A reducer data query method based on multi-modal data
By using a multimodal data clustering and similarity matching method for speed reducer data query, the problems of low query efficiency and inaccurate matching are solved, and efficient and accurate query result display and interactive operation are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HANGZHOU JIE DRIVE TECH
- Filing Date
- 2022-08-22
- Publication Date
- 2026-06-09
Smart Images
Figure CN115455513B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of information technology for speed reducers in engineering machinery equipment, specifically, to a speed reducer data query method based on multimodal data. Background Technology
[0002] Mechanical design is an arduous and meticulous task. Traditional manual methods of mechanical design suffer from low efficiency, large calculation errors, and wasted human and material resources. While advanced CAD technology can address these issues, its geographical limitations hinder remote data transmission and synchronous data sharing, requiring further improvement. Web-based design systems leverage the efficiency and interactivity of the internet to achieve rapid design task completion, unrestricted by geographical or time constraints.
[0003] Currently, product information is primarily communicated between companies and users through product catalogs. However, due to the comprehensive functions and diverse types of general-purpose products such as speed reducers, and especially with the ever-accelerating pace of product updates, traditional catalogs often cannot keep up with product upgrades, resulting in untimely information updates. Furthermore, with the increasing number and variety of similar products, users often feel overwhelmed by the sheer volume of catalogs and frequently have to sift through numerous catalogs to purchase a single product, wasting time and effort.
[0004] Meanwhile, due to the diversified demands of various parties for speed reducers, there are significant problems in matching various performance parameters with speed reducers, and the use of various speed reducer parameters is also insufficient. Therefore, the existing technology suffers from low speed reducer query efficiency. Summary of the Invention
[0005] To address the current issues of mismatched and inaccurate matching in the field of speed reducers, this invention seeks protection for a speed reducer data query method based on multimodal data, characterized by comprising:
[0006] Obtain metadata of the speed reducers to be put into storage;
[0007] Perform multimodal data clustering decomposition on the metadata of the speed reducer to be queried;
[0008] The metadata of the speed reducer after the multimodal data is clustered and decomposed is then entered into the database.
[0009] Retrieve the required data for the speed reducer to be queried;
[0010] Perform multimodal data clustering on the required speed reducer data to be queried;
[0011] The query data on reducer requirements obtained from the multimodal data clustering is compared with the reducer metadata obtained from the multimodal data clustering decomposition in the database.
[0012] Once the speed reducer query results are obtained, the speed reducer query results are pushed to the querying user.
[0013] Specifically, obtaining the metadata of the speed reducer to be put into storage also includes:
[0014] The gearbox entry terminal receives a metadata extraction request message for the gearbox to be entered into the warehouse from the node to be entered into the warehouse. The metadata extraction request message for the gearbox to be entered into the warehouse includes the identification information of the first metadata, which is the multimodal dataset of the gearbox to be entered into the warehouse.
[0015] The gearbox entry terminal determines the metadata information of the second metadata based on the metadata information of the first metadata. The second metadata is a gearbox multimodal dataset that has an access association with the first metadata.
[0016] The reducer receiving end sends the metadata information of the first metadata and the metadata information of the second metadata to the node to be received, so that the node to be received adds the metadata information of the first metadata and the metadata information of the second metadata to its cache.
[0017] Specifically, the multimodal data clustering decomposition of the metadata of the speed reducer to be queried further includes:
[0018] The metadata information of the first metadata and the metadata information of the second metadata in the cache of the node to be added to the database are clustered by category;
[0019] Based on the attribute information of the metadata information, the metadata information is divided into first multimodal data, second multimodal data, and third multimodal data.
[0020] Specifically, the step of storing the speed reducer metadata after multimodal data clustering and decomposition into the database also includes:
[0021] Multiple multimodal data partitions are established in the database, and the multimodal data partitions are used to store the metadata of the speed reducer after clustering and decomposition.
[0022] The metadata of the speed reducer after clustering and decomposition is stored in the corresponding multimodal data partition.
[0023] Specifically, obtaining the required data for the speed reducer to be queried also includes:
[0024] Receive user requests for data on the required speed reducer;
[0025] The type of query content in the request to obtain the required data of the speed reducer is determined. If the query content is one or more keywords, the ordered group of keywords is obtained directly based on the query content. If the query content is one or more sentences, the sentences are processed one by one to obtain the ordered group of keywords.
