A method and system for constructing a ceramic membrane material database and a computer readable medium

Abstract

The application relates to a ceramic membrane material database construction method, system and computer readable medium, and belongs to the technical field of big data. The method comprises the following steps: step 1, classifying the technology in the research direction of the ceramic membrane material, and determining the keywords in each classification; step 2, according to the determined keywords, retrieving and obtaining the bibliographic information and full-text information of the literature through a database; step 3, constructing a literature classifier through a machine learning method, classifying the literature obtained in step 2 according to the classification in step 1, and indexing the data information; step 4, analyzing the data obtained in step 3, and giving an analysis result. The patent can classify the technical literature in the ceramic membrane field, obtain the literature information in each research direction, classify and analyze the literature through a machine learning method, give an analysis result, help to summarize the existing literature data, and improve the research and development efficiency.

Description

TECHNICAL FIELD

[0001] The application relates to a ceramic membrane material database construction method and system and a computer readable medium, and belongs to the technical field of big data. BACKGROUND

[0002] Reaction and separation are the core of the petrochemical industry, and how to improve the efficiency of the reaction and separation process has always been a hot and cutting-edge topic in the petrochemical industry and research field. Membrane separation technology with high-performance membrane materials as the core has the characteristics of high separation efficiency and low energy consumption, and has become one of the key supporting technologies for the green development of the petrochemical industry. High-performance membrane materials are typical new chemical materials, among which ceramic membrane materials are porous membrane materials formed by high-temperature sintering of inorganic materials such as aluminum oxide, titanium oxide and zirconium oxide, and have the characteristics of acid and alkali resistance, solvent resistance, high-temperature resistance, high mechanical strength and long service life, and are an ideal separation material for process industries such as petroleum chemical industry.

[0003] The research on ceramic membranes began in the 1940s and was initially used for the separation of uranium isotopes in the nuclear industry. In the 1980s, it began to be used in civilian fields, mainly for the separation of liquids such as milk, wine and beer.

[0004] With the development of material genetic engineering and data technology, the research and development of new materials has entered the fourth paradigm of "data-driven". Applying the research method of material genetic engineering, through "data-driven", it is expected to become a new breakthrough point in the field of ceramic membrane materials. With the continuous development of ceramic membrane technology, its preparation technology and application technology data are continuously accumulated. How to collect a large amount of diversified data from literature and process the data with unified structured rules to complete the construction of the ceramic membrane material database is a basic problem to be solved in the data-driven material research and development mode.

[0005] At the same time, with the continuous accumulation of research results, the number of literature on ceramic membranes and their application technology is also increasing. In the vast amount of literature, there are many research directions and research results. For example, to solve the anti-pollution performance of ceramic membranes, a large number of studies have been conducted in the current literature, and the solutions include at least surface organic modification, doping modification, adjustment of membrane pore size, etc. The performance of the ceramic membranes prepared by the above methods is also quite different. In the subsequent research process, if a better research approach is needed, a large amount of technical literature needs to be read. Therefore, how to analyze important research directions and research approaches from a large number of technical data can significantly improve the efficiency of scientific research and promote the generation of innovative achievements. SUMMARY

[0006] The technical problem to be solved by the present application is how to establish a database according to the existing literature on ceramic membranes, and further analyze the research results of materials according to the database information and give the results.

[0007] A method for constructing a ceramic membrane material database, comprising the following steps:

[0008] Step 1, classify the technology in the research direction of ceramic membrane materials, and determine the keywords in each classification;

[0009] Step 2, according to the determined keywords, retrieve and obtain the title and full text information of the literature through the database;

[0010] Step 3, construct a literature classifier by machine learning method, classify the literature obtained in step 2 according to the classification in step 1, and index the data information;

[0011] Step 4, analyze the data obtained in step 3 and give the analysis results.

[0012] In step 1, the classification obtained includes ceramic membrane preparation technology, ceramic membrane element, and ceramic membrane application technology; each classification can be a first-level classification, or can include multiple levels of classification.

