MBSE-based SysML parametric diagram model calculation method

Abstract

The present invention discloses a parametric model calculation method based on MBSE, comprising the following steps: obtaining a SysML parametric diagram model constructed in MBSE; parsing the SysML parametric diagram model into value attributes and constraint attributes, converting the attribute models corresponding to the value attributes into non-computational nodes, and converting the constraint models corresponding to the constraint attributes into computational nodes; constructing binding chains between nodes, and solving the value attributes and constraint attributes through the binding chains. The method can analyze the constraint relationships between model elements of a complex SysML parametric diagram model, and can also perform error analysis on the SysML parametric model.

Description

Technical Field

[0001] The present invention belongs to the technical field of software development, and in particular relates to a SysML parameter diagram model calculation method based on MBSE. Background Art

[0002] With increasing demands for usability, comfort, and safety in industrial products across industries like aviation, aerospace, and automotive, the electrification and intelligence of these products are increasing, and product complexity is soaring. Model-Based Systems Engineering (MBSE) can model all aspects of the system design process, improving design efficiency, ensuring system quality, and facilitating the maintenance and upgrade of design solutions.

[0003] System modeling requires a modeling language. SysML (Systems Modeling Language), a standard of the Object Management Group (OMG), extends the Unified Modeling Language (UML) to meet the needs of system modeling. Currently, SysML is widely used in MBSE practices.

[0004] The SysML language specification defines how to define constraint models and the relationships between constraints and models, but it doesn't address constraint evaluation and resolution. In MBSE, SysML modeling yields a SysML parametric diagram model, which documents the constraint relationships between model elements. By performing parameter analysis on the SysML parametric diagram model, the constraint relationships between model elements can be more clearly understood, enabling model analysis—a crucial requirement for MBSE modeling.

[0005] In existing parameter engines, the computational complexity of SysML parameter models is too large, the dependencies between constraint attributes in SysML parameter models are very complex, the development difficulty and development cycle are very long, and the relationship context of SysML parameter models is disordered and chaotic, making it difficult to solve the parameters of SysML parameter diagram models.

[0006] Of course, there are existing parameter engine calculations that can realize parameter solution of SysML parametric diagram models. However, before the calculation, it is necessary to execute and analyze the direct dependencies of the parameter constraint attributes involved in the calculation to identify and analyze the execution order between the parameter constraints. In addition, it only supports expressions written in the SysML language, which greatly limits the application of parameter solution. Summary of the Invention

[0007] In view of the above, the present invention provides a SysML parameter diagram model calculation method based on MBSE, which provides a basis for analyzing the correctness of the parameter model by calculating the constraint relationship between model elements of the parameter model.

[0008] To achieve the above-mentioned object of the invention, an embodiment provides a parameter model calculation method based on MBSE, comprising the following steps:

[0009] Get the SysML parametric diagram model built in MBSE;

[0010] Parse the SysML parameter diagram model into value attributes and constraint attributes, convert the attribute model corresponding to the value attribute into a non-calculated node, and convert the constraint model corresponding to the constraint attribute into a calculated node;

[0011] Build binding chains between nodes and solve value attributes and constraint attributes through the binding chains.

[0012] Preferably, the non-computational node consists of an attribute model and an attribute default value, wherein the attribute default value is obtained from a default value attribute of the attribute model;

[0013] The computing node is composed of an input module, an output module and an expression parser, wherein the input module corresponds to the input constraint parameters of the expression, the output module corresponds to the output constraint parameters of the expression, and the expression parser corresponds to the constraints of the constraint model.

[0014] Preferably, the binding chain constructed between nodes includes a binding chain between value attributes and value attributes, a binding chain between value attributes and constraint parameters, and a binding chain between constraint parameters and constraint parameters, wherein the constraint parameters include input constraint parameters and output constraint parameters.

[0015] Preferably, when building a binding chain, first, determine the type attributes of the two end models corresponding to the two end nodes. When the type attributes of the two end models are the same, or the type attribute of one end model is a subclass of the type attribute of the other end model, build a binding chain between the two end nodes.

