Method for constructing a knowledge graph computing framework for building roof photovoltaic potential assessment and query, as well as framework, equipment and storage medium

Abstract

The present invention discloses a method for constructing a knowledge graph computing framework for building roof photovoltaic potential assessment and query integration, a framework, a device and a storage medium, and the method includes: constructing an ontology model of the knowledge graph; constructing a building roof photovoltaic potential data base; establishing a building roof photovoltaic potential assessment model based on expert knowledge; and semanticizing the building roof photovoltaic potential assessment model and the query model. The present invention can effectively integrate multi-dimensional heterogeneous data with expert knowledge in the field of building photovoltaic potential assessment, and based on remote sensing image information extraction technology, realize dynamic assessment of building roof photovoltaic potential and support multi-level and multi-scale flexible query retrieval of assessment results.

Description

Technical Field

[0001] The present invention belongs to the field of knowledge graph artificial intelligence application technology in the direction of new energy, and in particular relates to a method for constructing an integrated knowledge graph computing framework for building rooftop photovoltaic potential assessment and query, as well as a framework, equipment and storage medium. Background Art

[0002] Early photovoltaic potential assessments were low-cost, but the results were often rough, and the data update process relied on human resources, which was slow and complicated, and the results lacked geographical reference value. In complex urban environments, more timely and accurate spatial assessments are needed to provide decision support for the selection of distributed photovoltaic installation addresses. Geographic Information Systems (GIS) have introduced geospatial information into photovoltaic potential assessments because they can use the geometric, geographical and attribute characteristics of geographic elements for calculations and analysis. In order to further improve the accuracy of assessments, technologies such as laser detection and ranging (LIDAR), photogrammetry and laser point clouds that can accurately obtain three-dimensional building information in cities have been applied to the field of building photovoltaic potential assessment. However, these methods are costly and require huge amounts of calculations, which limits their application in large-scale photovoltaic potential assessments. Remote sensing images carry rich information about ground objects and can be used as a source of inexpensive and high-quality data. Using remote sensing information extraction technology, ground object information such as buildings can be obtained from remote sensing images at low cost and high efficiency. Spatial information extracted based on remote sensing images is not only comprehensive, but also can update ground object information based on the time when the image was taken, showing obvious advantages in multi-time scale assessment needs. Although remote sensing information technology provides an effective method for calculating urban buildings, in the assessment of building photovoltaic potential, in addition to the extracted building information, multidimensional factors such as solar radiation and photovoltaic panel specifications need to be considered.

[0003] When faced with the problem of assessing the photovoltaic potential of buildings, it is not enough to simply consider the geometric, geographical and attribute characteristics of the building, but also to comprehensively consider multiple factors such as the performance parameters of the photovoltaic equipment and solar radiation conditions. Traditional calculation methods have certain limitations in the fusion calculation and fast query of these multi-source heterogeneous data. For fast result query, traditional calculation results are usually stored in large-capacity storage hard disks in various data formats such as Excel, GeoJSON, Shp, etc. When it is necessary to find the photovoltaic calculation potential of a building in a certain large or small target area, it often takes a lot of manpower and time to find the required result data from the complex storage results.

[0004] In the implementation of existing building rooftop photovoltaic potential assessment technology, a core defect is that it relies too much on manual multi-step operations. The correctness of the results is related to the business capabilities of practitioners. The direct cause of this defect is the lack of fusion and coordination of multi-heterogeneous data. In addition, with the acceleration of urbanization, the update and change of building information occurs frequently, but the existing technology lacks an effective mechanism to capture these changes in a timely manner, making it difficult to ensure the timeliness of the assessment results. The lack of query efficiency and flexibility further highlights the backwardness of existing data management methods and the imperfection of query tools. In practical applications, photovoltaic companies often need to quickly query and retrieve the photovoltaic potential of a specific area anytime and anywhere to support decision-making and resource allocation. However, the existing technology usually stores the detailed assessment results in multiple formats in large-capacity storage hard disks, which is inefficient when retrieving the photovoltaic potential of a specific area. It requires a lot of manpower and time for data screening and analysis, which limits the application efficiency of the assessment results. In addition, the decentralized architecture and isolated functional modules of the existing technology lead to the separation of the assessment and query processes, and lack of a unified and coherent user experience. These shortcomings restrict the application efficiency and accuracy of existing technologies in the evaluation of photovoltaic potential on building roofs, demonstrating the need for an innovative method to achieve more efficient and accurate evaluation and query, which is exactly the problem that the present invention aims to solve by constructing an integrated method for calculating and querying photovoltaic potential on building roofs based on a knowledge graph. Summary of the invention