[0026] Process the ordered groups of keywords to obtain the ordered groups of intermediate keywords;
[0027] Multimodal data clustering is performed based on the ordered groups of intermediate keywords.
[0028] Specifically, the multimodal data clustering of the speed reducer demand data to be queried further includes: performing category clustering on the metadata information of the speed reducer demand data to be queried;
[0029] Based on the attribute information of the metadata information, the metadata information is divided into first multimodal data, second multimodal data, and third multimodal data.
[0030] Specifically, the step of comparing the query speed reducer requirement data obtained from the multimodal data clustering with the speed reducer metadata obtained from the multimodal data clustering decomposition in the database further includes:
[0031] The similarity of the query reducer requirement data after clustering the multimodal data is compared with the reducer metadata after clustering and decomposition of the multimodal data in the database for the corresponding modal data.
[0032] If the similarity between the first multimodal data of the gearbox metadata after multimodal data clustering and decomposition in the database and the first multimodal data of the gearbox demand data to be queried after multimodal data clustering is greater than a first threshold, then the gearbox corresponding to the gearbox metadata after multimodal data clustering and decomposition in the database will be used as the query result.
[0033] If the similarity between the first multimodal data of the gearbox metadata after multimodal data clustering and decomposition in the database and the first multimodal data of the gearbox demand data to be queried after multimodal data clustering is not greater than a first threshold, then the similarity between the second multimodal data of the gearbox metadata after multimodal data clustering and decomposition in the database and the second multimodal data of the gearbox demand data to be queried after multimodal data clustering is further compared to whether it is greater than a second threshold. If it is greater than the second threshold, then the gearbox corresponding to the gearbox metadata after multimodal data clustering and decomposition in the database is taken as the query result.
[0034] If the similarity is not greater than the second threshold, then the similarity between the third multimodal data of the gearbox metadata after multimodal data clustering and decomposition in the database and the third multimodal data of the gearbox demand data to be queried after multimodal data clustering is further compared to whether it is greater than the third threshold. If it is greater than the third threshold, then the gearbox corresponding to the gearbox metadata after multimodal data clustering and decomposition in the database is taken as the query result.
[0035] If the result is not greater than the third threshold, no query results are returned.
[0036] Specifically, obtaining the speed reducer query results and pushing the speed reducer query results to the querying user also includes:
[0037] If a query result exists, the speed reducer corresponding to the speed reducer metadata obtained after multimodal data clustering and decomposition in the database will be used as the query result.
[0038] Specifically, when the multimodal data is clustered and decomposed and then stored in the database, a mapping relationship is established between the speed reducer metadata and the speed reducer. The mapping relationship between the speed reducer metadata and the speed reducer is obtained, and based on the mapping relationship, the speed reducer corresponding to the speed reducer metadata is obtained as the query result.
[0039] The speed reducer is displayed to the querying user in a multimodal manner.
[0040] This invention utilizes modal-based clustering decomposition of speed reducer parameters. It compares the query speed reducer requirement data from multimodal data clustering with the corresponding multimodal data clustered metadata of the speed reducer in the database. Based on the similarity matching of data under each modality and the matching priority of each modality type, it rationally analyzes the user's speed reducer requirements and provides query results that meet the user's query needs. This solution relies on multimodal data processing technology to creatively decompose and cluster the metadata of transmission machinery structures such as speed reducers. In the subsequent release of query results, it achieves dynamic display of the product model, allowing users to easily view the product's two-dimensional or three-dimensional model during the query process. Users can perform interactive operations such as zooming, moving, and rotating to observe the product from different angles and also to observe local details. Attached Figure Description
[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0042] Figure 1 This is a flowchart illustrating the workflow of a speed reducer data query method based on multimodal data, as described in this invention.
[0043] Figure 2 This is a flowchart illustrating a first embodiment of a speed reducer data query method based on multimodal data, as described in this invention.
[0044] Figure 3 This is a schematic diagram of a second embodiment of a speed reducer data query method based on multimodal data involved in the present invention. Detailed Implementation
[0045] The illustrative embodiments of this application include, but are not limited to, a method for querying speed reducer data based on multimodal data.