[0013] In step 2, the database includes at least one or more of journal database, degree thesis database, conference paper database, and patent database.

[0014] In step 3, the classifier is selected from support vector machine, neural network, K nearest neighbor, naive Bayes, deep learning, etc.

[0015] In step 4, the process of analyzing the data includes: classifying the literature under a classification according to the data indexing information, obtaining the research direction classification; extracting the improvement results of each research direction from the full text of the literature, and outputting the analysis results according to the improvement result degree.

[0016] The research direction includes one or more of ceramic membrane anti-pollution performance, ceramic membrane strength, interception performance, and membrane defect repair.

[0017] For the analysis of the ceramic membrane anti-pollution performance of the ceramic membrane, the following steps are further included:

[0018] Step 4-1, obtain the full text and image data in the literature, and identify the membrane filtration flux change curve;

[0019] Step 4-2, binarize the curve and identify the horizontal and vertical coordinate axes and multiple flux curves.

[0020] Step 4-3, identifying the maximum and minimum points on the ordinate axis, calculating the scale R between the coordinate value and the pixel value according to the character position of the maximum value and the minimum value;

[0021] Step 4-4, according to the scale R and the pixel point coordinate, obtaining the ordinate axis value of the first point and the last point of each curve in the abscissa axis direction;

[0022] Step 4-5, calculating the performance improvement rate R of the technical improvement, wherein:

[0023]

[0024] V0 l and Vf l are the ordinate coordinates of the first point and the last point of the curve in the lower position in the image, V0 u and Vf u are the ordinate coordinates of the first point and the last point of the curve in the higher position in the image;

[0025] Step 4-6, sorting the performance improvement rates obtained by the research and development of each research direction, and giving the analysis results.

[0026] In step 4-6, it further includes the step of data classification of the performance improvement rates obtained by the research and development of each research direction:

[0027] Step 4-6-1, obtaining the text content below the curve graph, and identifying the drawing number corresponding to the curve graph;

[0028] Step 4-6-2, searching for the text position in the full text of the literature which refers to the drawing, performing semantic recognition on the context text content of the text position, and finding the words related to the ceramic membrane application system;

[0029] Step 4-6-3, indexing the literature with the obtained application system words, clustering the literature data in step 4-6 according to the application system words, sorting the performance improvement rates of each class, and giving the analysis results.

[0030] The application also provides a ceramic membrane material database construction system, comprising:

[0031] A keyword input module is used for entering the keywords in each category classified and determined by the research direction of the ceramic membrane material;

[0032] A literature retrieval and acquisition module is used for retrieving and acquiring the title and full text information of the literature through the database according to the entered keywords;

[0033] The literature classification module is used for constructing a literature classifier through a machine learning method, classifying the obtained literature according to a technical classification, and indexing data information;

[0034] The analysis module is used for analyzing the data obtained by the literature classification module and giving an analysis result.

[0035] The application further provides a computer readable medium, which records a program capable of running the construction method of the ceramic membrane material database.

[0036] The application further provides a computer device, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor realizes the construction method of the ceramic membrane material database when executing the program.

[0037] The ceramic membrane material database can classify the technical literature in the field of ceramic membranes, obtain literature information in each research direction, classify and analyze the literature through a machine learning method, give an analysis result, help to summarize the existing literature data, and improve the research and development efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 It is the construction method flowchart of the database in the patent;

[0039] Figure 2 It is the visual display effect of the literature database;

[0040] Figure 3 It is the visual display effect of the literature database;

[0041] Figure 4 It is some typical membrane flux curve graphs;

[0042] Figure 5 It is a schematic diagram of the image recognition process of the membrane flux curve graph;

[0043] Figure 6 It is the construction flowchart of the method of the application;

[0044] Figure 7 It is the computer system diagram of the application. DETAILED DESCRIPTION

[0045] In order to make the purposes, technical schemes and advantages of the application clearer, the application is further described in detail below with reference to the drawings and examples. It should be understood that the specific examples described herein are only used to explain the application, and are not used to limit the application.