[0016] Preferably, when building a binding chain, a method for transferring attribute values through the binding chain is also determined, including:

[0017] (1) Transfer according to the definition order of both ends of the binding chain. If the model at one end of the binding chain is created before the model at the other end and has default attribute values, the attribute values are copied from the end of the model created earlier to the end of the model created later during initialization.

[0018] (2) Transfer according to the change of attribute value. If the attribute value at one end of the binding chain changes, the attribute value is transferred from the end where the attribute value changes to the other end where the attribute value remains unchanged.

[0019] Preferably, when the attribute value is passed through the binding chain, the attribute value is copied according to the multiplicity of the attribute or constraint parameter. The receiving end of the binding chain will update the attribute value according to the definition of the multiplicity of its own model. If the multiplicity of the receiving end is less than that of the sending end, the receiving end can only receive the same number of attribute values; if the multiplicity of the receiving end is greater than that of the sending end, the receiving end only updates the attribute value of the number of the multiplicity of the sending end, and the attribute value remains unchanged thereafter.

[0020] Preferably, solving the value attribute and the constraint attribute through the binding chain includes:

[0021] Taking the non-computational node assigned the default attribute value as the starting point, the attribute value of the starting point is assigned to the computational node that has a binding relationship with the starting point through the binding chain;

[0022] The compute node obtains the constraint expression through the constraint attribute, and obtains the input constraint parameter and output constraint parameter through the constraint expression. Then, through the binding chain, the value of the value attribute of the non-compute node is assigned to the corresponding input constraint parameter. The constraint expression is then executed using the expression parser. After execution, the output constraint parameter name is used to obtain the value of the output constraint parameter from the expression parser, completing the assignment of the output constraint parameter.

[0023] Another node associated with the computation node through a binding chain copies the attribute value from the output constraint parameter and continues the next solution. The other node is another non-computation node or another computation node.

[0024] Preferably, when the other node is another non-computational node, the value attribute of the other non-computational node is copied from the output constraint parameter and used as the starting point of a new round for the next round of solution calculation, and the process is iterated until there are no more available starting points.

[0025] When the other node is another computing node, the other computing node obtains another constraint expression through the constraint attribute, and obtains another input constraint parameter and another output constraint parameter through the other constraint expression, and then assigns the value of the output constraint parameter of the computing node to the corresponding other input constraint parameter through the binding chain; then the expression parser is used to execute the other constraint expression. After the execution is completed, the value of the other output constraint parameter is obtained from the expression parser by outputting the name of the other constraint parameter, thereby completing the assignment of the other output constraint parameter.

[0026] Preferably, after obtaining the constraint expression, the constraint expression content and expression language type are obtained according to the definition of the constraint specification, and then a suitable expression parsing module is selected according to the expression language model to solve the constraint expression.

[0027] Preferably, when solving constraint expressions, constraint expressions written in three languages ​​(Python, Matlab and JavaScript) are supported.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] Based on parsing the SysML parametric diagram model into value attributes and constraint attributes, by constructing non-computational nodes and computational nodes, and building binding chains between the nodes, when calculating expressions, there is no need to sort out the direct dependencies of each expression, and there is no need to pay attention to the dependencies between complex constraint attributes, which greatly reduces the development difficulty and shortens the development cycle.

[0030] The present invention uses the binding chain relationship between value attributes, input constraint parameters and output constraint parameters. During the calculation process, starting from a non-computational node, the calculation path is determined through the input constraint parameters and output constraint parameters of the calculation node. The complex relationship context of the SysML parameter diagram model is sorted out and it is executed iteratively in an orderly manner. The logic of traversing and evaluating the dependency relationships between calculation nodes before calculation is eliminated, and the calculation process is simplified.

[0031] The present invention supports expressions written in Python, Matlab, and JavaScript, and supports third-party languages to evaluate expressions in models. Compared with existing technologies, it has higher flexibility and applicability.

[0032] The method provided by the present invention supports solving constraints according to the expressions in the constraints during simulation initialization and when related variables are changed, added, or deleted. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0034] Figure 1 It is a flow chart of the MBSE-based parameter model calculation method provided in the embodiment. DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not limit the scope of protection of the present invention.