[0005] The present invention proposes a method for constructing an integrated knowledge graph computing framework for building rooftop photovoltaic potential assessment and query, as well as a framework, equipment and storage medium, which can effectively integrate multi-dimensional heterogeneous data with expert knowledge in the field of building photovoltaic potential assessment, support flexible expansion of the computing framework and dynamic updating of the data base, and realize dynamic evaluation of building photovoltaics and multi-level and multi-scale flexible query and retrieval of evaluation results.

[0006] The specific technical solutions are as follows:

[0007] A method for constructing a knowledge graph computing framework for building rooftop photovoltaic potential assessment and query integration, the method comprising:

[0008] Construct an ontology model of the knowledge graph; realize the fusion and collaborative analysis of multi-heterogeneous data in the evaluation of building photovoltaic potential by combing and integrating the relevant concepts of key data and expert knowledge, and construct a knowledge expression framework for building photovoltaic potential evaluation;

[0009] Construct a data base for photovoltaic potential on building roofs; realize digital management of assessment data by instantiating building and other relevant geographic entity information, so as to provide a data basis for accurate photovoltaic potential assessment;

[0010] Establish a building rooftop photovoltaic potential assessment model based on expert knowledge; determine the optimal installation parameters by quantifying the multi-level shading influencing factors, and achieve scientific assessment of the building's photovoltaic potential;

[0011] The building rooftop photovoltaic potential assessment model and query model are semanticized; based on the reasoning mechanism, rule extraction is used to realize the automation and intelligence of the assessment and retrieval process, which is used to improve the efficiency and accuracy of the assessment and retrieval.

[0012] The present invention also provides a knowledge graph computing framework for building rooftop photovoltaic potential assessment and query integration, the framework comprising:

[0013] Ontology model of knowledge graph; used to integrate and express key concepts, entities and their interrelationships in the field of building photovoltaic potential assessment, and provide structured knowledge representation for data fusion and collaborative analysis;

[0014] Building rooftop photovoltaic potential data base; used to store and manage multivariate heterogeneous data, including building geometry, geography and attribute information, as well as photovoltaic equipment performance parameters, solar radiation conditions, etc., to provide data support for the evaluation model;

[0015] Expert knowledge-based building rooftop photovoltaic potential assessment model; used to achieve scientific calculation of building rooftop photovoltaic potential, including the assessment of key indicators such as available area, installed capacity and first-year power generation, to ensure the accuracy and reliability of the assessment results;

[0016] The semantic module of the building rooftop photovoltaic potential assessment model and query model is used to associate the assessment model, query model and related information and data in the knowledge graph, realize the automation and intelligence of the assessment process and query process, and improve the operational efficiency of the assessment and result retrieval.

[0017] The present invention also proposes a computer device, comprising: at least one memory and at least one processor; the memory is used to store one or more programs; when the one or more programs are executed by the at least one processor, the at least one processor implements the above method.

[0018] The present invention also provides a computer program product, comprising a computer program / instruction, which implements the above method when executed by a processor.

[0019] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program implements the above method when executed by a processor.

[0020] The ontology model of the present invention sorts out and integrates the concepts and relationships of evaluation objectives, expert knowledge, and key data and knowledge including building characteristics, photovoltaic technology, sunshine conditions, etc., to achieve multi-heterogeneous data fusion and collaborative analysis in building photovoltaic potential assessment. By designing a module-to-module coupling, the independent knowledge graph computing framework structure in the module improves the flexibility and scalability of the computing framework. From the low layer to the surface layer, they are the ontology model, data base, algorithm model and evaluation application. The modules are constrained by established rules to be standardized, and there are multiple independent parts inside the module. This structural design allows the framework to easily adapt to new business needs and data sources while maintaining the stability of the existing system. Further, the design of the reasoning mechanism realizes the communication and exchange between the modules, and the built-in "data instantiation" action function enables the data base to adapt to the dynamic development and scalability of urban data, ensuring the timeliness of photovoltaic potential assessment. In addition, based on the expert knowledge in related fields, the building photovoltaic potential assessment model and result retrieval model are constructed, and the scientificity, systematicness and repeatability of the assessment process are ensured through formalization. These models are associated with the relevant information and data in the knowledge graph using semantic technology to improve the intelligent level of the building photovoltaic potential assessment and retrieval process.