[0046] Reference to“an embodiment” herein means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase“in one embodiment” in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily all referring to a common set of embodiments, although they can. It is explicitly contemplated that embodiments described herein can be combined to create modifications that are not explicitly presented.

[0047] The flow of the construction data of the patent is shown as Figure 1 The present application is based on dynamic container data storage technology, and different data templates are established for different ceramic membrane data sources and data types. According to the characteristics of ceramic membrane materials, process characteristics and application scenarios, the multi-source data of ceramic membrane materials are classified and collected. According to the data template, the uploading and modeless storage of standard data are realized. Based on the knowledge framework and correlation of ceramic membrane materials, through big data and machine learning technology, the collected ceramic material data are analyzed, the high-order logical relationship between the data is established, and the key support for ceramic membrane material data annotation and knowledge graph is formed. Combined with the semantic and image recognition technology of artificial intelligence, the data of ceramic membrane material preparation process, performance and process application are efficiently extracted from professional literature such as ceramic membrane material related journals, patents, books and ceramic membrane material product public data, and the real-time update and supplement of ceramic membrane material database are realized.

[0048] In the construction method of the patent, first, the literature related to ceramic membranes needs to be searched and processed to obtain initial data. In this step, the technical route and keywords are determined. The technical links related to the technology are mainly divided into: ceramic membrane preparation technology (subdivided into membrane materials and preparation methods), ceramic membrane application, and ceramic membrane element and membrane process.

[0049] After preliminary collection and analysis of the literature materials in the field of ceramic membrane materials, the typical materials, preparation process and application fields of ceramic membrane materials are technically subdivided, and the technical decomposition table as shown in Table 1 is formed.

[0050] Table 1 Technical decomposition table of ceramic membrane preparation and application

[0051]

[0052]

[0053] "Ceramic membrane" is the core keyword of the search, containing "ceramic" and "membrane" two key elements. But in the field of ceramic membrane research and application, often use the specific name instead of "ceramic" element, such as: alumina membrane, titanium oxide membrane, etc. If the "ceramic membrane" is directly used as the keyword search, will cause serious missed. Thus, according to the decomposition of ceramic membrane preparation and application technology in table 1, the retrieval key words are combed, as shown in table 2.

[0054] Table 2 ceramic membrane search element table

[0055]

[0056] For material elements, "ceramic" is the core keyword, but not a necessary condition. It should be used "or" (+, or) as the search operator between "alumina", "titanium oxide" and other keywords. Because in Chinese environment often use chemical formula to express materials, so in the search formula also need to increase "Al2O3", "TiO2" and other keywords. At the same time, considering that the material elements are usually reflected in the paper title (TI) or keywords (KY).

[0057] The precision and recall rate are two key indicators in literature retrieval, improving the precision can reduce the influence of useless information (noise) on subsequent analysis, and obtaining high recall rate can guarantee the comprehensiveness of the result analysis, there is a dialectical relationship between the two. The selection and refinement strategy of subject headings is as follows:(1) for the core keyword of the search "ceramic membrane", reference the subject classification of the literature database search results, classify the literature and the corresponding technology, collect the keywords with higher frequency.(2) in the field of ceramic membrane research and application, often use the specific name of a ceramic material, such as: alumina membrane, titanium oxide membrane, etc. In order to improve the recall rate, the commonly used material name and chemical formula are supplemented in the keywords.(3) in Chinese context, "membrane" is also often used to express the name of non separation function thin film materials. In order to exclude this kind of interference, the function characteristics of separation membrane materials are used to limit "membrane", and the limiting keywords such as separation, filtration and purification are added.(4) some non ceramic membrane materials often use ceramic membrane as carrier, or add ceramic materials to enhance the performance of the membrane. By introducing the name of typical non ceramic membrane materials, the search results are refined, and the exclusion keywords such as organic membrane, PVDF, PTFE and PVA are added.(5) in this search, perovskite membrane and molecular sieve membrane are not included in the analysis object. Related feature keywords are not included in the professional search formula, or are excluded in the professional search formula. According to the above search strategy, the keywords (KY) and title (TI) are selected as the search area, and the keywords are combined by logical operators and (and), or (or) and not (not) to obtain the following professional search formula:

