Digital twinborn body data processing method, device, equipment, medium and program product

Abstract

The invention discloses a digital twin data processing method and device, equipment, a medium and a program product, and the method comprises the steps: obtaining historical equipment data which comprises equipment specification data, and twin hierarchy label data and equipment attribute label data corresponding to the equipment specification data; inputting the equipment standard data into the digital twinborn prediction model for encoding operation and decoding operation to obtain an equipment level prediction result and an equipment attribute prediction result; and performing model training based on the equipment level prediction result, the equipment attribute prediction result, the twinborn body level label data and the equipment attribute label data to generate a digital twinborn body prediction model. According to the embodiment of the invention, the encoding and decoding processing is carried out on the equipment specification data through the digital twinborn prediction model, the hierarchy and attribute relationship between different equipment is obtained, the automatic construction of the digital twinborn is realized, and the construction accuracy and efficiency are improved.

Description

Technical Field

[0001] This application belongs to the technical field of digital twins, and particularly relates to a method, apparatus, equipment, medium and program product for processing digital twin data. Background Technology

[0002] Digital twins fully utilize data such as physical models, sensor updates, and operational history to integrate multi-disciplinary, multi-physical, multi-scale, and multi-probability simulation processes, completing mapping in virtual space to reflect the entire lifecycle of the corresponding physical equipment.

[0003] Digital twins require the collection and analysis of large amounts of data. Traditional methods involve manual collection, analysis, modeling, and simulation testing, which consume a lot of human and material resources, and the resulting digital twins have poor generalization ability and flexibility.

[0004] Therefore, how to develop an intelligent method for constructing digital twins is a technical problem that urgently needs to be solved. Summary of the Invention

[0005] This application provides a digital twin data processing method, apparatus, device, medium, and program product, which can realize the automated construction of digital twins and improve the accuracy and efficiency of construction.

[0006] In a first aspect, embodiments of this application provide a digital twin data processing method based on intelligent algorithms, comprising: acquiring historical device data corresponding to a target scenario, wherein the historical device data includes device specification data and twin-level label data and device attribute label data corresponding to the device specification data; inputting the device specification data into a digital twin prediction model, encoding the device specification data based on feature extraction from the digital twin prediction model to obtain device specification encoded data; decoding the device specification encoded data based on the decoding layer of the digital twin prediction model to obtain device-level prediction results and device attribute prediction results; and training a model based on the device-level prediction results, the device attribute prediction results, the twin-level label data, and the device attribute label data to generate a digital twin prediction model.

[0007] In one possible implementation, after generating the digital twin prediction model, the method further includes: inputting target device data into the digital twin prediction model, outputting the hierarchical relationship results and attribute relationship results of the target device data; and sending the hierarchical relationship results and attribute relationship results to the digital twin platform to generate visualized digital twin data.

[0008] In one possible implementation, sending the hierarchical relationship results and the attribute relationship results to a digital twin platform to generate visualized digital twin data includes: determining the structured text data and structured table data corresponding to the hierarchical relationship results and the attribute relationship results; sending the structured text data and the structured table data to the digital twin platform; rendering the digital twin through data mapping; and generating visualized digital twin data of the target device.

[0009] In one possible implementation, before obtaining the historical device data corresponding to the target scenario, the method further includes: cleaning the historical device data; and performing data annotation processing on the text data and table data in the cleaned historical device data according to the device KKS code data to obtain twin-level tag data and device attribute tag data.

[0010] In one possible implementation, the step of encoding the device specification data based on the feature extraction of the digital twin prediction model to obtain device specification encoded data includes: performing location encoding processing on the device specification data in the historical device data to obtain location encoding information, and performing token encoding processing on the device specification data to obtain token encoding information; integrating the location encoding information and the token encoding information to obtain target embedding information; and inputting the target embedding information into the BERT network model for feature extraction to obtain device specification encoded data.

[0011] In one possible implementation, the step of inputting the target embedding information into a BERT network model for feature extraction to obtain device specification encoding data includes: inputting the target embedding information into a BERT network model for feature extraction, performing encoding operations based on the encoding layer of the BERT network model, and obtaining corresponding digital twin hierarchical encoding data and digital twin attribute encoding data.

[0012] In one possible implementation, the decoding operation of the device specification encoded data based on the decoding layer of the digital twin prediction model to obtain device-level prediction results and device attribute prediction results includes: inputting the digital twin-level encoded data and the digital twin attribute encoded data into the decoding layer of the BERT network model for decoding to obtain digital twin-level decoded data and digital twin attribute decoded data; and performing probability mapping processing on the digital twin-level decoded data and the digital twin attribute decoded data respectively to obtain the corresponding device-level prediction results and device attribute prediction results.