[0021] The present invention has the following beneficial effects:

[0022] The present invention provides a method and framework for constructing an integrated knowledge graph computing framework for building rooftop photovoltaic potential assessment and query, which can effectively integrate multi-dimensional heterogeneous data with expert knowledge in the field of building photovoltaic potential assessment, and based on remote sensing image information extraction technology, realize dynamic assessment of building rooftop photovoltaic potential and support multi-level and multi-scale flexible query and retrieval of assessment results.

[0023] The present invention realizes the effective fusion and collaborative computing of multi-dimensional heterogeneous data through the knowledge graph computing framework, supports dynamic evaluation in building data update scenarios, provides multi-level and multi-scale building photovoltaic potential evaluation result retrieval function, and constructs an integrated evaluation and query framework, which greatly improves the efficiency of photovoltaic potential evaluation and improves the utilization rate of evaluation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A flow chart of a method for constructing a knowledge graph computing framework for building rooftop photovoltaic potential assessment and query integration according to the present invention;

[0025] Figure 2 It is a schematic diagram of entity attributes and relationships between entities in the ontology model;

[0026] Figure 3 This is a schematic diagram of the GeoSPARQL spatial ontology structure;

[0027] Figure 4This is a schematic diagram of the reasoning ontology structure;

[0028] Figure 5 Semantic structure diagram for building photovoltaic potential assessment model;

[0029] Figure 6 Semantic structure diagram of building photovoltaic potential query model. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical scheme and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in each embodiment of the present invention described below can be combined with each other as long as they do not conflict with each other. To achieve the above-mentioned purpose, the present invention adopts the following technical scheme.

[0031] The present invention proposes a method for constructing an integrated knowledge graph computing framework for building rooftop photovoltaic potential assessment and query, as well as a framework, device and storage medium. A knowledge graph computing framework with building rooftop outline information and spatial geographic location information as the core and coordinated with other data required for photovoltaic resource assessment is constructed, aiming to provide technical support for the assessment and result retrieval of building rooftop photovoltaic potential. The knowledge graph uses the Protégé tool to establish a relevant ontology model, uses the OWL (Web Ontology Language) ontology language to semantically represent knowledge, implements semantic search based on GeoSPARQL query syntax, and uses the GraphDB graph database combined with the Key-Value database as a storage carrier to store triples of spatial information. A method for constructing an integrated knowledge graph computing framework for building rooftop photovoltaic potential assessment and query, such as Figure 1 As shown, including:

[0032] (1) Constructing the ontology model of the knowledge graph;

[0033] The ontology model of the knowledge graph of the present invention includes the building photovoltaic potential ontology, space ontology and reasoning ontology. The building photovoltaic potential ontology involves relevant concepts that appear in the evaluation and query process. The attributes of each concept entity and the relationship between entities are as follows: Figure 2As shown in the figure. The main conceptual entities of the ontology include "building", "building type", "administrative division", "photovoltaic equipment", "photovoltaic panel specifications", "installation parameters", and "solar conditions", which provide a data basis for the evaluation. In addition, the "photovoltaic potential" entity in the ontology allows the evaluation results to be updated synchronously in the knowledge graph. The "target evaluation scope" and "target query scope" entities serve as the entry points for evaluation and query, defining the spatial boundaries of evaluation and query operations, allowing users to analyze photovoltaic potential and retrieve data for specific geographical or logical areas.

