Engineering data processing method based on generative artificial intelligence and related equipment

Abstract

The invention provides an engineering data processing method and related equipment based on generative artificial intelligence, and relates to the technical field of electrical digital data processing.The method comprises the steps that multiple pieces of historical engineering data of different engineering environments and design parameters are obtained, engineering feature data are extracted after preprocessing, and the engineering feature data are stored in a database; engineering scheme design parameters and engineering environment data are covered. And then, constructing an engineering data prediction model by using a generative AI algorithm and the labeled historical engineering data, simulating real environment data distribution, obtaining target design parameters in a current engineering environment, and inputting the target design parameters and the current engineering environment into the model to obtain a predicted value. And comparing the predicted value with an actual observed value to obtain a comparison result, and finally evaluating and adjusting the engineering data prediction model according to the comparison result. Through the method, the accuracy and reliability of engineering data processing can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of electrical digital data processing, and in particular to an engineering data processing method and related equipment based on generative artificial intelligence. Background Art

[0002] In today's engineering field, whether it is construction engineering, mechanical engineering or electronic engineering, accurate data processing and prediction play a vital role in engineering planning, implementation and subsequent maintenance. With the continuous increase in the scale and complexity of projects, traditional empirical engineering data processing methods can no longer meet the needs. Engineering decisions need to be based on in-depth analysis and accurate prediction of various engineering data to reduce costs, improve engineering quality and ensure engineering safety.

[0003] At present, many engineering data processing relies on traditional statistical analysis methods. After obtaining historical engineering data, engineers will use statistical software to perform simple descriptive statistics on the data, calculate statistical quantities such as mean and standard deviation, so as to understand the central tendency and dispersion degree of the data. For different variables in engineering data, such as engineering design parameters and environmental parameters, correlation analysis will be used to find the linear relationship between them. When making predictions, linear regression models are often used to construct regression equations based on the relationship between variables in historical data, and then substitute the relevant parameters of the current project into the equation to obtain the predicted values. In order to improve the accuracy of the prediction, engineers will try to collect more historical data, filter the data, and remove some obviously abnormal data points.

[0004] However, engineering data is often highly nonlinear and complex, and traditional linear regression models cannot accurately capture the complex relationships between data, so traditional methods find it difficult to accurately process and predict engineering data. Summary of the invention

[0005] The present application provides an engineering data processing method and related equipment based on generative artificial intelligence for accurately processing and predicting highly nonlinear and complex engineering data.

[0006] In the first aspect, the present application provides an engineering data processing method based on generative artificial intelligence, which is applied to a server, and the method includes: obtaining multiple historical engineering data with different design parameters under different engineering environments; after preprocessing the historical engineering data, performing feature extraction on the historical engineering data to obtain engineering feature data, and the engineering feature data at least includes engineering scheme design parameters and engineering environment data, and the engineering environment data refers to data related to the environment in which the engineering activities are located; combining the generative AI algorithm and the historical engineering data annotated with the engineering feature data to construct an engineering data prediction model, and the engineering data prediction model is used to simulate the data distribution under a real engineering environment; after obtaining the target design parameters under the current engineering environment, the target design parameters and the current engineering environment are input into the engineering data prediction model to obtain the current engineering data prediction value; comparing the current engineering data prediction value with the actual observation value to obtain a comparison result; and evaluating and adjusting the engineering data prediction model according to the comparison result.

[0007] By adopting the above technical solutions, multiple historical engineering data of different engineering environments and design parameters are obtained to provide rich samples for subsequent analysis. Preprocessing and feature extraction obtain engineering feature data containing engineering scheme design parameters and engineering environment data, which can accurately extract key information. Combining generative AI algorithms and annotated historical engineering data to build a model, the advantages of the algorithm can be used to learn data features and simulate the data distribution in a real engineering environment. The target design parameters and environment of the current project are input into the model to obtain the predicted values, and the model is compared with the actual observed values ​​to evaluate and adjust the model, so as to achieve dynamic optimization of engineering data, improve the accuracy and reliability of predictions, and provide a more scientific basis for engineering decision-making.

[0008] In combination with some embodiments of the first aspect, in some embodiments, after the step of constructing an engineering data prediction model by combining a generative AI algorithm and historical engineering data annotated with the engineering feature data, it also includes: receiving a text description input by a user, and identifying target design parameters using natural language processing technology; identifying a first current engineering environment from an engineering environment image uploaded by a user using image recognition technology; determining a second current engineering environment through environmental sensors in a plurality of engineering activity set areas; and determining a final current engineering environment by combining the first current engineering environment and the second current engineering environment.

[0009] By adopting the above technical solution, the user input text description is received and the target design parameters are identified using natural language processing technology, which facilitates users to express their needs and improves the convenience of interaction. The first current engineering environment is identified from the image uploaded by the user, and the second current engineering environment is determined using environmental sensors. Environmental information is collected through multiple channels to ensure the comprehensiveness of the information. The two are combined to determine the final current engineering environment, and data from different sources are integrated to make the current engineering environment data obtained more accurate and real, thereby providing a more realistic input for the engineering data prediction model and improving the adaptability and prediction accuracy of the prediction model to the actual situation of the current engineering.

[0010] In combination with some embodiments of the first aspect, in some embodiments, after the step of determining the final current engineering environment by combining the first current engineering environment and the second current engineering environment, it also includes: using geographic information system data to analyze the topography, geological structure and infrastructure distribution information in the preset area of ​​the area where the current project is located; after fusing the topography, the geological structure, the infrastructure distribution information and the final current engineering environment data, the final current engineering environment data is adjusted.

