CN117421596A - Energy storage container fault early warning system and energy storage system - Google Patents
Energy storage container fault early warning system and energy storage system Download PDFInfo
- Publication number
- CN117421596A CN117421596A CN202311414751.3A CN202311414751A CN117421596A CN 117421596 A CN117421596 A CN 117421596A CN 202311414751 A CN202311414751 A CN 202311414751A CN 117421596 A CN117421596 A CN 117421596A
- Authority
- CN
- China
- Prior art keywords
- energy storage
- data
- cloud
- classifier
- fault early
- 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
- 238000004146 energy storage Methods 0.000 title claims abstract description 121
- 238000012549 training Methods 0.000 claims abstract description 31
- 238000013508 migration Methods 0.000 claims abstract description 16
- 230000005012 migration Effects 0.000 claims abstract description 16
- 230000006870 function Effects 0.000 claims description 13
- 238000013507 mapping Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 6
- 238000002580 esophageal motility study Methods 0.000 claims description 4
- 238000003062 neural network model Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 238000013526 transfer learning Methods 0.000 abstract description 8
- 238000000034 method Methods 0.000 description 13
- 238000004422 calculation algorithm Methods 0.000 description 6
- 230000006855 networking Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application provides an energy storage container trouble early warning system and energy storage system, wherein, energy storage container trouble early warning system includes: cloud AI platform, edge AI equipment and big data platform; the cloud AI platform trains and updates a fault early warning offline model based on historical fault data of the energy storage containers corresponding to all the overall energy storage projects; the large data platform counts the local running state characteristic data of the energy storage container corresponding to each energy storage project in real time; when an energy storage project to which the edge AI equipment belongs newly joins the system, training an online classifier based on a fault early warning offline model; and the local running state features are subjected to migration learning and fault early warning through an offline classifier and an online classifier. Cloud-edge integrated fault early warning based on online transfer learning can fully utilize cloud big data and data of local edge equipment, and ensure fault early warning efficiency and stability.
Description
Technical Field
The application relates to the technical field of energy storage, in particular to an energy storage container fault early warning system and an energy storage system.
Background
Most of fault early warning implementation schemes of the energy storage container belong to a local edge scheme, namely, a data driving scheme utilizes local BMS data to perform fault recognition and fault early warning. Few schemes belong to cloud schemes, and cloud-edge integrated schemes for fault early warning of the energy storage container are few.
The local edge scheme can only use the data of the local limited energy storage container, and cannot timely utilize the data of other energy storage containers. The cloud scheme has the problem that the cloud scheme cannot be used when no network exists locally.
Disclosure of Invention
An object of the application is to provide an energy storage container trouble early warning system and energy storage system, based on integrative trouble early warning of cloud limit of online migration study, can make full use of high in the clouds data and the data of local edge equipment, guarantee trouble early warning efficiency and stability.
In a first aspect, the present application provides an energy storage container fault pre-warning system, the system comprising: cloud AI platform, edge AI equipment and big data platform; the cloud AI platform and the big data platform are respectively connected with EMSs corresponding to the energy storage containers; the cloud AI platform trains and updates a fault early-warning offline model based on historical fault data of the energy storage containers corresponding to all global energy storage projects, and updates the latest fault early-warning offline model to all edge AI devices; the large data platform is used for counting the local running state characteristic data of the energy storage container corresponding to each energy storage project in real time and sending the local running state characteristic data to the edge AI equipment corresponding to the belonging energy storage project in real time; when an energy storage project to which the edge AI equipment belongs newly joins the system, training an online classifier for a fault early warning offline model by adopting a certain number of marked running state characteristic data of the energy storage project to which the edge AI equipment belongs; and the original offline classifier and online classifier of the self fault early warning offline model are adopted, and meanwhile, the local running state characteristics sent by the big data platform in real time are subjected to migration learning and fault early warning.
Further, the cloud AI platform includes: the cloud energy storage module and the cloud AI training module; the cloud energy storage module is communicated with the EMS of all the overall energy storage projects and is used for collecting marked running state characteristic data of the energy storage containers corresponding to all the overall energy storage projects; and the cloud AI training module is used for training and updating the current fault early warning offline model based on the marked running state characteristic data which are stored in the cloud energy storage module in an increment mode.