[0013] In one possible implementation, the step of training the model based on the device-level prediction result, the device attribute prediction result, the twin-level label data, and the device attribute label data to generate a digital twin prediction model includes: performing data matching between the device-level prediction result and the twin-level label data, and performing data matching between the device attribute prediction result and the device attribute label data to obtain a matching result; and training the model if the matching result meets the model training requirements to generate the digital twin prediction model.

[0014] Secondly, embodiments of this application provide a digital twin data processing device based on intelligent algorithms. The device includes: a preprocessing module for acquiring historical device data corresponding to a target scenario, the historical device data including device specification data and corresponding twin-level label data and device attribute label data; an encoding module for inputting the device specification data into a digital twin prediction model, and encoding the device specification data based on feature extraction from the digital twin prediction model to obtain device specification encoded data; a decoding module for decoding the device specification encoded data based on the decoding layer of the digital twin prediction model to obtain device-level prediction results and device attribute prediction results; and a model building module for training a model based on the device-level prediction results, the device attribute prediction results, the twin-level label data, and the device attribute label data to generate a digital twin prediction model.

[0015] Thirdly, embodiments of this application provide a digital twin data processing device based on intelligent algorithms, the device comprising: a processor and a memory storing computer program instructions; the processor, when executing the computer program instructions, implements the digital twin data processing method based on intelligent algorithms as described in any one of the first aspects.

[0016] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the digital twin data processing method based on intelligent algorithms as described in any one of the first aspects.

[0017] Fifthly, embodiments of this application provide a computer program product, wherein instructions in the computer program product, when executed by a processor of an electronic device, cause the electronic device to perform the digital twin data processing method based on intelligent algorithms as described in any one of the first aspects.

[0018] This application provides a digital twin data processing method, apparatus, device, medium, and program product, which can acquire historical device data corresponding to a target scenario, input device specification data into a digital twin prediction model, and obtain device level prediction results and device attribute prediction results by encoding and decoding the device specification data. Based on the device level prediction results, device attribute prediction results, twin level label data, and device attribute label data, the model is trained to generate a digital twin prediction model. The digital twin prediction model is used to encode and decode the device specification data to obtain the hierarchical and attribute relationships between different devices, thereby realizing automated construction of digital twins and improving the accuracy and efficiency of construction. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating a digital twin data processing method based on intelligent algorithms provided in an embodiment of this application; Figure 2 This is a flowchart illustrating another digital twin data processing method based on intelligent algorithms provided in an embodiment of this application; Figure 3 This is a flowchart illustrating another digital twin data processing method based on intelligent algorithms provided in the embodiments of this application; Figure 4 This is a schematic diagram of a process for determining historical device data provided in an embodiment of this application; Figure 5 This is a flowchart illustrating a process for determining device specification coding data, provided in an embodiment of this application. Figure 6 This is a flowchart illustrating a method for determining device level prediction results and device attribute prediction results, as provided in an embodiment of this application. Figure 7 This is a flowchart illustrating another digital twin data processing method based on intelligent algorithms provided in an embodiment of this application; Figure 8 This is a flowchart illustrating the encoding process of a BERT network model provided in an embodiment of this application; Figure 9 This is a flowchart illustrating the encoding process of another BERT network model provided in this application embodiment; Figure 10 This is a schematic diagram of a process for generating a digital twin prediction model provided in an embodiment of this application; Figure 11 This is a schematic diagram of the structure of a digital twin data processing device based on intelligent algorithms provided in an embodiment of this application; Figure 12 This is a schematic diagram of the structure of a digital twin data processing device based on intelligent algorithms provided in an embodiment of this application. Detailed Implementation

[0021] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0022] It should be noted that, in this document, relational terms such as "first" and "second" are used merely 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 limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0023] KKS codes serve as the equipment identification cards for power plants, uniquely identifying equipment, systems, and even small parts using a combination of letters and numbers for convenient lifecycle management. It consists of three parts: a functional identifier (e.g., ABB for the fuel delivery system), a location identifier (e.g., 10 for Boiler No. 1 area), and an equipment identifier (e.g., AP001 for Feedwater Pump No. 1). This coding system is strictly defined; the meaning and value of each digit are rigorously regulated, conforming to international standards and allowing for direct computer processing.

[0024] The construction of a digital twin requires the collection and analysis of a large amount of data. The conventional method is to collect and analyze the data manually and determine the logical relationships between the data manually. This involves a certain degree of experience and error. On this basis, modeling and simulation testing require a lot of manpower and resources, and the resulting digital twin has poor generalization ability and flexibility.