[0011] By adopting the above technical solutions, the GIS data is used to analyze the topography, geological structure and infrastructure distribution information, and the engineering environment data is supplemented from a macro-geographic perspective. By integrating this information with the final current engineering environment data, multiple factors can be considered comprehensively. Adjusting the fused data can optimize the environmental data to more accurately reflect the complex environmental conditions in which the project is actually located. The adjusted environmental data can make the engineering data prediction model more comprehensive and accurate when simulating the real engineering environment, thereby improving the scientificity and reliability of the model prediction.

[0012] In combination with some embodiments of the first aspect, in some embodiments, after the step of constructing an engineering data prediction model by combining a generative AI algorithm and historical engineering data annotated with the engineering feature data, it also includes: obtaining real-time data during the implementation of the current project, the real-time data at least including engineering progress and equipment operating status data; when the project progress is delayed or the equipment operating status data is abnormal, adjusting the parameters in the engineering data prediction model according to preset rules based on the severity of the delay or abnormality.

[0013] By adopting the above technical solutions, real-time data during the project implementation process, such as project progress and equipment operation status data, can be obtained to reflect the progress of the project in real time. When the project progress is delayed or the equipment is operating abnormally, the model parameters are adjusted according to the preset rules, and the model can be corrected in time according to the actual situation. Since the actual situation of the project will affect the project data, this adjustment can make the model better adapt to the dynamic changes of the project, maintain the accuracy of the project data prediction, and avoid the disconnection between the model prediction and the actual situation due to sudden project conditions.

[0014] In combination with some embodiments of the first aspect, in some embodiments, when the progress of the project is delayed or the equipment operation status data is abnormal, after the step of adjusting the parameters in the engineering data prediction model according to preset rules based on the severity of the delay or abnormal situation, it also includes: establishing an engineering data prediction model sharing platform for different engineering types according to the engineering type of the current project; when the engineering data prediction model is adjusted and optimized, uploading the adjusted model parameters and optimization experience to the sharing platform; after receiving a case viewing instruction sent by the user end, obtaining optimization cases similar to the current project from the sharing platform, and optimizing the current engineering data prediction model again in combination with the characteristics of the current project.

[0015] By adopting the above technical solutions, a sharing platform is established according to the project type, providing a channel for data and experience exchange for different projects. After the model is adjusted and optimized, parameters and experience are uploaded to accumulate industry knowledge. When a user-side case viewing instruction is received, similar optimization cases are obtained and the model is optimized again in combination with the characteristics of the current project, so that the model can learn from the successful experience of other projects and further improve it in combination with its own characteristics. This method can continuously improve the performance of the model, improve the accuracy and efficiency of engineering data prediction, and promote the common progress of the entire engineering industry in data prediction technology.

[0016] In combination with some embodiments of the first aspect, in some embodiments, after the step of comparing the current engineering data prediction value with the actual observation value to obtain a comparison result, it also includes: if the comparison result shows that the error between the current engineering data prediction value and the actual observation value exceeds a preset error range, and the current engineering data prediction value is associated with a preset engineering safety key indicator, then the emergency response mechanism is automatically started; an alarm message is sent to the management end, and at the same time, virtual reality technology is used in combination with a hazard database to simulate suspected dangerous scenes; and the suspected dangerous scenes are sent to the management end for display.

[0017] By adopting the above technical solution, after comparing the predicted value with the actual observed value, when the error exceeds the preset range and is associated with the key safety indicator, the emergency response mechanism is automatically activated to respond quickly to potential risks. Sending alarm information to the management end and using virtual reality technology combined with the hazard database to simulate suspected dangerous scenes can allow managers to understand potential dangerous situations intuitively and quickly. Sending the simulated scene to the management end for display provides a strong basis for managers to make scientific decisions in a timely manner, which helps to quickly take effective measures to deal with risks.

[0018] In combination with some embodiments of the first aspect, in some embodiments, after the step of sending an alarm message to the management end and simulating a suspected dangerous scene using virtual reality technology combined with a danger database, it also includes: using blockchain technology to record the alarm message and the suspected dangerous scene.

[0019] By adopting the above technical solutions, the tamper-proof and traceable characteristics of blockchain can ensure that the data is authentic and reliable. These records can provide accurate original data for subsequent analysis of abnormal situations and facilitate the review of the causes of accidents.

[0020] In a second aspect, the present application provides a server, comprising: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the server to execute the method described in the first aspect and any possible implementation of the first aspect.

[0021] In a third aspect, the present application provides a computer-readable storage medium comprising instructions, which, when executed on a server, causes the server to execute the method described in the first aspect and any possible implementation of the first aspect.

[0022] In a fourth aspect, the present application provides a computer program product, which, when executed on a server, enables the server to execute the method described in the first aspect and any possible implementation manner of the first aspect.

[0023] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. The prediction model is built based on the collection, preprocessing and feature extraction of historical engineering data, combined with a generative AI algorithm, and the target design parameters and environment of the current project are input into the model, and compared with the actual observation values ​​for evaluation and adjustment. Therefore, the technical problem of inaccurate prediction of traditional engineering data and the inability to dynamically adjust to adapt to different engineering environments and design parameter changes is effectively solved, thereby achieving dynamic and accurate prediction of engineering data, providing a more scientific and reliable basis for engineering decision-making.