[0058] ((((KY or TI) = ceramic + alumina + aluminum oxide + corundum + titanium oxide + titanium dioxide + zirconium oxide + zirconium dioxide + silicon oxide + silicon dioxide + yttrium oxide + yttrium trioxide + hafnium oxide + hafnium dioxide + iron oxide + iron trioxide + aluminum nitride + boron nitride + silicon nitride + silicon carbide + titanium carbide + mullite + cordierite + kaolin + bentonite + bauxite + attapulgite + fly ash + clay + AI2O3 + TiO2 + ZrO2 + SiO2 + Y2O3 + HfO2 + Fe2O3 + AlN + BN + Si3N4 + SiC + MXene) and ((KY or TI) = membrane) and ((KY or TI or AB) = separation + filtration + purification + impurity removal + concentration + recovery + treatment + permeation + flux + contactor + reactor)) or ((KY or TI) = ceramic membrane)) not (((Ky or TI) = polyvinylidene fluoride + polytetrafluoroethylene + polyether + polysulfone + cellulose acetate + polyamide + polyimide + polyvinyl alcohol + polylactic acid + polyethylene + polypropylene + polyurethane + polyacrylic acid + polydimethylsiloxane + glutaraldehyde + chitosan + cellulose + exchange + bipolar + fiber + graphene + polymer + macromolecule + resin + organic + permeation membrane + hybrid + blending + PVDF + PTFE + PES + PI + PAA + PVC + PLA + PE + PP + PAN + PET + PDMS + GO + CNT + CS + AC + nanotube + steel + palladium + alloy + metal + molecular sieve + zeolite + carbon + carbon + ZIF + COF + MOF + pervaporation + nanocomposite + biofilm + biofilm formation + film coating) not ((KY or TI) = ceramic membrane)) not ((KY or TI) = coating + optical + film coating + film coating + aluminum alloy + magnesium alloy + titanium alloy) not ((AB = thin film) not (AB = ceramic membrane)) not (AB = conversion film + chemical film + semiconductor + tooth + sensor + membrane electrode + antireflection film + high reflection film + antireflection film + reflection film + fresh-keeping + friction + micro-arc oxidation + cathodic arc + sputtering + infiltration + porcelain + electrodeposition + electrolytic deposition + blow molding + stretching + corrosion prevention)

[0059] As an example, we used CNKI as the database, and removed the check mark of the "Chinese-English expansion" option. We used the professional search formula to search the database. The time range was from the recorded date to 2021, and 2427 search results were obtained. After removing the newspaper, conference abstract, and review, 1774 target literatures were obtained. It is worth noting that when using the export function of CNKI, we need to select the appropriate result output format to match the subsequent analysis software. VOSviewer software supports the direct import of Refworks and Endnote files; while Citespace needs to convert the Refworks or Endnote format file to WOS format through the "Import / Export" function in the "Data" menu, and then import the WOS format file. The results of the preliminary data are shown in Figure 2 .

