Process industry process modeling method under cross-platform interaction condition

Abstract

The invention discloses a process industry process modeling method under a cross-platform interaction condition. The process industry process modeling method under the cross-platform interaction condition covers standardized algorithm development logic, a fault injection mechanism and a unified API interface design. Specifically, a universal mechanism model base class is constructed based on the digital twinning characteristic of the technological process, and flexible development and application of the model are supported through inheriting and expanding; two fault injection mechanisms based on performance failure and environmental disturbance are designed, and dynamic behaviors of the system under different working conditions are simulated; and through the design based on a unified API interface, the consistency of data interaction and function calling among different platforms is ensured, and cross-platform efficient cooperation and data sharing are realized.

Description

Technical Field

[0001] The present invention relates to a process industry process modeling method, in particular to a process industry process modeling method under cross-platform interaction conditions. Background Art

[0002] With the continuous development of industrial automation and information technology, the production process steps in process industries such as petroleum, chemical, and electric power are becoming more coupled and the mechanisms are becoming more complex. In this context, process modeling, as an important part of the process industry, has become an important tool for improving production efficiency, optimizing resource allocation, and ensuring production safety.

[0003] Traditional process modeling methods mainly rely on tools and technologies on a single platform, making it difficult to achieve data and information sharing between different systems and different equipment, resulting in many challenges in the process industry in practical applications. First, process models are usually developed based on a single platform or specific tools, which limits their application in a cross-platform environment. For large process companies, it is often necessary to share and collaborate on data and models across multiple platforms, but existing modeling tools and methods often lack good cross-platform compatibility, leading to interaction barriers and data transmission problems between different platforms. Secondly, the process modeling process is complex and involves a large amount of real-time data, which places higher requirements on the computing power, data processing capabilities, and adaptability of modeling tools to actual process flows. However, existing technologies often cannot effectively solve many problems in cross-platform collaboration and real-time data processing. Therefore, how to achieve cross-platform interaction between process modeling and downstream businesses in the process industry is of great significance for solving the mechanism modeling needs of complex process flows.

[0004] Although in the prior art, the process modeling method for the process industry based on cross-platform interaction conditions has been applied in some fields and has made certain progress, there are still problems such as insufficient cross-platform collaboration capabilities, insufficient process modeling accuracy and complexity, and lack of unified standards and interfaces. Therefore, the present invention proposes a process modeling method for the process industry under cross-platform interaction conditions. It aims to achieve efficient collaboration and data sharing between multiple platforms through unified cross-platform data interfaces and standards, combined with advanced modeling algorithms, to improve the modeling accuracy and real-time performance of the process flow, and has broad application prospects. Through the method of the present invention, flexible and efficient process flow modeling can be achieved between different platforms, thereby providing strong support for the digital transformation of the process industry. Summary of the invention

[0005] In order to solve the shortcomings of existing process industry process modeling methods in cross-platform interaction, the present invention proposes a process industry process modeling method under cross-platform interaction conditions. This method realizes efficient collaboration and data sharing between multiple platforms through a unified cross-platform data interface and standard, combined with a modeling method, and effectively improves the accuracy and real-time performance of process flow modeling. Compared with the traditional single-platform modeling method, the present invention can achieve cross-platform collaboration, reduce data transmission delays, improve system compatibility and flexibility, and further promote the digital transformation of the process industry.

[0006] The process industry process modeling method under cross-platform interaction conditions enhances the reliability and flexibility of the system by realizing a unified data interface and a standardized modeling framework, combined with a fault injection mechanism and an API definition.

[0007] The fault injection module is used to simulate various fault conditions that may occur when exchanging data and collaborating models between different platforms, so as to ensure that the system can still operate efficiently in different environments and conditions. At the same time, the design of the API interface definition ensures the standardization of data interaction and function calls between different platforms, so that the process modeling tools of each platform can work together through a unified interface, thereby achieving cross-platform interoperability and data sharing. These operations not only ensure the real-time transmission and synchronization of data and information between multiple platforms, but also optimize the adaptability of the model, and can be adjusted in real time according to different operating environments.