[0024] 2. By using the geographic information system to analyze the topography, geological structure and infrastructure distribution information of the project area and integrating and adjusting it with the current project environment data, the technical problem of incomplete and inaccurate consideration of project environment information in existing project data processing has been effectively solved, thereby achieving the technical effect of improving the simulation capability of the project data prediction model for complex environmental conditions and providing more comprehensive and accurate environmental data support for the planning and implementation of project construction.

[0025] 3. Due to the adoption of a technical means of establishing an engineering data prediction model sharing platform according to the engineering type, uploading and sharing the optimized model parameters and experience, and re-optimizing the model in combination with the current engineering characteristics after receiving the user-side case viewing instructions, it effectively solves the technical problem that a single engineering data prediction model is difficult to continuously optimize and cannot fully utilize the experience of similar cases in the industry, thereby achieving the technical effect of improving model performance, improving the accuracy and efficiency of engineering data prediction, and promoting the common progress of the engineering industry in data prediction technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a flow chart of the engineering data processing method based on generative artificial intelligence in an embodiment of the present application; Figure 2 is another flow chart of the engineering data processing method based on generative artificial intelligence in an embodiment of the present application; Figure 3 It is a schematic diagram of a physical device structure of a server in an embodiment of the present application. DETAILED DESCRIPTION

[0027] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification and appended claims of the present application, the singular expressions "one", "a kind of", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to and includes any or all possible combinations of one or more listed items.

[0028] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0029] For ease of understanding, the following is a description of the process of the method provided by this implementation. Figure 1 , which is a flow chart of the engineering data processing method based on generative artificial intelligence in an embodiment of the present application.

[0030] S101, acquiring a plurality of historical engineering data with different design parameters under different engineering environments; The server will be connected to various engineering databases. These databases may come from different engineering fields, such as the industry standard database in the field of construction engineering, which stores a large amount of detailed data of completed construction projects, including building design parameters (such as building structure type, building area, number of floors, etc.) and corresponding engineering environment data (such as the climate conditions, geological conditions, surrounding environment, etc. of the project location); the internal database of the enterprise in the field of mechanical engineering contains the design parameters of various mechanical equipment (such as the size, material, transmission mode, etc. of mechanical parts) and its operation data under different working conditions.

[0031] The server uses a standardized data interface protocol to establish connections with these databases and select historical engineering data that meets the requirements according to pre-set query rules. For example, for a bridge construction project data acquisition task, the server will send a query instruction to the relevant traffic engineering database, requesting the design parameters (such as span, beam height, pier form, etc.) of bridges of different regions and different structural types in the past ten years, as well as the engineering environment data during construction (such as earthquake intensity, wind speed, river flow, etc. in the region).

[0032] S102, after preprocessing the historical engineering data, extracting features from the historical engineering data to obtain engineering feature data, the engineering feature data at least including engineering scheme design parameters and engineering environment data, the engineering environment data refers to data related to the environment in which the engineering activities are located; After acquiring multiple historical engineering data with different design parameters in different engineering environments, the server needs to preprocess these data to improve data quality. The preprocessing at least includes data cleaning and data standardization.

[0033] After preprocessing, the server performs feature extraction. For engineering design parameters, taking bridge engineering as an example, the server uses the principles of analytical geometry and structural mechanics to extract geometric parameters such as bridge span, beam height, and pier spacing from the digital files of bridge design drawings; by reading the material list and design instructions, it obtains material parameters such as bridge structure material type and strength grade; using logical analysis methods, it sorts out functional parameters such as bridge structure systems (such as beam bridges, arch bridges, cable-stayed bridges, etc.), construction processes (such as cantilever casting method, jacking method, etc.).

[0034] In terms of extracting engineering environmental data, for meteorological environmental data, the server establishes an interface with the meteorological data platform to obtain historical meteorological data of the project location, including temperature, humidity, wind speed, precipitation, etc., and uses data fusion technology to integrate multi-source meteorological data to improve data accuracy. For geological environmental data, the server extracts from the digital documents of geological exploration reports, uses natural language processing technology to parse the report text, and identifies the stratigraphic structure and geotechnical parameters (such as the elastic modulus and Poisson's ratio of the rock and soil); uses geographic information system (GIS) technology to obtain geological structural information of the project area, such as fault distribution and fold morphology. For surrounding environmental data, the server uses image recognition technology to analyze satellite images or aerial images of the area surrounding the project to identify geographical elements such as buildings, roads, and rivers; and uses web crawler technology to obtain infrastructure distribution data around the project from relevant geographic information websites, such as the location and scale of substations and communication base stations.

[0035] S103, constructing an engineering data prediction model by combining a generative AI algorithm and historical engineering data annotated with the engineering feature data, where the engineering data prediction model is used to simulate data distribution in a real engineering environment; When building an engineering data prediction model, the server will select cutting-edge generative AI algorithms that are adapted to the characteristics of engineering data, such as a hybrid algorithm of a variational autoencoder (VAE) combined with a generative adversarial network (GAN). This is because of the complexity and nonlinearity of engineering data, and it is difficult for traditional algorithms to effectively capture its inherent laws, while this hybrid algorithm can take advantage of both. VAE learns the potential distribution of data by encoding and decoding data, maps engineering feature data to a low-dimensional space, extracts key features, reduces data redundancy, and enables the model to focus on core information. GAN generates data samples that are closer to the real data distribution through adversarial training of the generator and the discriminator, thereby improving the generalization ability of the model.