Further, the cloud AI module carries out full-scale learning training on the neural network model by adopting historical fault data at the initial stage of construction of the fault early-warning model to obtain a fault early-warning offline model at the initial stage of construction of the fault early-warning model.
Further, when the edge AI equipment is connected with the network, the updating processing of the cloud AI platform to the fault early warning offline model operated by the cloud AI platform is accepted, and on-line transfer learning and fault early warning are carried out based on the local operation state characteristic data counted by the big data platform in real time; and when offline, performing online migration learning and fault early warning based on the local running state characteristic data counted by the big data platform in real time.
Further, the big data platform is used for carrying out data statistics on the local running state characteristic data of the energy storage container; the data statistics method comprises the following steps: dimension statistics, dimensionless statistics and multidimensional statistics.
Further, the edge AI device is further configured to predict, for each piece of data to be detected in the local running state feature data, the piece of data to be detected through an offline classifier and the online classifier in the fault early-warning offline model, so as to obtain a first prediction result and a second prediction result; and carrying out weighted summation on the first prediction result and the second prediction result to obtain a fault early warning result corresponding to the data to be detected.
Further, the update formulas of the offline classifier and the online classifier weights are as follows:
s t (u)＝exp{-ηl(∏(u T x t ),∏(y t ))},l(z,y)＝(z-y) 2 ；
wherein w is t An online classifier representing the time t; v denotes an offline classifier; w (w) 1,t Representing the weight corresponding to the online classifier at the time t; w (w) 1,t+1 Representing the weight corresponding to the online classifier function at the time t+1; w (w) 2,t Representing the weight corresponding to the off-line classifier at the time t; w (w) 2,t+1 Representing the weight corresponding to the offline classifier at the time t+1; s is(s) t () Representing a gaussian kernel function; s is(s) t (v) Representing the value of the loss of the offline classifier at the time t after Gaussian mapping; s is(s) t (w t ) The value of the loss of the online classifier at the time t after Gaussian mapping is represented; x is x t A sample instance representing time t; y is t A real label of a sample at the time t is represented; (z-y) 2 For the loss function, η is the learning rate and n () is the mapping function.
Further, the edge AI device is further configured to perform incremental training on the online classifier by using a certain number of marked newly added operation state feature data of the energy storage items to which the edge AI device belongs, so as to update the online classifier.
Further, when the fault early warning offline model is updated and trained, new energy storage project data with new project labels are continuously added, and incremental training is performed at fixed time intervals.
In a second aspect, the present application further provides an energy storage system, which comprises the energy storage container fault early warning system according to the first aspect.
In energy storage container trouble early warning system and energy storage system that this application provided, energy storage container trouble early warning system includes: cloud AI platform, edge AI equipment and big data platform; the cloud AI platform and the big data platform are respectively connected with EMSs corresponding to the energy storage containers; the cloud AI platform trains and updates a fault early-warning offline model based on historical fault data of the energy storage containers corresponding to all global energy storage projects, and updates the latest fault early-warning offline model to all edge AI devices; the large data platform is used for counting the local running state characteristic data of the energy storage container corresponding to each energy storage project in real time and sending the local running state characteristic data to the edge AI equipment corresponding to the belonging energy storage project in real time; when an energy storage project to which the edge AI equipment belongs newly joins the system, training an online classifier for a fault early warning offline model by adopting a certain number of marked running state characteristic data of the energy storage project to which the edge AI equipment belongs; and the original offline classifier and online classifier of the self fault early warning offline model are adopted, and meanwhile, the local running state characteristics sent by the big data platform in real time are subjected to migration learning and fault early warning. Cloud-edge integrated fault early warning based on online transfer learning can fully utilize cloud-edge big data and data of local edge equipment, and fault early warning efficiency and stability can be guaranteed no matter in an offline state or a networking state.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a block diagram of a fault early warning system of an energy storage container according to an embodiment of the present application;
FIG. 2 is a block diagram of another energy storage container fault early warning system according to an embodiment of the present disclosure;
fig. 3 is a schematic diagram of an online migration learning process according to an embodiment of the present application;
fig. 4 is a three-terminal overall flowchart provided in an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all, of the embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.
