CN109902379A - A kind of screening system and its screening technique of ternary solar battery active layer material - Google Patents
A kind of screening system and its screening technique of ternary solar battery active layer material Download PDFInfo
- Publication number
- CN109902379A CN109902379A CN201910140001.9A CN201910140001A CN109902379A CN 109902379 A CN109902379 A CN 109902379A CN 201910140001 A CN201910140001 A CN 201910140001A CN 109902379 A CN109902379 A CN 109902379A
- Authority
- CN
- China
- Prior art keywords
- data
- machine learning
- solar battery
- sample data
- active layer
- 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
Landscapes
- Photovoltaic Devices (AREA)
Abstract
The invention discloses the screening systems and its screening technique of a kind of ternary solar battery active layer material, it is related to organic semiconductor thin-film solar cell technical field, the present invention includes data acquisition unit: being used for sampler sample data, the sample data includes donor material in organic solar batteries, acceptor material and third component material characteristics information；Data analysis unit: carrying out data analysis to sample data collected, establishes simultaneously optimal prediction model；Program prediction unit: feasible experimental program is filtered out with the prediction model optimized, the present invention combines machine learning method with traditional experimental designs, its feature can be carried out intelligent classification by machine learning algorithm, and provide preliminary feasible scheme, more convenient and fast method is provided to the selection of ternary solar battery active layer material, play the role of Computer Aided Design in the experiment concept phase, intelligent recommends more rationally various experimental design point to experimental design person.
Description
Technical field
The present invention relates to organic semiconductor thin-film technical field of solar batteries, more particularly to a kind of ternary sun
The screening system and its screening technique of energy battery-active layer material.
Background technique
Twice after the industrial revolution, the fossil energies such as coal, petroleum, natural gas are used extensively, but are also exactly this quasi-fossil
A large amount of burnings of the energy bring a series of resource exhaustions, environmental degradation, atmosphere pollution etc. and seriously affect people's life and health
The problem of, therefore, develop a kind of reserves it is huge, cleaning, free of contamination renewable energy have become the extensive of today's society
Common recognition, research solar energy power generating, which solves energy crisis, becomes the emphasis and hot spot of field of renewable energy research.
In the research of solar battery, ternary organic solar batteries, because its manufacture craft is simple, device is frivolous,
The advantages that cost of manufacture is cheap receives significant attention in recent years, most important to be exactly in the research of ternary solar battery
Select correct active layer material.
In previous research work, researchers by read relative literature, then understand it is various can
After the characteristics of dopant material, feasible experimental program is tested on last Choice Theory, however, this research method,
Research cycle is very long, and selected experimental program is not necessarily exactly the side for keeping ternary solar battery active layer performance best
Case.
Summary of the invention
It is an object of the invention to: scientific research work is needed in order to solve the selection of current ternary solar battery active layer material
Authors read lot of documents, and workload is various, and the problem that research cycle is too long, and the present invention provides a kind of ternary solar energy
The screening system and its screening technique of battery-active layer material utilize the method screening ternary solar battery activity of machine learning
Layer material operation is simpler efficiently, facilitates experimental design procedure well.
The present invention specifically uses following technical scheme to achieve the goals above:
A kind of screening system of ternary solar battery active layer material, comprising:
Data acquisition unit: be used for sampler sample data, the sample data include in organic solar batteries to
Body material, acceptor material and third component material characteristics information；
Data analysis unit: carrying out data analysis to sample data collected, establishes simultaneously optimal prediction model；
Program prediction unit: feasible experimental program is filtered out with the prediction model optimized.
Further, the organic solar batteries include ternary solar battery and non-ternary solar battery, data
Donor material, acceptor material and third component material characteristics packet in acquisition unit organic solar batteries collected
It includes: lumo energy, HOMO energy level, absorption spectrum ranges, emission spectrum range, electron mobility and hole mobility.
Further, the data analysis unit includes:
Data sorting unit: for the collected sample data of institute to be divided into training dataset and test data set；
Data cleansing unit: for cleaning the sample data of training dataset and test data concentration；
Machine learning unit: machine is carried out to the sample data after cleaning as prediction model using GBDT machine learning model
Device study, and model evaluation is carried out to GBDT machine learning model.
Further, the sample data ratio of the training dataset and test data set is 9:1；What training data was concentrated
Sample data optimizes GBDT machine learning model parameter for training GBDT machine learning model；The sample that test data is concentrated
Data are used to test the precision of the GBDT machine learning model after training.
