An intelligent design and simulation system for engineering technology research and development

Abstract

This invention relates to the field of intelligent design and simulation technology, specifically to an intelligent design and simulation system for engineering technology research and development. It includes an evolution trend extraction module, a dependency structure construction module, an instruction offset identification module, a working condition segment location module, and a matching relationship screening module. The system extracts the magnitude, rate, and span of changes from the design records, groups them to generate trend labels, constructs continuous paths, identifies behavioral offset segments, locates high-density disturbance segments, filters matching paths with consistent directions, and outputs simulation results. This invention, by introducing data processing of direction changes, rate fluctuations, and boundary spans, completes behavioral trend grouping and path construction. Combined with the angle matching between the target direction and the path direction, it identifies structural offset segments and high-density disturbance segments, extracts matching paths with continuous directions and consistent responses, and improves the mapping accuracy and coherence between design and simulation.

Description

Technical Field

[0001] This invention relates to the field of intelligent design and simulation technology, and in particular to an intelligent design and simulation system for engineering technology research and development. Background Technology

[0002] The field of intelligent design and simulation technology in engineering technology research and development encompasses several key areas, including computer-aided design-based modeling technology, parametric modeling and visualization analysis technology, and integrated multidisciplinary simulation analysis technology. The core of this technology lies in achieving virtual reconstruction and predictive evaluation of the engineering product design process through intelligent analysis of engineering design parameters, geometric structures, and functional requirements, combined with multiple dimensions such as physical simulation, structural analysis, and thermal calculation. Overall, the field of intelligent design and simulation technology is model-driven, using computational reasoning and rule systems to provide end-to-end assistance from design scheme generation to performance evaluation. It is widely used in engineering practices such as mechanical manufacturing, electronic engineering, aerospace, and civil engineering, and continues to develop and integrate with fields such as artificial intelligence, digital twins, and knowledge engineering.

[0003] One of the patented technologies, an intelligent design and simulation system for engineering technology research and development, refers to a system solution that integrates design logic construction and engineering environment parameter simulation to address the need for collaborative advancement of design modeling and performance simulation during the development of complex engineering products. The technical aspects covered by this patent include: firstly, generating product geometric configurations through semantic rule parsing and engineering constraint logic calculation for multi-dimensional engineering parameter design inputs; secondly, importing material properties, load types, and process conditions such as working boundary conditions based on modeling to establish a simulation solution framework based on finite element analysis; and thirdly, constructing an iterative calculation relationship between design results and simulation feedback to form a basis for design adjustments. This system typically integrates and deploys modeling and simulation tasks through embedded inference rule sets, engineering database calling mechanisms, and visual operation interfaces.

[0004] Existing technologies, based on static modeling and rule-driven approaches, lack a mechanism for extracting the changing trends of phased design behavior. In scenarios where design behavior evolves continuously or targets are adjusted, they cannot dynamically identify the coupling relationship between offset paths and response periods. This leads to problems such as ambiguous feedback positioning and difficulty in interpreting abnormal responses in complex simulations. For example, in the process of multi-stage design adjustments, it is difficult to track the coherent changes in structural behavior and simulation response, affecting the efficiency of design correction and the accuracy of performance verification. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing an intelligent design and simulation system for engineering technology research and development.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent design and simulation system for engineering technology research and development, the system comprising:

[0007] The evolution trend extraction module acquires stage data in the engineering technology research and development process, extracts the directional change amplitude, change rate and boundary span of continuous stages, judges the consistency of direction and the degree of fluctuation, groups them into trend segment labels, and generates a stage behavior change combination table.

[0008] Based on the trend label combination in the stage behavior change combination table, the module determines the directional consistency and fluctuation frequency segment overlap of adjacent label groups, establishes connectable channels, filters out the span of channel discontinuities and removes short jump segments, and generates a structural behavior progression path map.

[0009] The instruction offset recognition module extracts the design target direction based on the distribution of path segments in the structural behavior progression path map, determines the degree of angular offset between the path and the target direction, filters path segments with low overlap rate, and generates a set of functional behavior offset segments.