[0060] For the English journal data, a corresponding strategy can be applied, setting a search expression as follows: (AB=membrane AND TS=(ceramic OR aluminum oxide OR aluminium oxide OR corundum OR zirconia OR zirconium oxide OR zirconium dioxide OR zircite OR titania OR Titanium oxide OR titanium dioxide OR silica OR silicon oxide OR yttria OR yttrium oxide OR hafnia OR haf-nium oxide OR ferric oxide OR iron oxide OR aluminium nitride OR boron nitride OR borazon OR silicon nitride OR carborundum OR silicon carbide OR carbofrax OR nicalon OR titanium carbide OR mullite OR cordierite OR silicon OR kaolin OR bentonite OR bauxite OR attapulgite OR portland OR apatite OR quartz OR flay ash OR clay OR Al2O3 OR TiO2 OR ZrO2 OR SiO2 OR Y2O3 OR HfO2 OR Fe2O3 OR AlN OR BN OR Si3N4 OR SiC OR TiC OR MXene) AND AB=(microfiltration OR ultrafiltration OR nanofiltration OR diafiltration OR MF OR UF OR NF pore OR porous OR porosity OR separate* OR filtrat* OR purificat* OR concentrat* OR recycl* OR flux OR contactor OR reactor)) NOT (TI=(PVDF OR PTFE OR PES OR PI OR PVA OR PAA OR PVC OR PLA OR PE OR PP ORPAN OR PET OR PDMS OR Pd membrane OR zeolite OR ZIF*OR COF*OR MOF*OR hybrid membrane OR pervaporation OR mixed matrix)OR TS=(PVDF OR PTFE OR PES OR PEI OR PI OR PVA OR PAA OR PVC OR PLA OR PE OR PP OR PAN OR PET OR PDMS OR Pd membrane OR molecular sieve OR zeolite OR ZIF*OR COF*OR MOF*OR hybrid membrane OR pervaporation OR mixed matrix)) AND (TI=membrane AND TI=(ceramic OR aluminum oxide OR aluminium oxide OR corundum OR zirconia OR zirconium oxide OR zirconium dioxide OR zircite OR tita-nia OR titanium oxide OR titanium dioxide OR silica OR silicon oxide OR yttria OR yttrium oxide OR hafnia OR hafnium oxide OR ferric oxide OR iron oxide OR aluminium nitride OR boron nitride OR borazon OR silicon nitride OR carborundum OR silicon carbide OR carbofrax OR nicalon OR titanium carbide OR mullite OR cordierite OR silicon OR kaolin OR bentonite OR bauxite OR attapulgite OR portland OR apatite OR quartz OR flay ash OR clay OR Al2O3 OR TiO2 OR ZrO2 OR SiO2 OR Y2O3 OR HfO2 OR Fe2O3 OR AlN OR BN OR Si3N4 ORSiC OR TiC OR MXene) NOT TI=(PVDF OR PTFE OR PES OR PI OR PVA OR PAA OR PVC OR PLA OR PE OR PP OR PAN OR PET OR PDMS OR Pd membrane OR zeolite OR ZIF* OR COF* OR MOF* OR hybrid membrane OR pervaporation OR mixed matrix OR film OR hybrid membrane OR pervaporation OR mixed matrix)).

[0061] In an embodiment, after obtaining the corresponding search information, the obtained literature information can be preliminarily classified, and preliminary data statistics can be performed in terms of application amount, applicant classification, and technical category. The software also has a display function of analysis and statistical information, which is used to display the preliminary statistical results of the literature title information in a graphical manner, as shown in FIGS. 1 to 3. Figure 2 , Figure 3

[0062] In an embodiment, for the above literature data search process, an existing literature database can be selected for searching and data acquisition. The literature database can be CNKI, ISI Web of Science, Tongfang, and Elsevier ScienceDirect database.

[0063] In an embodiment, after obtaining the above data, the data needs to be classified. The classification can be performed by modeling and classifying the text data of the title and abstract. Based on some basic methods of natural language processing, the existing machine learning model can be used for processing. That is, the obtained literature information can be classified according to the research content. For example, a support vector machine, a neural network, a K nearest neighbor, a naive Bayes, and a deep learning classifier can be used for processing. After the training set data is used to train the model and the classification effect is verified by the verification data, the classifier model is constructed when the acceptable accuracy is reached. The construction of the classifier can be performed by using the existing technology. For example, reference can be made to the related prior art literature (Wang Hao, Ye Peng, Deng Sanhong. Application of machine learning in automatic classification of Chinese journal papers [J]. Data analysis and knowledge discovery, 2014, 30(3): 80-87, Ye Peng. Automatic classification of Chinese journal papers based on machine learning. Nanjing University.).

