Method for model fusion based on petrochemical equipment, training method and related equipment thereof

Abstract

The application provides a training method based on model fusion of petrochemical equipment and related equipment thereof, including: in response to selection of the operation of the current petrochemical equipment, determining the current petrochemical equipment; according to the current petrochemical equipment, determining the task required to be executed by the current petrochemical equipment and a target model; according to the current petrochemical equipment, the task required to be executed by the current petrochemical equipment, the target model and target data, determining a target fusion scheme required to be executed; obtaining the target data; inputting the task required to be executed by the current petrochemical equipment and the target data into the target model according to the target fusion scheme, outputting the result corresponding to the target fusion scheme; analyzing the result; if the error between the result and the target data is less than a preset threshold, determining that the training of the target model is completed; and storing the target model, the current petrochemical equipment, the task required to be executed by the current petrochemical equipment, the target fusion scheme and the target data in association. The method can realize the execution of the task of the current equipment by multiple model fusion.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence, and in particular to a method and training method for model fusion based on petrochemical equipment and related equipment. Background Technology

[0002] Currently, with the development of AI technology in the petrochemical industry, a number of AI algorithm models have been developed for customized applications in specific areas such as optimization, prediction, monitoring and early warning, and fault warning. However, most AI products originate from laboratories and face challenges such as a lack of practical application scenarios, a disconnect between actual data and model data, limited functionality that cannot cover all needs in different scenarios, and high barriers to product iteration, deployment, and maintenance. These challenges result in difficulties in the implementation of AI products and poor application effects.

[0003] Furthermore, in current practical applications, various models (such as asset models, device mechanism models, AI algorithm models, etc.) are scattered and basically exclusive in actual use (i.e., applied independently in devices and scenarios). Therefore, how to simulate and realize the integration of multiple technologies such as asset models, device mechanism models, and AI algorithm models to improve the intelligence level of devices and scenarios in different devices and scenarios through experimental methods is a technical problem that urgently needs to be solved in the existing technology. Summary of the Invention

[0004] This application provides a training method, apparatus, and computer-readable storage medium for model fusion based on petrochemical equipment, which helps to solve the problem of being unable to fuse multiple models to handle the tasks required by the current equipment.

[0005] To achieve the above objectives, this application adopts the following technical solution:

[0006] Firstly, a training method for model fusion based on petrochemical equipment is provided. The method includes: responding to an operation selecting a current petrochemical equipment, determining the current petrochemical equipment; determining, based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment and a target model for performing the task, wherein the target model includes at least two models; determining a target fusion scheme to be executed based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment, the target models, and the data; acquiring target data, including monitoring data, real-time operating data, historical operating data, and maintenance data of the current petrochemical equipment; inputting the task to be performed by the current petrochemical equipment and the target data into the target model according to the target fusion scheme, and outputting the result corresponding to the target fusion scheme; analyzing the result based on the evaluation function of the target model and the target data; if the error between the result and the target data is less than a preset threshold, determining that the training of the target model is complete; and storing the target model, the current petrochemical equipment, the task to be performed by the current petrochemical equipment, the target fusion scheme, and the target data in an associated manner.

[0007] As can be seen from the method described in the first aspect, during the training process, the target fusion scheme of the target model is determined by the current petrochemical equipment, the task of the current petrochemical equipment, and the task of executing the current petrochemical equipment. Then, the accuracy of the fusion scheme is verified by the target data. This solves the problem that multiple models cannot be fused to handle the execution required by the current equipment, thereby improving the accuracy of the target fusion scheme in practical applications.

[0008] In conjunction with the first aspect, in one possible design scheme, after analyzing the results based on the evaluation function of the target model and the target data, the process includes: if the error between the result and the actual result is not less than a preset threshold, then repeating the process of determining the target fusion scheme to be executed based on the current petrochemical equipment, the task to be executed by the current petrochemical equipment, the target model, and the target data, until the error between the result and the target data is less than the preset threshold. According to this possible design scheme, this embodiment, when the result output by the target model is not less than the preset threshold, repeatedly determines the target fusion scheme to be executed based on the current petrochemical equipment, the task to be executed by the current petrochemical equipment, the target model, and the target data, until the result output by the target model is less than the preset threshold, thereby improving the accuracy of the target model's processing based on the target fusion scheme.

[0009] In conjunction with the first aspect, in one possible design scheme, before inputting the task to be executed by the current device and the target data into the target model according to the target fusion scheme, and outputting the result corresponding to the target fusion scheme, the method includes: starting an experimental container for the operating scenario of the current petrochemical equipment on the Kubernetes cluster, and executing the target fusion scheme within the experimental container for the operating scenario of the current petrochemical equipment. According to this possible design scheme, the target models of each petrochemical device are simulated and run within their respective containers, achieving environmental isolation and preventing interference between them. This makes the output results of the target models more accurate and improves the accuracy of the target fusion scheme.

[0010] In conjunction with the first aspect, in one possible design scheme, the evaluation function of the target model includes evaluating the convergence of the loss function and evaluating the inference accuracy of the target model. The analysis of the results based on the evaluation function of the target model and the historical operating data includes: inputting the results into the loss function of the target model and determining whether the loss function has converged; and analyzing the results based on the convergence of the loss function. According to this possible design scheme, using the loss function to evaluate the processing accuracy of the target model is more convincing.

[0011] In conjunction with the first aspect, in one possible design scheme, acquiring the target data includes: sending associated membership data and fields of the task required to be performed by the current petrochemical equipment; responding to the user's selected target associated membership data and target fields; and acquiring the target data based on the target associated membership data and target fields.