In view of the problems that in the prior art, a local edge scheme can only use local limited energy storage container data, the data of other energy storage containers cannot be timely utilized, and a cloud scheme has the problem that the local energy storage container data cannot be used when no network exists. The embodiment of the application provides an energy storage container fault early warning system and energy storage system, cloud side integrated fault early warning based on online transfer learning can make full use of cloud big data and data of local edge equipment, and guarantee fault early warning efficiency and stability.
For the sake of understanding the present embodiment, first, a fault early warning system for an energy storage container disclosed in the present embodiment will be described in detail.
Fig. 1 is a block diagram of a fault early warning system for an energy storage container according to an embodiment of the present application, where the system includes: the cloud AI platform 11, the edge AI device 12 and the big data platform 13; the cloud AI platform 11 and the big data platform 13 are respectively connected with EMSs corresponding to the energy storage containers; the cloud AI platform 11 trains and updates the fault early-warning offline model based on the historical fault data of the energy storage containers corresponding to all the global energy storage projects, and updates the latest fault early-warning offline model to all the edge AI devices 12. The big data platform 13 is used for counting the local running state characteristic data of the energy storage container corresponding to each energy storage project in real time and sending the local running state characteristic data to the edge AI equipment 12 corresponding to the belonging energy storage project in real time; when the energy storage project to which the edge AI device 12 belongs newly joins the system, an online classifier is trained on the self fault early warning offline model by adopting a certain number of marked running state characteristic data of the energy storage project to which the edge AI device belongs; and the original offline classifier and online classifier of the self fault early warning offline model are adopted, and meanwhile, the local running state characteristics sent by the big data platform in real time are subjected to migration learning and fault early warning.
The fault prediction offline model also realizes incremental learning based on the marked running state characteristic data of all global energy storage items, and updates the fault early warning offline model which completes incremental learning each time into all edge AI equipment 12; when the fault early warning offline model is updated and trained, new energy storage project data with new project labels are continuously added, and incremental training is carried out at fixed time intervals.
Referring to a preferred system embodiment shown in fig. 2, the cloud AI platform includes: the cloud energy storage module and the cloud AI training module; the cloud energy storage module is communicated with the EMS of all the overall energy storage projects and is used for collecting marked running state characteristic data of the energy storage containers corresponding to all the overall energy storage projects; and the cloud AI training module is used for training and updating the current fault early warning offline model based on the marked running state characteristic data which are stored in the cloud energy storage module in an increment mode.
The energy storage container is a carrier for storing and converting energy in an energy storage project and internally comprises BMS and EMS which are responsible for data transmission; the energy storage cloud platform is used for storing data of relevant equipment in the energy storage container of all energy storage projects of an enterprise; the cloud AI is responsible for training a machine learning model based on big data of the energy storage cloud platform; the big data platform is responsible for the statistics of the local running state characteristic data; the edge AI device is responsible for training and fault pre-warning of local algorithm models (e.g., online classifiers for the current project).
In the implementation, the data of a plurality of projects in the cloud are utilized, the data volume is large in general, for a data driving model, model parameters are required to be obtained based on enough and diversified training data, and BMS data in the energy storage containers of the projects are uploaded to an energy storage cloud platform through EMS; then, the cloud AI trains the offline model, which needs to be carried out on a cloud workstation, and the offline model realizes incremental learning along with the continuous increase of the data volume of the energy storage cloud platform, so that the full-volume data training model is avoided, the consumption of computing resources is reduced, and meanwhile, the accuracy of the model can be continuously improved.
In other words, the cloud AI module performs full-scale learning training on the neural network model by adopting historical fault data in the early stage of construction of the fault early-warning model, so as to obtain a fault early-warning offline model in the early stage of construction of the fault early-warning model.
Further, when the edge AI equipment is connected with the network, the updating processing of the cloud AI platform to the fault early warning offline model operated by the cloud AI platform is accepted, and on-line transfer learning and fault early warning are carried out based on the local operation state characteristic data counted by the big data platform in real time; and when offline, performing online migration learning and fault early warning based on the local running state characteristic data counted by the big data platform in real time.