[0064] ​In one embodiment, after the literature retrieval and the title information reading, the full text of the literature can be further obtained. The relevant PDF format file can be automatically obtained, and the relevant text and image data in the PDF format literature can be obtained by using the prior art method, for example, the pdf file is parsed by using the PDF Parser, the object parsed by the PDF Parser is saved by using the PDF Document, and the page content of the parsed document is processed by using the PDF Page Interpreter.

[0065] In one embodiment, in order to further analyze the research results, the patent also extracts and processes the content of the obtained literature text part. In the research process of ceramic membranes, various technical routes in different directions are reported, for example, repairing the large pore defects of ceramic membranes, improving the surface hydrophilicity of ceramic membranes, improving the surface hydrophobicity of ceramic membranes, improving the molecular weight cut-off of ceramic membranes, modifying intelligent groups to membranes, improving the strength of ceramic membranes, etc. In each direction, there are also more research routes. Taking the improvement of the anti-pollution performance of ceramic membranes as an example, the improvement can be carried out in various ways, including the modification of the surface of ceramic membranes, the addition of some nano materials in the preparation of the separation layer of ceramic membranes, and the control of the surface roughness to improve the anti-pollution performance of ceramic membranes, etc. After obtaining these statistical data, in order to save the time input of researchers and obtain the research direction with significant performance improvement as soon as possible, the patent also analyzes the technical achievements in different directions and gives the better research ideas.

[0066] More specific data analysis and processing methods are as follows:

[0067] For the anti-pollution performance of ceramic membranes, the filtration flux data is usually used for characterization test, and the flux attenuation curve is usually present in the literature, and the performance curve of the ceramic membrane prepared by the improved method is given in the graph, and the performance curve of the ceramic membrane prepared by the conventional method is also given, such as Figure 4As shown in Figure a (Li Kun, Xin Jiaqi, Ding Ziyao, et al. Antifouling performance of ceramic ultrafiltration membranes modified with photocatalysts [J]. Journal of Nanchang University: Science Edition, 2021, 45(5): 9.), the figure contains two flux decay curves. Since the filtration flux curve has high image characteristics, it is easy to distinguish it from other pictures in the article. The article in this journal used ceramic ultrafiltration membranes modified with photocatalysts to improve antifouling performance. Similarly, in some other literature, roughness control methods were used to solve the antifouling performance of ceramic membranes in filtering oil-water mixtures (Zhang Bingbing, Zhong Zhaoxiang, Xing Weihong. Influence of surface roughness of ceramic membrane on the filtration performance of oily wastewater [J]. Membrane Science and Technology, 2011, 31(4): 6.), and there are also methods to improve the antifouling performance of ceramic membranes by modifying the surface of ceramic membranes with titanium oxide nanoparticles (Li Kun. Study on the performance of ceramic membrane surface modification based on titanium oxide material for membrane fouling control [D]. Beijing University of Chemical Technology, 2012.). The figures are as follows. Figure 4 As shown in b and c.

[0068] The above figures have their own distinctive characteristics, such as Figure 5 As shown, in one embodiment, after acquiring the image from the PDF file, the throughput data image can be identified using a common image classifier. Next, by performing edge detection on the binarized image, the horizontal and vertical coordinate axes can be identified, and the edge lines within the coordinate axis range can be identified, allowing for the simultaneous identification of multiple curves in the image. Next, the maximum and minimum values ​​on the vertical coordinate axis are identified using an OCR module, and the scale R between the coordinate values ​​and pixel values ​​can be calculated based on the character positions of the maximum and minimum values. Then, the first point Point0 on the horizontal axis of each curve is calculated. i And the last point Pointf i The vertical coordinate (x0) i y0 i ), (xf i yf i ), where i represents the curve number, and Point0 can be calculated based on the scale R. i and Pointf i The value V0 on the vertical axis i and Vf i This leads to the flux improvement rate R;

[0069]