[0025] To address the aforementioned technical problems, embodiments of this application provide a digital twin data processing method, apparatus, device, medium, and program product. By determining the device specification data, a digital twin prediction model is used to encode and decode the device specification data to obtain the hierarchical and attribute relationships between different devices. A method for automatically identifying, constructing, and optimizing the digital twin structure and its attributes is developed through automatic analysis of the input device specification data. This method can automatically complete the construction of the twin's hierarchy and attributes, and supports subsequent manual fine-tuning, reducing labor costs and improving construction efficiency. The digital twin data processing method based on intelligent algorithms provided in this application embodiment is described below.

[0026] Figure 1 This is a flowchart illustrating a digital twin data processing method based on intelligent algorithms provided in an embodiment of this application. Figure 1 As shown, the specific steps of the digital twin data processing method based on intelligent algorithms include: S101~S104.

[0027] S101. Obtain historical device data corresponding to the target scenario. The historical device data includes device specification data, twin-level tag data and device attribute tag data corresponding to the device specification data.

[0028] The target scenario mentioned here can be understood as the application scenario for building a digital twin prediction model, such as a power plant environment, hospital environment, or school environment. Historical equipment data includes the equipment name, location information, functional information, and attribute information of various devices existing in the target scenario. Equipment specification data represents the industry standard data of equipment in the current target scenario; through equipment specification data, the unique identification information of each device can be determined. Twin-level tag data and equipment attribute tag data represent the reference level and reference attribute relationships of the devices.

[0029] Furthermore, in order to determine the relationship between each device in the selected target scenario, historical device data of multiple devices in the target scenario is first collected to provide basic data for determining the relationship and attribute relationship between different devices through historical device data in the next step.

[0030] S102. Input the equipment specification data into the digital twin prediction model, and encode the equipment specification data based on the feature extraction of the digital twin prediction model to obtain the equipment specification coded data.

[0031] Digital twin prediction models are used to construct the relationships between different devices in a target scene. The model outputs the device hierarchy and attribute detection results.

[0032] To further determine the relationships between multiple devices, the device specification data of historical devices is used as the processing data input into the digital twin prediction model. Feature extraction is performed on the device information contained in the device specification data. First, the device specification data is encoded to obtain the device specification encoded data.

[0033] S103. The decoding layer based on the digital twin prediction model performs decoding operations on the equipment specification encoded data to obtain the equipment level prediction results and equipment attribute prediction results.

[0034] After encoding the input equipment specification data, the encoded equipment specification data is further decoded to obtain the output results of the digital twin prediction model. This leads to the equipment level prediction results and equipment attribute prediction results corresponding to the historical equipment data, providing training results for the next step of building a digital twin prediction model for historical equipment data.

[0035] S104. Based on the device level prediction results, device attribute prediction results, twin level label data, and device attribute label data, train the model to generate a digital twin prediction model.

[0036] Based on the prediction results of the digital twin prediction model corresponding to historical device data, the model is matched with the hierarchical label data of the twin and the device attribute label data to determine the training results of the current model. Using the same method, different historical device data corresponding to different devices are used as training samples and input into the digital twin prediction model for the same encoding and decoding processing. After matching with the hierarchical label data of the twin and the device attribute label data, the training results corresponding to different devices are obtained. All the training results are used to generate a digital twin prediction model. By constructing digital twin prediction models between different devices through device specification data, the hierarchical and attribute relationships between different devices are obtained, realizing the automated construction of digital twins and improving the accuracy and efficiency of construction.

[0037] This application provides a digital twin data processing method that can acquire historical device data corresponding to a target scenario, input device specification data into a digital twin prediction model, and obtain device level prediction results and device attribute prediction results by encoding and decoding the device specification data. Based on the device level prediction results, device attribute prediction results, twin level label data, and device attribute label data, the model is trained to generate a digital twin prediction model. The digital twin prediction model is then used to encode and decode the device specification data to obtain the hierarchical and attribute relationships between different devices, thereby achieving automated construction of digital twins and improving the accuracy and efficiency of the construction.

[0038] Figure 2This is a flowchart illustrating another digital twin data processing method based on intelligent algorithms provided in this application embodiment. Figure 2 This description is based on the above embodiments. Figure 2 The provided diagram illustrates the steps of the digital twin data processing method based on intelligent algorithms, specifically including: S201~S202.

[0039] S201. Input the target device data into the digital twin prediction model and output the hierarchical relationship results and attribute relationship results of the target device data.

[0040] S202. Send the hierarchical relationship results and attribute relationship results to the digital twin platform to generate visualized digital twin data.

[0041] The target device data refers to newly added devices within the target scenario. A digital twin platform can be understood as a platform for calling and displaying the predictive model of the digital twin.