[0034] The concept entity "building" has its attributes: "has roof area" points to the data type "xsd:string"; it has its relationships: "has type" points to the entity "building type", "is located in" points to the entity "administrative division", "has photovoltaic potential" points to the entity "photovoltaic potential", and "has photovoltaic equipment" points to the entity "photovoltaic equipment". The concept entity "building type" has its attributes: "has type number" and "has reduction factor", which point to the data types "xsd:int" and "xsd:double" respectively. The concept entity "administrative division" contains the entities "provincial administrative division", "prefectural administrative division", "county shape division", and "street administrative division", and has its relationships: "has photovoltaic potential" points to the entity "photovoltaic potential". The concept entity "provincial administrative division" has its attributes: "has name" and "has installation angle", which point to the data types "xsd:string" and "xsd:double" respectively; it has relationships: "subordinate to", which points to the entity "prefectural administrative division". The concept entity "prefecture-level administrative division" has its attributes: "has a name" and "has equivalent operating hours", which point to the data types "xsd:string" and "xsd:double" respectively; it has a relationship: "subordinate to", which points to the entity "district-level administrative division". The concept entity "district-level administrative division" has its attributes: "has a name" points to the data type "xsd:string"; it has a relationship: "subordinate to", which points to the entity "street-level administrative division". The entity "street-level administrative division" has its attributes: "has a name" points to the data type "xsd:string". The concept entity "photovoltaic equipment" has its relationships: "has specifications" points to the entity "photovoltaic panel specifications", "has installation parameters" points to the entity "installation parameters", and "has solar conditions" points to the entity "solar conditions". The concept entity "photovoltaic panel specifications" has its attributes: "has a single block area" and "has rated output power", which point to the data types "xsd:double" and "xsd:double" respectively. The concept entity "installation parameters" has its attributes: "has installation inclination" and "has installation spacing", which point to the data types "xsd:double" and "xsd:double" respectively. The concept entity "solar conditions" has its attributes: "has solar altitude angle" and "has solar azimuth", which point to the data types "xsd:double" and "xsd:double" respectively. The concept entity "photovoltaic potential" has its attributes: "has available area", "has installed capacity", and "has first-year power generation", which point to the data types "xsd:double", "xsd:double", and "xsd:double" respectively. The concept entity "target evaluation scope" and "target query scope" have a relationship: "located in" points to the entity "administrative division".

[0035] In order to express geographical entities with spatial characteristics such as administrative divisions and buildings, this paper introduces GeoSPARQL proposed by OGC (OpenGeospatial Consortium) as the spatial ontology. Figure 3 . Spatial objects have two subclasses: features and geometries. Features have their own geometric features. Geometries can usually be divided into points, polylines, and polygons according to the geometric form of geographic features, which in turn provides support for spatial operations such as distance calculation, spatial intersection and union, and geographic buffers. Geographic entity data instances correspond to spatial objects in the spatial ontology, realizing the effective association between geographic entities and ontology concepts, and building a knowledge expression framework that is conducive to spatial operations.

[0036] The reasoning ontology is the rule constraint for automatically executing the reasoning program on the knowledge graph, and is the basis for the knowledge graph to carry out reasoning. The structure is as follows Figure 4 As shown in the figure. An inference rule contains event objects and action objects. Event objects are the triggering conditions of inference rules, and action objects are the corresponding actions executed after the inference rules are triggered. The concept of event objects is divided into independent events and event combinations according to the independence of events. Action objects are also divided into independent and combined. Inference rules are associated with each other through specific action objects "notifying the next rule" and event objects "receiving notification from the previous rule".

[0037] (2) Constructing a data base on the photovoltaic potential of building roofs;

[0038] The data base is composed of instantiated source data. Instantiation refers to mapping specific data to entities and relationships in the knowledge graph through extraction and fusion. The present invention reads the building information and administrative division information extracted based on remote sensing images in the data file one by one, and instantiates the spatial and attribute information of geographic entities such as buildings and administrative divisions in the form of triples through knowledge extraction and knowledge fusion, thereby constructing a digital base for evaluating the photovoltaic potential of buildings.

[0039] Knowledge extraction refers to the process of automatically identifying and extracting valuable information, relationships and other knowledge from data. In the present invention, geographic entity data is stored in GeoJSON format before entering the knowledge graph data base. Geometry in GeoJSON is mapped to the predicate of the spatial information triple, and the Geometry value is automatically mapped to the object of the spatial information triple that conforms to the GeoSPARQL format specification; the key in the attribute set is automatically mapped to the predicate of the corresponding type of geographic entity attribute triple, and the attribute value is automatically mapped to the object of the corresponding predicate in the triple.

[0040] The triples are reorganized into multiple key-value pairs according to the following rules and stored in the Key-Value database:

[0041] 1) "Subject + Predicate": "Object";

[0042] 2) "object + predicate": "subject";

[0043] 3) "Geographic entity type + system timestamp of the instance": "Subject";

[0044] 4) “Geographic entity type + geocode to which the instance belongs”: “Subject”.