[0012] In existing technologies, there is no AI model training platform specifically designed for configuring, training, predicting, and saving models using petrochemical equipment as the primary key. Existing AI platforms often require users to customize data tables, fields, and models for configuration. However, due to the unique characteristics of the petrochemical industry—many plants have numerous pieces of equipment, each with multiple models in operation, and different equipment and tasks require different model types and data collection methods—using commonly available AI platforms does not improve the efficiency of modeling personnel and lacks organic integration with other internal platforms to provide more comprehensive model capabilities. This embodiment, while possessing the general capabilities of common AI platforms, strengthens its close integration with the petrochemical industry, combining with data lakes and knowledge bases to enable modeling personnel to quickly configure and train target models, thereby improving work efficiency.

[0013] Secondly, a method for model fusion based on petrochemical equipment is provided. This method includes: obtaining a list for selecting petrochemical equipment and tasks; responding to the operation of selecting the list of petrochemical equipment and tasks, determining the current petrochemical equipment and the tasks to be performed by the current petrochemical equipment; determining a target model for executing the tasks based on the current petrochemical equipment and the tasks to be performed by the current petrochemical equipment, wherein the target model includes at least two models; obtaining target data, including monitoring data, real-time operating data, historical operating data, and maintenance data of the current petrochemical equipment, wherein the target data originates from server database reading, local database reading, NFS mounting, and reading from other system APIs; determining a target fusion scheme to be executed based on the current petrochemical equipment, the tasks to be performed by the current petrochemical equipment, the target models, and the target data; and inputting the tasks to be performed by the current petrochemical equipment and the target data into the target model according to the target fusion scheme, and outputting the result corresponding to the target fusion scheme. Based on this possible design scheme, it can be seen that this embodiment adopts a target fusion scheme that is most suitable for the current petrochemical equipment and the tasks to be executed by the single-signature petrochemical equipment, based on the petrochemical equipment and tasks selected by the user. This enables the integration of multiple technologies such as asset models, petrochemical equipment mechanism models, and AI algorithm models to jointly improve the intelligence level of petrochemical equipment and scenarios under different petrochemical equipment and tasks.

[0014] In conjunction with the second aspect, in one possible design scheme, determining the fusion scheme based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment, the target model, and the target data includes: if the required fusion scheme cannot be determined based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment, the target model, and the target data, then selecting the fusion scheme closest to the required fusion scheme from the set of fusion schemes as the target fusion scheme. According to this possible design scheme, selecting the fusion scheme closest to the required fusion scheme as the target fusion scheme can avoid the situation where the current petrochemical equipment cannot operate when the target fusion scheme cannot be determined.

[0015] Thirdly, a training device for model fusion based on petrochemical equipment is provided. The device includes: a first response module, a first determination module, a target fusion scheme determination module, a first acquisition module, a first running module, an analysis module, a training completion determination module, and a storage module.

[0016] The first response module is used to respond to the operation of selecting the current petrochemical equipment and to determine the current petrochemical equipment;

[0017] The first determining module is used to determine, based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment and the target model for performing the task, wherein the target model includes at least one;

[0018] The target fusion scheme determination module is used to determine the target fusion scheme to be executed based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment, and the target model;

[0019] The first acquisition module is used to acquire target data, which includes monitoring data, real-time operation data, historical operation data, and maintenance data of the current petrochemical equipment.

[0020] The first operation module is used to input the tasks to be performed by the current petrochemical equipment and the target data into the target model according to the target fusion scheme, and output the result corresponding to the target fusion scheme;

[0021] An analysis module is used to analyze the results based on the evaluation function of the target model;

[0022] The training completion determination module is used to determine that the training of the target model is complete if the error between the result and the historical data is less than a preset threshold.

[0023] The storage module is used to associate and store the target model, the current petrochemical equipment, the tasks to be performed by the current equipment, the target fusion scheme, and the target data.

[0024] Optionally, the acquisition module of the training device based on model fusion of petrochemical equipment described in the third aspect is specifically used for: sending associated membership data and fields of the task required to be executed by the current petrochemical equipment; responding to the user's selected target associated membership data and target fields; and acquiring the target data according to the target associated membership data and target fields.

[0025] Optionally, the analysis module of the training device for model fusion based on petrochemical equipment described in the third aspect is specifically used for: inputting the result into the loss function of the target model, determining whether the loss function converges, and analyzing the result based on the convergence of the loss function.

[0026] Optionally, the underwater petrochemical equipment-based model fusion training device described in the third aspect may further include a loop module and a startup module.

[0027] The loop module is used to repeat the steps after obtaining the target data if the error between the result and the target data is not less than a preset threshold, until the error between the result and the target data is less than the preset threshold.

[0028] The startup module is used to start an experimental container for the current petrochemical equipment's operating scenario on the Kubernetes cluster, and the target model executes the target fusion scheme within the experimental container for the current petrochemical equipment's operating scenario.

[0029] Furthermore, the technical effects of the training device based on petrochemical equipment model fusion described in the third aspect can be referred to the technical effects of the training method based on petrochemical equipment model fusion described in the first aspect, and will not be repeated here.

[0030] Fourthly, a model fusion device based on petrochemical equipment is provided. This model fusion device includes: a list acquisition module, a second response module, a target model determination module, a second acquisition module, a second determination module, and a second execution module.

[0031] The list retrieval module is used to retrieve a list of petrochemical equipment and tasks for selection.

[0032] The second response module is used to respond to the operation of selecting a list of petrochemical equipment and tasks, and to determine the current petrochemical equipment and the tasks that the current petrochemical equipment needs to perform;

[0033] A target model determination module is used to determine a target model for performing the task based on the current petrochemical equipment and the task that the current petrochemical equipment needs to perform, wherein the target model includes at least two models;

[0034] The second acquisition module is used to acquire target data, which includes monitoring data, real-time operation data, historical operation data, and maintenance data of the current petrochemical equipment. The target data is read from the server database, read from the local database, mounted via NFS, and read from other system APIs.