[0036] When preparing annotated historical engineering data with engineering feature data, the server will first deeply annotate the previously extracted engineering feature data. For engineering design parameters, the annotation not only includes the specific values ​​of the parameters, but also indicates the engineering subsystem to which it belongs, the design standard specification basis, and the ideal range under different working conditions. For example, in bridge engineering, when annotating the bridge span, it will be clear that it belongs to the bridge superstructure design parameter, based on the "General Specifications for Highway Bridge and Culvert Design", and give the reasonable value range of different types of bridge spans. For engineering environmental data, the annotation will cover the data collection time, location, accuracy, and degree of association with the key links of the project. For example, meteorological data will be marked with the specific time node of the collection, accurate to the minute, the longitude and latitude coordinates of the collection location, the measurement accuracy of the data, such as temperature accurate to 0.1°C, and explain its impact on the construction progress of the project (such as high temperature affecting the quality of concrete pouring and thus affecting the construction progress) and the long-term stability of the project (such as the impact of strong winds on the stress of the bridge structure).

[0037] During the model training process, the server uses powerful computing resources and adopts a distributed training strategy. The historical engineering data is divided into multiple subsets according to certain rules and assigned to multiple computing nodes for parallel training. The training tasks on each computing node are relatively independent, and they collaborate through a parameter synchronization mechanism. For example, using the parameter server architecture, the computing nodes regularly upload the gradient information during the training process to the parameter server. The parameter server updates the model parameters based on these gradient information and sends the updated parameters to the computing nodes, thereby accelerating the training process, improving training efficiency, and reducing training time. At the same time, the server will also use transfer learning technology to draw on the parameters and experience of models that have been trained in other similar engineering fields. For example, some parameters and training experience related to structural mechanics analysis from the model that has been successfully applied in the field of construction engineering are migrated to the bridge engineering data prediction model to accelerate the convergence of the model, improve the training effect of the model, and enable it to simulate the data distribution in the real engineering environment faster and better.

[0038] In some embodiments, during the process of project advancement, the server will continue to obtain real-time data during the current project implementation process, which at least covers the project progress and equipment operation status data. For project progress data, the server obtains information such as the actual completion time and planned completion time of each stage of the project in real time by connecting to the project management system, so as to accurately grasp the advancement rhythm of the project. In construction projects, the server can obtain progress data of various links such as foundation construction and main structure construction from the project management software. For equipment operation status data, the server is connected to the monitoring system of various engineering equipment to collect the operating parameters of the equipment in real time, such as the speed, temperature, vibration amplitude of mechanical equipment, and the voltage and current of electrical equipment. These data can reflect the working status of the equipment in a timely manner and help staff to discover potential problems in advance. When the server detects that the project progress is delayed or the equipment operation status data is abnormal, it will adjust the parameters in the engineering data prediction model according to the severity of the delay or abnormality according to the preset rules. The server will pre-set different delay and abnormality levels. In terms of project progress, if the actual progress lags behind the planned progress by less than 10%, it is a minor delay, 10%-30% is a moderate delay, and more than 30% is a severe delay; in terms of equipment operation status, the abnormal level is divided according to the degree to which the key parameters of the equipment deviate from the normal range. For minor problems, the server may only make fine adjustments to some minor parameters in the model, such as adjusting the work efficiency parameters related to the equipment; for severe problems, the model structure or key parameters may be significantly adjusted, and the impact of equipment failures on the progress and quality of the entire project may be re-evaluated, and key parameters such as relevant weight coefficients in the model may be adjusted accordingly to make the model more in line with the actual situation of the project and ensure the accuracy of the prediction.

[0039] In order to promote the exchange of data and experience between different projects, the server will establish a sharing platform for engineering data prediction models of different types according to the engineering type of the current project. On this platform, valuable resources of various types of projects in the process of data prediction model optimization are stored. The sharing platform of construction engineering type will store the model optimization experience of different building structures and different scales of construction projects; the sharing platform of mechanical engineering type focuses on the relevant content of operation data processing and model optimization of various mechanical equipment. In this way, project teams of the same or similar engineering types can share and obtain useful information on the platform to promote the technological progress of the entire industry. When the engineering data prediction model is adjusted and optimized, the server will upload the adjusted model parameters and optimization experience to the sharing platform. These uploaded contents become important knowledge reserves of the sharing platform and provide reference for other projects. When the server receives the case viewing instruction sent by the user end, it will filter out optimization cases similar to the current project from the sharing platform. By comparing the characteristics of the current project, such as differences in project scale, construction environment, and materials used, the selected cases are analyzed for adaptability. Combined with these analysis results, the server will perform secondary optimization on the current engineering data prediction model. We may learn from the model structure adjustment methods in similar cases, or refer to their parameter optimization strategies to further improve the performance of the current model, so that it can more accurately predict engineering data and provide more reliable support for engineering decision-making.

[0040] S104, after obtaining the target design parameters under the current engineering environment, input the target design parameters and the current engineering environment into the engineering data prediction model to obtain the current engineering data prediction value; The server obtains the target design parameters in the current engineering environment mainly through a variety of interactive methods. On the one hand, the server provides a friendly user interface to support users to enter the target design parameters in the form of structured tables or free text. For structured table input, the server will preset the corresponding parameter template according to different engineering types. Users only need to fill in the specific values ​​according to the template, and the server will perform format verification and data legitimacy checks in real time. For example, when entering architectural engineering design parameters, for parameters such as building area and number of floors, the server will set a reasonable value range according to the architectural design specifications. If the user input exceeds the range, the system will immediately prompt an error and guide the user to correct it. For free text input, the server uses natural language processing technology to first perform semantic analysis on the input text, extract key information, and then match and associate it with the engineering parameter knowledge base to convert the natural language description into precise design parameters. On the other hand, the server can also be integrated with the design software system. Many engineering design works will use professional design software. The server realizes data interaction with these design software by developing corresponding interface programs. When the designer completes the design in the software, the server can automatically extract the target design parameters from the database or file of the design software to ensure the accuracy and consistency of the parameters and avoid errors that may be caused by manual input.