[0036] The embodiment provides a parameter model calculation method based on MBSE, which is implemented by parameter engine calculation. The parameter engine calculation mainly solves how to solve the constraint expression associated with the constraint attribute in the SysML parameter diagram model and associate the solution result with the corresponding value attribute.

[0037] In an embodiment, parameter engine calculation is designed to satisfy the constraint relationships between model elements expressed in SysML. The main application scenario of parameter engine calculation at the model level is the SysML parameter diagram model. However, it can also be used to solve constraints on any UML class. Using parameter engine calculation, mathematical expressions or logical expressions can be defined as constraints on blocks to limit their attribute values. If the expression is an equation, parameter engine calculation will solve the constraint expression and use the result of the evaluation to update the attribute value. If the expression is a logical expression, parameter engine calculation will use the expression to verify the attribute value. The parameter engine is able to solve the parameter constraint relationships created in the parameter diagram model in SysML.

[0038] In the embodiment, the parameter engine calculation will perform constraint solving based on the expression in the constraint when the simulation is initialized and the relevant variables are changed, added, or deleted. The parameter engine calculation only supports expressions written in the syntax it supports. By default, the parameter engine calculation uses the JavaScript parser to solve the expression. It is also possible to write expressions in specially supported scripting languages. Currently, the supported languages are Python, Matlab, and JavaScript. The parameter engine calculation allows the use of third-party languages to evaluate expressions in the SysML parametric diagram model. The expression is understood as an opaque expression and corresponds to a constraint expression under the constraint attribute. Therefore, any language supported by the external calculator can be used in the body of the expression.

[0039] like Figure 1 As shown, the parameter model calculation method based on MBSE performed by the parameter engine includes the following steps:

[0040] Step 1: Obtain the SysML parametric diagram model built in MBSE.

[0041] The SysML parametric diagram model is a model designed based on the parametric diagram coded in the SysML language in MBSE. During use, users only need to drag module properties into the SysML parametric diagram and connect business-related properties through connectors to obtain the SysML parametric diagram model. The SysML parametric diagram model contains various value attributes and their corresponding attribute models, constraint attributes and their corresponding constraint models.

[0042] Step 2: Parse the SysML parameter diagram model into value attributes and constraint attributes, convert the attribute model corresponding to the value attribute into a non-computational node, and convert the constraint model corresponding to the constraint attribute into a computational node.

[0043] In this embodiment, parsing the SysML parametric diagram model forms the basis for subsequent SysML parametric diagram solution calculations. The core of this parsing process is to convert the defined SysML parametric diagram model into simulation model data. Based on this parsing core, the SysML parametric diagram model is parsed to obtain value attributes and their corresponding attribute models, constraint attributes and their corresponding constraint models, and component attributes. The associations between these attributes are also obtained, and connectors are used to determine the connectivity and nested relationships between attributes.

[0044] In this embodiment, after parsing, non-computational nodes and computational nodes are created. The non-computational nodes correspond to the attribute model corresponding to the value attribute. Specifically, the non-computational nodes consist of the attribute model and the attribute default value, where the attribute default value is obtained from the default value attribute of the attribute model. During subsequent iterative solving processes, the attribute value of the value attribute is further updated based on the binding chain assignment.

[0045] Compute nodes correspond to constraint models corresponding to constraint attributes. Specifically, a compute node consists of an input module, an output module, and an expression parser. The input module corresponds to the input constraint parameters of the expression, the output module corresponds to the output constraint parameters of the expression, and the expression parser corresponds to the constraints of the constraint model. After parsing is complete, the SysML parametric diagram model is converted into a network topology diagram consisting of non-compute nodes and compute nodes.

[0046] Step 3: Build a binding chain between nodes.

[0047] In this embodiment, the association relationship between attributes is established by constructing binding chains between nodes. Since the nodes include non-computational nodes corresponding to value attributes and computational nodes corresponding to constraint attributes, the binding chains constructed between the nodes include binding chains between value attributes, binding chains between value attributes and constraint parameters, and binding chains between constraint parameters, where the constraint parameters include input constraint parameters and output constraint parameters.