[0008] The characteristics of the present invention are:

[0009] (1) According to the characteristics of the digital twin of the process flow, a universal mechanism model base class is constructed, which provides a good specification for process modeling in the process industry, thereby effectively ensuring the consistency of the integration and calling methods of various process models.

[0010] (2) According to the characteristics of common failure mechanisms in the process flow, two universal fault injection mechanisms corresponding to internal performance degradation and external environmental disturbances are constructed, which effectively reduces the difficulty of developing process models for fault simulation requirements.

[0011] (3) Through a unified API interface and standardized data exchange mechanism, the consistency of data interaction and function calls between different platforms is ensured, seamless collaboration and data sharing between different platforms are achieved, the application limitations of traditional process modeling methods in a cross-platform environment are solved, and the scalability and ease of use of the system are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 Modeling the entire process cycle for process industries

[0013] Figure 2Design and inheritance diagram for the base class

[0014] Figure 3 Schematic diagram of the fault injection mechanism process

[0015] Figure 4 Schematic diagram for API management and interface definition DETAILED DESCRIPTION

[0016] A process industry process modeling method under cross-platform interactive conditions provided by the present invention is described in detail below with reference to the accompanying drawings.

[0017] The present invention provides a process industry process modeling method under cross-platform interaction conditions. The process industry process modeling method is oriented to industries such as petroleum, chemical industry, and electric power, and is aimed at application requirements such as complex algorithm development, insufficient system fault tolerance, and poor data interaction compatibility in cross-platform collaboration. Figure 1 As shown in the figure, by introducing standardized algorithm development logic, fault injection mechanism and unified API interface, the development difficulty of process modeling is reduced, the robustness and stability of the system in a multi-platform environment are enhanced, the consistency of data sharing and function calls between different platforms is achieved, and certain support is provided for the efficient collaboration of complex industrial systems.

[0018] The core of the process industry process modeling method under cross-platform interaction conditions is to use the standardized algorithm development logic of class calls, fault injection mechanism and unified API interface to solve the problems of repeated development, high fault tolerance requirements and inter-platform collaboration limitations of traditional process modeling methods in cross-platform environments.

[0019] 1. Standardize algorithm development logic

[0020] Based on the process modeling requirements of the process industry, the present invention constructs a universal mechanism model base class to form a standardized design and calling framework. Figure 2 As shown, the present invention realizes the standardization of process model parameters and methods through base class design, supports the consistency and universality of the model, and allows flexible expansion of new functions through inheritance mechanism, ensuring the rapid development and efficient collaboration of diversified models.

[0021] The invention encapsulates the core parameters of the process model into standardized attributes through the design of the base class, ensuring the universality and consistency of the parameters. In the design of the base class, the core parameters and methods of the model provide the basic capabilities for building and expanding the process model through standardized encapsulation. The parameter set of the base class is divided into two parts: default parameters and dynamic configuration parameters: the default parameters are used to define the physical properties of the model, including common basic indicators in process modeling such as heat transfer, flow, density, etc., representing the initial state and standardized characteristics of the model; the dynamic configuration parameters allow users to adjust these physical properties according to needs, providing flexibility for the model to adapt to different working conditions. In addition, the base class also contains attributes for managing time steps and an internal cache for recording status to support dynamic updates and efficient operation of the model.

[0022] The method system of the base class covers functions such as parameter configuration, time management, state cleaning, single-step deduction and multi-step simulation. The parameter configuration function modifies the core physical parameters of the model according to the adjustment factors input by the user to meet the needs of different modeling scenarios. The time management function calculates and adjusts the time step so that the model responds to dynamic input changes at different times. The state cleaning function resets the internal cache at the beginning of each simulation to ensure the independence and accuracy of the calculation process. The single-step calculation function realizes the process recursion based on the current state variable and calculates the state at the next moment. The multi-step simulation function completes the entire simulation process through iteration on the basis of single-step calculation and generates state data in a continuous time period. During the simulation process, the model realizes complete modeling of the process flow through the fusion of these methods.