[0041] There are also multiple channels for obtaining current engineering environment data. In addition to obtaining data from user-uploaded images and environmental sensors as mentioned above, the server will also be connected to a professional environmental monitoring platform. For example, for meteorological environmental data, the server establishes a data connection with the official monitoring platform of the meteorological department to obtain high-precision and high-frequency meteorological data in real time, including temperature, humidity, wind speed, wind direction, air pressure, etc. For geological environmental data, the server connects to the database of the geological exploration agency to obtain the latest geological exploration report data, including stratigraphic structure, rock and soil mechanics parameters, geological structure information, etc.

[0042] After obtaining the complete target design parameters and current engineering environment data, the server inputs these data into the established engineering data prediction model. During the input process, the server will pre-process the data, such as normalization, and uniformly map data of different magnitudes and units to a specific interval to improve the efficiency and stability of model training and prediction. After the model receives the data, it predicts and calculates the current engineering data based on the rules and patterns learned in training. The predicted values ​​output by the model may be in various forms, such as continuous values ​​(such as stress and strain values ​​of engineering structures, engineering cost estimates), discrete categories (such as possible types of engineering failures, engineering quality levels), etc. The server will further process and display the prediction results according to the type of predicted values ​​and actual application requirements.

[0043] S105, comparing the current engineering data prediction value with the actual observation value to obtain a comparison result; First, for different types of engineering data, the server will develop a special data comparison strategy. For numerical data, such as the stress and strain values ​​of engineering structures and the physical performance parameters of engineering materials, the server will use a combination of absolute error and relative error for calculation. The absolute error can intuitively reflect the difference between the predicted value and the actual observed value, while the relative error can reflect the proportion of this difference in the actual observed value, thereby more comprehensively evaluating the accuracy of the prediction.

[0044] For non-numerical data, such as engineering failure types, engineering quality levels and other classified data, the server will use indicators such as classification accuracy, recall and F1 value for evaluation. Classification accuracy is used to measure the proportion of correctly predicted samples in the total number of samples; recall focuses on the proportion of samples that are actually of a certain category and are correctly predicted; F1 value is the harmonic mean of comprehensive consideration of accuracy and recall, which can more balancedly reflect the performance of the model in classification tasks. When comparing, the server will compare the prediction results with the actual observed classification labels one by one, and then calculate the values ​​of various indicators.

[0045] During the data comparison process, the server will also consider the time series characteristics of the data. Engineering data often has time correlation, and the changes in data at different time points reflect the dynamic development process of the project. Therefore, the server will use time series analysis methods to compare the changing trends of the predicted values ​​and actual observed values ​​in the time dimension. By calculating statistics such as the autocorrelation function and partial autocorrelation function of the time series, the server can determine whether the predicted value can accurately capture the changing trend of the actual data. If the actual observed value shows seasonal fluctuations and the predicted value fails to reflect this trend, even if the error is small at a single time point, the overall prediction effect is not ideal.

[0046] In some embodiments, when the predicted value deviates from the actual observed value and involves key safety indicators, a series of measures will be triggered to deal with potential risks. Specifically, during the monitoring and analysis of engineering data, the server continuously compares the predicted value of the current engineering data with the actual observed value. Once the comparison result shows that the error between the two exceeds the preset error range, and the predicted value is associated with the preset engineering safety key indicators, the system will immediately and automatically start the emergency response mechanism. These key safety indicators may include structural stress limits in construction projects, bearing capacity of key components of equipment in mechanical engineering, etc. In construction, if the predicted stress value of the building structure exceeds the error range of the actual observed value, and the stress value is close to or exceeds the designed safety limit, the start-up conditions are met. After the emergency response mechanism is activated, the server sends an alarm message to the management end on the one hand to attract the attention of managers. These alarm messages contain detailed risk prompts, such as the specific location of the project, the key safety indicators involved, the error situation, etc., so that managers can quickly understand the severity and general situation of the problem. At the same time, the server uses virtual reality (VR) technology combined with the hazard database to simulate suspected dangerous scenes. The hazard database stores detailed data and simulation models of dangerous situations that may be encountered in various projects, covering chain reactions caused by fire, collapse, and equipment failure. Through VR technology, these data are converted into realistic three-dimensional scenes to show the development process of potential dangers and the possible consequences. In order to ensure the authenticity, reliability and traceability of the data, the server uses blockchain technology to record this information after sending alarm information and suspected dangerous scenes to the management end. Blockchain is a distributed ledger technology with the characteristics of being tamper-proof and traceable. The server stores the content of the alarm information, sending time, receiving object, and simulation data and key parameters of suspected dangerous scenes in encrypted form on multiple nodes of the blockchain. This recording method prevents the data from being maliciously tampered with, and any access and modification to the data will leave traces. In the subsequent investigation or review of the accident, relevant personnel can accurately obtain the alarm and simulation scene data at the time through the blockchain, clearly understand the situation when the incident occurred, and analyze the cause of the accident. S106: Evaluate and adjust the engineering data prediction model according to the comparison result.