[0070] In the formula, V0 l and Vf lV0 represents the horizontal coordinates of the first and last points on the lower curve in the image, along with their vertical coordinates. u and Vf u These are the horizontal coordinates of the first point and the vertical coordinates of the last point on the curve that is higher in the image. Since the degree of membrane fouling can be judged by the magnitude of the curve's descent, the curve representing good antifouling performance will be relatively flat and located at a higher position in the image. Conversely, for ceramic membranes with severe fouling, the curve is located closer to the bottom of the image and has a larger descent. The R-value can be calculated to measure the degree to which flux attenuation is suppressed after technological improvements. A larger R-value indicates less attenuation and better antifouling performance. The above image processing can be implemented using existing OpenCV tools. The attached figure can contain only two curves; generally, the higher curve represents the improved membrane material performance, and the lower curve represents the unimproved membrane material performance. Alternatively, the attached figure can contain multiple curves. Similarly, the highest-positioned curve can be defined as the optimal membrane performance obtained under this method, without considering other curves besides the highest and lowest, to calculate the membrane performance improvement rate obtained by this research method.

[0071] In one embodiment, after processing the literature on improving the antifouling performance of ceramic membranes obtained from the database into text and images according to the above method, and after extracting and analyzing the text and image data, the improvement rate of the antifouling performance of ceramic membranes obtained by different improvement approaches can be obtained. The improvement rate of ceramic membrane performance by each method is related to the main approach of the article. Further clustering of the data can also improve the presentation of the analyzed data. This clustering can be performed using NLP methods or based on a pre-defined terminology classification table. For example, organic grafting modification improvement schemes can be grouped into one category, and nanoparticle doping technologies can be grouped into another. This patent does not limit this approach.

[0072] Furthermore, since the properties of the liquid used in ceramic membrane filtration experiments also have a certain impact on membrane fouling, in one embodiment, the filtration system data in the literature can be further indexed. Specifically, when identifying the flux decay curve image in the literature, the text below the image is identified, and the figure number is extracted. Since the figure number is usually located below the image and begins with "...", this approach can be used to further index the filtration system data in the literature. Figure 1"Fig.1", "Figure.1", and the like, so it is easier to extract the figure number from the habit of using the figure number, and then search for the text position of the figure number in the text. In scientific literature, the experimental part of the figure is usually analyzed and reviewed, so the position of the image number is found in the upper and lower paragraphs by semantic recognition. The context here can be 50, 100, 150 words before and after the figure number position, which can be adjusted manually according to the recognition results. This patent does not make specific limitations. The name of the system can reflect the filter system in the experiment corresponding to the image. After obtaining the system name, the relevant literature data is labeled and indexed, which can further classify the anti-pollution performance and improve the analysis accuracy of the results. For example, some specific keywords in the context can be extracted, such as "bovine milk protein", "BSA", "dye wastewater", "biochemical effluent", etc. These words can be obtained from the vocabulary or manually set.

[0073] Based on the above method, other related research directions in ceramic membrane literature can also be data, image, and text grabbing and analysis. For example, the performance of the membrane in different application scenarios can be classified, sorted and analyzed, and the improvement points in different membrane processes can be given to analyze the research results of improving the performance of the membrane. The research results of the membrane equipment and components can also be sorted.

[0074] The above-mentioned various modules in the ceramic membrane database construction device can be realized by software, hardware and their combinations. The above-mentioned various modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so that the processor can call and execute the operations of the above-mentioned various modules.

[0075] As Figure 7As shown, in one embodiment, a computer device is provided, which can be a terminal. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The network interface of the computer device is used to communicate with external terminals through network connection. The computer program is executed by the processor to implement the construction method of the ceramic membrane database. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc. Those skilled in the art can understand that the above computer system structure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0076] Based on the above examples, in one embodiment, a computer device is also provided, which includes a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to implement the construction method of the ceramic membrane material database of any of the above embodiments.

[0077] Those skilled in the art can understand that all or part of the processes in the above embodiments can be completed by a computer program instructing related hardware. The program can be stored in a non-volatile computer readable storage medium, such as a storage medium of a computer system, and executed by at least one processor in the computer system to implement the processes of the above embodiments for the construction method of the ceramic membrane database.

[0078] Accordingly, in one embodiment, a computer storage medium computer readable storage medium is also provided, which stores a computer program, wherein the program is executed by a processor to implement the construction method of the ceramic membrane database of any of the above embodiments.

