A method and system for converting data tables in the power generation industry

Abstract

A method and system for converting data tables in the power generation industry provided by the present invention include the following steps: obtaining existing non-standard tables a of different data types in the power industry, and constructing a table conversion library from all the obtained non-standard tables a; using the obtained table conversion library to convert the non-standard tables to be converted; the present invention fully considers the professional, special and habitual characteristics of data tables in power generation industry reports, presentations, documents, etc., and artificially standardizes and unifies all commonly used data tables to facilitate subsequent data matching, identification and extraction; then, the common (human habitual) table forms are compared and analyzed with the standard tables to establish a table conversion library, and at the same time, an identification standard, conversion algorithm and conversion path are designed for each table stored in the library, so as to conveniently convert various tables step by step into standard tables according to the "path" in the table conversion library.

Description

Technical Field

[0001] The present invention belongs to the field of the power industry, and particularly relates to a method and system for converting data tables in the power generation industry. Background Art

[0002] The power industry is the basic energy industry of the national economy. With the development of the social economy, the demand for electricity in various industries has increased significantly, and the requirements for power supply reliability and power quality have been rising day by day. In recent years, as the goal of China's power industry has shifted to the production, development, transmission, and consumption of clean energy, it is necessary to rely on intelligent means to build smart grids and smart power plants, and continuously improve their flexibility, security, and defense capabilities. And intelligence and informatization complement each other. Informatization is the basis of intelligence and also the result of intelligence.

[0003] Currently, innovative technologies such as cloud computing, big data, and artificial intelligence are setting off a new upsurge, driving the industry to accelerate digital transformation. The energy industry, including oil and gas, thermal power, wind power, photovoltaic, hydropower, and nuclear power, is no exception. The main problem faced by the digital transformation of the energy industry is to ensure enterprise informatization management and production to achieve quality improvement and efficiency increase. In the process of digital transformation, the most important thing is digital assets. How to handle the surging data volume is an issue that any enterprise cannot avoid.

[0004] At present, many power enterprises only abstractly convert information into electronic "data" without exploring the value of "information" and "data", which is just the low-level stage of informatization. With the deepening of informatization, many originally paper-preserved charts, archives and other materials have begun to be converted into electronic versions for storage. However, these unstructured data have only been digitized but not yet reached the level of digitalization and cannot be mined and utilized. Among these materials and documents, the data carried in the form of tables has the largest amount of information and is the most valuable. However, as an unstructured form, the table can be designed into various forms to carry data according to factors such as personal habits, preferences, and data types. This has resulted in the fact that although the table can conveniently and intuitively reflect and carry data, it was designed from the beginning for human cognition and recognition. People can well understand the meaning represented by each item in the table, but this is very inapplicable to the situation of processing data through various programs and algorithms in the big data era. Currently, artificial intelligence cannot well "understand and identify" a table and the meaning of the data therein.

[0005] Therefore, how to conveniently and quickly identify the data in the table is the most concerned issue for many data extraction algorithms and programs. Currently, there are already many programs that can extract data from "specific" or "standardized" tables, but these programs cannot identify and extract data from "disordered and non-standardized" tables. Each year in the power industry, a large amount of information and data are presented, statistically analyzed, and tabulated. A large number of data tables need to be extracted and analyzed. However, the forms of the tables are diverse and cannot be unified. Therefore, it is very necessary and meaningful to convert all the tables into "standardized" ones so that "general" algorithms can be used to identify these "standardized" tables. Summary of the Invention

[0006] The object of the present invention is to provide a method and system for converting data tables in the power generation industry, which solves the above-mentioned deficiencies in the prior art.

[0007] To achieve the above object, the technical solution adopted by the present invention is:

[0008] A method for converting data tables in the power generation industry provided by the present invention includes the following steps:

[0009] Obtain existing non-standard tables a of different data types in the power industry, and construct a table conversion library from all the obtained non-standard tables a.

[0010] Use the obtained table conversion library to convert the non-standard tables to be converted.