[0048] In the embodiment, the values between the two attributes connected by the connector should remain the same and consistent. Any modification of the value at one end will be passed to the other end through the connector. Based on this, the binding chain determines whether the two ends can be connected through the value attribute or the type attribute of the constraint parameter. If the types corresponding to their type attributes are consistent, it means that the values between them can be copied to each other. When constructing a binding chain, first, determine the type attributes of the two end models corresponding to the two end nodes. When the type attributes of the two end models are the same, or the type attribute of the model at one end is a subclass of the type attribute of the model at the other end, a binding chain is constructed between the two end nodes.

[0049] In the embodiment, when constructing a binding chain, a method for transferring attribute values through the binding chain is also determined, including: (1) transferring according to the definition order of the two ends of the binding chain. If the model at one end of the binding chain is created before the model at the other end and has a default attribute value, then during initialization, the attribute value is copied from the end of the model created earlier to the end of the model created later; (2) transferring according to the change of the attribute value. If the attribute value at one end of the binding chain changes, the attribute value is transferred from the end where the attribute value changes to the other end where the attribute value does not change. Specifically, when the attribute value is transferred through the binding chain, the attribute value is copied according to the multiplicity of the attribute or constraint parameter. The receiving end of the binding chain will update the attribute value according to the definition of the multiplicity of its own model. The multiplicity is understood as the number of values of a single attribute. If the multiplicity of the receiving end is less than that of the sending end, the receiving end can only receive the same number of attribute values; if the multiplicity of the receiving end is greater than that of the sending end, the receiving end only updates the attribute values of the number of the sending end's multiplicity, and the remaining attribute values remain unchanged.

[0050] Step 4: Solve the value attributes and constraint attributes through the binding chain.

[0051] Based on the binding chain built between nodes, the value attributes and constraint attributes are solved through the binding chain. Specifically, the output constraint parameters are solved based on the input constraint parameters through the calculation node. This process is the core function of the parameter engine calculation. The input constraint parameters are used as the input parameters of the expression, and then the output constraint parameters are calculated according to the language type specified by the expression. The values of the input constraint parameters are assigned through the binding chain.

[0052] In the embodiment, the expression is generally an opaque expression, which is divided into two types: (1) assignment expression, and (2) logical expression. Among them, the assignment expression assigns the result to other value attributes through the output constraint parameter. Through the binding chain, the value of the output constraint parameter will be passed to the corresponding attribute, and then the non-computational node corresponding to the attribute with the new value will participate in the parameter engine calculation as the starting point of the next round. The result of the logical expression will determine whether the constraint attribute meets the constraint condition. If the calculation result is correct (true), it means that this constraint attribute and other adjacent attributes meet the constraint condition, otherwise the constraint condition is not met.

[0053] In the embodiment, when the parameter engine calculates and solves the value attribute and the constraint attribute, the following steps are included:

[0054] (a) Taking a non-computational node assigned a default attribute value as the starting point, the attribute value of the starting point is assigned to the computational nodes that have a binding relationship with the starting point through the binding chain.

[0055] (b) The computing node obtains the constraint expression through the constraint attribute, and obtains the input constraint parameter and output constraint parameter through the constraint expression. Then, through the binding chain, the value of the value attribute of the non-computing node is assigned to the corresponding input constraint parameter.

[0056] In this embodiment, the computation node obtains the constraint expression from the constraint attribute, and from the constraint expression, obtains the input constraint parameters and output constraint parameters. By binding the value attribute to the input constraint parameters, the input constraint parameters are assigned a value before the parameter engine performs calculations. This operation copies the attribute value from the associated value attribute to the input constraint parameter, and the value of the input constraint parameter is used as the input parameter for the expression calculation. This completes the preparations for calculating the constraint expression.

[0057] (c) The constraint expression is executed using an expression parser. After the execution is completed, the value of the output constraint parameter is obtained from the expression parser by outputting the constraint parameter name, thereby completing the assignment of the output constraint parameter.