Further, the data cleansing unit is to shadows such as outlier, exceptional value, null value and repetition values in sample data
The abnormal data for ringing GBDT machine learning model precision is purged.
Further, the machine learning unit according to training data concentrate sample data training AUC/ROC value with
And accuracy rate, rate of precision and the recall rate of sample data test are concentrated to test data, GBDT machine learning model is carried out
Model evaluation.
Further, evaluation criteria is established to the GBDT machine learning model, according to evaluation criteria optimizing regulation GBDT
Machine learning model parameter obtains optimal fitting model.
Further, the program prediction unit is according to the feasibility of the preset experimental program of optimal fitting model evaluation.
A kind of screening technique of ternary solar battery active layer material, includes the following steps:
S1: ternary solar battery and non-ternary solar battery active layer material are obtained respectively using data acquisition unit
Each characteristic information as sample data；
S2: merging all sample datas and obtain data set, is pressed the sample data in data set using data sorting unit
Training dataset and test data set are randomly divided into according to ratio 9:1；
S3: it is cleaned, is removed using the sample data that data cleansing unit concentrates training dataset and test data
Abnormal data；
S4: machine learning is carried out to the training dataset sample data after cleaning using GBDT machine learning model, constantly
Adjust GBDT machine learning model parameter, iteration GBDT machine learning model, until the sample that the test data after cleaning is concentrated
The precision for the GBDT machine learning model that data test obtains is optimal, and obtains optimal fitting model；
S5: preset experimental program data are imported into optimal fitting model and carry out machine learning, filter out feasible experiment
Scheme.
Further, each characteristic information of ternary solar battery active layer material constitutes positive sample data in the S1
Each characteristic information of collection, non-ternary solar battery active layer material constitutes negative sample data set, the positive sample data set
It is identical as negative sample data set size, keep the harmony of positive negative sample.
Beneficial effects of the present invention are as follows:
1, the present invention combines machine learning method with traditional experimental designs, is ternary solar battery activity
The selection of layer material provides more convenient and fast system and method, plays good Computer Aided Design in the experiment concept phase
More rationally various experimental program is intelligently recommended in effect to experimental design person.
2, the present invention introduces GBDT machine learning mould when screening to ternary solar battery active layer material
Type, GBDT (Gradient Boosting Decision Tree) is a kind of decision Tree algorithms based on iteration, generalization ability
By force, accuracy rate is high, can be used for several scenes；Utilize the sample data composing training in positive sample data set and negative sample data set
Data set and test data set are trained GBDT machine learning model by training dataset, constantly update Optimized model
Parameter improves the reliability of selected experimental program.
Detailed description of the invention
Fig. 1 is system block diagram of the invention.
Fig. 2 is method flow schematic diagram of the invention.
Specific embodiment
In order to which those skilled in the art better understand the present invention, with reference to the accompanying drawing with following embodiment to the present invention
It is described in further detail.
Embodiment 1
As shown in Figure 1, the present embodiment provides a kind of screening system of ternary solar battery active layer material, including data
Acquisition unit, data analysis unit and program prediction unit:
Data acquisition unit: be used for sampler sample data, the sample data include in organic solar batteries to
Body material, acceptor material and third component material characteristics information；
Data analysis unit: carrying out data analysis to sample data collected, establishes simultaneously optimal prediction model；
Program prediction unit: feasible experimental program is filtered out with the prediction model optimized；
The organic solar batteries include ternary solar battery and non-ternary solar battery, data acquisition unit institute
In the organic solar batteries of acquisition donor material, acceptor material and third component material characteristics information include: lumo energy,
HOMO energy level, absorption spectrum ranges, emission spectrum range, electron mobility, hole mobility；
The data analysis unit includes:
Data sorting unit: for the collected sample data of institute to be divided into training dataset and test data set；
Data cleansing unit: for cleaning the sample data of training dataset and test data concentration, specifically to sample
The abnormal data that outlier, exceptional value, null value and repetition values in data etc. influence GBDT machine learning model precision carries out clear
It removes；
Machine learning unit: machine is carried out to the sample data after cleaning as prediction model using GBDT machine learning model
Device study, and model evaluation is carried out to GBDT machine learning model；
The sample data ratio of the training dataset and test data set is 9:1；The sample data that training data is concentrated
For training GBDT machine learning model, optimize GBDT machine learning model parameter；The sample data that test data is concentrated is used for
The precision of GBDT machine learning model after test training.