[0035] The second determining module is used to determine the target fusion scheme to be executed based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment, the target model, and the target data;

[0036] The second execution module is used to input the task to be performed by the current petrochemical equipment and the target data into the target model according to the target fusion scheme, and output the result corresponding to the target fusion scheme.

[0037] Optionally, the target fusion scheme determination module of the petrochemical equipment model fusion device described in the fourth aspect is specifically used for: if the fusion scheme to be executed cannot be determined based on the current petrochemical equipment, the task to be executed by the current petrochemical equipment, the target model, and the target data, then the fusion scheme closest to the one to be executed is selected from the fusion scheme set as the target fusion scheme.

[0038] Fifthly, embodiments of this application provide an electronic device. The electronic device includes a processor and a memory; the memory stores a computer program, which, when executed by the processor, causes the electronic device to perform the method described in any implementation of the first aspect, or to perform the method described in the second aspect.

[0039] In a sixth aspect, embodiments of this application provide a computer-readable storage medium, including: a computer program or instructions; when the computer program or instructions are run on a computer, the computer causes the computer to perform the method described in any possible implementation of the first aspect, or to perform the method described in the second aspect. Attached Figure Description

[0040] Figure 1 A schematic diagram illustrating the specific process of the training method for model fusion based on petrochemical equipment provided in this application embodiment;

[0041] Figure 2 A schematic diagram illustrating the specific process of the model fusion method based on petrochemical equipment provided in this embodiment of the application;

[0042] Figure 3 A schematic diagram of the YOLOv4 network mentioned in the underwater target detection method provided in the embodiments of this application;

[0043] Figure 4 A schematic diagram of the structure of a training device for model fusion based on petrochemical equipment provided in an embodiment of this application;

[0044] Figure 5 This is a schematic diagram of the structure of a model fusion device based on petrochemical equipment provided in an embodiment of this application;

[0045] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0046] Reference numerals: 400 - Training device for model fusion based on petrochemical equipment; 410 - First response module; 420 - First determination module; 430 - Target fusion scheme determination module; 440 - First acquisition module; 450 - First running module; 460 - Analysis module; 470 - Training completion determination module; 480 - Storage module; 500 - Model fusion device based on petrochemical equipment; 510 - List acquisition module; 520 - Second response module; 530 - Target model determination module; 540 - Second acquisition module; 550 - Second determination module; 560 - Second execution module; 2000 - Electronic device; 2001 - Processor; 2002 - Memory; 2003 - Transceiver; 2004 - Processor. Detailed Implementation

[0047] The technical solution in this application will now be described with reference to the accompanying drawings.

[0048] In the embodiments of this application, words such as "exemplarily" and "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as an "example" in this application should not be construed as being better or more advantageous than other embodiments or design options. Specifically, the use of the word "example" is intended to present the concept in a concrete manner. Furthermore, in the embodiments of this application, the meaning expressed by "and / or" can be both, or it can be either one or the other.

[0049] In the embodiments of this application, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0050] In the embodiments of this application, sometimes the subscript such as W1 may be mistakenly written as a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0051] Currently, there is no AI model training platform specifically designed for configuring, training, predicting, and saving models using petrochemical equipment as the primary key. Existing AI platforms often require users to customize data tables, fields, and models for configuration. However, due to the unique characteristics of the petrochemical industry—many plants have numerous pieces of equipment, each with multiple models in operation—and different equipment and tasks require different model types and data collection methods. Current common AI platforms do not improve the efficiency of modeling personnel and are not organically integrated with other platforms, failing to provide more comprehensive model capabilities.

[0052] The following will combine Figure 1 This application provides a specific illustration of a training method for model fusion based on petrochemical equipment, as illustrated in the embodiments. For example, Figure 1 This is a flowchart illustrating the training method for model fusion based on petrochemical equipment provided in this application embodiment, specifically the training process for model fusion based on petrochemical equipment.

[0053] like Figure 1 As shown, the training method for model fusion based on petrochemical equipment includes the following steps:

[0054] S110, in response to the operation of selecting the current petrochemical equipment, determine the current petrochemical equipment.

[0055] S120, based on the current petrochemical equipment, determine the task that the current petrochemical equipment needs to perform and the target model for performing the task, wherein the target model includes at least two.

[0056] By utilizing industry algorithms and model libraries, asset models, equipment mechanism models, and AI algorithm models applicable to current petrochemical equipment can be determined.

[0057] In some embodiments, the tasks that current petrochemical equipment needs to perform may include optimization, prediction, monitoring and early warning, and fault early warning. Optimization problems refer to finding the optimal solution or parameter values ​​among numerous options or values ​​under certain conditions, so as to achieve the optimal performance of one or more functional indicators, or to maximize or minimize certain performance indicators of the system. Optimization methods are mathematically based applied techniques used to solve various optimization problems. In the energy industry, optimization problems are widely present in many fields such as signal processing, image processing, production scheduling, task allocation, pattern recognition, automatic control, and mechanical design. Optimization problems are mainly solved through optimization algorithms, which are mainly divided into evolutionary algorithms and swarm intelligence algorithms. Evolutionary algorithms include genetic algorithms, evolutionary programming, and evolutionary strategies, which have wide applications in many fields such as function optimization, pattern recognition, machine learning, neural network training, and intelligent control. Among them, genetic algorithms are simulation evolution optimization algorithms with widespread influence in evolutionary computation; swarm intelligence algorithms are computational algorithms based on the behavioral patterns of biological groups, used to solve distributed problems. Swarm intelligence algorithms provide a new approach to finding solutions to complex distributed problems without centralized control or a global model. Swarm intelligence is a novel approach that can effectively solve most global optimization problems. In recent years, numerous algorithms have emerged in the field of swarm intelligence theory, such as: ant colony optimization, particle swarm optimization, fish swarm optimization, bee swarm optimization, cat swarm optimization, wolf swarm optimization, chicken swarm optimization, bird swarm optimization, culture optimization, weed optimization, bat optimization, cuckoo optimization, fruit fly optimization, frog jumping optimization, bacteria foraging optimization, firefly optimization, and fireworks optimization, etc. Among these, ant colony optimization and particle swarm optimization are the two most important swarm intelligence algorithms.

