司 | Method and device for constructing component prediction model |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20200151014A1 (en) * | 2018-11-09 | 2020-05-14 | Solecular, Inc. | Methods and systems for forecasting upcoming resource requirement |
-
2020
- 2020-12-22 CN CN202011524755.3A patent/CN112598173B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111415027A (en) * | 2019-01-08 | 2020-07-14 | 顺丰科技有限公司 | Method and device for constructing component prediction model |
CN110225075A (en) * | 2019-03-25 | 2019-09-10 | 北京快电科技有限公司 | A kind of building energy internet wisdom operation cloud operating system |
Also Published As
Publication number | Publication date |
---|---|
CN112598173A (en) | 2021-04-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104639626A (en) | Multi-level load forecasting and flexible cloud resource configuring method and monitoring and configuring system | |
CN112862253B (en) | Electric-gas-thermal energy storage collaborative planning method and system considering wind power uncertainty | |
CN110874665B (en) | Control device and method for wind generating set | |
CN116227757A (en) | Comprehensive energy management and control method and system based on intelligent cloud gateway | |
CN112261120A (en) | Cloud-side cooperative task unloading method and device for power distribution internet of things | |
CN115314343A (en) | Source-load-storage resource aggregation control gateway device and load and output prediction method | |
CN114003387A (en) | Micro-service load balancing and elastic expansion and contraction method based on reinforcement learning | |
CN114036825A (en) | Collaborative optimization scheduling method, device, equipment and storage medium for multiple virtual power plants | |
CN116418064A (en) | Charging control method, system, electronic device and computer readable storage medium | |
CN113452025B (en) | Model-data hybrid driven power grid expected fault assessment method and system | |
CN112598173B (en) | Self-organizing learning modeling method for time sequence data of energy storage system under cloud platform | |
CN112926762A (en) | Clearing method and device based on safety constraint unit combination model | |
CN111799793A (en) | Source-grid-load cooperative power transmission network planning method and system | |
CN115549204A (en) | Micro-grid cluster double-layer distributed cluster peak shaving method and device | |
CN110649652A (en) | New energy sending-out system phase modulator configuration method and device based on risk quantitative evaluation | |
CN115577996A (en) | Risk assessment method, system, equipment and medium for power grid power failure plan | |
CN115528750A (en) | Data model hybrid drive unit combination method for power grid safety and stability | |
CN103873388A (en) | Network content control method and network equipment | |
CN109857590B (en) | Data backup system of power dispatching center | |
CN114861995A (en) | Demand prediction method and device for power supplies and computer equipment | |
CN109063859B (en) | Power grid equipment maintenance optimization processing method and device | |
CN114683908A (en) | Battery replacement capability evaluation method and system for battery replacement station, electronic device and medium | |
CN111783487A (en) | Fault early warning method and device for card reader equipment | |
CN112491067A (en) | Active power distribution network capacity configuration method based on composite energy storage | |
Qin et al. | Cnn-gru-attention based short-term load forecasting of distribution networks |
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 | ||
CB02 | Change of applicant information |
Address after: 226100 No.888, Jiuhua Road, high tech Zone, Nantong City, Jiangsu Province Applicant after: Wotai Energy Co.,Ltd. Address before: No. 888, Jiuhua Road, high tech Zone, Nantong City, Jiangsu Province, 226399 Applicant before: NEOVOLTAIC ENERGY NANTONG Co.,Ltd. |
|
CB02 | Change of applicant information | ||
GR01 | Patent grant | ||
GR01 | Patent grant |