[0011] Preferably, the specific method for constructing the table conversion library is:

[0012] S11, set corresponding standard tables according to different data types and data requirements to obtain an original table conversion library. Among them, the table conversion library includes N-level hierarchical tables, where the N-level hierarchical tables are the 0th layer, the 1st layer, the 2nd layer, and the Nth layer respectively, and the standard table is used as the 0th layer of the original table conversion library.

[0013] S12, add all the obtained non-standard tables a to the original table conversion library as template tables for the corresponding hierarchical tables; and set the conversion algorithms corresponding to each hierarchical table to obtain the table conversion library.

[0014] Preferably, in S12, when adding all the obtained non-standard tables a to the original table conversion library as template tables for the corresponding hierarchical tables; and setting the conversion algorithms corresponding to each hierarchical table, the specific method is:

[0015] S121, set the table conversion rules for each standard table A.

[0016] S122. Calculate the distance N between each non-standard form a and the corresponding standard form A according to the form conversion rules of each standard form A.

[0017] S123. Determine whether there is a template form in the Nth-level form in the form conversion library that corresponds to the data type of the non-standard form a and has the same form. If so, end; otherwise, proceed to S124.

[0018] S124. Determine whether there is a template form in the (N - 1)th-level form in the form conversion library that corresponds to the data type of the non-standard form a, has the same form, and the distance between the two is 1, where:

[0019] If so, record the non-standard form a as a template form and add the conversion algorithm for the template form at the same time.

[0020] If not, use the non-standard form a as the template form corresponding to the Nth-level form in the form conversion library and add the conversion algorithm for the template form at the same time. Use this conversion algorithm to convert the non-standard form a to generate a new form, and use the new form as the template form corresponding to the (N - 1)th-level form in the form conversion library.

[0021] S125. Search in the (N - 2)th-level form in the form conversion library to determine whether there is a template form that corresponds to the data type of the new form, has the same form, and the distance between the two is 1, where:

[0022] If so, record the new form as a template form and add the conversion algorithm for the template form at the same time.

[0023] If not, use the new form as the template form corresponding to the (N - 1)th-level form in the form conversion library and add the conversion algorithm for the template form at the same time. Use this conversion algorithm to convert the new form to generate a new form, and use the new form as the template form corresponding to the (N - 2)th-level form in the form conversion library.

[0024] S126. Repeat S125 until there is a template form in the [specific level]th-level form in the form conversion library that corresponds to the data type of the new form, has the same form, and the distance between the two is 1.

[0025] Preferably, the specific method of the conversion algorithm is:

[0026] Perform row and column conversion on the non-standard form a. The content of the row and column conversion includes row-column interchange, deletion of useless columns or rows in the form, supplementation of form column information, and normalization of the form N*M.

[0027] Preferably, the table conversion rule means that in the process of converting a non-standard table into the form of a required standard table, each step of conversion only targets one type of content for conversion.

[0028] Preferably, calculate the distance N between each non-standard table a and the corresponding standard table A. The specific method is as follows:

[0029] Count the number of steps required to convert the non-standard table into the form of the required standard table, and use the obtained number of steps as the distance N between each non-standard table a and the corresponding standard table A.

[0030] Preferably, use the obtained table conversion library to convert the non-standard table to be converted. The specific method is as follows:

[0031] S21. Match the non-standard table to be converted with the table conversion library according to the data type and data requirements. If the match fails, use the non-standard table to be converted to improve the table conversion library; if the match is successful, proceed to the next step;

[0032] S22. Find the template table corresponding to the non-standard table to be converted in the table conversion library;

[0033] S23. Determine whether the corresponding template table is a standard table. Among them, if the corresponding template table is a standard table, convert the non-standard table to be converted into the form of the required standard table according to the corresponding conversion rule; otherwise, proceed to the next step;

[0034] S24. According to the conversion rule of the corresponding template table, convert the non-standard table to be converted into the form of the corresponding template table at the next level of the corresponding table, and repeat S22 until the non-standard table to be converted is converted into the form of a standard table.