[0058] In the energy industry, real-time monitoring systems collect vast amounts of data, some of which exhibit periodicity and other temporal characteristics, and are thus termed time series. Extracting the information contained within these time series data to assist in human decision-making, or even to automate it, can significantly improve production and decision-making efficiency. Examples include production forecasting, logistics scheduling, and oil storage scheduling. Common forecasting models include ARIMA models, neural network models, and Prophet models. ARIMA, short for Autoregressive Integrated Moving Average Model, is a general term for a class of models. ARIMA models perform stationarity checks on time series data. If the check fails, appropriate transformations such as logarithmic transformations and differencing are applied to transform it into a stationary sequence. After passing the stationarity check, white noise is detected. If the sequence is not white noise, a suitable ARIMA model can be selected for fitting. If the error value passes the white noise detection, the fitted model can be used to predict the time series data. Neural network models, such as RNNs (Recurrent Neural Networks) and LSTMs (Long Short-Term Memory), incorporate the time factor, ensuring that the output of a neuron is not only related to the current time but also incorporates the output from the previous time point in the model's training, ultimately forming a neural network that propagates over time. The Prophet model's overall framework consists of four parts: Modeling (building a time series model), Forecast Evaluation (model evaluation using historical data simulations; when model parameters are uncertain, multiple attempts can be made, and the most suitable model is determined based on the simulation results), Surface Problems (if the model's overall performance remains unsatisfactory after trying various parameters, potential causes of large errors can be presented, allowing analysts to analyze the problems), and Visually Inspect Forecasts (providing visual feedback on the prediction results; after feedback to analysts, they consider further adjustments and model building). This entire process is a cyclical system combining analysts and automated processes. It can predict the output value for the desired time period based on sequence data from previous periods, combined with historical data, achieving good training results with relatively low model training costs.

[0059] The monitoring and early warning functions in AI applications are mainly reflected in visual intelligence. The monitoring deployment in the energy and chemical industry is comprehensive, but the monitoring content lacks intelligence and requires manual identification, which is still a major deficiency in terms of intelligence. It is impossible to provide early warning and evidence collection after the event, and it also consumes a lot of manpower. At present, the visual intelligence in the energy and chemical industry should be mainly reflected in the application of monitoring videos, which may include: (1) Basic safety management: such as oil and gas leaks, perimeter security, dangerous area behavior (smoking, falling, fire, making or receiving phone calls), labor protection equipment (safety helmet, work clothes); (2) Safe production: high-altitude observation of fire, personnel on-duty detection (whether they are sleeping, absent from their post, or running); (3) Operational safety: hot work operation specifications, confined space operation specifications, flammable and explosive scene operation specifications, high-altitude operation specifications, hoisting operation specifications; (4) Hazardous chemical transportation: specific license plates, vehicle models, explosion-proof signs, hazardous chemical storage and handling; (5) Safety inspection: instrument readings, instrument panel reading comparison.

[0060] To address the above needs, some fundamental algorithms of deep learning that excel in image processing, such as CNN and YOLO, can be applied to create a series of specialized algorithm models for energy and chemical engineering. Examples include: intrusion detection, safety helmet detection, tanker truck detection, license plate recognition, smoke detection, face detection, crowd density detection, running detection, fall detection, absenteeism detection, liquid leak detection, smoking detection, goods handling detection, instrument detection and reading, and so on.

[0061] The energy industry has a large number of large-scale equipment units (such as refining and coking furnaces, reciprocating machines, compressors, etc.). The equipment operation status is mostly monitored by SCADA systems, which store the collected data in a real-time database. The system mainly uses preset alarm values ​​and sets a certain fluctuation range for the current value (a single parameter). If the value exceeds the fluctuation range, an alarm is displayed.

[0062] Human intelligence focuses on the following aspects in the energy field: (1) Mechanism model approach: integrating and analyzing process parameters to form a mechanism model for multi-parameter joint alarm; (2) Feature engineering approach: after establishing a case library, analyze parameter features, extract data features, and then use machine learning algorithms (decision tree, Naive Bayes, KNN, neural network, SVM) and regression algorithms (logistic regression) to train sample data to establish a multi-factor abnormal alarm model; (3) Deep learning approach: through recurrent neural network technology, wavelet transform analysis, short-time Fourier analysis, Wenger distribution analysis and other enhanced processing of various waveform parameters in the equipment (such as waveform data of voltage, current, flow rate, pressure, etc.), non-stationary vibration data is converted into information that can be analyzed by the system, thereby realizing the quantitative standard of data, automatically discovering the equipment operation characteristics under different working conditions, and realizing intelligent evaluation and management of equipment operation status through pattern recognition technology.

[0063] S130, determine the target fusion scheme to be executed based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment, the target model, and the target data.

[0064] The target fusion scheme refers to a scheme in which models used to perform the task selected by the current device can be used simultaneously. The target fusion scheme can also solve the problem that multiple models can only be used individually, and can also solve the exclusivity of model use.