[0045] For geographic entities with spatial characteristics, in addition to the attribute information constrained in the ontology, multi-scale geographic coding index information is added. The present invention calculates the geographic code based on the Web Mercator projection grid division method and indexes the geographic entity in the spatiotemporal knowledge graph. The triple representation of the geographic coding index information is: "geographic entity hasTileCode tile code". The subject "geographic entity" is connected to its corresponding object "tile code" using the predicate "hasTileCode". The tile code is in the string format as follows: "z00_x0000_y0000", z represents the grid division level, the value is in the range of 0-26, and x and y represent the horizontal and vertical coordinates of the grid division space. A geographic entity has a series of tile codes of different scales, and the tile code is a string data type. For example, "z12_x3420_y1672" means that the Mercator projection grid division space has a horizontal coordinate of 3420, a vertical coordinate of 1672, and a grid code division level of 12. Multi-scale tile codes are inserted into different types of geographic entities in the process of building the data base, and matching is performed based on the tile codes associated with the geographic entities to support multi-scale spatial retrieval of spatiotemporal elements.

[0046] (3) Establish a model for assessing the photovoltaic potential of building roofs based on expert knowledge;

[0047] Based on the expert knowledge of photovoltaic potential assessment, the present invention establishes a building photovoltaic assessment model, which includes the calculation of available area, installed capacity and first-year power generation. In the real world, roofs are often blocked by surrounding buildings, vegetation, etc., as well as structural obstacles such as parapets, air conditioners and elevator shafts on the roof. These factors limit the effective laying area of ​​​​photovoltaic panels. In order to more accurately assess the photovoltaic potential of the roof, the present invention uses a reduction factor that integrates the impact of the city, block and individual building levels to quantify the impact of shading factors on the actual available area. Buildings are divided into low-rise residential buildings, high-rise residential buildings, medium-rise buildings and industrial buildings according to their floor height and use. The corresponding reduction coefficient is calculated by formula (1), where is the urban factor reduction coefficient, is the block factor reduction coefficient, is the reduction coefficient of the single building factor.

[0048] (1)

[0049] In the northern hemisphere, the maximum power generation can be obtained when the solar photovoltaic panels face due south. The determination of the optimal installation inclination angle needs to consider the local geographical information and specific installation conditions. The present invention adopts a traversal search method to establish an annual light radiation model on the inclined surface and traverse the photovoltaic panel inclination angle of 0~90° to obtain the optimal installation inclination angle of the photovoltaic panel. Due to site limitations, it is usually necessary to install several rows of photovoltaic modules at a certain distance from front to back. The design of the spacing needs to comprehensively consider the local latitude, solar altitude angle, azimuth, array orientation and installation inclination angle and other parameters. The basic principle of the design is to ensure that the front and rear rows do not block each other during the period of 09:00-15:00 on the most unfavorable day of light throughout the year (i.e., the winter solstice). The optimal spacing can be calculated according to formula (2), where θ is the optimal installation inclination angle, α is the solar azimuth angle, h is the solar altitude angle, and l is the length of the photovoltaic panel.

[0050] (2)

[0051] According to the formula, the photovoltaic panel area can be deduced ( ) and roof usable area ( ), where is the width of the photovoltaic panel, as shown below:

[0052] (3)

[0053] Define the fill factor ( ) is the installed photovoltaic panel area ( ) and roof usable area ( ), the calculation formula is as follows.

[0054] (4)

[0055] Since installed capacity provides a fixed performance indicator that is not affected by weather fluctuations, this paper uses it as the core parameter for evaluating the potential of building rooftop photovoltaics. The installed capacity is calculated based on formula (5), and the corresponding power generation is calculated based on formula (6).

[0056] (5)

[0057] (6)

[0058] Where: is the peak power generation of photovoltaic panels per unit area; It is the power generation of the photovoltaic system on the roof of the building within a certain period of time; Indicates the length of the time period; Indicates the inverter efficiency; Total installed photovoltaic capacity; Indicates the amount of solar radiation; Indicates the amount of solar radiation under standard working conditions; Indicates heat loss efficiency; Indicates the operating temperature of the photovoltaic cell; Indicates the operating temperature of photovoltaic cells under standard working conditions.