[0035] A data table conversion system for the power generation industry, including:

[0036] A conversion library construction unit, which is used to obtain existing non-standard tables a of different data types in the power industry, and construct a table conversion library from all the obtained non-standard tables a;

[0037] A table conversion unit, which is used to convert the non-standard table to be converted by using the obtained table conversion library.

[0038] Compared with the prior art, the beneficial effects of the present invention are:

[0039] A method for converting tables in the power generation industry provided by the present invention converts non-standardized tables into standardized tables. During the conversion process, the data in the original tables is correspondingly converted into the standard tables. A standard table library and a non-standard table library are established, and a conversion algorithm is established according to the differences between non-standard tables and standard tables, increasing the types of tables that can be converted and the conversion efficiency. At the same time, the present invention fully considers the professionalism, particularity, and usage habits of data tables in power generation industry reports, presentations, documents, etc., and artificially standardizes, normalizes, and unifies all commonly used data tables, facilitating later data matching, recognition, and extraction. Then, the common (human-habitual) table forms are compared and analyzed with the standard tables to establish a table conversion library. At the same time, an identification standard, conversion algorithm, and conversion path are designed for each table stored in the library, facilitating the conversion of various tables into standard tables step by step according to the "path" in the table conversion library. It has the following advantages:

[0040] First: By standardizing, normalizing, and unifying the design of commonly used data tables, other similar data tables are converted into the same standard table. For this type of data table, only one data matching and extraction algorithm or program for this type of standard table needs to be developed, thus forming a unified standard, reducing the development work for extracting data from data tables, and facilitating development and implementation.

[0041] Second: The table conversion method has scalability and universality. According to the "distance" conversion principle, only "1 step" distance needs to be converted each time, so that any complex and non-standard table can be simplified and converted into a standard table "step by step". Therefore, in theory, any table can be converted into the corresponding standard table.

[0042] Third: By establishing a table conversion library, not only can various and complex data tables be standardized into multiple unified standard tables, but also "intermediate" conversion tables, conversion algorithms, identification standards, and conversion paths can be supplemented and added during the establishment process, facilitating the conversion of any non-standard table into a standard table through "specific" paths and algorithms.

[0043] Furthermore, using the table identification standard and conversion algorithm, a unified and standardized table conversion path is established. Through manual and automatic conversion, identification and comparison, the table identification standard and conversion algorithm are increased and improved, and then the table conversion library is updated regularly to provide the correctness, feasibility, and uniqueness of table conversion.

[0044] Furthermore, the data conversion algorithm and identification standard from each standard table to the next lower layer are set, providing a standard for type identification of the table to be converted and an algorithm and path for conversion towards the standard table. Brief Description of the Drawings

[0045] Figure 1Schematic diagram of the table conversion library;

[0046] Figure 2 Flow chart for the extension of the table conversion library;

[0047] Figure 3 Flow chart for table conversion. Detailed implementation manners

[0048] The present invention will be further described in detail below with reference to the accompanying drawings.

[0049] A table conversion method for the power generation industry provided by the present invention converts non-standard tables into standard tables, and at the same time, during the conversion process, the data in the original table is correspondingly converted into the standard table; a standard table library and a non-standard table library are established, and according to the differences between the non-standard table and the standard table, a conversion algorithm and a conversion path are established; the conversion algorithm and the conversion path are added or optimized to increase the types of tables that can be converted and the conversion efficiency.

[0050] A data table conversion method for the power generation industry according to the present invention includes creating an extended table conversion library and table conversion, and the specific details of each part are as follows:

[0051] Step 1, create an extended table conversion library.

[0052] First, formulate various unified standard data tables corresponding to different data applications and use them as the final goal for the conversion of all other tables;

[0053] Secondly, on the basis of the standard table, various non-standard tables, human habit customized tables, and simplified or cumbersome tables are sorted and added, and according to the conversion algorithm principle, intermediate conversion tables in the process from these tables to the standard table are added, and the conversion algorithms, conversion paths (association relationships) between each layer of tables and the recognition criteria for each type of table are added in sequence (from the Nth layer to the 0th layer).