[0065] Complex fusion tasks can be accomplished using the provided low-code development tools, by writing scripts to define the relationships between multiple models during fusion, or by dragging and dropping execution logic scripts using a pipeline tool.

[0066] S140, acquire target data, the target data including the monitoring data, real-time operation data, historical operation data, and maintenance data of the current petrochemical equipment.

[0067] In some embodiments, the target data can be read from a server database, a local database, an NFS mount, or other system APIs. Through various data access methods, it is possible to access monitoring data, maintenance data (including fault causes, fault data, etc.), historical operating data, and real-time operating data (various parameters during operation, such as voltage, temperature, sound, etc.) from different devices and production tasks. This data can be used to experiment with different fusion schemes and ultimately form an efficient and usable overall fusion task solution.

[0068] In some embodiments, data required by different models can be accessed. For example, after selecting a pyrolysis furnace, the data menu can list the associated data and fields related to that model of pyrolysis furnace. Users can select tables and fields by checking boxes to access simulation or production data such as tube wall temperature, open / closed status, and feeding status.

[0069] Optionally, step S140 may include: sending associated membership data and fields of the task required to be performed by the current petrochemical equipment; responding to the user-selected target associated membership data and target fields; and obtaining the target data based on the target associated membership data and target fields.

[0070] S150, according to the target fusion scheme, input the task to be performed by the current petrochemical equipment and the target data into the target model, and output the result corresponding to the target fusion scheme.

[0071] In some embodiments, before performing step S150, the method includes: starting an experimental container for the current petrochemical equipment's operating scenario on a Kubernetes cluster, and executing the target fusion scheme within the experimental container for the current petrochemical equipment's operating scenario.

[0072] Because different models require different runtime environments (for example, some models require GPU acceleration), and deep learning frameworks also differ (such as TensorFlow, PyTorch, and PaddlePaddle), AI model operation demands a lot of computing power, so it is necessary to allocate resources for each model and then run it in a containerized manner.

[0073] S160, Analyze the results based on the evaluation function of the target model and the target data.

[0074] During training, simulated data or real production data are input, and the model evaluation functions under each task (such as model analysis reports, parameters, accuracy, etc.) are used to analyze the model's inference accuracy, time, etc., thereby determining the model's training results.

[0075] Optionally, S160 may specifically include: inputting the result into the loss function of the target model, determining whether the loss function converges, and analyzing the result based on the convergence of the loss function.

[0076] A loss function is used to evaluate the degree to which a model's predicted values ​​differ from the actual values. A better loss function generally indicates better model performance. Different models typically use different loss functions. Loss functions are divided into empirical risk loss functions and structural risk loss functions. The empirical risk loss function refers to the difference between the predicted and actual results, while the structural risk loss function is the empirical risk loss function plus a regularization term. The loss function of the target model can be set according to actual needs; no specific limitations are imposed here.

[0077] S170, if the error between the result and the target data is less than a preset threshold, then the training of the target model is determined to be complete.

[0078] During training, if the error between the result and the target data is less than the error threshold, the training of the target model can be considered complete, indicating that the target fusion scheme is feasible. The error threshold can be set according to actual needs and is not specifically limited here.

[0079] S180, the target model, the current petrochemical equipment, the tasks to be performed by the current petrochemical equipment, the target fusion scheme, and the target data are associated and stored.

[0080] Once the feasibility of the target integration solution is confirmed or a satisfactory integration solution is obtained, users can save the current equipment scenario and the selected tasks such as optimization, prediction, monitoring and early warning, and fault early warning as knowledge points in the knowledge base product. The content includes the asset model, device mechanism model, and AI algorithm model configured for each task; the user-edited logical relationship scripts between these models; and the data requirements for model execution, such as the corresponding data tables, data fields, and data preprocessing methods.

[0081] Because petrochemical companies have a wide range of business operations, a large number of equipment models, and complex data sources, whenever a solution for integrating the goals of a certain equipment and task is successfully implemented, the solution needs to be stored as a template in the knowledge base product to accumulate knowledge for similar equipment.

[0082] In some embodiments, a database can be pre-set to store the target model, the current petrochemical equipment, the tasks to be performed by the current petrochemical equipment, the target fusion scheme, and the target data. In specific applications, after determining the petrochemical equipment and the tasks to be performed by the petrochemical equipment, the target fusion scheme can be directly matched in the database, thereby improving work efficiency.

[0083] Optionally, after step S160, the method further includes: if the error between the result and the target data is not less than a preset threshold, then repeating the step of determining the target fusion scheme to be executed based on the current petrochemical equipment, the task to be executed by the current petrochemical equipment, the target model, and the target data, until the error between the result and the target data is less than the preset threshold.

[0084] When the error between the output result and the target data is not less than a preset threshold, it can be determined that the target fusion scheme cannot be executed or the target fusion scheme is not the best scheme. Thus, step S130 is repeated so that a better target fusion scheme can be obtained in the next iteration.

[0085] Figure 2This is a block diagram illustrating a training process for model fusion based on petrochemical equipment, according to an embodiment of this application. In this application's solution, a list of petrochemical equipment and a task list are accessed through various methods. An algorithm model library with a petrochemical industry background is used to determine asset models, device mechanism models, and AI algorithm models suitable for various petrochemical production equipment and different tasks. Then, through multiple data access methods, historical and real-time data from different equipment and production scenarios can be accessed for experimentation with different fusion schemes, ultimately forming an efficient and usable overall fusion scenario solution. Finally, a one-stop platform for model configuration, data import, model fusion, model prediction, and model iteration addresses the shortcomings of petrochemical enterprises in using AI models, providing end-to-end model training and prediction functions for petrochemical business scenarios, thereby obtaining model fusion schemes for various petrochemical equipment in different tasks.