Updating an offline model by the edge AI equipment in a networking state, and performing online learning and transfer learning; if the edge equipment is not networked, only online learning and transfer learning are performed; in the online learning process, a big data platform is required to provide local running state characteristic data, namely real-time data flow, wherein the data comprises dimension statistics, dimensionless statistics and multidimensional statistics.
Referring to fig. 3, the process of performing online migration learning by the edge AI device to perform fault early warning is as follows:
(1) Acquiring statistical local running state characteristic data from a big data platform;
(2) And training an online classifier for the fault early warning offline model by adopting a certain number of marked running state characteristic data of the energy storage projects to which the online classifier belongs.
In specific implementation, the edge AI device is further used for performing incremental training on the online classifier by adopting a certain number of marked newly-added running state characteristic data of the energy storage items to which the edge AI device belongs so as to update the online classifier.
(3) Aiming at each piece of data to be detected in the local running state characteristic data, predicting the data to be detected through an offline classifier and an online classifier in a fault early warning offline model to obtain a first prediction result and a second prediction result; the data to be predicted is online time sequence data in the local running state characteristic data, and the online time sequence data comprises data which are acquired correspondingly by a new project, a current scene or a current new energy storage project.
(4) And carrying out weighted summation on the first prediction result and the second prediction result to obtain a fault early warning result corresponding to the data to be detected.
The online migration learning and core part is the updating of the weights of the offline classifier and the online classifier, and the updating formulas of the weights of the offline classifier and the online classifier are as follows:
s t (u)＝exp{-ηl(∏(u T x t ),∏(y t ))},l(z,y)＝(z-y) 2 ；
wherein w is t An online classifier representing the time t; v denotes an offline classifier; w (w) 1,t Representing the weight corresponding to the online classifier at the time t; w (w) 1,t+1 Representing the weight corresponding to the online classifier function at the time t+1; w (w) 2,t Representing the weight corresponding to the off-line classifier at the time t; w (w) 2,t+1 Representing the weight corresponding to the offline classifier at the time t+1; s is(s) t () Representing a gaussian kernel function; s is(s) t (v) Representing the value of the loss of the offline classifier at the time t after Gaussian mapping; s is(s) t (w t ) The value of the loss of the online classifier at the time t after Gaussian mapping is represented; x is x t A sample instance representing time t; y is t A real label of a sample at the time t is represented; (z-y) 2 For the loss function, η is the learning rate and n () is the mapping function. The above formula indicates that the weight at time t+1 is determined by the proportion of the loss value of the classifier at the previous time to the total loss value.
It should be noted that, after the edge AI device updates the offline model, the weights of the offline classifier and the online classifier remain before the update.
The cloud side end integrated architecture combines a cloud AI platform comprising a cloud energy storage module and a cloud AI training module, and edge AI equipment and a big data platform, so that the cloud and the edge equipment work cooperatively to provide data and calculation support for a fault early warning algorithm model.
Referring to the overall flowchart shown in fig. 4, the specific operation is as follows:
step 1, the cloud AI platform collects historical fault data of multiple items, performs characteristic engineering and data sampling, and is required to train a fault early warning offline model by using the full data in the initial operation, and incremental learning is required in the operation process, so that the performance of the offline model is continuously improved.
And 2, acquiring real-time data of a plurality of local energy storage containers by the big data platform, and carrying out dimension, dimensionless statistics and space-time dimension statistical analysis.
And step 3, the edge AI equipment acquires edge statistical data from the big data platform and performs feature engineering, and then decides a model training action in the next step according to the networking state of the energy storage container. Under the condition that an algorithm is operated for the first time and is networked, an edge AI device needs to acquire an offline model from a cloud; and in the process of algorithm operation and under the networking condition, the updating of the offline model is completed. In the case of non-networking, the most recently updated offline model will participate in subsequent online migration learning.
And step 4, finally, training the online migration learning model by utilizing the edge statistical data obtained in the step 2 and the offline model obtained in the step 3.