[0054] Finally, a conversion chain, conversion algorithm from non-standard tables to standard tables and the recognition criteria for each layer of tables are formed, which becomes a table conversion library.

[0055] Among them, the standard table refers to an N*M data table with a title row. The "rows" of the table (except the first row) represent a parameter, device, item, etc., indicating a kind of thing; the "columns" of the table (except the first column, unit column, etc.) represent different situations where the thing is located (such as time, operating conditions, conditions); the item at the intersection of "row" and "column" is the performance, value, state, etc. of the thing under a certain condition.

[0056] Step 2, table conversion.

[0057] Obtain the table x to be converted. According to the table recognition standard, identify the table type and the corresponding table Xn, and then through the conversion chain of the standard table X corresponding to Xn, convert the table step by step according to the conversion algorithms of each layer of the conversion chain, and finally obtain the table x0 in the form of the corresponding standard table X.

[0058] As Figures 1 to 3 shown, specifically, a method for constructing a table conversion library in the power generation industry provided by the present invention includes the following steps:

[0059] Step 1, according to the data analysis and requirements in the power industry, customize different standardized tables for different data types and data requirements, design recognition criteria according to the characteristics of the tables, and add these standard tables to the table conversion library;

[0060] Construct an N-level hierarchical table in the table conversion library, where the N-level hierarchical tables are the 0th layer, the 1st layer, the 2nd layer, and the Nth layer respectively; and use the standard tables in the table conversion library as the 0th layer;

[0061] Step 2, obtain the existing non-standard tables a of different data types in the power industry, and add all the obtained non-standard tables a to the table conversion library as template tables. The specific method of addition is:

[0062] Find the corresponding standard table A in the table conversion library according to the data type of each non-standard table a;

[0063] Set the table conversion rules for each standard table A;

[0064] Calculate the distance N between the non-standard table a and the corresponding standard table A according to the table conversion rules of each standard table A;

[0065] Step 3, according to the distance N between the non-standard table a and the corresponding standard table A in Step 2, check whether there is a corresponding template table in the Nth layer hierarchical table in the table conversion library that has the same form as the non-standard table a. If it exists, it means that this type of non-standard table a has been added to the table conversion library, and go to Step 7; if not, go to Step 4;

[0066] Step 4, check whether there is a template table A in the (N - 1)th layer hierarchical table in the table conversion library that has the same data type as the non-standard table a, the same form, and a distance of 1 between the two n-1 , where, if it exists, record this non-standard table a as the template table A n , and at the same time add the conversion algorithm of the template table A n ; if not, use this non-standard table a as the template table A corresponding to the Nth layer hierarchical table in the table conversion library n, while adding template table A n 's conversion algorithm; use this conversion algorithm to convert the non-standard table a to generate a new table A n-1 , take the new table A n-1 as the template table A corresponding to the table at the (N - 1)-th layer in the table conversion library n-1 ,;

[0067] Step 5, check in the table at the (N - 2)-th layer in the table conversion library to see if there exists a template table A n-1 whose data type corresponds to that of the new table A and the distance between them is 1 n-2 , where if it exists, record the new table A n-1 as the template table A n-1 , while adding the conversion algorithm of template table A n-1 ;

[0068] If it does not exist, take the new table A n-1 as the template table A corresponding to the table at the (N - 1)-th layer in the table conversion library n-1 , while adding the conversion algorithm of template table A n-1 ; use this conversion algorithm to convert the new table A n-1 to generate a new table A n-2 , take the new table A n-2 as the template table A corresponding to the table at the (N - 2)-th layer in the table conversion library n-2 ;

[0069] Step 6, repeat Step 5 until there exists a template table A n-x in the table at the (N - y)-th layer in the table conversion library whose data type corresponds to that of the new table A and the distance between them is 1 n-y , where y = x + 1;

[0070] Step 7, according to Steps 2 and 3, expand the table conversion library by incorporating common standard tables and common non-standard tables into the table conversion library, including the recognition criteria, conversion algorithms, and conversion paths of the corresponding tables;

[0071] Step 8, use the data tables in various reports, presentations, documents, etc. in the power generation industry to train the table conversion library, and continuously adjust the rules in the table conversion library through the comparison of manual and automatic recognition and conversion, so as to improve the correctness of the table conversion library and the number of tables.