The machine learning unit is according to the AUC/ROC value for concentrating sample data training to training data and to test number
According to accuracy rate, rate of precision and the recall rate for concentrating sample data test, model evaluation is carried out to GBDT machine learning model, it is right
The GBDT machine learning model establishes evaluation criteria, according to evaluation criteria optimizing regulation GBDT machine learning model parameter, obtains
To optimal fitting model, program prediction unit is according to the feasibility of the preset experimental program of optimal fitting model evaluation.
As shown in Fig. 2, the present embodiment also provides a kind of screening technique of ternary solar battery active layer material, including such as
Lower step:
Step 1: acquisition data
S1: ternary solar battery and non-ternary solar battery active layer material are obtained respectively using data acquisition unit
Each characteristic information as sample data；
Each characteristic information of ternary solar battery active layer material constitutes positive sample data set, non-ternary solar-electricity
Each characteristic information of pond active layer material constitutes negative sample data set, includes that ternary solar battery is given in positive sample data set
Body material, the lumo energy of acceptor material and third component material, HOMO energy level, absorption spectrum ranges, emission spectrum range,
The characteristic informations such as electron mobility, hole mobility, the sample data in negative sample data set is according to non-ternary solar battery
Donor material, the lumo energy of acceptor material and third component material, HOMO energy level, absorption spectrum ranges, emission spectrum model
It encloses, electron mobility, hole mobility generate at random, independent assortment collocation data, so that negative sample data set and positive sample number
It is corresponding according to the feature of concentration, and positive sample data set is identical as negative sample data set size, keeps the equilibrium of positive negative sample
Property；
Step 2: analysis data
S2: merging all sample datas and obtain data set, is pressed the sample data in data set using data sorting unit
Training dataset and test data set are randomly divided into according to ratio 9:1；
S3: it is cleaned, is removed using the sample data that data cleansing unit concentrates training dataset and test data
Abnormal data；
S4: machine learning is carried out to the training dataset sample data after cleaning using GBDT machine learning model, constantly
Adjust GBDT machine learning model parameter, iteration GBDT machine learning model, until the sample that the test data after cleaning is concentrated
The precision for the GBDT machine learning model that data test obtains is optimal, and obtains optimal fitting model；
Step 3: program prediction
S5: the various experimental program data by preset hope as active layer material are fabricated to and concentrate with training data
The same format of characteristic information imports optimal fitting model and carries out machine learning, filters out feasible experimental program, assistant experiment
The proportion of the ternary organic solar batteries active layer material of suitable multiplicity is recommended in the design of scheme to experimenter.
The above, only presently preferred embodiments of the present invention, are not intended to limit the invention, patent protection model of the invention
It encloses and is subject to claims, it is all to change with equivalent structure made by specification and accompanying drawing content of the invention, similarly
It should be included within the scope of the present invention.
Claims (10)
1. a kind of screening system of ternary solar battery active layer material characterized by comprising
Data acquisition unit: it is used for sampler sample data, the sample data to include donor material in organic solar batteries
Material, acceptor material and third component material characteristics information；
Data analysis unit: carrying out data analysis to sample data collected, establishes simultaneously optimal prediction model；
Program prediction unit: feasible experimental program is filtered out with the prediction model optimized.
2. a kind of screening system of ternary solar battery active layer material according to claim 1, it is characterised in that: institute
Stating organic solar batteries includes ternary solar battery and non-ternary solar battery, and data acquisition unit is collected organic
In solar battery donor material, acceptor material and third component material characteristics information include: lumo energy, HOMO energy level,
Absorption spectrum ranges, emission spectrum range, electron mobility and hole mobility.
3. a kind of screening system of ternary solar battery active layer material according to claim 1, which is characterized in that institute
Stating data analysis unit includes:
Data sorting unit: for the collected sample data of institute to be divided into training dataset and test data set；
Data cleansing unit: for cleaning the sample data of training dataset and test data concentration；
Machine learning unit: engineering is carried out to the sample data after cleaning as prediction model using GBDT machine learning model
It practises, and model evaluation is carried out to GBDT machine learning model.
4. a kind of screening system of ternary solar battery active layer material according to claim 3, it is characterised in that: institute
The sample data ratio for stating training dataset and test data set is 9:1；The sample data that training data is concentrated is for training
GBDT machine learning model optimizes GBDT machine learning model parameter；The sample data that test data is concentrated is for testing training
The precision of GBDT machine learning model afterwards.
5. a kind of screening system of ternary solar battery active layer material according to claim 3, it is characterised in that: institute
Stating data cleansing unit influences GBDT machine learning model to outlier, exceptional value, null value and the repetition values etc. in sample data
The abnormal data of precision is purged.