[0059] (4) Semanticization of photovoltaic potential model on building roofs

[0060] The above-mentioned building photovoltaic potential assessment model is semantically modified to extract building photovoltaic potential assessment and result retrieval rules. The present invention uses the geographical entity of the "target assessment scope" as one of the triggering conditions of the assessment rule. When the attribute information of the "target assessment scope" meets the concept definition in the ontology model, the reasoning program will match the rule object that can be triggered. When the triggering condition of the rule is met, the reasoning program will execute the corresponding action, which is the photovoltaic potential assessment model formalized in computer language. The reasoning program automatically obtains the reasoning rules associated with the target calculation scope object based on the SPARQL query language, and triggers the action objects in sequence according to the event objects that meet the conditions according to the rules to complete the photovoltaic potential assessment, and output the calculation results according to the format requirements. The semantic evaluation rules are as follows: Figure 5 As shown in the figure, after the "target evaluation scope" in GeoJSON format is input into the inference program, its corresponding inference rule "input target evaluation scope rule" is matched; it is determined whether there is a spatial intersection between the "target evaluation scope" and the data of the "county-level administrative division" in the data base; if there is an intersection, the action object "check whether the county-level administrative divisions within the target calculation scope have been calculated" is executed; when the previous action is completed, the "notify the next rule" is executed; if all county-level administrative divisions have been calculated, the "output building photovoltaic potential result rule" is notified to start execution, and the list of "county-level administrative divisions" that have been calculated and the corresponding calculation results are passed to the rule; if there are county-level administrative divisions that have not been fully calculated, the "calculate building photovoltaic potential rule" is notified, and the list of "county-level administrative divisions" that have not been calculated is passed to the rule. After receiving the notification, the "rule for calculating the photovoltaic potential of buildings" determines whether the "buildings", "provincial administrative divisions", "prefectural administrative divisions", and "street-level administrative divisions" under the "county-level administrative divisions" in the data base are not all empty; if the requirements are met, the actions "calculate the installation spacing of the county-level administrative divisions", "calculate the photovoltaic potential of buildings", "aggregate or allocate the calculation results to the corresponding administrative divisions", and "notify the next rule" are executed in sequence; the "buildings" with the photovoltaic potential assessment results are passed to the "rule for outputting the photovoltaic potential results of buildings". After receiving the notification, the "rule for outputting the photovoltaic potential results of buildings" executes the action "aggregate and output the assessment results" to output the photovoltaic potential assessment results of buildings within the target assessment range in GeoJSON format.

[0061] Similarly, the geographic entity of the "target query range" is used as one of the triggering conditions of the query rule. Figure 6 As shown in the figure, the "target query range" can be any range, such as a city, a district, or even a community. When the "target query range" has a spatial intersection with the "Building PV Potential Results" entity in the data base, that is, there are buildings in the query range that have completed PV potential assessment, the query rule will be triggered to execute the action "summarize and output the photovoltaic potential of buildings in the target range", count the photovoltaic potential of buildings in the target query range, and output it in GeoJSON format. The semantic result retrieval rules are as follows: After the "target query range" in GeoJSON format is input into the inference program, the corresponding inference rule "query building photovoltaic potential rule" is matched; determine whether there is a spatial intersection between the "target query range" and the data of the "building photovoltaic potential results" in the data base; if there is an intersection, execute the action object "summarize and output the target query range content building photovoltaic potential assessment results", and output the building photovoltaic potential assessment results in the target query range in GeoJSON format.