[0086] The following will combine Figure 3 This application provides a specific illustration of the model fusion method based on petrochemical equipment. For example, Figure 3 This is a flowchart illustrating the method for model fusion based on petrochemical equipment provided in this application embodiment, specifically the process of model fusion based on petrochemical equipment.

[0087] like Figure 3 As shown, the model fusion method based on petrochemical equipment includes the following steps:

[0088] S210, Obtain a list for selecting petrochemical equipment and tasks.

[0089] S220, in response to the operation of selecting a list of petrochemical equipment and tasks, determine the current petrochemical equipment and the tasks to be performed by the current petrochemical equipment.

[0090] S230, determine a target model for performing the task based on the current petrochemical equipment and the task to be performed by the current petrochemical equipment, wherein the target model includes at least two.

[0091] S240, acquire target data, the data including the current petrochemical equipment's monitoring data, real-time operation data, historical operation data, and maintenance data, the target data being read from the server database, read from the local database, mounted via NFS, and read from other system APIs.

[0092] S250, determine the target fusion scheme to be executed based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment, the target model, and the target data.

[0093] Optionally, step S250 includes: if the required fusion scheme cannot be determined based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment, the target model, and the target data, then the fusion scheme closest to the required fusion scheme is selected from the set of fusion schemes as the target fusion scheme.

[0094] In some embodiments, a set can be pre-set to store the various fusion schemes generated during training, so that the target fusion scheme can be quickly determined according to the device and task in subsequent applications.

[0095] S260, the target model inputs the task to be performed by the current petrochemical equipment and the target data into the target model according to the target fusion scheme, and outputs the result corresponding to the target fusion scheme.

[0096] This application's solution, while possessing the common capabilities of typical AI platforms, strengthens its close integration with the petrochemical industry. By combining with internal data lakes and knowledge bases, it enables modeling personnel to quickly configure, train, and use models for prediction, thereby improving work efficiency, quickly locating problems, and rapidly resolving business needs.

[0097] The following will describe in detail the virtual device corresponding to the training method for model fusion based on petrochemical equipment provided in the embodiments of this application, namely the training device for model fusion based on petrochemical equipment, and the virtual device corresponding to the model fusion method for petrochemical equipment provided in the embodiments of this application, namely the model fusion device based on petrochemical equipment.

[0098] For example, Figure 4 This is a schematic diagram of the structure of the training device 400 for model fusion based on petrochemical equipment provided in an embodiment of this application. Figure 4 As shown, the training 400 based on model fusion of petrochemical equipment includes: a first response module 410, a first determination module 420, a target fusion scheme determination module 430, a first acquisition module 440, a first running module 450, an analysis module 460, a training completion determination module 470, and a storage module 480.

[0099] For ease of explanation, Figure 4 Only the main components of the training device 400 for model fusion based on petrochemical equipment are shown.

[0100] The first response module 410 is used to respond to the operation of selecting the current petrochemical equipment and to determine the current petrochemical equipment;

[0101] The first determining module 420 is used to determine, based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment and the target model for performing the task, wherein the target model includes at least two.

[0102] The target fusion scheme determination module 430 is used to determine the target fusion scheme to be executed based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment, and the target model;

[0103] The first acquisition module 440 is used to acquire target data, which includes the monitoring data, real-time operation data, historical operation data, and maintenance data of the current petrochemical equipment.

[0104] The first running module 450 is used to input the task to be performed by the current petrochemical equipment and the target data into the target model according to the target fusion scheme, and output the result corresponding to the target fusion scheme;

[0105] Analysis module 460 is used to analyze the results based on the evaluation function of the target model;

[0106] Training completion determination module 470 is used to determine that the training of the target model is complete if the error between the result and the historical data is less than a preset threshold.

[0107] The storage module 480 is used to store the target model, the current petrochemical equipment, the tasks to be performed by the current equipment, the target fusion scheme, and the target data.

[0108] Optionally, the acquisition module 410 of the training device 400 based on petrochemical equipment model fusion is specifically used for: sending associated membership data and fields of the task required to be performed by the current petrochemical equipment; responding to the user's selected target associated membership data and target fields; and acquiring the target data according to the target associated membership data and target fields.

[0109] Optionally, the analysis module 460 of the petrochemical equipment-based model fusion training device 400 is specifically used for: inputting the results into the loss function of the target model, determining whether the loss function converges, and analyzing the results based on the convergence of the loss function.

[0110] Optionally, the training device 400 for model fusion based on petrochemical equipment may also include a loop module and a startup module.

[0111] The loop module is used to repeat the steps after obtaining the target data if the error between the result and the target data is not less than a preset threshold, until the error between the result and the target data is less than the preset threshold.

[0112] The startup module is used to start an experimental container for the current petrochemical equipment's operating scenario on the Kubernetes cluster, and the target model executes the target fusion scheme within the experimental container for the current petrochemical equipment's operating scenario.

[0113] Furthermore, the technical effect of the training device 400 based on petrochemical equipment model fusion can be referenced from the technical effect of any of the aforementioned underwater target detection methods, and will not be elaborated here.

[0114] For example, Figure 5 This is a schematic diagram of the structure of the training device 500 provided in an embodiment of this application. Figure 5 As shown, the model fusion device 500 based on petrochemical equipment includes: a list acquisition module 510, a second response module 520, a target model determination module 530, a second acquisition module 540, a second determination module 550, and a second execution module 560.

[0115] For ease of explanation, Figure 5 Only the main components of the petrochemical equipment-based model fusion device 500 are shown.