[0047] Based on the comparison results obtained by S105, the server will conduct a comprehensive evaluation of the engineering data prediction model from multiple dimensions. First, from the perspective of error indicators, the server will set different error thresholds. If the absolute error or relative error exceeds the preset threshold, it indicates that the model has a large deviation when predicting this part of the data. For classified data, if the classification accuracy, recall rate or F1 value is lower than the set standard, it also means that the model needs to be improved in classification performance. The server will judge the severity of the model deviation based on the degree of deviation of these indicators.

[0048] Secondly, the server will analyze the performance of the model under different engineering environments and design parameter combinations. By comparing a large amount of historical engineering data and current engineering data in groups, the prediction accuracy of the model under specific environmental conditions (such as high temperature, high humidity, strong earthquake zones, etc.) or within specific design parameter ranges (such as long-span bridge design, super high-rise building design, etc.) is observed. If it is found that the model performs poorly in certain specific situations, it means that the model is not adaptable enough to these special situations and needs to be optimized in a targeted manner.

[0049] During the model adjustment phase, the server will adopt different adjustment strategies based on the evaluation results. If the prediction deviation is caused by unreasonable model parameter settings, the server will use the optimization algorithm to readjust the model parameters. For example, the stochastic gradient descent algorithm is used to iteratively update the model parameters according to the back propagation of the error, so that the model is optimized in the direction of reducing the error. During the adjustment process, the server will continuously monitor the changes in the error indicators, and stop parameter adjustment when the error converges to a certain degree or reaches the set optimization goal. In addition, the server will continue to collect new engineering data and retrain the model regularly. With the continuous accumulation of engineering practice, the new data will contain more information and change trends. The server will merge the new data with the historical data, re-perform data preprocessing, feature extraction and model training, so that the model can adapt to the dynamic changes of engineering data and maintain good prediction performance.

[0050] In the embodiment of the present application, a series of technical means are adopted, such as obtaining historical engineering data from multiple sources, finely preprocessing and feature extraction of data, building models in combination with cutting-edge generative AI algorithms, collecting current engineering data from multiple channels for prediction, and dynamically optimizing models based on comparison results, so that the potential value and complex relationships in engineering data can be fully explored. It effectively solves the problems existing in traditional engineering data processing methods, such as low prediction accuracy, difficulty in adapting to dynamic changes in engineering environment and design parameters, lack of in-depth mining of engineering environment information, and lack of industry experience sharing in model optimization, thereby achieving accurate and dynamic processing and prediction of engineering data and improving the quality and safety of engineering construction.

[0051] After combining the above content, the following is a more detailed description of the process of the method provided by this implementation. Figure 2 , is another flow chart of the engineering data processing method based on generative artificial intelligence in an embodiment of the present application.

[0052] S201, receiving a text description input by a user, and identifying target design parameters using natural language processing technology; The server first builds a natural language processing (NLP) system that integrates a variety of advanced NLP technologies. The server pre-processes the text input by the user, uses lexical analysis technology, and divides the text into individual words or tokens through a word segmentation algorithm.

[0053] In the semantic understanding stage, the server uses a semantic understanding model based on deep learning. The model is pre-trained on large-scale text data and learns rich language knowledge and semantic representation. The server inputs the preprocessed text into these models, and the model understands the intention and key information of the text by deeply analyzing the semantic relationship between words in the text. In order to more accurately identify the target design parameters, the server builds and maintains a terminology knowledge base dedicated to the engineering field. This knowledge base covers professional terms of various engineering types, parameter names and their definitions, possible value ranges, and other information. During the recognition process, the server matches and associates the key information output by the semantic understanding model with the terminology knowledge base.

[0054] S202, identifying a first current engineering environment from an engineering environment image uploaded by a user using image recognition technology; The server preprocesses the images uploaded by users, including image denoising, normalization and other operations, to improve the quality and consistency of the images and facilitate subsequent feature extraction. In the feature extraction process, the server uses a pre-trained CNN model, which is trained on a large-scale image dataset and can learn rich image feature representations. The server inputs the pre-processed engineering environment image into the pre-trained model, which extracts low-level features such as edges, textures, and colors in the image through convolution operations, and then extracts higher-level and more abstract features through the combination of multi-layer networks. These features can represent various engineering environment elements in the image.

[0055] In order to accurately identify specific elements in the engineering environment, the server built an engineering environment element classifier, which is trained based on a large amount of labeled engineering environment image data and can identify common engineering environment elements such as buildings, terrain, vegetation, roads, water bodies, etc. The server inputs the features extracted by the CNN model into the classifier, and the classifier outputs the engineering environment element category corresponding to each area in the image through feature analysis and judgment.

[0056] The server then uses target detection algorithms to locate and identify specific engineering environment targets in the image. These algorithms can quickly detect the location and category of targets in the image, and the server can use them to accurately identify key engineering environment targets in the image, such as large equipment at the construction site and special buildings in the surrounding area.

[0057] S203, determining a second current engineering environment through a plurality of environmental sensors in the engineering activity setting area; The server establishes a stable and reliable communication connection with multiple environmental sensors in the set area of ​​the engineering activity. These sensors may include meteorological sensors (such as temperature sensors, humidity sensors, wind speed sensors, air pressure sensors, etc.), geological sensors (such as earthquake sensors, soil moisture sensors, geotechnical pressure sensors, etc.), and environmental monitoring sensors (such as noise sensors, air quality sensors, water quality sensors, etc.). The server uses standardized communication protocols to interact with sensors to ensure the stability and efficiency of data transmission. The server also synchronizes the time of different types of sensor data to ensure that all data has the same time reference, which is convenient for subsequent comprehensive analysis.