[0072] In Step 4, the conversion algorithm refers to "transforming" the rows and columns of the table without changing the information carried by the table, including row-column swapping of the table, deletion of useless columns / rows of the table, supplementation of column information of the table, normalization of the table N*M, etc.

[0073] The conversion rule is to convert a non-standard table step by step into the form of a standard table. During the conversion process, each step of the conversion only involves one type of "transformation". For example, deletion means deleting all the useless columns / rows that should be deleted. Row-column interchange transposes the entire table diagonally (rows become columns, and columns become rows). On the basis of not changing the information carried by the table, completing one type of transformation represents one step of conversion.

[0074] The "distance" mentioned above is the "number of steps" or the number of types of transformations required to convert from a non-standard table to a standard table. During the conversion process, the order of transformation is specified, and all conversion chains must be carried out in accordance with this transformation order. Therefore, the data conversion method for each layer is fixed, which ensures the uniqueness of each standard table conversion chain.

[0075] A method for converting power generation industry data tables based on a table conversion library includes the following steps:

[0076] Step 1: Extract the table in the report, denoted as table a. According to the table recognition standard of the table conversion library, identify the table type. If the recognition fails, use table a to improve the table conversion library; if the recognition is successful, find the corresponding table Ax in the table conversion library through the table type and proceed to the next step.

[0077] Step 2: Determine whether table Ax is a standard table. If table Ax is a standard table, proceed to Step 3; if table Ax is not a standard table, then according to the associated table Ax-1 of table Ax and the conversion algorithm, convert table a into the same type as table Ax-1, denoted as table ax. Then replace table Ax with table Ax-1 and replace table a with table ax, and repeat Step 2.

[0078] Step 3: Through one or more times of Step 2, finally convert table a into the corresponding table a0 of the standard table A type.

[0079] Step 4: Through the above steps, the standard table form table a0 corresponding to table a is obtained. Then, through other table data extraction methods, the data in the table can be matched and extracted.

[0080] A power generation industry data table conversion system includes:

[0081] A conversion library construction unit for obtaining existing non-standard tables a of different data types in the power industry and constructing a table conversion library from all the obtained non-standard tables a.

[0082] A table conversion unit for converting the non-standard table to be converted by using the obtained table conversion library.

[0083] The method described in the present invention can simply and conveniently convert various non-standard, irregular, and cumbersome data tables into various standard tables with unified general specifications on the basis of a table conversion library, which is conducive to the subsequent matching extraction and analysis of data. According to the type of data table, the method creates a standard table, and by analyzing the gap between the non-standard and standard tables, designs a step-by-step conversion algorithm and the recognition criteria for intermediate transition tables to form a unified table conversion library. After identifying the data table to be converted by using the table conversion library, the method performs step-by-step table conversion through a conversion chain, and finally converts the table to be converted into the corresponding standard table form without losing the information carried in the original table during the conversion process.

[0084] The method described in the present invention can manually expand, update, and optimize the table conversion library, and theoretically can cover all types of tables to achieve the standardized conversion of all data tables.

[0085] The present invention has been applied to the data table extraction system in the power generation industry. The system uses the table conversion library and table conversion method in the present invention, can convert a large number of non-standard data tables into various standardized table forms, and outputs them in the form of standard data table files, providing strong support for the matching extraction of table data.