Claims

1. A method for constructing an integrated knowledge graph computing framework for building rooftop photovoltaic potential assessment and query, characterized in that: The method comprises: Construct an ontology model of the knowledge graph; realize the fusion and collaborative analysis of multi-heterogeneous data in the evaluation of building photovoltaic potential by combing and integrating the relevant concepts of key data and expert knowledge, and construct a knowledge expression framework for the evaluation of building photovoltaic potential; Construct a data base for photovoltaic potential on building roofs; realize digital management of assessment data by instantiating building and other relevant geographic entity information, so as to provide a data basis for accurate photovoltaic potential assessment; Establish a building rooftop photovoltaic potential assessment model based on expert knowledge; determine the optimal installation parameters by quantifying the multi-level shading influencing factors, and achieve scientific assessment of the building's photovoltaic potential; Semanticization of building rooftop photovoltaic potential assessment model and query model; automation and intelligence of assessment and retrieval process are achieved through rule extraction based on reasoning mechanism, which is used to improve the efficiency and accuracy of assessment and retrieval; The ontology model of the knowledge graph includes the building photovoltaic potential ontology, spatial ontology and reasoning ontology; the building photovoltaic potential ontology involves relevant conceptual entities that appear in the evaluation and query process. The conceptual entities include buildings, administrative divisions, solar conditions, photovoltaic potential, photovoltaic equipment, target evaluation scope and target query scope. The administrative divisions include provinces, prefecture-level cities, districts, cities, counties and streets. Photovoltaic equipment is further connected to photovoltaic panel specifications, installation parameters and solar conditions. The photovoltaic potential conceptual entity allows the evaluation results to be updated synchronously in the knowledge graph. The conceptual entities of target evaluation scope and target query scope serve as the entry point for evaluation and query, defining the spatial boundaries of evaluation and query operations, which are used by users to analyze photovoltaic potential and retrieve data for specific geographical or logical areas. The building rooftop photovoltaic potential assessment model includes the calculation of the available area, installed capacity and first-year power generation. The reduction coefficient of the comprehensive city, block and individual building level is used to quantify the impact of shading factors on the actual available area. Buildings are divided into low-rise residential buildings, high-rise residential buildings, medium-rise buildings and industrial buildings according to their floor height and use. The corresponding reduction coefficient is calculated by formula (1), where is the urban factor reduction coefficient, is the block factor reduction coefficient, is the reduction factor of the single building factor; （1） The traversal search method is used to establish an annual sunlight radiation model on the inclined surface, and the 0-90° photovoltaic panel inclination angle is traversed to obtain the optimal installation inclination angle of the photovoltaic panel. The design of the photovoltaic module spacing comprehensively considers the local latitude, solar altitude angle, azimuth, array orientation and installation inclination parameters to ensure that the front and rear rows do not block each other during the period of 09:00-15:00 on the winter solstice, the most unfavorable day of the year. The optimal spacing is calculated according to formula (2), where θ is the optimal installation inclination angle, α is the solar azimuth angle, h is the solar altitude angle, and l is the length of the photovoltaic panel; （2） The photovoltaic panel area is derived according to the formula and roof area relationship, is the width of the photovoltaic panel, as shown below: （3） Defining the Fill Factor The area of ​​installed photovoltaic panels and roof area The ratio is calculated as follows: （4） The installed capacity is calculated based on formula (5), and the corresponding power generation is calculated based on formula (6). （5） （6） Where: is the peak power generation of photovoltaic panels per unit area; It is the power generation of the photovoltaic system on the roof of the building within a certain period of time; Indicates the length of the time period; Indicates the inverter efficiency; Total installed photovoltaic capacity; Indicates the amount of solar radiation; Indicates the amount of solar radiation under standard working conditions; Indicates heat loss efficiency; Indicates the operating temperature of the photovoltaic cell; Indicates the operating temperature of photovoltaic cells under standard working conditions.

2. The method according to claim 1, characterized in that GeoSPARQL is used as the spatial ontology to express administrative divisions, buildings and other geographic entities with spatial characteristics. The spatial ontology includes spatial objects, which include two subclasses: features and geometry. Features have geometric characteristics. Geometric features are divided into points, polylines and polygons according to the geometric shapes of geographic features. They are used to calculate spatial operations including distance, spatial intersection and union, and geographic buffer zone. Geographic entity data instances correspond to spatial objects in the spatial ontology, realizing the effective association between geographic entities and ontology concepts, and constructing a knowledge expression framework that is conducive to spatial operations.

3. The method according to claim 2, characterized in that The reasoning ontology is the reasoning rule constraints for the automatic execution of reasoning programs on the knowledge graph. The reasoning rules contain event objects and action objects. The event objects are the triggering conditions of the reasoning rules, and the action objects are the corresponding actions executed after the reasoning rules are triggered. Event objects are divided into independent events and event combinations according to the independence of events, and action objects are divided into independent actions and combined actions. The reasoning rules are associated with each other through specific action objects notifying the next rule and event objects receiving notification from the previous rule.

4. The method according to claim 1, characterized in that: The data base for photovoltaic potential on building roofs is composed of instantiated source data. Instantiation refers to mapping specific data to entities and relationships in the knowledge graph through extraction and fusion. By reading the building information and administrative division information extracted based on remote sensing images in the data files one by one, and through knowledge extraction and knowledge fusion, the spatial and attribute information of each geographic entity of buildings, provinces, prefecture-level cities, districts, counties, and streets is instantiated in the form of triples, thus building a digital base for evaluating the photovoltaic potential of buildings.