[0058] In this embodiment, after obtaining a constraint expression, the constraint expression content and expression language type are obtained according to the definition of the constraint specification. Then, an appropriate expression parser is selected based on the expression language model to solve the constraint expression. When solving the constraint expression, constraint expressions written in three languages ​​(Python, Matlab, and JavaScript) are supported. That is, expression parsers supporting JavaScript, Python, and Matlab are used for solution. After execution is completed, the output constraint parameter name is output and the value of the output constraint parameter is obtained from the expression parser, completing the assignment of the output constraint parameter.

[0059] (d) Another node associated with the computation node through a binding chain copies the attribute value from the output constraint parameter and continues the next solution. The other node is another non-computation node or another computation node.

[0060] In the embodiment, since the binding chain establishes a connection relationship between the computing node and the non-computing node, when the other node is another non-computing node, the value attribute of the other non-computing node copies the attribute value from the output constraint parameter and serves as the starting point of a new round for the next round of solution calculation, and iterates until there is no available starting point.

[0061] In an embodiment, the binding chain also establishes a connection relationship between the computing node and the non-computing node. When the other node is another computing node, the other computing node obtains another constraint expression through the constraint attribute, and obtains another input constraint parameter and another output constraint parameter through the other constraint expression. Then, through the binding chain, the value of the output constraint parameter of the computing node is assigned to the corresponding other input constraint parameter; then the expression parser is used to execute the other constraint expression. After the execution is completed, the value of the other output constraint parameter is obtained from the expression parser by outputting the name of the other constraint parameter, thereby completing the assignment of the other output constraint parameter.

[0062] In this embodiment, by binding output constraint parameters to attributes, the parameter engine will assign a value to the output constraint parameter after calculation. After the assignment is completed, the value attribute associated with the output constraint parameter will copy the value from the output constraint parameter. The value attribute is then assigned to other associated value attributes through the binding chain.

[0063] The method provided in the embodiment does not need to pay attention to the dependencies between constraint attributes, and during calculation, there is no need to consider the computational dependencies between constraint attributes. Based on the autonomous iteration method between non-computational nodes and computational nodes, after the initialization and parsing of the SysML parameter diagram model is completed, the non-computational nodes assigned default values are first identified, and these non-computational nodes are used as starting points. The values of the starting points are then assigned to the input constraint parameters with a binding chain, so that the corresponding computational nodes can be triggered for calculation, and the results of the calculations are then passed to other non-computational nodes or computational nodes through the output constraint parameters, and the cycle repeats until there are no available starting points. This method gets rid of the work of sorting out the computational dependencies between constraint attributes, and simplifies the complexity and maintenance cost of the parameter calculation engine.

[0064] The method provided in the embodiment utilizes attribute and connector models to generate corresponding non-computational nodes (value attributes) and computational nodes (constraint attributes), and the nodes are connected through binding chains. The binding chain consists of two binding chain ends, which are abstractions of attributes and constraint parameters, including the values above the attributes and constraint parameters. This method transfers attribute values between non-computational nodes and computational nodes. Attribute values will be passed to other attributes through the binding chain. When the constraint attributes are calculated, the attribute values can be passed to the input constraint parameters or the output constraint parameter values can be passed to the attributes. The value changes can be quickly diffused to the entire network through the value binding chain.

[0065] The specific implementation methods described above provide a detailed description of the technical solutions and beneficial effects of the present invention. It should be understood that the above is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, supplements and equivalent substitutions made within the scope of the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A parameter model calculation method based on MBSE, characterized in that: The following steps are involved: Get the SysML parametric diagram model built in MBSE; Parse the SysML parameter diagram model into value attributes and constraint attributes, convert the attribute model corresponding to the value attribute into a non-calculated node, and convert the constraint model corresponding to the constraint attribute into a calculated node; Build binding chains between nodes and solve value attributes and constraint attributes through binding chains, including: Taking the non-computational node assigned the default attribute value as the starting point, the attribute value of the starting point is assigned to the computational node that has a binding relationship with the starting point through the binding chain; The compute node obtains the constraint expression through the constraint attribute, and obtains the input constraint parameter and output constraint parameter through the constraint expression. Then, through the binding chain, the value of the value attribute of the non-compute node is assigned to the corresponding input constraint parameter. The constraint expression is then executed using the expression parser. After execution, the output constraint parameter value is obtained from the expression parser through the output constraint parameter name, completing the assignment of the output constraint parameter. Another node associated with the compute node through the binding chain copies the attribute value from the output constraint parameter and continues the next solution. This other node can be another non-compute node or another compute node. When the other node is a non-computational node, the value attribute of the other non-computational node is copied from the output constraint parameter and used as the starting point for the next round of solution calculation. This iteration continues until there are no available starting points. When the other node is another computing node, the other computing node obtains another constraint expression through the constraint attribute, and obtains another input constraint parameter and another output constraint parameter through the other constraint expression, and then assigns the value of the output constraint parameter of the computing node to the corresponding other input constraint parameter through the binding chain; then the expression parser is used to execute the other constraint expression. After the execution is completed, the value of the other output constraint parameter is obtained from the expression parser by outputting the name of the other constraint parameter, thereby completing the assignment of the other output constraint parameter.