The energy storage container fault early warning system provided by the embodiment of the application is a fault early warning scheme for realizing the energy storage container based on online migration learning, and can ensure the real-time performance and accuracy of early warning. Different algorithm models are executed aiming at the networking state of the local energy storage container, so that the safety of the energy storage container is ensured. Along with the continuous increase of cloud data volume, the offline model realizes incremental learning, avoids using a full-volume data training model, so as to reduce the consumption of computing resources, and simultaneously can continuously improve the accuracy of the model. And the online transfer learning and the cloud edge are integrated to realize fault early warning under the cooperation of the cloud edge, so that the reliability of the fault early warning module is improved. The cloud side end integrated architecture combines a cloud AI platform comprising a cloud energy storage module and a cloud AI training module, and edge AI equipment and a large data platform, so that the cloud and the edge equipment work cooperatively, data and calculation force support is provided for a fault early warning algorithm model, and safe operation of an energy storage system is really guaranteed.
Based on the above embodiments, the present application further provides an energy storage system, which includes the energy storage container fault early warning system as described above.
The energy storage system provided in the embodiments of the present application has the same implementation principle and technical effects as those of the embodiments of the foregoing system, and for a brief description, reference may be made to corresponding matters in the embodiments of the foregoing system where no reference is made to the description of the embodiments of the energy storage system.
The embodiment of the present application further provides a computer readable storage medium, where a computer executable instruction is stored, where the computer executable instruction, when being called and executed by a processor, causes the processor to implement the foregoing method, and the specific implementation may refer to the foregoing method embodiment and is not described herein.
The method, the apparatus and the computer program product of the electronic device provided in the embodiments of the present application include a computer readable storage medium storing program codes, where the instructions included in the program codes may be used to execute the method described in the foregoing method embodiment, and specific implementation may refer to the method embodiment and will not be described herein.
The relative steps, numerical expressions and numerical values of the components and steps set forth in these embodiments do not limit the scope of the present application unless specifically stated otherwise.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a non-volatile computer readable storage medium executable by a processor. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In the description of the present application, it should be noted that the directions or positional relationships indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. are based on the directions or positional relationships shown in the drawings, are merely for convenience of description of the present application and to simplify the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and thus should not be construed as limiting the present application. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
Finally, it should be noted that: the foregoing examples are merely specific embodiments of the present application, and are not intended to limit the scope of the present application, but the present application is not limited thereto, and those skilled in the art will appreciate that while the foregoing examples are described in detail, the present application is not limited thereto. Any person skilled in the art may modify or easily conceive of the technical solution described in the foregoing embodiments, or make equivalent substitutions for some of the technical features within the technical scope of the disclosure of the present application; such modifications, changes or substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application, and are intended to be included in the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (10)
1. An energy storage container fault pre-warning system, the system comprising: cloud AI platform, edge AI equipment and big data platform; the cloud AI platform and the big data platform are respectively connected with EMSs corresponding to the energy storage containers;
the cloud AI platform trains and updates a fault early-warning offline model based on historical fault data of the energy storage containers corresponding to all global energy storage projects, and updates the latest fault early-warning offline model to all edge AI devices;
the large data platform is used for counting the local running state characteristic data of the energy storage container corresponding to each energy storage project in real time and sending the local running state characteristic data to the edge AI equipment corresponding to the belonging energy storage project in real time;
when the energy storage project of the edge AI equipment newly joins the system, training an online classifier for the fault early warning offline model by adopting a certain number of marked running state characteristic data of the energy storage project of the edge AI equipment; and the original offline classifier and the online classifier of the self fault early warning offline model are adopted, and meanwhile, the local running state characteristics sent by the big data platform in real time are subjected to migration learning and fault early warning.
2. The system of claim 1, wherein the cloud AI platform comprises: the cloud energy storage module and the cloud AI training module; the cloud energy storage module is communicated with the EMS of all the global energy storage projects and is used for collecting marked running state characteristic data of the energy storage containers corresponding to all the global energy storage projects; the cloud AI training module is used for training and updating the current fault early warning offline model based on the marked running state characteristic data which are stored in the cloud energy storage module in an increment mode.
3. The system of claim 2, wherein the cloud AI module performs full-scale learning training on the neural network model using the historical fault data at an initial stage of construction of the fault early-warning model to obtain the fault early-warning offline model at the initial stage of construction of the fault early-warning model.