[0058] The server uses data fusion technology to fuse different types of data collected by multiple sensors. The server uses the weighted average method to assign different weights to the data of each sensor according to the accuracy and reliability of the sensor, and then calculates the weighted average as the fusion result. For some complex environmental parameters, the server will use more advanced data fusion algorithms, such as the Kalman filter algorithm, to optimally estimate the data of multiple sensors by establishing a state space model, and obtain more accurate and comprehensive second current engineering environment data.

[0059] S204, determining a final current engineering environment by combining the first current engineering environment and the second current engineering environment; The server uses data fusion technology when combining the first current engineering environment and the second current engineering environment. In the data fusion process, the server uses a data fusion algorithm based on weight distribution. Different weights are assigned to the first current engineering environment data and the second current engineering environment data according to the accuracy, reliability and relevance of the data to engineering decisions. For data closely related to engineering safety and key decisions, such as geological conditions and the location of large equipment, if the sensor data has higher accuracy and reliability in these aspects, a higher weight is assigned to the sensor data; while for some macro-environmental features, such as the general layout of surrounding buildings, image recognition data may have more advantages and its weight will be increased accordingly. The two types of data are integrated by weighted averaging, weighted fusion, etc. For example, for the terrain data of a certain area, image recognition provides the general outline information of the terrain, and the sensor provides the height data of the local terrain. The server will fuse the two according to the weight to obtain a more accurate description of the terrain.

[0060] S205, using geographic information system data to analyze the topography, geological structure and infrastructure distribution information in the area where the current project is located; The server first connects to professional geographic information system (GIS) databases, which store high-resolution topographic data, detailed geological structure information, and comprehensive infrastructure distribution data. When obtaining topographic information, the server uses the spatial analysis function of GIS to process the digital elevation model (DEM) data of the current project area. By calculating parameters such as slope, slope direction, and terrain undulation, the changing characteristics of the terrain are accurately analyzed. Using contour line generation technology, the ups and downs of the terrain are intuitively displayed, providing basic terrain information for project planning.

[0061] For geological structure information, the server extracts geological layer data from the GIS database, including information such as stratum distribution, fault orientation, and fold morphology. Using geological modeling technology, a three-dimensional geological model is constructed to intuitively present the spatial distribution characteristics of geological structures. The server will also combine the principles of geomechanics to conduct a preliminary assessment of the stability of geological structures and analyze the potential impact of different geological structures on engineering construction, such as the risk of earthquakes near faults and the unstable stratigraphic structure that may affect the foundation stability of buildings.

[0062] When obtaining the distribution information of infrastructure in the preset area, the server retrieves various types of infrastructure in the area, such as roads, bridges, water and electricity facilities, communication base stations, etc., through the network analysis and spatial query functions of GIS. For road facilities, the server analyzes the road's grade, width, traffic capacity and other parameters to evaluate its impact on the transportation of engineering materials and the flow of personnel; for water and electricity facilities, the server obtains information such as their location, capacity, and supply range to determine whether they can meet the water and electricity needs during the construction and operation of the project; for communication base stations, the server determines their coverage and signal strength to ensure smooth communication at the project site.

[0063] The server uses the buffer analysis function of GIS to create buffers of different radius with the construction area as the center, and analyzes the distribution density and importance of infrastructure within the buffer. For infrastructure that is close to the project and has a greater impact on the project, such as nearby substations and water supply pipelines, key marking and analysis are carried out to provide detailed infrastructure information for subsequent project planning and environmental data integration.

[0064] S206: After integrating the topography, the geological structure, the infrastructure distribution information and the final current engineering environment data, the final current engineering environment data is adjusted.

[0065] The server uses data assimilation technology to deeply integrate topography, geological structure, infrastructure distribution information with the final current engineering environment data. When integrating topography data, the server integrates the terrain parameters (such as slope and slope direction) obtained from GIS with the terrain-related information in the first and second current engineering environment data. Through spatial interpolation and data matching, the accuracy and integrity of terrain data are improved. For geological structure data, the server combines the stratigraphic information and fault data in the geological model with the geological sensor data at the engineering site to correct and improve the geological environment data to ensure that the geological data can accurately reflect the actual geological conditions of the engineering area.

[0066] When integrating infrastructure distribution information, the server associates the location, attributes and other data of various infrastructure with the surrounding environmental elements in the project environment data. Combine road information with project transportation route planning, optimize transportation plans based on road capacity and project material transportation needs; match hydropower facility information with project hydropower supply needs, evaluate whether existing hydropower facilities can meet project needs, and formulate corresponding solutions if there are deficiencies.

[0067] After the data fusion is completed, the server adjusts the final current engineering environment data after fusion according to the actual needs and characteristics of the project. For the engineering construction phase, the server focuses on data related to construction safety and progress. Adjust the layout plan of the construction site according to the topographic data to avoid setting up important construction facilities in areas with complex or unstable terrain; optimize the basic construction plan based on geological structure data, and strengthen foundation reinforcement measures for areas with poor geological conditions. For the engineering operation phase, the server adjusts the focus of environmental data based on infrastructure distribution data and the functional requirements of the project. If the project is a large commercial building, the server will focus on analyzing the impact of surrounding transportation facilities and public service facilities on engineering operations, and make corresponding adjustments to the environmental data based on the analysis results to provide more targeted data support for engineering operation management.