[0116] List retrieval module 510 is used to retrieve a list for selecting petrochemical equipment and tasks;

[0117] The second response module 520 is used to respond to the operation of selecting a list of petrochemical equipment and tasks, and to determine the current petrochemical equipment and the tasks that the current petrochemical equipment needs to perform.

[0118] The target model determination module 530 is used to determine a target model for performing the task based on the current petrochemical equipment and the task that the current petrochemical equipment needs to perform, wherein the target model includes at least two models.

[0119] The second acquisition module 540 is used to acquire target data, which includes the monitoring data, real-time operation data, historical operation data, and maintenance data of the current petrochemical equipment. The target data is read from the server database, read from the local database, mounted via NFS, and read from other system APIs.

[0120] The second determining module 550 is used to determine the target fusion scheme to be executed based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment, the target model, and the target data;

[0121] The second execution module 560 is used to input the task to be performed by the current petrochemical equipment and the target data into the target model according to the target fusion scheme, and output the result corresponding to the target fusion scheme.

[0122] Optionally, the second determining module 550 of the petrochemical equipment-based model fusion device 500 is specifically used to: if the fusion scheme to be executed cannot be determined based on the current petrochemical equipment, the task to be executed by the current petrochemical equipment, the target model, and the target data, then select the fusion scheme closest to the one to be executed from the fusion scheme set as the target fusion scheme.

[0123] Furthermore, the technical effects of the model fusion device 500 based on petrochemical equipment can be referred to the description of the aforementioned related embodiments, and will not be repeated here.

[0124] Optionally, embodiments of this application also provide a computer-readable storage medium, which includes a computer program or instructions that, when executed on a computer, cause the underwater target detection method or training method provided in any embodiment of this application to be executed.

[0125] Optionally, embodiments of this application also provide an electronic device for executing the underwater target detection method or apparatus provided in any embodiment of this application, or for executing the training method or apparatus provided in any embodiment of this application.

[0126] like Figure 6 As shown, electronic device 2000 may include processor 2001.

[0127] Optionally, the electronic device 2000 may also include a memory 2002 and / or a transceiver 2003.

[0128] The processor 2001 is coupled to the memory 2002 and the transceiver 2003, which can be connected via a communication bus.

[0129] The following is combined with Figure 6 A detailed introduction to each component of the electronic device 2000:

[0130] The processor 2001 is the control center of the electronic device 2000. It can be a single processor or a collective term for multiple processing elements. For example, the processor 2001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement the embodiments of this application, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0131] Optionally, the processor 2001 can perform various functions of the electronic device 2000 by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0132] In a specific implementation, as one example, the processor 2001 may include one or more CPUs, for example... Figure 6 CPU0 and CPU1 are shown in the diagram.

[0133] In a specific implementation, as one example, the electronic device 2000 may also include multiple processors, for example... Figure 6 The processors 2001 and 2004 are shown. Each of these processors can be a single-core processor or a multi-core processor. Here, "processor" can refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0134] The memory 2002 is used to store the software program that executes the solution of this application, and is controlled by the processor 2001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0135] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 2002 may be integrated with the processor 2001 or may exist independently and be accessible through the interface circuit of the electronic device 2000. Figure 6 (Not shown in the image) is coupled to processor 2001, and this embodiment does not specifically limit this.

[0136] Transceiver 2003 is used for communication with other electronic devices. For example, if electronic device 2000 is an underwater robot, transceiver 2003 can be used to communicate with network devices or with another terminal device. As another example, if electronic device 2000 is a network device, transceiver 2003 can be used to communicate with terminal devices or with another network device.

[0137] Alternatively, transceiver 2003 may include a receiver and a transmitter. Figure 6 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the sending function.

[0138] Optionally, the transceiver 2003 can be integrated with the processor 2001, or it can exist independently and be connected via the interface circuit of the electronic device 2000. Figure 6 (Not shown in the image) is coupled to processor 2001, and this embodiment does not specifically limit this.

[0139] It should be noted that, Figure 6 The structure of the electronic device 2000 shown does not constitute a limitation on the electronic device. Actual electronic devices may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0140] Furthermore, the technical effects of the electronic device 2000 can be referred to the technical effects of the underwater target detection method described in the above method embodiments, and will not be repeated here.

[0141] It should be understood that the processor in the embodiments of this application can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0142] It should also be understood that the memory in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0143] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0144] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0145] In this application, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or multiple items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0146] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0147] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0148] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0149] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0150] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0151] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0152] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0153] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A training method for model fusion based on petrochemical equipment, characterized in that, The method includes: In response to the operation of selecting the current petrochemical equipment, the current petrochemical equipment is determined; Based on the current petrochemical equipment, determine the tasks that the current petrochemical equipment needs to perform and the target models for performing the tasks, wherein the target models include at least two types; wherein the tasks to be performed include: optimization, prediction, monitoring and early warning, and fault early warning; Based on the current petrochemical equipment, the tasks that the current petrochemical equipment needs to perform, and the target model, determine the target fusion scheme to be executed; Acquire target data, which includes monitoring data, real-time operating data, historical operating data, and maintenance data of the current petrochemical equipment; Start an experimental container for the current petrochemical equipment's operating scenario on the Kubernetes cluster, and execute the target fusion scheme within the experimental container for the current petrochemical equipment's operating scenario; According to the target fusion scheme, the task to be performed by the current petrochemical equipment and the target data are input into the target model, and the result corresponding to the target fusion scheme is output. The results are analyzed based on the evaluation function of the target model and the target data; If the error between the result and the target data is less than a preset threshold, then the training of the target model is determined to be complete. The target model, the current petrochemical equipment, the tasks to be performed by the current petrochemical equipment, the target fusion scheme, and the target data are stored in a knowledge base product in the form of knowledge points. The target model includes at least two of the following: asset model, device mechanism model, and AI algorithm model. The stored content also includes logical relationship information between the target models and data configuration information corresponding to the operation of the target models.