4. The system of claim 3, wherein the edge AI device, when being networked, accepts update processing of a fault early warning offline model operated by the cloud AI platform, and performs online migration learning and fault early warning based on local operation state feature data counted by the big data platform in real time; and when offline, performing online migration learning and fault early warning based on the local running state characteristic data counted by the big data platform in real time.
5. The system of claim 4, wherein the big data platform is configured to perform data statistics on local operational status characteristic data of the energy storage container; the data statistics mode comprises the following steps: dimension statistics, dimensionless statistics and multidimensional statistics.
6. The system of claim 5, wherein the edge AI device is further configured to predict, for each data to be detected in the local operational state feature data, the data to be detected by an offline classifier and the online classifier in the fault early-warning offline model, respectively, to obtain a first prediction result and a second prediction result; and carrying out weighted summation on the first prediction result and the second prediction result to obtain a fault early warning result corresponding to the data to be detected.
7. The system of claim 6, wherein the update formulas for the offline classifier and the online classifier weights are as follows:
s t (u)＝exp{-ηl(∏(u T x t ),∏(y t ))},l(z,y)＝(z-y) 2 ；
wherein w is t An online classifier representing the time t; v denotes an offline classifier; w (w) 1, Representing the weight corresponding to the online classifier at the time t; w (w) 1, Representing the weight corresponding to the online classifier function at the time t+1; w (w) 2, Representing the weight corresponding to the off-line classifier at the time t; w (w) 2, Representing the weight corresponding to the offline classifier at the time t+1; s is(s) t () Representing a gaussian kernel function; s is(s) t (v) Representing the value of the loss of the offline classifier at the time t after Gaussian mapping; s is(s) t (w t ) The value of the loss of the online classifier at the time t after Gaussian mapping is represented; x is x t A sample instance representing time t; y is t A real label of a sample at the time t is represented; (z-y) 2 As a loss function, η is a learning rate, n () is a mappingA function.
8. The system of claim 1, wherein the edge AI device is further configured to incrementally train the online classifier with a number of noted newly added operational state characteristic data for the stored energy item to which it belongs to update the online classifier.
9. The system of claim 1, wherein the fault pre-warning offline model is updated with new stored energy item data with new item labels added continuously for incremental training at regular time intervals.
10. An energy storage system, characterized in that the energy storage system comprises an energy storage container fault pre-warning system according to any one of claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311414751.3A CN117421596A (en) | 2023-10-27 | 2023-10-27 | Energy storage container fault early warning system and energy storage system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311414751.3A CN117421596A (en) | 2023-10-27 | 2023-10-27 | Energy storage container fault early warning system and energy storage system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117421596A true CN117421596A (en) | 2024-01-19 |
Family
ID=89522577
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311414751.3A Pending CN117421596A (en) | 2023-10-27 | 2023-10-27 | Energy storage container fault early warning system and energy storage system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117421596A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190325350A1 (en) * | 2018-04-20 | 2019-10-24 | International Business Machines Corporation | Transfer learning without local data export in multi-node machine learning |
| CN110719210A (en) * | 2019-12-05 | 2020-01-21 | 赣江新区智慧物联研究院有限公司 | Industrial equipment predictive maintenance method based on cloud edge cooperation |
| CN111507489A (en) * | 2020-04-20 | 2020-08-07 | 电子科技大学中山学院 | Cloud-edge-coordinated amusement equipment fault prediction and health management system and method |
| CN113744888A (en) * | 2021-09-02 | 2021-12-03 | 深圳万海思数字医疗有限公司 | Regional epidemic trend prediction early warning method and system |
| US20220092407A1 (en) * | 2020-09-23 | 2022-03-24 | International Business Machines Corporation | Transfer learning with machine learning systems |