[0068] The server will also continue to monitor changes in the project environment. When changes are found in the topography, infrastructure, etc., the corresponding data will be updated in a timely manner, and the data will be re-integrated and adjusted to ensure that the final current project environment data can always accurately reflect the actual environmental conditions of the project.

[0069] In the embodiment of the present application, due to the use of natural language processing, image recognition, sensor data fusion, geographic information system analysis, and multi-source data deep fusion and dynamic adjustment and other technical means, it is possible to obtain more comprehensive, accurate and practical current engineering environment data, and provide high-quality input data for the engineering data prediction model, while improving the model's adaptability to complex environments and the reliability of prediction. It effectively solves the problems of incomplete and inaccurate environmental data acquisition, poor quality of model input data, difficulty in comprehensively considering the complex geographical environment around the project, and inability to dynamically optimize environmental data according to the actual needs of the project in traditional engineering data processing, thereby achieving improved accuracy and reliability of engineering data prediction.

[0070] The following describes the server in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , is a schematic diagram of a physical device structure of a server in an embodiment of the present application.

[0071] It should be noted that Figure 3 The structure of the server shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0072] like Figure 3 As shown, the server includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method described in the above embodiment. In RAM 303, various programs and data required for system operation are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0073] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, a button switch, etc.; an output section 307 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0074] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.

[0075] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0076] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.

[0077] Specifically, the server of this embodiment includes a processor and a memory, and a computer program is stored in the memory. When the computer program is executed by the processor, the engineering data processing method based on generative artificial intelligence provided by the above embodiment is implemented.

[0078] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the server described in the above embodiment; or may exist independently without being assembled into the server. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the server, the server implements the engineering data processing method based on generative artificial intelligence provided in the above embodiment.

[0079] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0080] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0081] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A method for processing engineering data based on generative artificial intelligence, applied to a server, characterized in that: The method comprises: Obtain multiple historical engineering data with different design parameters in different engineering environments; After preprocessing the historical engineering data, feature extraction is performed on the historical engineering data to obtain engineering feature data, wherein the engineering feature data at least includes engineering scheme design parameters and engineering environment data, wherein the engineering environment data refers to data related to the environment in which the engineering activities are located; Building an engineering data prediction model by combining a generative AI algorithm with historical engineering data annotated with the engineering feature data, wherein the engineering data prediction model is used to simulate data distribution in a real engineering environment; After obtaining the target design parameters under the current engineering environment, the target design parameters and the current engineering environment are input into the engineering data prediction model to obtain the current engineering data prediction value; Comparing the predicted value of the current engineering data with the actual observed value to obtain a comparison result; The engineering data prediction model is evaluated and adjusted according to the comparison results.

2. The method according to claim 1, characterized in that After the step of building an engineering data prediction model by combining the generative AI algorithm and the annotated historical engineering data with the engineering feature data, the method further includes: Receive the text description input by the user and identify the target design parameters using natural language processing technology; Using image recognition technology to identify a first current engineering environment from an engineering environment image uploaded by a user; Determine a second current engineering environment through multiple environmental sensors in the engineering activity setting area; The final current engineering environment is determined by combining the first current engineering environment and the second current engineering environment.

3. The method according to claim 2, characterized in that After the step of determining the final current engineering environment by combining the first current engineering environment and the second current engineering environment, the method further includes: Use geographic information system data to analyze the topography, geological structure and infrastructure distribution information in the preset area where the current project is located; After the topography, the geological structure, the infrastructure distribution information and the final current engineering environment data are integrated, the final current engineering environment data is adjusted.

4. The method according to claim 1, characterized in that: After the step of building an engineering data prediction model by combining the generative AI algorithm and the annotated historical engineering data with the engineering feature data, the method further includes: Acquire real-time data during the current project implementation process, wherein the real-time data at least includes project progress and equipment operation status data; When the project progress is delayed or the equipment operation status data is abnormal, the parameters in the project data prediction model are adjusted according to the severity of the delay or abnormality and in accordance with preset rules.

5. The method according to claim 4, characterized in that When the project progress is delayed or the equipment operation status data is abnormal, after the step of adjusting the parameters in the project data prediction model according to the preset rules based on the severity of the delay or abnormal situation, the method further includes: Establish a sharing platform for engineering data prediction models of different engineering types according to the engineering type of the current project; When the engineering data prediction model is adjusted and optimized, the adjusted model parameters and optimization experience are uploaded to the sharing platform; After receiving the case viewing instruction sent by the user, the optimization case similar to the current project is obtained from the sharing platform, and the current project data prediction model is optimized again in combination with the characteristics of the current project.

6. The method according to claim 1, characterized in that After the step of comparing the predicted value of the current engineering data with the actual observed value to obtain a comparison result, the method further includes: If the comparison result shows that the error between the current engineering data prediction value and the actual observation value exceeds the preset error range, and the current engineering data prediction value is associated with the preset engineering safety key indicator, the emergency response mechanism is automatically started; Send alarm information to the management end, and use virtual reality technology combined with the danger database to simulate suspected dangerous scenes; The suspected dangerous scene is sent to the management end for display.

7. The method according to claim 6, characterized in that After sending the alarm information to the management end and simulating the suspected dangerous scene by using virtual reality technology combined with the danger database, it also includes: The alarm information and the suspected dangerous scene are recorded using blockchain technology.

8. A server, characterized in that: The server includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the server to execute the method described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a server, the server is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is run on a server, the server is caused to execute the method according to any one of claims 1 to 7.

Images