[0023] When inheriting the base class to build a new model, it is necessary to operate according to the established framework and standardized process. First, the default parameters and configuration mechanism of the base class should be inherited to ensure that the basic physical properties and configuration logic of the new model are consistent with the system. Secondly, according to the specific needs of the new model, the single-step calculation logic can be rewritten to implement a calculation method that is more in line with the actual process. At the same time, if new functions need to be introduced, exclusive methods or properties can be added, but compatibility with the overall structure of the base class should be ensured. In addition, the data input and output format of the base class must be strictly followed to ensure that the new model can be seamlessly integrated into the existing system, and data exchange and collaboration can be achieved with other functional modules through standard interfaces. Through this inheritance and extension method, a variety of mechanism models can be quickly developed on a standardized basis to provide support for complex process modeling in the process industry and avoid repeated development work.

[0024] 2. Fault injection mechanism

[0025] The present invention constructs a universal fault injection mechanism for common faults in process flow, such as Figure 3As shown in the figure, the mechanism includes two main forms: performance failure injection and environmental disturbance injection. The performance failure injection and environmental disturbance injection both adopt the injection configuration specification of the system, as shown in the following table. Among them, the fault time is used to configure the relative time range of fault injection, written as "fault injection start frame position: fault injection end frame position". If the start frame position is missing, the fault is injected from the first frame by default; if the end frame position is missing, the fault is injected continuously until the end of the last frame by default. The parameters involved are used to configure the model position of fault injection, which are respectively reflected as disturbances at the model parameter and simulation signal levels in the performance failure injection and environmental disturbance injection. The injection mode represents the method of fault injection, and provides two modes: "propagation mechanism injection" and "sensing mechanism injection". The former means that the injected fault will change the value of the parameter in the simulation process, thereby affecting the subsequent signal propagation process; the latter means that the injected fault only stays in the sensor signal, and still uses the normal value to participate in the subsequent simulation process. The fault expression represents the specific injection form of the fault, and the fault mechanism is quantitatively characterized according to the fault expression specification.

[0026] Table 1 Fault injection format

[0027] Serial number Failure time Involved parameters Injection Mode Fault Expression ... ... ... ... ...

[0028] The fault expression is mainly defined by the combination of specified functions, and the main function logic is shown in the following table.

[0029] Table 2 Main logic format of fault injection functions

[0030]

[0031]

[0032] 2.1 Performance Failure Injection

[0033] The performance failure injection form described is a mechanism to simulate the performance degradation or inherent failure of the process system itself, aiming to reproduce the impact of equipment aging, material performance degradation or internal parameter deviation on system operation. This mechanism reflects the performance degradation that may occur in the process flow under long-term operation or overload conditions, and is the core part of fault simulation. Performance failure injection usually focuses on changes in key system parameters, such as heat exchange efficiency, fluid viscosity or density, which directly affect the calculation process and output results of the process model.

[0034] The key to achieving performance failure injection lies in the dynamic adjustment of the core parameters of the model. First, it is necessary to identify the key parameters in the model that affect performance. These parameters usually correspond to the core physical properties of the device or the key operating conditions of the system. During the simulation process, the injection mechanism gradually adjusts these parameters through specific rules, and these fault expressions are passed in the configuration interface of the model for setting. These settings will then affect the results of each calculation during the simulation process, which is directly reflected in the changes in the model output results. Users can inject corresponding faults into the parameters in the base class and set specific fault configurations, including fault types and fault injection levels, to achieve refined fault simulation. For example, the parameters in the above base class can be set to simulate performance degradation in a linear decrease, nonlinear offset, or periodic fluctuation manner, including decreased thermal efficiency, increased flow resistance, etc., as shown in the following table.