[0042] When new device data is obtained, the target device data is input into the trained digital twin prediction model following the steps for processing historical device data. The model outputs the hierarchical relationship results and attribute relationship results of the target device data, thereby determining the association between the newly added target device data and historical device data. The obtained hierarchical relationship results and attribute relationship results are then sent to the digital twin platform for display through visualized digital twin data.

[0043] For example, the raw KKS code data of newly added equipment in a power plant (such as the newly added BMS-BF-FE-015 feedwater pump and TPS-TU-ST-025 sensor) is acquired. After data preprocessing, the KKS codes are input into a trained digital twin prediction model, which automatically extracts features to obtain the corresponding hierarchical and attribute results for the new equipment. The hierarchical and attribute results of the new equipment are then output to the digital twin platform to render the digital twin. Based on the actual scenario requirements, the output results are locally adjusted to form the final twin data.

[0044] This application embodiment inputs the newly added device into the constructed digital twin prediction model to obtain the hierarchical and attribute relationships of the digital twin prediction model of the newly added device. After data mapping processing, the model output results are sent to the digital twin platform for rendering to obtain the twin data of the newly added device. This realizes the automated generation process of the hierarchical and attribute relationships of any device and improves the output efficiency of the digital twin.

[0045] Figure 3 This is a flowchart illustrating another digital twin data processing method based on intelligent algorithms provided in this application embodiment. Figure 3This is a further explanation of the process of generating visualized digital twin data in step S202. Figure 3 This is based on the previous embodiment. Figure 3 The provided diagram illustrates the steps of the digital twin data processing method based on intelligent algorithms, specifically including: S301~S302.

[0046] S301. Determine the structured text data and structured table data corresponding to the hierarchical relationship results and attribute relationship results.

[0047] S302. Send structured text data and structured table data to the digital twin platform, render the digital twin through data mapping, and generate visualized digital twin data of the target device.

[0048] Structured text data is represented as JSON files, and structured tabular data is represented as CSV files.

[0049] Furthermore, the hierarchical and attribute relationship results of the newly added devices are output to the digital twin platform as structured JSON text data and structured CSV table data. The digital twin platform renders a digital twin based on the hierarchical relationship and attributes of the results. By rendering the digital twin through data mapping, a visualized digital twin data of the target device is generated. The output results are then locally adjusted based on the actual scenario requirements to form the final digital twin data.

[0050] This application embodiment inputs newly added devices into the constructed digital twin prediction model to obtain the hierarchical and attribute relationships of the digital twin prediction model for the newly added devices. After data mapping processing according to structured JSON text data and structured CSV table data, the model output results are sent to the digital twin platform for rendering to obtain the twin data of the newly added devices. This realizes the automated generation process of hierarchical and attribute relationships of any device, improving the output efficiency of digital twins.

[0051] Figure 4 This is a schematic diagram of a process for determining historical device data provided in an embodiment of this application. Figure 4 This is a further explanation of the process of determining historical device data in step S101. Figure 4 Is Figure 1 The description is based on the corresponding embodiments. Figure 4 The provided diagram shows the specific steps for determining historical device data, including: S401~S402.

[0052] S401. Perform data cleaning on historical equipment data.

[0053] S402. The text and tabular data in the cleaned historical equipment data are annotated according to the equipment KKS code data to obtain twin-level label data and equipment attribute label data.

[0054] Based on the construction of a digital twin prediction model, relevant data on digital twins from different scenarios and fields within a power plant environment were collected. Historical equipment data for different devices within each scenario was obtained, and this data was cleaned to ensure its suitability. The cleaned historical equipment data was then labeled according to the device KKS codes, with the labeling files being JSON text data and CSV table data. For different information data, corresponding digital twin levels and attributes were labeled, resulting in twin level label data and device attribute label data, providing training data for the next step of training the BERT network model.

[0055] This application embodiment preprocesses the collected historical device data, and ensures the qualification of the dataset through data cleaning and labeling. It obtains the digital twin hierarchy and attributes of different historical device data, providing training data for the next step of training the BERT network model and improving the accuracy of data processing.

[0056] Figure 5 This is a schematic diagram of a process for determining device specification coding data provided in an embodiment of this application. Figure 5 This is a further explanation of the equipment specification coding data process in step S102. Figure 5 Is Figure 1 The description is based on the corresponding embodiments. Figure 5 The provided diagrams and the steps for encoding the equipment specifications specifically include: S501~S504.

[0057] S501. Perform location encoding processing on the equipment specification data in the historical equipment data to obtain location encoding information, and perform token encoding processing on the equipment specification data to obtain token encoding information.