[0046] It is understood that, as used herein, the terms "module" and "unit" may refer to or include, or be part of, application-specific integrated circuits (ASICs), electronic circuits, processors (shared, dedicated, or grouped) and / or memory that execute one or more software or firmware programs, combinational logic circuits, and / or other suitable hardware components that provide the described functionality.
[0047] It is understood that in the various embodiments of this application, the processor may be a microprocessor, a digital signal processor, a microcontroller, etc., and / or any combination thereof. According to another aspect, the processor may be a single-core processor, a multi-core processor, etc., and / or any combination thereof.
[0048] It is understood that the speed reducer data query method based on multimodal data provided in this application can be implemented on various electronic devices, including but not limited to servers, distributed server clusters composed of multiple servers, mobile phones, tablets, laptops, desktop computers, wearable devices, head-mounted displays, mobile email devices, portable game consoles, portable music players, e-readers, personal digital assistants, virtual reality or augmented reality devices, and televisions with one or more processors embedded or coupled thereto.
[0049] See attached document Figure 1 This invention claims protection for a method for querying speed reducer data based on multimodal data, characterized by comprising:
[0050] Obtain metadata of the speed reducers to be put into storage;
[0051] Perform multimodal data clustering decomposition on the metadata of the speed reducer to be queried;
[0052] The metadata of the speed reducer after the multimodal data is clustered and decomposed is then entered into the database.
[0053] Retrieve the required data for the speed reducer to be queried;
[0054] Perform multimodal data clustering on the required speed reducer data to be queried;
[0055] The query data on reducer requirements obtained from the multimodal data clustering is compared with the reducer metadata obtained from the multimodal data clustering decomposition in the database.
[0056] Once the speed reducer query results are obtained, the speed reducer query results are pushed to the querying user.
[0057] Specifically, obtaining the metadata of the speed reducer to be put into storage also includes:
[0058] The gearbox entry terminal receives a metadata extraction request message for the gearbox to be entered into the warehouse from the node to be entered into the warehouse. The metadata extraction request message for the gearbox to be entered into the warehouse includes the identification information of the first metadata, which is the multimodal dataset of the gearbox to be entered into the warehouse.
[0059] The gearbox entry terminal determines the metadata information of the second metadata based on the metadata information of the first metadata. The second metadata is a gearbox multimodal dataset that has an access association with the first metadata.
[0060] The reducer receiving end sends the metadata information of the first metadata and the metadata information of the second metadata to the node to be received, so that the node to be received adds the metadata information of the first metadata and the metadata information of the second metadata to its cache.
[0061] The speed reducer includes friction belt drive and meshing belt drive. According to the cross-sectional shape of the transmission belt, it is divided into: flat belt, V belt, multi-ribbed belt, round belt, and synchronous belt. The design principle of belt drive is: to transmit the specified power without slipping, and at the same time have sufficient fatigue strength and a certain service life.
[0062] This invention employs a special method to implement graphical query tasks and is implemented programmatically. This method requires tabulating the graphics, that is, representing the numbers in the graphics using a table, so that the program can automatically search for results. The program uses arrays and a series of loop statements to complete the entire query task.
[0063] Specifically, as illustrated below, the first metadata includes at least the power data of the reducer, and the second metadata includes at least the speed data of the small pulley of the reducer. The first metadata and the second metadata have an image correspondence relationship.
[0064] The main parameters of speed reducer products include motor power, transmission ratio, speed, torque, installation dimensions, working load, working environment, and impact strength. However, the parameters vary slightly between different types. Taking NGW as an example, the detailed parameters are as follows: NGW series: model, serial number, transmission ratio, number of transmission stages, power, speed, price, etc.; NGW-LDF series: model, serial number, transmission ratio, number of transmission stages, power, speed, price, etc.; NGW-S series: model, serial number, transmission ratio, number of transmission stages, power, assembly method, speed, price, etc.; NGW-Z series: model, serial number, transmission ratio, number of transmission stages, power, assembly method.
[0065] The main external components of the speed reducer include the body, housing, rear cover, and front cover; the main internal components include the low-speed shaft, high-speed shaft, gear shaft, gears, and bushings.