[0035] Table 3 Internal performance fault injection

[0036] Serial number Failure time Involved parameters Injection Mode Fault Expression 1 0:15 Thermal_efficiency Propagation Mechanism Injection X*(1-0.02*t) 2 15: Flow_rate Propagation Mechanism Injection X+(-0.1*t**2) 3 :30 Density Propagation Mechanism Injection X+(-0.005*exp(t))

[0037] In each simulation iteration, the performance failure injection mechanism updates the parameters according to the preset rules and feeds back to the model calculation in real time. For example, the reduction of heat conduction efficiency may directly change the thermal conductivity in the heat transfer equation, resulting in abnormal temperature distribution; the deviation of fluid density will affect the fluid dynamics equation, thereby changing the calculation results of flow and pressure.

[0038] 2.2 External environment injection

[0039] The external environment injection form is a mechanism to simulate the interference of external environment changes on the operation of the process system, and is used to reproduce the impact of environmental factors on the dynamic behavior of the system. Its implementation can cover two modes: propagation mechanism injection and sensing mechanism injection, which are respectively suitable for different fault simulation requirements.

[0040] In the propagation mechanism injection mode, the external disturbance signal directly changes the core environmental parameters used in the model calculation and propagates to the subsequent state through the internal mechanism of the process model. For example, when simulating a sudden rise in the external temperature, the heat transfer equation in the model will adjust the heat conduction parameters in real time, thereby affecting the heat distribution and state changes of the entire system. According to actual needs, the propagation mechanism injection can be subdivided into three scenarios: short-term shocks, periodic fluctuations, and random fluctuations. Short-term shocks simulate instantaneous changes in external conditions, such as a sudden increase in external temperature; periodic fluctuations reproduce regular environmental changes, such as periodic changes in conditions caused by the temperature difference between day and night; random fluctuations are used to simulate irregular disturbances, such as occasional pressure fluctuations or sudden abnormal interference.

[0041] In the sensing mechanism injection mode, the injected fault only affects the parameter value recorded by the sensor, but the core model calculation still uses the undisturbed real parameter value. This mode is used to simulate the impact of sensor abnormality or failure on system monitoring and fault diagnosis. For example, a temperature sensor may record a temperature higher than the actual value due to a fault, but the heat transfer calculation in the model is still based on the real ambient temperature. The sensing mechanism injection mechanism allows the definition of various forms of sensor deviations, including fixed offset values, periodic errors or random disturbances, so as to evaluate the impact of sensor failures on system reliability and operational stability.

[0042] The key to realize external environment injection lies in the generation of external disturbance signals. In specific operations, the user transmits the type, range and injection mode of the disturbance signal to the system, specifies the fault variable, mathematical expression of the disturbance signal and the action position, so as to realize the comprehensive simulation of various complex environmental interferences, as shown in the following table.

[0043] Table 4 External environment fault injection

[0044] Serial number Failure time Involved parameters Injection Mode Fault Expression 1 15 Ambient_temp Propagation Mechanism Injection X+(5*sin(0.5*PI*t)) 2 :60 Humidity Propagation Mechanism Injection X*(random()*0.3+0.7) 3 15:30 Pressure Sensing mechanism injection X+(101325+2*sin(0.2*t)) 30:60 FlowRate Sensing mechanism injection X+(5+0.1*(t-10))

[0045] In each simulation iteration, the disturbance signal is dynamically updated according to the preset rules and acts on the parameter calculation process of the model, affecting the instantaneous state of the system. By comparing and analyzing the change rules of the disturbance signal and the model output response, the specific impact of environmental disturbances on the system operation state can be evaluated, including process performance fluctuations, abnormal propagation paths, and system recovery capabilities.