| CN115271164A (en) * | 2022-06-20 | 2022-11-01 | 浙江工业大学 | Photovoltaic power generation system power prediction method based on information fusion online migration |
| CN115392113A (en) * | 2022-08-03 | 2022-11-25 | 北京九天翱翔科技有限公司 | Cross-working condition complex electromechanical system residual life prediction system and method |
-
2023
- 2023-10-27 CN CN202311414751.3A patent/CN117421596A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190325350A1 (en) * | 2018-04-20 | 2019-10-24 | International Business Machines Corporation | Transfer learning without local data export in multi-node machine learning |
| CN110719210A (en) * | 2019-12-05 | 2020-01-21 | 赣江新区智慧物联研究院有限公司 | Industrial equipment predictive maintenance method based on cloud edge cooperation |
| CN111507489A (en) * | 2020-04-20 | 2020-08-07 | 电子科技大学中山学院 | Cloud-edge-coordinated amusement equipment fault prediction and health management system and method |
| US20220092407A1 (en) * | 2020-09-23 | 2022-03-24 | International Business Machines Corporation | Transfer learning with machine learning systems |
| CN113744888A (en) * | 2021-09-02 | 2021-12-03 | 深圳万海思数字医疗有限公司 | Regional epidemic trend prediction early warning method and system |
| CN115271164A (en) * | 2022-06-20 | 2022-11-01 | 浙江工业大学 | Photovoltaic power generation system power prediction method based on information fusion online migration |
| CN115392113A (en) * | 2022-08-03 | 2022-11-25 | 北京九天翱翔科技有限公司 | Cross-working condition complex electromechanical system residual life prediction system and method |
Non-Patent Citations (3)
| Title |
|---|
| YONGJIE ZHANG: "A Transfer Learning-Based High Impedance Fault Detection Method Under a Cloud-Edge Collaboration Framework", 《 IEEE ACCESS》, vol. 8, 8 September 2020 (2020-09-08), XP011809854, DOI: 10.1109/ACCESS.2020.3022639 * |
| 吴乃星: "《基于移动通信大数据的城市计算》", 31 January 2020, 华中科技大学出版社, pages: 77 * |
| 张文田;凌卫新;: "基于相似度的神经网络多源迁移学习算法", 科学技术与工程, no. 15, 28 May 2019 (2019-05-28) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20230094389A1 (en) | Quantum computing based deep learning for detection, diagnosis and other applications | |
| CN107766929B (en) | Model analysis method and device | |
| CN107067025A (en) | A kind of data automatic marking method based on Active Learning | |
| Huang et al. | Machine fault detection for intelligent self-driving networks | |
| Zhang et al. | A hybrid learning framework for imbalanced stream classification | |
| CN116628220A (en) | Electric power work ticket generation method and system based on relation graph convolution neural network | |
| CN113283602B (en) | Offline knowledge learning and online dynamic correction-based power grid system fault diagnosis method | |
| CN110441478A (en) | A kind of river ecological environmental data on-line monitoring method, system and storage medium | |
| CN112860918B (en) | Sequential knowledge graph representation learning method based on collaborative evolution modeling | |
| CN111860971B (en) | Method and device for predicting residual life of turnout switch machine | |
| CN116501444B (en) | Abnormal cloud edge collaborative monitoring and recovering system and method for virtual machine of intelligent network-connected automobile domain controller | |
| CN113825165B (en) | 5G slice network congestion early warning method and device based on time diagram network | |
| CN112149866A (en) | Intelligent manufacturing workshop anomaly prediction and control method based on edge cloud cooperation | |
| CN116402352A (en) | Enterprise risk prediction method and device, electronic equipment and medium | |
| Wang et al. | Partition cost-sensitive CART based on customer value for Telecom customer churn prediction | |
| CN108846489A (en) | A kind of large scale equipment health Evaluation method based on SVM | |
| CN117421596A (en) | Energy storage container fault early warning system and energy storage system | |
| Pesala et al. | Incremental learning vector auto regression for forecasting with edge devices | |
| CN115081483B (en) | Hydraulic generator rotor fault diagnosis method based on feature selection and GWO-BP | |
| CN116415958B (en) | Abnormal data detection method and device based on quantum technology and storage medium | |
| CN112801815B (en) | Power communication network fault early warning method based on federal learning | |
| Yu et al. | A particle filter and SVM integration framework for fault-proneness prediction in robot dead reckoning system | |
| Datta | Predicting bike-share usage patterns with machine learning | |
| CN115438190A (en) | Power distribution network fault decision-making assisting knowledge extraction method and system | |
| CN115630979A (en) | Day-ahead electricity price prediction method and device, storage medium and computer equipment |
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 |