5. The method according to claim 4, characterized in that Knowledge extraction refers to the process of automatically identifying and extracting valuable information and relationships from data. Geographic entity data is stored in GeoJSON format before entering the knowledge graph data base. Geometry in GeoJSON format is mapped to the predicate of spatial information triples, and Geometry values ​​are automatically mapped to the objects of spatial information triples that conform to the GeoSPARQL format specification; the keys in the attribute set are automatically mapped to the predicates of the corresponding type of geographic entity attribute triples, and the attribute values ​​are automatically mapped to the objects of the corresponding predicates in the triples, and the triples are reorganized into multiple key-value pairs according to the following rules and stored in the Key-Value database: 1) "Subject + Predicate": "Object"; 2) "object + predicate": "subject"; 3) "geographic entity type + system timestamp of the instance": "subject"; 4) "geographic entity type + geocode to which the instance belongs": "subject".

6. The method according to claim 5, characterized in that For geographic entities with spatial characteristics, in addition to the attribute information constrained in the ontology, multi-scale geocoding index information is added; the geocoding is calculated based on the Web Mercator projection gridding method and the geographic entity is indexed in the spatiotemporal knowledge graph. The triple representation of the geocoding index information is: geographic entity hasTileCode tile code, subject "geographic entity", using the predicate "hasTileCode" to connect its corresponding object "tile code", the tile code is a string format as follows: "z00_x0000_y0000", z represents the gridding level, the value is in the range of 0-26, x and y represent the horizontal and vertical coordinates of the gridding space, a geographic entity has a series of tile codes of different scales, different types of geographic entities are inserted into multi-scale tile codes in the process of building the data base, and matching is performed based on the tile codes associated with the geographic entity to support multi-scale spatial retrieval of spatiotemporal elements.

7. The method according to claim 6, characterized in that The semanticization of the photovoltaic potential model on the roof of a building includes: the geographic entity of the target calculation range is used as the trigger condition of the evaluation rule. When the attribute information of the target calculation range meets the concept definition in the ontology model, the reasoning program matches the triggerable rule object. When the trigger condition of the rule is met, the reasoning program will execute the corresponding action, which is the photovoltaic potential evaluation model formalized using computer language.

8. The method according to claim 7, characterized in that The semanticization of the building rooftop photovoltaic potential model also includes: the geographic entity of the target query range is used as the trigger condition of the query rule, and the target query range is an arbitrary range; when the target query range has a spatial intersection with the building photovoltaic potential result entity in the data base, that is, there are buildings in the query range that have completed photovoltaic potential assessment, the query rule will be triggered to execute the action to summarize and output the photovoltaic potential of the buildings within the target range, count the photovoltaic potential of the buildings within the target query range, and output it in GeoJSON format.

9. A building rooftop photovoltaic potential assessment query integrated knowledge graph computing system based on the method described in any one of claims 1 to 8, characterized in that: include: Ontology model of knowledge graph; used to integrate and express key concepts, entities and their interrelationships in the field of building photovoltaic potential assessment, and provide structured knowledge representation for data fusion and collaborative analysis; Building rooftop photovoltaic potential data base; used to store and manage multivariate heterogeneous data, including building geometry, geography and attribute information, as well as photovoltaic equipment performance parameters, solar radiation conditions, etc., to provide data support for the evaluation model; Expert knowledge-based building rooftop photovoltaic potential assessment model; used to achieve scientific calculation of building rooftop photovoltaic potential, including the assessment of key indicators such as available area, installed capacity and first-year power generation, to ensure the accuracy and reliability of the assessment results; The semantic module of the building rooftop photovoltaic potential assessment model and query model is used to associate the assessment model, query model and related information and data in the knowledge graph, realize the automation and intelligence of the assessment process and query process, and improve the operational efficiency of the assessment and result retrieval.

10. A computer device, characterized in that: include: at least one memory and at least one processor; The memory is used to store one or more programs; When the one or more programs are executed by at least one processor, the at least one processor implements the method according to any one of claims 1 to 8.

11. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method according to any one of claims 1 to 8 is implemented.

12. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images