[0046] 3. Unified API interface definition

[0047] The present invention realizes efficient collaboration and data sharing of process industry process modeling in a cross-platform environment through a modular API management mechanism and a unified API interface design. Figure 4 As shown, the invention develops a modular API management mechanism, designs a unified API interface and realizes the interaction between the two.

[0048] 3.1 Modular API Management Mechanism

[0049] The API management mechanism is managed by the model manager module, which dynamically manages process model instances and provides an independent operating environment for each model. Its core functions include model registration, model acquisition, model reset and model deletion.

[0050] The model registration function receives the parameter configuration passed in by the user, creates a new process model instance, and assigns it a unique identifier (UUID) to ensure the independence and traceability of the model instance in a multi-user, multi-task environment. The model acquisition function accurately accesses a specific model instance through an identifier (UUID), supports user queries and operations on the model, and facilitates centralized management of tasks. The model reset function provides a fast clearing and initialization function for the model state to support the execution of multiple rounds of simulation tasks, ensuring that the model always maintains data consistency and computational independence during use. The model deletion function releases the resources of a specific model instance after the task is completed, improves system operation efficiency, and avoids redundant instances from occupying too much memory. The above operations establish a dynamic and efficient model instance lifecycle management mechanism for the manager, ensures task independence and avoids operation conflicts, and provides support for complex modeling tasks.

[0051] 3.2 Unified API interface design

[0052] The API interface is the core bridge for users to interact with the system, providing a series of standardized functions to achieve orderly connection between data transmission and model operation. Its logic mainly includes simulation task construction, fault injection, step simulation and state management.

[0053] The simulation task construction interface allows users to upload the structured configuration of the model (such as JSON format), and the system maps these configurations to the dynamic attributes of the model. After receiving the configuration, the manager module will call the model registration function to create a new model instance, and generate a universally unique identifier (UUID) and return it to the user. In the implementation, the interface parses the configuration file submitted by the user, maps the input parameters to the dynamic attributes of the model, and verifies and processes them according to the predefined parameter mapping rules to ensure the standardization of the data format and content.

[0054] The fault injection interface supports users to adjust the dynamic parameters of the model in real time to simulate complex process failure scenarios. Users specify the fault injection location (such as a certain parameter) and the injection degree (such as offset value or change rate) through the interface. After the system parses the user input, it locates the target model instance through the manager and calls its configuration method to dynamically modify the model parameters. In the implementation, the fault injection data is transmitted in the form of standardized expressions. The interface updates the parameters in real time according to the expression type (such as fixed value, linear change or periodic fluctuation), and directly applies the injection results to the next calculation of the model.

[0055] The step-by-step simulation interface allows the user to control the dynamic operation of the model based on the time series. The user passes in the timestamp and data frame through the interface. After receiving the timestamp, the interface first parses the timestamp and calls the time management function of the model to dynamically adjust the time step of the simulation. Subsequently, the parameters in the data frame are mapped to the dynamic input of the model, and the single-step calculation method is called to generate the simulation results for the next moment. The system organizes the calculation results into a structured data format (such as JSON) and returns them to the user, clearly showing the dynamic behavior of the system. In implementation, the interface ensures high consistency between time series input and model response through a unified time synchronization mechanism and dynamic data mapping rules, thereby improving the accuracy and availability of simulation results.

[0056] The state management interface includes two functions: model reset and deletion. The reset interface is used to clean up the internal cache of the model and restore the default parameter configuration to provide a consistent initial state for multiple rounds of simulation tasks; the deletion interface is used to release the model instance and the system resources it occupies. In the implementation, the reset interface calls the internal cleanup method of the model to ensure the independence of the model state; the deletion interface completely clears the model instance and its associated data through the manager module to avoid redundant occupation of system memory. The two functions ensure the effective management of the life cycle of the model instance through standardized interface design and support the efficient operation of multi-task parallel operations.