[0058] Location coding information is obtained by performing location coding on the input equipment specification data. For example, when a digital twin prediction model is applied to a power plant environment, assuming the KKS code corresponding to the historical equipment is BMS-BF-FE-003, this KKS code is encoded by a location coding layer to obtain a location vector P1. This vector encodes location features such as "feedwater pump No. 003 belongs to the combustion system auxiliary equipment" and "it is the 3rd in the feedwater pump group", avoiding the model from confusing it with "feedwater pump of the turbine system".

[0059] Furthermore, the equipment specification data is then tokenized to obtain token-encoded information, thus achieving the encoding of the equipment specification data. This provides parameter data for the next step of decoding. For example, when a digital twin prediction model is applied to a power plant environment, assuming the KKS code corresponding to the historical equipment is BMS-BF-FE-003, this KKS code, after being encoded by the token encoding layer, is split into multiple semantic tokens: BMS: Boiler Combustion Management System; BF: Boiler auxiliary equipment; FE: Water pump; 003: Equipment number; These token vectors are integrated to obtain token encoding information. The model can quickly identify the device's "system affiliation, device type, and core functions" through the token encoding information.

[0060] S502. Based on the location encoding information and token encoding information, the target embedding information is obtained by integrating and processing them.

[0061] Based on the location encoding information and token encoding information of the corresponding device specification data in the historical device data, the location encoding information and token encoding information are integrated and processed to obtain target embedding information, which provides reference data for the next step of feature extraction.

[0062] In one possible scenario, the digital twin prediction model is applied to a power plant environment. Assuming the KKS code corresponding to the historical equipment is BMS-BF-FE-003, the target embedding information is obtained after the equipment specification data is integrated by location encoding and token encoding. Through the target embedding information, the digital twin prediction model can not only determine the hierarchical relationship of the equipment in the digital twin prediction model (such as "power plant → boiler workshop → combustion system → auxiliary unit → feedwater pump group → feedwater pump No. 003"), but also automatically associate the attribute relationships of the equipment (such as rated flow 200m³ / h, rated head 1200kPa, operating temperature ≤150℃, associated sensor number BMS-BF-ST-045, etc.), and finally output the core information of the digital twin that meets the needs of the power plant.

[0063] S503. Input the target embedding information into the BERT network model for feature extraction to obtain the device specification coding data.

[0064] Based on the target embedding information, in order to further determine the hierarchical and attribute relationships between the device specification data and other devices, the target embedding information is input into the BERT network model for feature extraction. The hierarchical and attribute relationships of the target embedding information are further extracted to obtain the device specification coding data.

[0065] S504. Input the target embedding information into the BERT network model for feature extraction, and perform encoding operations based on the encoding layer of the BERT network model to obtain the corresponding digital twin hierarchical encoding data and digital twin attribute encoding data.

[0066] Based on the target embedding information, how to obtain the equipment specification coding data corresponding to the association and attribute relationships between the equipment and other equipment? Specifically, the target embedding information needs to be input into the BERT network model for feature extraction. The hierarchical and attribute relationships in the target embedding information are then encoded to obtain the corresponding digital twin hierarchical coding data and digital twin attribute coding data. For example, applying the digital twin prediction model to a power plant environment, assuming the historical equipment's KKS code is BMS-BF-FE-003, the equipment hierarchical association is extracted from the target embedding information, and the features of other feedwater pump equipment in the same group are associated. The implicit connections between "feedwater pump No. 003" and other equipment are explored (such as the "feedwater pipeline", "pressure sensor (KKS code BMS-BF-ST-045)", "frequency converter control cabinet", etc., which need to be connected to the combustion system. These associations are not directly written in the KKS code, but the coding layer will extract them through historical data and semantic features). Next, perform preliminary attribute feature analysis: initially associate the core operating attributes of the water pump equipment (such as "rated flow rate 200m³ / h" and "rated head 1200kPa"). These attributes are further refined from the semantic features of the token-encoded information.

[0067] This application embodiment inputs equipment specification data into the BERT network model for encoding and feature extraction, obtaining the hierarchical association and attribute relationships of the equipment specification data, which serve as the hierarchical encoding data and attribute encoding data of the digital twin. This initially determines the hierarchical and attribute relationships of the digital twin prediction model, providing training data for further training of the digital twin prediction model and further improving the training accuracy.

[0068] Figure 6 This is a flowchart illustrating a process for determining device level prediction results and device attribute prediction results, provided in an embodiment of this application. Figure 6 This is a further explanation of the process of determining the equipment level prediction result and the equipment attribute prediction result in step S103. Figure 6 Is Figure 1 The description is based on the corresponding embodiments. Figure 6 The provided diagram shows the steps for determining the equipment level prediction results and equipment attribute prediction results, specifically including: S601~S602.