[0066] Calculate the rated power based on the input parameters, and the selected power should be less than the nominal input power. The formula for calculating the rated power is: Selected power = Actual input power + Usage coefficient x Starting coefficient x Reliability coefficient < Nominal input power.
[0067] Thermal balance verification is performed to determine whether a cooling device needs to be installed. The formula for permissible thermal balance power is: actual input power x ambient temperature coefficient x operating cycle coefficient x power utilization coefficient < thermal balance power.
[0068] Users can find relevant products that meet their needs by simply entering the relevant parameters (multiple products may meet the conditions); if no relevant products are found, the system will automatically record the details, and the design department can quickly design the product that the user expects by querying the design system.
[0069] The input method of this invention employs an intelligent interface implemented using HTML and JavaScript. It achieves a user-friendly window interface and an intelligent navigation system, allowing users to easily understand the system architecture and usage. Clients access the system through a client-side HTML page. To achieve the dynamic effects of a Windows interactive page, HTML alone is insufficient; therefore, JavaScript is used to develop the interface, maximizing its Windows-like functionality.
[0070] The main program, implemented in JSP and Java, primarily handles the logical concepts of belt drive design, automatically retrieving data from the backend database, and performing calculations. Since the belt drive design process involves numerous diagrams and data, a backend database is essential for its reliability. The middleware layer accesses the backend database via JDBC and communicates with the browser via Sockets or HTTP protocols. The design also includes extensive calculations and form extraction, utilizing object-oriented Java and JSP to complete the entire calculation and data processing process. This includes tasks such as belt drive queries, data interaction between web pages, data table filtering, and belt drive validation.
[0071] The backend database is built using Oracle and stores the chart data. It includes a management database for recording information such as channels, groups, and users. The backend also includes a separate Java application—System Manager—to help users manage the management database.
[0072] Specifically, the multimodal data clustering decomposition of the metadata of the speed reducer to be queried further includes:
[0073] The metadata information of the first metadata and the metadata information of the second metadata in the cache of the node to be added to the database are clustered by category;
[0074] Based on the attribute information of the metadata information, the metadata information is divided into first multimodal data, second multimodal data, and third multimodal data.
[0075] The first multimodal data is text data;
[0076] The second multimodal data is chart data;
[0077] The third multimodal data is image data;
[0078] Specifically, the step of storing the speed reducer metadata after multimodal data clustering and decomposition into the database also includes:
[0079] Multiple multimodal data partitions are established in the database, and the multimodal data partitions are used to store the metadata of the speed reducer after clustering and decomposition.
[0080] The metadata of the speed reducer after clustering and decomposition is stored in the corresponding multimodal data partition.
[0081] Obtain data partitioning rules, which are used to indicate the correspondence between multimodal data and partitions;
[0082] Based on the data partitioning rules and the correspondence between partitions and data partition identifiers, a correspondence between multimodal data and data partition identifiers is established.
[0083] Add corresponding data partition identifiers to data entries in the database according to the correspondence between multimodal data and data partition identifiers; and
[0084] The data entry is stored in the corresponding partition of the database based on its data partition identifier.
[0085] The step of adding the corresponding data partition identifier further includes:
[0086] Formulate a data query statement based on the aforementioned multimodal data;
[0087] Obtain the data entries corresponding to the multimodal data;
[0088] The matching table is queried to obtain the data partition identifier corresponding to the multimodal data; and the data partition identifier is added to the obtained data entries corresponding to the multimodal data.
[0089] Specifically, obtaining the required data for the speed reducer to be queried also includes:
[0090] Receive user requests for data on the required speed reducer;
[0091] The type of query content in the request to obtain the required data of the speed reducer is determined. If the query content is one or more keywords, the ordered group of keywords is obtained directly based on the query content. If the query content is one or more sentences, the sentences are processed one by one to obtain the ordered group of keywords.
[0092] Process the ordered groups of keywords to obtain the ordered groups of intermediate keywords;
[0093] Multimodal data clustering is performed based on the ordered groups of intermediate keywords.
[0094] Specifically, the multimodal data clustering of the speed reducer demand data to be queried further includes: performing category clustering on the metadata information of the speed reducer demand data to be queried;
[0095] Based on the attribute information of the metadata information, the metadata information is divided into first multimodal data, second multimodal data, and third multimodal data.