Claims

1. A method for converting data tables in the power generation industry, characterized in that, It includes the following steps: Obtain the existing non-standard table a of different data types in the power industry, and construct a table conversion library from all the obtained non-standard tables a; Use the obtained table conversion library to convert the non-standard table to be converted; The specific method of constructing the table conversion library is: S11, Set corresponding standard tables according to different data types and data requirements to obtain an original table conversion library. Among them, the table conversion library includes N-level hierarchical tables, where the N-level hierarchical tables are the 0th layer, the 1st layer, the 2nd layer, and the Nth layer respectively, and the standard table is used as the 0th layer of the original table conversion library; S12, Add all the obtained non-standard tables a to the original table conversion library as the template tables of the corresponding hierarchical tables; And set the conversion algorithm corresponding to each hierarchical table to obtain the table conversion library; In S12, add all the obtained non-standard tables a to the original table conversion library as the template tables of the corresponding hierarchical tables; And set the conversion algorithm corresponding to each hierarchical table. The specific method is: S121, Set the table conversion rules for each standard table A; S122, Calculate the distance N between each non-standard table a and the corresponding standard table A according to the table conversion rules of each standard table A; S123, Determine whether there is a template table in the Nth layer hierarchical table of the table conversion library that corresponds to the data type of the non-standard table a and has the same form. If so, end; Otherwise, go to S124; S124. Determine whether there is a template table in the (N - 1)-th level table of the table conversion library that corresponds to the data type of the non-standard table a, has the same form, and the distance between the two is 1. , where: If it exists, record the non-standard form a as a template form , and at the same time add a conversion algorithm for the template form ; If not, use the non-standard form a as the template form corresponding to the form at the Nth level in the form conversion library , and at the same time add a conversion algorithm for the template form ; use this conversion algorithm to convert the non-standard form a to generate a new form , and use the new form as the template form corresponding to the form at the (N-1)th level in the form conversion library ; S125, search in the hierarchical table at the (N - 2)-th layer of the table conversion library to check if there is a template table corresponding to the new table in terms of data type, having the same form, and with a distance of 1 between the two, where: If it exists, record the new table as a template table , and at the same time add a conversion algorithm for the template table ; If not, the new table is used as the template table corresponding to the (N - 1)-th level table in the table conversion library , and meanwhile, the conversion algorithm for the template table is added; the new table is converted using this conversion algorithm to generate a new table , and the new table is used as the template table corresponding to the (N - 2)-th level table in the table conversion library ; S126. Repeat S125 until reaching the hierarchical table in the table conversion library where there is a template table corresponding to the data type of the new table, with the same form and a distance of 1 between them up to this point, where ; Use the obtained table conversion library to convert the non-standard table to be converted. The specific method is: S21, Match the non-standard table to be converted with the table conversion library according to the data type and data requirements. If the match fails, use the non-standard table to be converted to improve the table conversion library; If the match is successful, go to the next step; S22, Find the template table corresponding to the non-standard table to be converted in the table conversion library; S23, Determine whether the corresponding template table is a standard table. If the corresponding template table is a standard table, convert the non-standard table to be converted into the form of the required standard table according to the corresponding conversion rules; Otherwise, go to the next step; S24, According to the conversion rules of the corresponding template table, convert the non-standard table to be converted into the form of the corresponding template table at the next level of the corresponding template table, and repeat S22 until the non-standard table to be converted is converted into the form of the standard table.

2. The method for converting a data table in the power generation industry according to claim 1, wherein The specific method of the conversion algorithm is: Perform row and column conversion on the non-standard table a. The content of the row and column conversion includes row-column interchange, deletion of useless columns or rows in the table, supplementation of table column information, and normalization of the table N*M.

3. A method for converting data tables in the power generation industry according to claim 2, characterized in that, The table conversion rule refers to that in the process of converting a non-standardized table into the form of the required standard table, each step of conversion only targets one kind of content for conversion.

4. A method for converting a data table in the power generation industry according to claim 2, characterized in that, Calculate the distance N between each non-standard table a and the corresponding standard table A. The specific method is: Count the number of steps required to convert the non-standardized table into the form of the required standard table, and use the obtained number of steps as the distance N between each non-standard table a and the corresponding standard table A.

5. A data table conversion system for the power generation industry, characterized in that, The method according to any one of claims 1 to 4, comprising: A conversion library construction unit, configured to obtain existing non-standard form a of different data types in the power industry, and construct a form conversion library from all the obtained non-standard forms a; A form conversion unit, configured to convert a to-be-converted non-standard form by using the obtained form conversion library.

Images