[0069] S601. Input the digital twin hierarchical encoded data and digital twin attribute encoded data into the decoding layer of the BERT network model for decoding to obtain the digital twin hierarchical decoded data and digital twin attribute decoded data.

[0070] S602. Perform probability mapping processing on the digital twin hierarchical decoding data and the digital twin attribute decoding data respectively to obtain the corresponding device hierarchical prediction results and device attribute prediction results.

[0071] After inputting equipment specification data into the BERT network model for encoding and feature extraction to obtain digital twin hierarchical encoding data and digital twin attribute encoding data, the digital twin prediction model is further trained. The digital twin hierarchical encoding data and digital twin attribute encoding data are decoded to obtain digital twin hierarchical decoded data and digital twin attribute decoded data. The decoded data is then input into the probability mapping layer to output the digital twin hierarchical level and attribute with the highest probability, thus obtaining the equipment hierarchical prediction result and the equipment attribute prediction result, thereby realizing the training of the digital twin prediction model. For example, when applying the digital twin prediction model to a power plant environment, assuming the KKS code corresponding to the historical equipment is BMS-BF-FE-003, the decoding and processing of the digital twin hierarchical encoding data and digital twin attribute encoding data will filter redundant features (such as removing the "misassociation between feedwater pump No. 003 and turbine system equipment" in the digital twin hierarchical encoding data and digital twin attribute encoding data) and correct fuzzy features (such as clarifying the attribute type of "parameter features" in the digital twin hierarchical encoding data and digital twin attribute encoding data, and labeling "flow rate 200m³ / h" as "rated flow rate") to obtain the equipment hierarchical prediction results and equipment attribute prediction results.

[0072] In one possible example scenario, Figure 7 This is a flowchart illustrating another digital twin data processing method based on intelligent algorithms provided in this application embodiment. According to... Figure 7 The diagram illustrates how to input the KKS code table data corresponding to historical device data into the digital twin prediction model, setting it as X=[x1,x2,x3,…,x] n ], where x n One column of data in the KKS code table is encoded using a position encoding layer to obtain position information P, represented by Equation 1. Simultaneously, it is encoded using a token embedding layer to obtain token information T, represented by Equation 2. P = position embedding(X) Equation 1 T = token embedding(X) Equation 2 The location encoding information and the token encoding information are integrated to obtain the overall embedded information E, which is specifically represented by Equation 3: E=concat(P,T) Equation 3 The overall embedded information E is input into the BERT model for feature extraction: First, it is input into the encoding layer for encoding, resulting in the encoded output F1, which is represented by Equation 4: F1 = encoder(E) Equation 4 After obtaining the encoded output F1, it is input to the decoding layer to obtain the decoded output F2, which is represented by Equation 5: F2 = decoder(F1) Equation 5 Finally, the probability mapping layer outputs the digital twin level and attribute output corresponding to the highest probability. The output result is represented by Equation 6: output = percent layer(F2) Equation 6 Based on the output results, the hierarchy and attribute relationships of the digital twin prediction model corresponding to the device are obtained, and then the digital twin prediction model is constructed.

[0073] In one possible example scenario, Figure 8 This is a flowchart illustrating the encoding process of a BERT network model provided in an embodiment of this application. Figure 9 This is a flowchart illustrating the encoding process of another BERT network model provided in this application embodiment. According to... Figure 8 The provided diagram illustrates that the Bidirectional Encoder Representation from Transformers (BERT) is a pre-trained language representation model. It integrates the input data (E1, E2, ..., E...). n After encoding, multiple coding units (Trm) are obtained. The output of these coding units is then decoded to obtain the final output feature (T1, T2, ..., T). n ).according to Figure 9 The provided diagram shows that the main structure of BERT is a transformer (i.e., Figure 8The TRM module in BERT (Transformer) is the basic structure of a BERT network model, which is the encoder part of a standard transformer structure. A transformer encoder unit is generated by stacking multi-head attention, layer normalization, feedforward, and layer normalization. Each layer of the BERT network model consists of one such encoder unit. Data encoding is performed through the encoder unit.

[0074] Figure 10 This is a schematic diagram of a process for generating a digital twin prediction model provided in an embodiment of this application. Figure 10 This is a further explanation of the process of generating a digital twin prediction model in step S104. Figure 10 Is Figure 1 The description is based on the corresponding embodiments. Figure 10 The provided diagram shows the specific steps for generating a digital twin prediction model, including: S1001~S1002.

[0075] S1001. Perform data matching between the equipment level prediction results and the twin level label data, and perform data matching between the equipment attribute prediction results and the equipment attribute label data to obtain the matching results.