[0096] Various model parameters include: New Standard Planetary Gear (NGW): model code, specifications, number of transmission stages, transmission ratio, assembly type, connection method, speed, and power; Old Standard Planetary Gear (NGW): model code, frame size, number of transmission stages, transmission ratio code, assembly type, speed, and power; Cylindrical Gear: model code, number of transmission stages, whether it is a hollow output, center distance between the last poles, ratio, assembly type, cooling method, whether it has a backstop, input shaft rotation direction, speed, and nominal transmission power; Bevel Cylindrical Gear: model code, number of transmission stages, whether it is a hollow output, center distance between the last poles, nominal transmission ratio, assembly type, input shaft rotation direction, cooling method, whether it has a backstop, speed, and power.
[0097] The first multimodal data is text data of the various model parameters mentioned above;
[0098] The second multimodal data consists of chart data of the various model parameters mentioned above;
[0099] The third multimodal data consists of image data of the various model parameters mentioned above;
[0100] If users are clear about the performance indicators and parameter values of the equipment they want to select, they can directly input these performance indicator parameter values. The system will then use calculations and reasoning based on knowledge rules to quickly and conveniently query the required equipment and its components. For example, for a speed reducer, the system calculates the parameters under the known operating conditions and the selected working medium. Based on the calculation results and the results of knowledge reasoning, the system obtains other important parameter values for the speed reducer. Then, based on these parameter values, it lists detailed information about the relevant speed reducer equipment. Users can then select the speed reducer equipment that meets their requirements from the detailed equipment information. Then, based on the main equipment, the system first performs matching rules to find several auxiliary equipment products that match this main product. Then, based on parameter calculations, it selects some matching auxiliary equipment products. The speed reducer query design is complete.
[0101] For details, please refer to the appendix. Figure 2 The step of comparing the query speed reducer requirement data obtained from the multimodal data clustering with the speed reducer metadata obtained from the multimodal data clustering decomposition in the database further includes:
[0102] The similarity of the query reducer requirement data after clustering the multimodal data is compared with the reducer metadata after clustering and decomposition of the multimodal data in the database for the corresponding modal data.
[0103] If the similarity between the first multimodal data of the gearbox metadata after multimodal data clustering and decomposition in the database and the first multimodal data of the gearbox demand data to be queried after multimodal data clustering is greater than a first threshold, then the gearbox corresponding to the gearbox metadata after multimodal data clustering and decomposition in the database will be used as the query result.
[0104] If the similarity between the first multimodal data of the gearbox metadata after multimodal data clustering and decomposition in the database and the first multimodal data of the gearbox demand data to be queried after multimodal data clustering is not greater than a first threshold, then the similarity between the second multimodal data of the gearbox metadata after multimodal data clustering and decomposition in the database and the second multimodal data of the gearbox demand data to be queried after multimodal data clustering is further compared to whether it is greater than a second threshold. If it is greater than the second threshold, then the gearbox corresponding to the gearbox metadata after multimodal data clustering and decomposition in the database is taken as the query result.
[0105] If the similarity is not greater than the second threshold, then the similarity between the third multimodal data of the gearbox metadata after multimodal data clustering and decomposition in the database and the third multimodal data of the gearbox demand data to be queried after multimodal data clustering is further compared to whether it is greater than the third threshold. If it is greater than the third threshold, then the gearbox corresponding to the gearbox metadata after multimodal data clustering and decomposition in the database is taken as the query result.
[0106] If the result is not greater than the third threshold, no query results are returned.
[0107] Specifically, for example, the first multimodal data is text data, that is, the keywords entered by the user are divided into text keywords and matched with the keyword text stored in the database for various model parameters. The first threshold is 80%. When the first threshold is not met, further comparison is made with chart data, which is a chart representation of the various model parameters, such as a line graph. Figure 3 The chart statistics of design power and small pulley speed are used to set a second threshold of 75% based on the slope comparison. When the second threshold is not met, the image data is further compared. The image data consists of images of each type of reducer in operation and in static display. A third threshold is set by comparing the image similarity, and the third threshold is set to 70%. The query result is obtained or the query result that does not exist is returned.
[0108] Select the input query parameters from the three categories of speed reducers: planetary gears (NGW), cylindrical gears, and bevel cylindrical gears. The parameters include installation method, motor power, and working condition.