[0079] Any technical features in the above-described embodiments can be combined in any manner, and for the sake of brevity, not all possible combinations are described above. It should be understood that the application encompasses all such possible combinations.

[0080] It should be noted that the terms "first", "second", and "third" in the embodiments of the present application are merely used to distinguish similar objects, and do not represent a specific order or sequence. It can be understood that the "first", "second", and "third" can be interchanged in a specific order or sequence as long as the interchanging does not cause contradiction. It should be understood that the objects distinguished by "first", "second", and "third" can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein.

[0081] The terms "comprise" and "have" and any variations thereof in the embodiments of the present application are intended to cover non-exclusive inclusion. For example, a process, method, device, product, or apparatus that includes a series of steps or modules is not limited to the listed steps or modules, but can optionally include other steps or modules not listed or inherent to the process, method, product, or apparatus.

[0082] The above-described embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the scope of the patent. It should be noted that for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A method of constructing a database of ceramic membrane materials, characterized by, It comprises the following steps: Step 1, classify the technology in the research direction of ceramic membrane material, and determine the keywords in each classification; Step 2, according to the determined keywords, through database retrieval and obtain the title and full text information of the literature; Step 3, the literature classifier is constructed by machine learning method, the literature obtained in step 2 is classified according to the classification in step 1, and the data information is indexed; Step 4, according to the data obtained in step 3, the analysis result is given; In step 1, the classification obtained includes ceramic membrane preparation technology, ceramic membrane element and ceramic membrane application technology; Each classification is a first level classification or contains multiple levels of classification; In step 4, the process of analyzing the data includes: classifying the literature under a classification according to the data indexing information to obtain the research direction classification; extracting the improvement results of each research direction from the full text of the literature, and outputting the analysis results according to the improvement result degree; The research direction includes one or more of the following: ceramic membrane anti-pollution performance, ceramic membrane strength, interception performance or membrane defect repair; For the analysis of the ceramic membrane anti-pollution performance of the ceramic membrane, the following steps are further included: Step 4-1, obtain the full text and image data in the literature, and identify the membrane filtration flux change curve; Step 4-2, binary processing is performed on the curve to identify the horizontal and vertical coordinate axes and multiple flux curves; Step 4-3, identify the maximum and minimum points on the vertical coordinate axis, and calculate the scale R between the coordinate value and the pixel value according to the character position of the maximum value and the minimum value; Step 4-4, according to the scale R and the pixel point coordinate, the vertical coordinate value of the first point and the last point of each curve in the horizontal coordinate axis direction is obtained; Step 4-5, calculate the performance improvement rate R of the technical improvement, wherein: ； V0 l and Vf l are the ordinate of the first and last points in the transverse direction of the lower curve in the image, V0 u and Vf u are the ordinate of the first and last points in the transverse direction of the upper curve in the image; Step 4-6, sort the performance improvement rate obtained by each research direction, and give the analysis result; In step 4-6, the step of data classification of the performance improvement rate obtained by each research direction is further included: Step 4-6-1, obtain the text content below the curve, and identify the figure number corresponding to the curve; Step 4-6-2, search for the text position in the full text of the literature which refers to the figure, perform semantic recognition on the text content of the context of the text position, and find the words related to the ceramic membrane application system; Step 4-6-3, index the literature with the obtained application system words, and cluster the literature data in step 4-6 according to the application system words, sort the performance improvement rate of each class, and give the analysis result.

2. The method of claim 1, wherein the ceramic membrane material database is constructed by: In step 2, the database at least contains one or more of the following: journal database, degree thesis database, conference paper database and patent database.

3. The method of claim 1, wherein the ceramic membrane material database is constructed by: In step 3, the classifier is selected from support vector machine, neural network, K nearest neighbor, naive Bayes or deep learning classifier.

4. A computer readable medium, characterized in that, The program recording the construction method of the ceramic membrane material database of any one of claims 1-3 is recorded.

Images