[0076] S1002. If the matching results meet the model training requirements, train the model to generate a digital twin prediction model.

[0077] Based on the obtained training results from multiple training iterations, the results are compared with pre-defined tag data to determine whether the training results meet the constraints, thereby optimizing the model training process. Data matching is performed between device-level prediction results and twin-level label data, and between device attribute prediction results and device attribute label data. When the matching requirements are met, the effectiveness of the training is guaranteed, and the digital twin prediction model is trained. Based on multiple training iterations on the training data, a pre-defined evaluation metric is used as the termination condition for model training to determine whether to end the training, thus obtaining a digital twin prediction model that meets the requirements.

[0078] This application embodiment uses equipment specification data as input data for a digital twin prediction model. After encoding and decoding the equipment specification data, the hierarchy and attribute results of the corresponding digital twin prediction model are obtained. Through multiple training iterations, a trained digital twin prediction model is obtained. This method automatically analyzes the input equipment specification data to automatically identify, construct, and optimize the digital twin structure and its attributes. It can automatically complete the construction of the twin's hierarchy and attributes and supports subsequent manual fine-tuning, significantly reducing labor costs and improving construction efficiency.

[0079] Figure 11 This is a schematic diagram of the structure of a digital twin data processing device based on intelligent algorithms provided in an embodiment of this application. Figure 11 The provided schematic diagram illustrates the specific structure of the digital twin data processing device 100 based on intelligent algorithms, which includes: Preprocessing module 1101 is used to obtain historical device data corresponding to the target scenario. The historical device data includes device specification data and twin-level tag data and device attribute tag data corresponding to the device specification data. The encoding module 1102 is used to input equipment specification data into the digital twin prediction model, and to encode the equipment specification data based on the feature extraction of the digital twin prediction model to obtain equipment specification encoded data. The decoding module 1103 is used to perform decoding operations on the device specification encoded data based on the decoding layer of the digital twin prediction model to obtain device-level prediction results and device attribute prediction results. The model building module 1104 is used to train the model based on the device level prediction results, device attribute prediction results, twin level label data, and device attribute label data to generate a digital twin prediction model.

[0080] The digital twin data processing device based on intelligent algorithms provided in this embodiment can be as follows: Figure 11 The digital twin data processing device 100 based on intelligent algorithms shown can achieve... Figures 1-10 The technical effects of the digital twin data processing method based on intelligent algorithms shown are described in detail in the relevant description. Figures 1-10 The corresponding explanation is concise and will not be elaborated upon here.

[0081] Figure 12 This is a schematic diagram of the structure of a digital twin data processing device based on intelligent algorithms provided in an embodiment of this application. The digital twin data processing device based on intelligent algorithms includes a processor 1201 and a memory 1202 storing computer program instructions.

[0082] Specifically, the processor 1201 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0083] Memory 1202 may include mass storage for data or instructions. For example, and not limitingly, memory 1202 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 1202 may include removable or non-removable (or fixed) media. Where appropriate, memory 1202 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 1202 is non-volatile solid-state memory.

[0084] Memory may include read-only memory (ROM), random access memory (RAM), disk storage media devices, optical storage media devices, flash memory devices, and electrical, optical, or other physical / tangible memory storage devices. Therefore, typically, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to the first aspect of this application.

[0085] The processor 1201 reads and executes computer program instructions stored in the memory 1202 to implement any of the digital twin data processing methods based on intelligent algorithms in the above embodiments.

[0086] In one example, the digital twin data processing device based on intelligent algorithms may further include a communication interface 1203 and a bus 1210. For example, Figure 12 As shown, the processor 1201, memory 1202, and communication interface 1203 are connected through bus 1210 and complete communication with each other.

[0087] The communication interface 1203 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0088] Bus 1210 includes hardware, software, or both, that couples components of an online data traffic metering device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 1210 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, any suitable bus or interconnect is contemplated herein.

[0089] Furthermore, in conjunction with the digital twin data processing method based on intelligent algorithms in the above embodiments, this application embodiment can provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the digital twin data processing methods based on intelligent algorithms in the above embodiments.

[0090] This application also provides a computer program product, including a computer program that, when executed by a processor, implements any of the intelligent algorithm-based digital twin data processing methods described in the above embodiments.