[0010] The working condition segment positioning module extracts the start and end time point data from the functional behavior offset segment set, performs density judgment on the load jump points, boundary switching points and direction reversal points in the simulation sequence during that period, filters the continuous jump areas and groups them according to the number of triggers, and generates a working condition disturbance response segment table.

[0011] As a further embodiment of the present invention, the stage behavior change combination table includes direction change angle labels, rate change level labels, and boundary span interval labels; the structural behavior progression path map includes trend channel links, fluctuation persistence statistics, and path segment sequence identifiers; the functional behavior offset segment set includes direction matching deviation segments, fluctuation distribution abnormal segments, and target behavior misalignment segments; and the working condition disturbance response segment table includes high-frequency load triggering segments, boundary turning point concentration segments, and response reversal dense segments.

[0012] As a further aspect of the present invention, the evolution trend extraction module includes a data change quantification submodule, a trend direction judgment submodule, and a fluctuation intensity classification submodule;

[0013] The data change quantification submodule acquires stage design record data in the intelligent design and simulation process of engineering technology research and development, detects various design record parameters in adjacent time periods, calculates the numerical differences of design parameters between adjacent time periods, calculates the change direction, change amplitude and change rate, monitors the range of design boundary parameters, quantifies the boundary span value, and establishes a parameter change feature set.

[0014] The trend direction judgment submodule calls the direction change amplitude value in the parameter change feature set, selects the direction change data of three adjacent time periods, calculates the angle between the three change directions, judges the degree of deviation between the angle and the benchmark value, divides the directional category according to the deviation range, performs directional classification operation on each group of three-segment trend, and generates directional trend identifier.

[0015] The fluctuation intensity classification submodule calculates the mean and standard deviation of the change rate for each time period based on the change rate value and boundary span value in the parameter change feature set, compares the deviation between the single time period value and the mean, determines the fluctuation intensity level, performs horizontal intensity comparison in combination with directional trend indicators, groups and generates trend label combinations, and establishes a stage behavior change combination table.

[0016] As a further embodiment of the present invention, the dependency structure construction module includes a trend continuity judgment submodule, a channel map construction submodule, and a disturbance filtering submodule;

[0017] The trend continuity judgment submodule detects the angle of change of fluctuation direction between adjacent trend label groups according to the trend label group affiliation relationship in the stage behavior change combination table, calculates the angle difference and compares it with the continuity judgment threshold, marks the start and end positions of the overlapping change frequency segment, judges the trend continuity status between label groups, performs a classification operation on the trend continuity label groups, and obtains a continuous label set.

[0018] The channel map construction submodule calls the data of the continuous label group in the continuous label set, establishes the coordinates of the connecting nodes between the label groups, constructs the connecting paths between the nodes, draws the spatial distribution of the connecting channels, determines the duration of the fluctuation at the channel connection, calculates the difference between the interval duration value and the set benchmark value, and generates the channel connection map.

[0019] The disturbance filtering submodule, based on the fluctuation duration interval values ​​in the channel connection graph, filters segments whose interval duration exceeds the short-period disturbance threshold, removes path nodes with intervals below the threshold, retains connection channels that meet the duration requirements, reconstructs the channel connection relationships after filtering, and establishes a structural behavior progressive path graph.

[0020] As a further aspect of the present invention, the instruction offset recognition module includes a target direction extraction submodule, a matching degree filtering submodule, and an offset segment recognition submodule;

[0021] The target direction extraction submodule obtains the target behavior direction parameters designed in the task book based on the path segment direction data in the structural behavior progressive path map, detects the target direction angle value and direction vector coordinates, calculates the vector angle between the path segment direction angle and the target direction angle, determines the angle deviation value between each path segment and the target direction, and establishes a set of direction matching degree values.

[0022] The matching degree filtering submodule calls the matching degree values ​​of each path segment in the direction matching degree value set, filters the path segments whose matching degree values ​​exceed the set matching threshold, counts the number of fluctuating segments in the path, calculates the overlap rate between the number of fluctuating segments and the total number of path segments, judges the overlap rate and compares it with the overlap judgment benchmark value, and obtains the overlap rate judgment identifier.