6. a kind of screening system of ternary solar battery active layer material according to claim 3, it is characterised in that: institute
Machine learning unit is stated according to the AUC/ROC value for concentrating sample data training to training data and sample is concentrated to test data
Accuracy rate, rate of precision and the recall rate of data test carry out model evaluation to GBDT machine learning model.
7. a kind of screening system of ternary solar battery active layer material according to claim 3, it is characterised in that: right
The GBDT machine learning model establishes evaluation criteria, according to evaluation criteria optimizing regulation GBDT machine learning model parameter, obtains
To optimal fitting model.
8. a kind of screening system of ternary solar battery active layer material according to claim 7, it is characterised in that: institute
Program prediction unit is stated according to the feasibility of the preset experimental program of optimal fitting model evaluation.
9. a kind of screening technique of ternary solar battery active layer material, which comprises the steps of:
S1: each of ternary solar battery and non-ternary solar battery active layer material is obtained respectively using data acquisition unit
A characteristic information is as sample data；
S2: merging all sample datas and obtain data set, using data sorting unit by the sample data in data set according to than
Example 9:1 is randomly divided into training dataset and test data set；
S3: being cleaned using the sample data that data cleansing unit concentrates training dataset and test data, is removed abnormal
Data；
S4: machine learning is carried out to the training dataset sample data after cleaning using GBDT machine learning model, is constantly adjusted
GBDT machine learning model parameter, iteration GBDT machine learning model, until the sample data that the test data after cleaning is concentrated
The precision for testing obtained GBDT machine learning model is optimal, and obtains optimal fitting model；
S5: preset experimental program data are imported into optimal fitting model and carry out machine learning, filter out feasible experimental program.
10. a kind of screening technique of ternary solar battery active layer material according to claim 9, it is characterised in that:
Each characteristic information of ternary solar battery active layer material constitutes positive sample data set, non-ternary solar-electricity in the S1
Each characteristic information of pond active layer material constitutes negative sample data set, the positive sample data set and negative sample data set size
It is identical.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910140001.9A CN109902379A (en) | 2019-02-26 | 2019-02-26 | A kind of screening system and its screening technique of ternary solar battery active layer material |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910140001.9A CN109902379A (en) | 2019-02-26 | 2019-02-26 | A kind of screening system and its screening technique of ternary solar battery active layer material |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109902379A true CN109902379A (en) | 2019-06-18 |
Family
ID=66945307
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910140001.9A Pending CN109902379A (en) | 2019-02-26 | 2019-02-26 | A kind of screening system and its screening technique of ternary solar battery active layer material |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109902379A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112466416A (en) * | 2020-11-03 | 2021-03-09 | 北京科技大学 | Material data cleaning method combined with prior knowledge of nickel-based alloy |
| CN113919576A (en) * | 2021-10-15 | 2022-01-11 | 厦门大学 | Method for predicting structural performance of solar cell |
| CN114094150A (en) * | 2020-06-29 | 2022-02-25 | 中国科学院大连化学物理研究所 | Method for predicting optimal operation condition of flow battery cell stack based on machine learning |
| CN114789557B (en) * | 2022-05-16 | 2023-07-25 | 马鞍山嘉兰智造科技有限公司 | Multi-nozzle printing method of 3d printer |
| CN117558380A (en) * | 2024-01-10 | 2024-02-13 | 中国科学院深圳先进技术研究院 | High-flux preparation method and system of magnetic micro-nano material based on artificial intelligence algorithm |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010182287A (en) * | 2008-07-17 | 2010-08-19 | Steven C Kays | Intelligent adaptive design |
| CN104755930A (en) * | 2012-03-26 | 2015-07-01 | 埃克塞尔柯尔有限责任公司 | Device and method for detection of analytes |
| US20160045841A1 (en) * | 2013-03-15 | 2016-02-18 | Transtar Group, Ltd. | New and improved system for processing various chemicals and materials |
-
2019
- 2019-02-26 CN CN201910140001.9A patent/CN109902379A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010182287A (en) * | 2008-07-17 | 2010-08-19 | Steven C Kays | Intelligent adaptive design |