[0109] Calculate the rated power using empirical formulas; Select power = actual input power x operating condition factor x safety factor, and the calculated result must be less than the nominal input power;
[0110] Verify the permissible thermal balance power; Actual input power × ambient temperature coefficient × hourly load duty cycle coefficient × nominal power utilization coefficient < thermal balance power;
[0111] Verify other power parameters, identify products that meet the requirements, and decide whether to install a cooler based on the permissible power data for thermal balance.
[0112] Specifically, obtaining the speed reducer query results and pushing the speed reducer query results to the querying user also includes:
[0113] If a query result exists, the speed reducer corresponding to the speed reducer metadata obtained after multimodal data clustering and decomposition in the database will be used as the query result.
[0114] Specifically, when the multimodal data is clustered and decomposed and then stored in the database, a mapping relationship is established between the speed reducer metadata and the speed reducer. The mapping relationship between the speed reducer metadata and the speed reducer is obtained, and based on the mapping relationship, the speed reducer corresponding to the speed reducer metadata is obtained as the query result.
[0115] The speed reducer is displayed to the querying user in a multimodal manner.
[0116] Using the ActiveX controls provided by CAD software systems for viewing 3D models allows for convenient display of 3D models. Taking SolidEdge CAD software as an example, it offers two ActiveX controls: SEPreview and SEPartX. These controls allow you to set the displayed 3D model through their PartFile property. Both controls support interactive operations such as rotation, translation, and scaling using the mouse. SEPreview is primarily used to display SolidEdge part models, while the SEPartX control can display SolidEdge Par (parts), ASM (assemblies), and DFT processes. All SolidEdge files (including those in Figure 1) and AutoCAD DWG and DXF files; the 3D models controlled and displayed by SEPreview and SEPartx are only applicable to SolidEdge model files. For 3D models from Pro / E, UG, and other systems, they can be opened directly in 501idEdge and then saved as 501idEdge files; for model files created by other CAD software that cannot be opened directly by 501idEdge, they can be converted to neutral files such as .GES or STEP in the original system first, and then Par or ASM files can be generated using SolidEdge;
[0117] It should be noted that all units / modules mentioned in the device embodiments of this application are logical units / modules. Physically, a logical unit / module can be a physical unit / module, a part of a physical unit / module, or a combination of multiple physical units / modules. The physical implementation of these logical units / modules themselves is not the most important factor; the combination of functions implemented by these logical units / modules is the key to solving the technical problems proposed in this application. Furthermore, to highlight the innovative aspects of this application, the above-described device embodiments of this application have not introduced units / modules that are not closely related to solving the technical problems proposed in this application. This does not mean that the above-described device embodiments do not contain other units / modules.
[0118] It should be noted that in the examples and description of this application, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0119] Although this application has been illustrated and described with reference to certain preferred embodiments thereof, those skilled in the art should understand that various changes in form and detail may be made thereto without departing from the spirit and scope of this application.