2. The method according to claim 1, characterized in that, The step of analyzing the results based on the evaluation function of the target model and the target data includes: If the error between the result and the target data is not less than a preset threshold, then the process of determining the target fusion scheme to be executed based on the current petrochemical equipment, the task to be executed by the current petrochemical equipment, and the target model is repeated until the error between the result and the target data is less than the preset threshold.

3. The method according to any one of claims 1-2, characterized in that, The evaluation function of the target model includes evaluating the convergence of the loss function and evaluating the inference accuracy of the target model. The analysis of the results based on the evaluation function of the target model and the target data includes: The results are input into the loss function of the target model to determine whether the loss function converges. The results are analyzed based on the convergence of the loss function.

4. The method according to claim 1, characterized in that, The acquisition of target data includes: Send the associated membership data and fields for executing the tasks required by the current petrochemical equipment; Respond to user-selected target associated data and target fields; The target data is obtained based on the target association and the target field.

5. A method for model fusion based on petrochemical equipment, characterized in that, The method includes: Retrieve a list for selecting petrochemical equipment and tasks; In response to the operation of selecting a list of petrochemical equipment and tasks, the current petrochemical equipment and the tasks to be performed by the current petrochemical equipment are determined; wherein, the tasks to be performed include: optimization, prediction, monitoring and early warning, and fault early warning; Based on the current petrochemical equipment and the tasks that the current petrochemical equipment needs to perform, a target model for performing the tasks is determined, and the target model includes at least two types; wherein, the target model includes at least two of the following: an asset model, a device mechanism model, and an AI algorithm model; Acquire target data, which includes monitoring data, real-time operation data, historical operation data, and maintenance data of the current petrochemical equipment. The target data is obtained from server database reading, local database reading, NFS mounting, and reading from other system APIs. Start the experimental container for the current petrochemical equipment operation scenario on the Kubernetes cluster; Based on the current petrochemical equipment, the tasks that the current petrochemical equipment needs to perform, and the target model, determine the target fusion scheme to be executed; The target model, within the experimental container of the current petrochemical equipment's operating scenario, inputs the tasks to be performed by the current petrochemical equipment and the target data into the target model according to the target fusion scheme, and outputs the result corresponding to the target fusion scheme.

6. A training device for model fusion based on petrochemical equipment, characterized in that, The device includes: The first response module is used to respond to the operation of selecting the current petrochemical equipment and to determine the current petrochemical equipment; The first determining module is used to determine, based on the current petrochemical equipment, the tasks that the current petrochemical equipment needs to perform and the target models for performing the tasks, wherein the target models include at least two types; wherein the tasks that need to be performed include: optimization, prediction, monitoring and early warning, and fault early warning; The target fusion scheme determination module is used to determine the target fusion scheme to be executed based on the current petrochemical equipment, the task to be performed by the current petrochemical equipment, and the target model; The first acquisition module is used to acquire target data, which includes monitoring data, real-time operation data, historical operation data, and maintenance data of the current petrochemical equipment. The startup module is used to start an experimental container for the current petrochemical equipment's operating scenario on a Kubernetes cluster, and the target model executes the target fusion scheme within the experimental container for the current petrochemical equipment's operating scenario. The first operation module is used to input the tasks to be performed by the current petrochemical equipment and the target data into the target model according to the target fusion scheme, and output the result corresponding to the target fusion scheme; An analysis module is used to analyze the results based on the evaluation function of the target model; The training completion determination module is used to determine that the training of the target model is complete if the error between the result and the target data is less than a preset threshold. The storage module is used to associate and store the target model, the current petrochemical equipment, the tasks to be performed by the current petrochemical equipment, the target fusion scheme, and the target data in the form of knowledge points in a knowledge base product; wherein, the target model includes at least two of the following: asset model, device mechanism model, and AI algorithm model, and the stored content also includes logical relationship information between the target models, as well as data configuration information corresponding to the operation of the target model.

7. A model fusion device based on petrochemical equipment, characterized in that, The device includes: The list retrieval module is used to retrieve a list of petrochemical equipment and tasks for selection. The second response module is used to respond to the operation of selecting a list of petrochemical equipment and tasks, and to determine the current petrochemical equipment and the tasks to be performed by the current petrochemical equipment; wherein, the tasks to be performed include: optimization, prediction, monitoring and early warning, and fault early warning; The target model determination module is used to determine the target model for performing the task based on the current petrochemical equipment and the task to be performed by the current petrochemical equipment. The target model includes at least two types. The target model includes at least two of the following: asset model, equipment mechanism model, and AI algorithm model. The second acquisition module is used to acquire target data, which includes monitoring data, real-time operation data, historical operation data, and maintenance data of the current petrochemical equipment. The target data is obtained from server database reading, local database reading, NFS mounting, and reading from other system APIs. The second startup module is used to start experimental containers for the current petrochemical equipment's operating scenario on the Kubernetes cluster. The second determining module is used to determine the target fusion scheme to be executed within the experimental container of the current petrochemical equipment's operating scenario, based on the current petrochemical equipment, the tasks to be performed by the current petrochemical equipment, and the target model. The second execution module is used to input the task to be performed by the current petrochemical equipment and the target data into the target model according to the target fusion scheme, and output the result corresponding to the target fusion scheme.

8. An electronic device, characterized in that, include: Processor, the processor being coupled to memory; The processor is configured to execute a computer program stored in the memory to cause the electronic device to perform the method as described in any one of claims 1-5.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a computer program or instructions that, when executed on a computer, cause the method as described in any one of claims 1-5 to be performed.

Images