2. The general parameter model calculation method based on MBSE according to claim 1, characterized in that: The non-computational node consists of an attribute model and an attribute default value, wherein the attribute default value is obtained from the default value attribute of the attribute model; The computing node is composed of an input module, an output module and an expression parser, wherein the input module corresponds to the input constraint parameters of the expression, the output module corresponds to the output constraint parameters of the expression, and the expression parser corresponds to the constraint attributes of the constraint model.

3. The general parameter model calculation method based on MBSE according to claim 2, characterized in that: The binding chains constructed between nodes include binding chains between value attributes, binding chains between value attributes and constraint parameters, and binding chains between constraint parameters. The constraint parameters include input constraint parameters and output constraint parameters.

4. The general parameter model calculation method based on MBSE according to claim 1 or 2, characterized in that: When building a binding chain, first determine the type attributes of the models at both ends corresponding to the two end nodes. When the type attributes of the models at both ends are the same, or the type attribute of the model at one end is a subclass of the type attribute of the model at the other end, build a binding chain between the two end nodes.

5. The general parameter model calculation method based on MBSE according to claim 1, characterized in that: When building a binding chain, the method for passing attribute values through the binding chain is also determined, including: (1) Transfer according to the definition order of both ends of the binding chain. If the model at one end of the binding chain is created before the model at the other end and has default attribute values, the attribute values are copied from the end of the model created earlier to the end of the model created later during initialization. (2) Transfer according to the change of attribute value. If the attribute value at one end of the binding chain changes, the attribute value will be transferred from the end where the attribute value changes to the other end where the attribute value remains unchanged.

6. The general parameter model calculation method based on MBSE according to claim 4, characterized in that: When building a binding chain, the method for passing attribute values through the binding chain is also determined, including: (1) Transfer according to the definition order of both ends of the binding chain. If the model at one end of the binding chain is created before the model at the other end and has default attribute values, the attribute values are copied from the end of the model created earlier to the end of the model created later during initialization. (2) Transfer according to the change of attribute value. If the attribute value at one end of the binding chain changes, the attribute value will be transferred from the end where the attribute value changes to the other end where the attribute value remains unchanged.

7. The general parameter model calculation method based on MBSE according to claim 5 or 6, characterized in that: When attribute values are passed through a binding chain, they are copied according to the multiplicity of the attribute or constraint parameter. The receiving end of the binding chain updates the attribute value according to the definition of its own model multiplicity. If the multiplicity of the receiving end is less than that of the sending end, the receiving end can only receive the same number of attribute values. If the multiplicity of the receiving end is greater than that of the sending end, the receiving end only updates the attribute value of the number of multiplicities of the sending end, and the attribute value remains unchanged.

8. The general parameter model calculation method based on MBSE according to claim 1, characterized in that: After obtaining the constraint expression, the constraint expression content and expression language type are obtained according to the definition of the constraint specification, and then the corresponding expression parsing module is selected according to the expression language model to solve the constraint expression.

9. The general parameter model calculation method based on MBSE according to claim 8, characterized in that: When solving constraint expressions, it supports constraint expressions written in three languages: Python, Matlab, and JavaScript.

Images