[0091] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0092] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0093] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0094] The aspects of this application have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by dedicated hardware performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0095] The above description is merely a specific implementation of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A method for processing digital twin data based on intelligent algorithm, characterized in that, The method comprises the following steps: obtaining historical equipment data corresponding to a target scene, wherein the historical equipment data comprises equipment specification data, twin body hierarchical label data corresponding to the equipment specification data, and equipment attribute label data; inputting the equipment specification data into a digital twin prediction model, performing encoding operation on the equipment specification data based on feature extraction of the digital twin prediction model, and obtaining equipment specification encoding data; performing decoding operation on the equipment specification encoding data based on a decoding layer of the digital twin prediction model, and obtaining equipment hierarchical prediction results and equipment attribute prediction results; performing model training based on the equipment hierarchical prediction results, the equipment attribute prediction results, the twin body hierarchical label data, and the equipment attribute label data, and generating a digital twin prediction model.

2. The method of claim 1, wherein, After the digital twin prediction model is generated, the method further comprises the following steps: inputting target equipment data into the digital twin prediction model, and outputting hierarchical relationship results and attribute relationship results of the target equipment data; sending the hierarchical relationship results and the attribute relationship results to a digital twin platform, and generating visualized digital twin data.

3. The method of claim 2, wherein, The step of sending the hierarchical relationship results and the attribute relationship results to the digital twin platform to generate visualized digital twin data comprises the following steps: determining structured text data and structured table data corresponding to the hierarchical relationship results and the attribute relationship results; sending the structured text data and the structured table data to the digital twin platform, rendering a digital twin through data mapping, and generating visualized digital twin data of the target equipment.

4. The method of claim 1, wherein, Before the step of obtaining historical equipment data corresponding to a target scene, the method further comprises the following steps: performing data cleaning on the historical equipment data; performing data annotation processing on text data and table data in the cleaned historical equipment data according to equipment KKS code data, and obtaining twin body hierarchical label data and equipment attribute label data.

5. The method of claim 1, wherein, The step of performing encoding operation on the equipment specification data based on feature extraction of the digital twin prediction model to obtain equipment specification encoding data comprises the following steps: performing position encoding processing on the equipment specification data in the historical equipment data to obtain position encoding information, and performing token encoding processing on the equipment specification data to obtain token encoding information; performing integration processing based on the position encoding information and the token encoding information to obtain target embedding information; inputting the target embedding information into a BERT network model for feature extraction to obtain equipment specification encoding data.

6. The method of claim 5, wherein, The step of inputting the target embedding information into the BERT network model for feature extraction to obtain equipment specification encoding data comprises the following steps: inputting the target embedding information into the BERT network model for feature extraction, performing encoding operation based on an encoding layer of the BERT network model to obtain corresponding digital twin hierarchical encoding data and digital twin attribute encoding data.

7. The method of claim 6, wherein, The decoding layer based on the digital twin prediction model decodes the device specification encoding data to obtain device level prediction results and device attribute prediction results, including: The digital twin level encoding data and the digital twin attribute encoding data are input into the decoding layer of the BERT network model for decoding operation to obtain digital twin level decoding data and digital twin attribute decoding data; The digital twin level decoding data and the digital twin attribute decoding data are respectively subjected to probability mapping processing to obtain corresponding device level prediction results and device attribute prediction results.

8. The method of claim 1, wherein, The device level prediction results, the device attribute prediction results, the twin level label data, and the device attribute label data are subjected to model training to generate a digital twin prediction model, including: The device level prediction results and the twin level label data are subjected to data matching, and the device attribute prediction results and the device attribute label data are subjected to data matching to obtain matching results; The matching results are subjected to model training to generate a digital twin prediction model when the matching results meet the model training requirements.

9. A smart algorithm based digital twin data processing apparatus, characterized in that, The device includes: The preprocessing module is configured to obtain historical device data corresponding to a target scene, wherein the historical device data includes device specification data, twin level label data corresponding to the device specification data, and device attribute label data; The encoding module is configured to input the device specification data into a digital twin prediction model, perform encoding operation on the device specification data based on feature extraction of the digital twin prediction model, and obtain device specification encoding data; The decoding module is configured to perform decoding operation on the device specification encoding data based on a decoding layer of the digital twin prediction model to obtain device level prediction results and device attribute prediction results; The model construction module is configured to perform model training based on the device level prediction results, the device attribute prediction results, the twin level label data, and the device attribute label data to generate a digital twin prediction model.

10. A smart algorithm based digital twin data processing device, characterized in that, The device includes a processor and a memory storing computer program instructions; The processor executes the computer program instructions to implement the digital twin data processing method based on intelligent algorithm according to any one of claims 1-8.

11. A computer readable storage medium, characterized in that, The computer readable storage medium stores computer program instructions, and the computer program instructions are executed by the processor to implement the digital twin data processing method based on intelligent algorithm according to any one of claims 1-8.

12. A computer program product, characterised in that, The instructions in the computer program product are executed by the processor of the electronic device to enable the electronic device to perform the digital twin data processing method based on intelligent algorithm according to any one of claims 1-8.

Images