[0057] 3.3 API management and interface collaboration mechanism

[0058] The API management mechanism and API interface form a highly coordinated workflow through the manager module. The manager module is responsible for the full life cycle management of process model instances, including the creation, acquisition, reset and deletion of instances, while the API interface serves as the entry point for user interaction, mapping user requests to manager operations. In the specific implementation, the interface calls the manager's method through the passed parameters. For example, when the user builds a simulation task, the interface passes the model configuration parameters to the manager's model registration method and generates a unique identifier (UUID) to identify the instance. In subsequent interface calls, such as fault injection or step simulation, the interface accurately locates the model instance through the UUID and calls related methods for operation.

[0059] The collaborative mechanism is manifested in the specific functions as the interface focuses on receiving user requests and data processing, while the manager is responsible for the specific operations of the model instance. For example, the step simulation interface receives the timestamp and input data frame, uses the manager to obtain the corresponding model instance, and calls the time step configuration method and calculation method of the instance to complete the simulation. At the same time, after the fault injection interface obtains the model instance through the manager, it calls the dynamic configuration method of the model to adjust the parameter value, thereby completing a variety of fault injection scenarios. This mechanism with clear division of labor ensures efficient collaboration between the interface and the manager, and uniformly manages the model instance through a unique identifier, ensuring the independence of operations and the consistency of data transmission.

Claims

1. A process modeling method for process industries under cross-platform interaction conditions, characterized by: The process industry process modeling method is aimed at the petroleum, chemical, electric power and other industries. It aims at the application limitations of the traditional single platform modeling method in the prior art in a cross-platform environment, and improves the adaptability of the traditional process modeling method in a cross-platform environment.

2. According to claim 1, the process industry process modeling method under cross-platform interaction conditions is characterized by: The process industry process modeling method includes three aspects: constructing a universal mechanism model base class and standardizing the process modeling process; designing a fault injection mechanism including performance failure injection and environmental disturbance injection; Implement a unified API interface and standardized data exchange mechanism to support cross-platform collaboration and data sharing.

3. According to claim 2, the process industry process modeling method under cross-platform interaction conditions is characterized by: The generalized mechanism model base class encapsulates the core parameters and methods of the process model and standardizes the design and calling framework of process modeling. It defines a unified input and output interface for the unified format of input data and the standardized display of output results of different types of process models; it supports the default setting and dynamic configuration of parameters, can automatically adjust the parameter configuration according to the actual operating environment, and can be manually adjusted according to user needs; through the inheritance mechanism, subclasses can inherit the general base class and extend specific modeling methods according to specific process requirements, ensuring the reusability and flexible extensibility of the model, and quickly adapting to different working conditions and industry needs.

4. According to claim 2, the process industry process modeling method under cross-platform interaction conditions is characterized by: The fault injection mechanism includes two forms: performance failure injection and environmental disturbance injection. Performance failure injection is used to simulate the inherent faults or performance degradation of the process system, specifically by dynamically adjusting the key parameters in the model to reflect the impact of factors such as equipment aging, corrosion, and wear on the system; environmental disturbance injection is used to simulate the interference of external environmental changes, and supports two injection modes: propagation mechanism injection, which analyzes the impact of external environmental changes on the process flow by simulating changes in environmental factors; sensing mechanism injection, which evaluates the interference of sensor failure or drift on system feedback control by simulating sensor errors or faults.

5. According to claim 2, the process industry process modeling method under cross-platform interaction conditions is characterized by: The API interface is implemented through a modular management mechanism, supporting simulation task construction, fault injection, step simulation and state management. Through a unified task scheduling framework, it can integrate different process models, fault injection and simulation modules into the same simulation process. The API interface includes functions such as model registration, model acquisition, model reset and deletion. It ensures efficient management of model instances through dynamic resource management, and can automatically update and clear outdated or no longer used model instances. The API interface supports cross-platform collaboration and data sharing, and adopts standardized data formats and protocols to realize data exchange between different platforms and systems and unified processing of simulation results.

Images