Claims
1. A method for querying speed reducer data based on multimodal data, characterized in that, include: Obtain metadata of the speed reducers to be put into storage; Perform multimodal data clustering decomposition on the metadata of the speed reducer to be queried; The metadata of the speed reducer after the multimodal data is clustered and decomposed is then entered into the database. Retrieve the required data for the speed reducer to be queried; Perform multimodal data clustering on the required speed reducer data to be queried; The query data on reducer requirements obtained from the multimodal data clustering is compared with the reducer metadata obtained from the multimodal data clustering decomposition in the database. Once the speed reducer query results are obtained, the speed reducer query results are pushed to the querying user. The step of performing multimodal data clustering on the speed reducer demand data to be queried further includes: performing category clustering on the metadata information of the speed reducer demand data to be queried; and dividing the metadata information into first multimodal data, second multimodal data, and third multimodal data according to the attribute information of the metadata information. The step of comparing the query reducer demand data obtained from the multimodal data clustering with the reducer metadata obtained from the multimodal data clustering decomposition in the database further includes: comparing the similarity of the corresponding modal data between the query reducer demand data obtained from the multimodal data clustering and the reducer metadata obtained from the multimodal data clustering decomposition in the database; if the similarity between the first multimodal data of the reducer metadata obtained from the multimodal data clustering decomposition in the database and the first multimodal data of the query reducer demand data obtained from the multimodal data clustering is greater than a first threshold, then the reducer corresponding to the reducer metadata obtained from the multimodal data clustering decomposition in the database is taken as the query result; if the similarity between the first multimodal data of the reducer metadata obtained from the multimodal data clustering decomposition in the database and the first multimodal data of the query reducer demand data obtained from the multimodal data clustering is greater than a first threshold, then the reducer corresponding to the reducer metadata obtained from the multimodal data clustering decomposition in the database is taken as the query result; If the similarity is not greater than the first threshold, then the similarity between the second multimodal data of the gearbox metadata after multimodal data clustering and decomposition in the database and the second multimodal data of the gearbox demand data after multimodal data clustering is greater than the second threshold; if it is greater than the second threshold, then the gearbox corresponding to the gearbox metadata after multimodal data clustering and decomposition in the database is used as the query result; if it is not greater than the second threshold, then the similarity between the third multimodal data of the gearbox metadata after multimodal data clustering and decomposition in the database and the third multimodal data of the gearbox demand data after multimodal data clustering is greater than the third threshold; if it is greater than the third threshold, then the gearbox corresponding to the gearbox metadata after multimodal data clustering and decomposition in the database is used as the query result; if it is not greater than the third threshold, then no query result is returned.
2. The method for querying reducer data based on multimodal data as described in claim 1, characterized in that, The process of obtaining the metadata of the speed reducer to be put into storage also includes: The gearbox entry terminal receives a metadata extraction request message for the gearbox to be entered into the warehouse from the node to be entered into the warehouse. The metadata extraction request message for the gearbox to be entered into the warehouse includes the identification information of the first metadata, which is the multimodal dataset of the gearbox to be entered into the warehouse. The gearbox entry terminal determines the metadata information of the second metadata based on the metadata information of the first metadata. The second metadata is a gearbox multimodal dataset that has an access association with the first metadata. The reducer receiving end sends the metadata information of the first metadata and the metadata information of the second metadata to the node to be received, so that the node to be received adds the metadata information of the first metadata and the metadata information of the second metadata to its cache.
3. The method for querying reducer data based on multimodal data as described in claim 2, characterized in that, The multimodal data clustering decomposition of the metadata of the speed reducer to be queried also includes: The metadata information of the first metadata and the metadata information of the second metadata in the cache of the node to be added to the database are clustered by category; Based on the attribute information of the metadata information, the metadata information is divided into first multimodal data, second multimodal data, and third multimodal data.
4. The method for querying reducer data based on multimodal data as described in claim 1, characterized in that, The step of storing the speed reducer metadata after multimodal data clustering and decomposition into a database also includes: Multiple multimodal data partitions are established in the database, and the multimodal data partitions are used to store the metadata of the speed reducer after clustering and decomposition. The metadata of the speed reducer after clustering and decomposition is stored in the corresponding multimodal data partition.
5. The method for querying reducer data based on multimodal data as described in claim 1, characterized in that, The process of obtaining the required data for the speed reducer to be queried also includes: Receive user requests for data on the required speed reducer; The type of query content in the request to obtain the required data of the speed reducer is determined. If the query content is one or more keywords, the ordered group of keywords is obtained directly based on the query content. If the query content is one or more sentences, the sentences are processed one by one to obtain the ordered group of keywords. Process the ordered groups of keywords to obtain the ordered groups of intermediate keywords; Multimodal data clustering is performed based on the ordered groups of intermediate keywords.
6. The method for querying reducer data based on multimodal data as described in claim 1, characterized in that, The step of obtaining the speed reducer query result and pushing the speed reducer query result to the querying user also includes: If a query result exists, the speed reducer corresponding to the speed reducer metadata obtained after multimodal data clustering and decomposition in the database will be used as the query result. Specifically, when the multimodal data is clustered and decomposed and then stored in the database, a mapping relationship is established between the speed reducer metadata and the speed reducer. The mapping relationship between the speed reducer metadata and the speed reducer is obtained, and based on the mapping relationship, the speed reducer corresponding to the speed reducer metadata is obtained as the query result. The speed reducer is displayed to the querying user in a multimodal manner.