| CN104755930A (en) * | 2012-03-26 | 2015-07-01 | 埃克塞尔柯尔有限责任公司 | Device and method for detection of analytes |
| US20160045841A1 (en) * | 2013-03-15 | 2016-02-18 | Transtar Group, Ltd. | New and improved system for processing various chemicals and materials |
Non-Patent Citations (3)
| Title |
|---|
| STEVEN A.LOPEZ等: ""Design Principles and Top Non-Fullerene Acceptor Candidates for Organic Photovoltaics"", 《JOULE》 * |
| YINGYING YUAN等: ""A thermally activated delayed fluorescence material for efficient ternary organic solar cells"", 《CONFERENCE: OPTOELECTRONIC MATERIALS AND DEVICES FOR SENSING AND IMAGING》 * |
| 程运华: ""硼、氮掺杂富勒烯的结构搜索"", 《中国优秀博硕士学位论文全文数据库(博士) 工程科技Ⅰ辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114094150A (en) * | 2020-06-29 | 2022-02-25 | 中国科学院大连化学物理研究所 | Method for predicting optimal operation condition of flow battery cell stack based on machine learning |
| CN114094150B (en) * | 2020-06-29 | 2023-11-17 | 中国科学院大连化学物理研究所 | Machine learning-based flow battery pile optimal operation condition prediction method |
| CN112466416A (en) * | 2020-11-03 | 2021-03-09 | 北京科技大学 | Material data cleaning method combined with prior knowledge of nickel-based alloy |
| CN112466416B (en) * | 2020-11-03 | 2024-04-12 | 北京科技大学 | Material data cleaning method combining nickel-based alloy priori knowledge |
| CN113919576A (en) * | 2021-10-15 | 2022-01-11 | 厦门大学 | Method for predicting structural performance of solar cell |
| CN114789557B (en) * | 2022-05-16 | 2023-07-25 | 马鞍山嘉兰智造科技有限公司 | Multi-nozzle printing method of 3d printer |
| CN117558380A (en) * | 2024-01-10 | 2024-02-13 | 中国科学院深圳先进技术研究院 | High-flux preparation method and system of magnetic micro-nano material based on artificial intelligence algorithm |
| CN117558380B (en) * | 2024-01-10 | 2024-04-09 | 中国科学院深圳先进技术研究院 | High-flux preparation method and system of magnetic micro-nano material based on artificial intelligence algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109902379A (en) | A kind of screening system and its screening technique of ternary solar battery active layer material | |
| Mirza et al. | High-efficiency swarm intelligent maximum power point tracking control techniques for varying temperature and irradiance | |
| CN109873610B (en) | Photovoltaic array fault diagnosis method based on IV characteristic and depth residual error network | |
| CN111008728B (en) | Prediction method for short-term output of distributed photovoltaic power generation system | |
| CN108983749B (en) | Photovoltaic array fault diagnosis method based on K-SVD training sparse dictionary | |
| CN104978613B (en) | A kind of photovoltaic output short term prediction method considering component temperature | |
| CN109711609A (en) | Photovoltaic plant output power predicting method based on wavelet transformation and extreme learning machine | |
| CN109635506A (en) | The photovoltaic cell model parameter identification method of adaptive chaos tree and seed algorithm | |
| CN112395289B (en) | Distributed photovoltaic data layered storage method and system | |
| CN111103477B (en) | Electrical load monitoring method and system based on 3D convolutional neural network | |
| CN105574612A (en) | Photovoltaic power generation prediction method based on data mining | |
| CN105512775A (en) | Method for power prediction of photovoltaic power generation system | |
| Merchaoui et al. | Improved fast particle swarm optimization based PV MPPT | |
| García-Vico et al. | Analysing concentrating photovoltaics technology through the use of emerging pattern mining | |
| CN107562992A (en) | A kind of Photovoltaic array maximum power tracking method based on SVM and particle cluster algorithm | |
| CN103293950B (en) | The control method that a kind of maximum photovoltaic power point based on LSSVM is followed the tracks of | |
| Mansour et al. | An efficient intelligent power detection method for photovoltaic system | |
| CN115185334A (en) | Photovoltaic MPPT method based on GRU neural network | |
| Senthilkumar et al. | Brief review on solar photovoltaic parameter estimation of single and double diode model using evolutionary algorithms | |
| Wang et al. | Digital twin based maximum power point estimation for photovoltaic systems | |
| CN113837434A (en) | Solar photovoltaic power generation prediction method and device, electronic equipment and storage medium | |
| Hadji et al. | Experimental analysis of genetic algorithms based MPPT for PV systems | |
| Ahmed et al. | Non-iterative MPPT Method: A Comparative Study | |
| CN111814399A (en) | Model parameter optimization extraction method and measurement data prediction method for solar photovoltaic cell system | |
| CN112084703A (en) | Photovoltaic cell system model parameter identification method based on variant shuffling frog leaping algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190618 |
|
| RJ01 | Rejection of invention patent application after publication |