[0023] The offset segment identification submodule, based on the overlap rate judgment identifier and matching degree filtering results, selects path segments with an overlap rate lower than the benchmark value and a matching degree that does not meet the requirements, marks the start and end positions of the offset segments, extracts the behavioral feature data of the offset segments, classifies the type attributes of the offset segments, and generates a set of functional behavioral offset segments.

[0024] As a further embodiment of the present invention, the working condition segment positioning module includes a trigger point positioning submodule, a density filtering submodule, and a statistical segmentation submodule.

[0025] The trigger point positioning submodule extracts the start and end time point data from the set of functional behavior offset segments, locates the load trigger point time in the corresponding time period in the simulation working condition sequence, detects the location of the boundary turning point, monitors the change time of the response reversal point, calculates the time interval of the three types of trigger points on the time line, records the time coordinates and type identifiers of each trigger point, and establishes the trigger point time series.

[0026] The density filtering submodule calls the time interval data of each trigger point in the trigger point time series, judges the comparison result of the interval duration of adjacent trigger points with the density filtering threshold, filters the trigger point group with an interval duration lower than the threshold, marks the start and end boundary positions of the trigger point group, calculates the number density value of trigger points in the group, and obtains the trigger point density distribution.

[0027] The statistical segmentation submodule, based on the trigger point group data in the trigger point density distribution, counts the number of load triggers, boundary turning times, and response reversals in each density segment, merges the counts of the same type of triggers, arranges the density segment sequence in chronological order, constructs the connection relationship between segments, and generates a table of operating condition disturbance response segments.

[0028] As a further aspect of the present invention, the system further includes:

[0029] The matching relationship screening module extracts the design target direction vector based on the directional distribution results of each trigger sequence in the working condition disturbance response segment table, performs a judgment on the coincidence of the start and end times of the response direction and the consistency comparison of the direction angle, screens out continuous directional consistent segments, and generates engineering structure simulation results.

[0030] The simulation results of the engineering structure include the directional continuous matching path, the target direction coverage segment, and the mapping relationship of the start and end segments.

[0031] As a further embodiment of the present invention, the matching relationship screening module includes a direction vector extraction submodule, a segment correspondence judgment submodule, and a path matching selection submodule;

[0032] The direction vector extraction submodule extracts the corresponding direction vector coordinates from the design behavior target set based on the direction distribution results of each trigger sequence in the working condition disturbance response segment table, calculates the response direction angle value and the target direction angle value, detects the start and end time nodes of overlapping path segments, measures the direction change amplitude value of the path segment, and establishes a direction matching dataset.

[0033] The segment correspondence judgment submodule calls the path segment time node data in the direction matching dataset to determine the time overlap range between the start and end segments of the response direction and the target direction segment, calculates the continuous change rate of the direction angle between adjacent path segments, compares the difference between the continuous change rate and the continuity judgment threshold, filters the path segments that meet the continuity requirements, and obtains the continuity evaluation result.

[0034] The path matching and selection submodule, based on the continuity evaluation results and directional consistency comparison data, selects path segments that satisfy both continuity and directional consistency conditions, marks matching path segment identifiers, constructs time series relationships of matching paths, summarizes path segment simulation parameters and response characteristics, and generates engineering structure simulation results.

[0035] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0036] In this invention, a data processing method for directional change amplitude, rate fluctuation, and boundary span is introduced into the design record to achieve structural grouping of stage behavior trends. By using the fluctuation continuity and frequency overlap characteristics between trends, a path link with correlation strength is constructed. By combining the target behavior direction vector and the path segment direction for angle matching and segment overlap judgment, structural offset segments can be identified from the deviation path, and high-density disturbance response segments can be further extracted to form a matching path with continuous direction and consistent response, effectively enhancing the structural mapping accuracy and process coherence between the design target and the simulation response. Attached Figure Description

[0037] Figure 1 This is a system flowchart of the present invention;

[0038] Figure 2 This is a flowchart illustrating the acquisition process of the evolution trend extraction module in this invention.

[0039] Figure 3 This is a flowchart illustrating the process of obtaining the dependent structure construction module of this invention;

[0040] Figure 4 This is a flowchart illustrating the acquisition process of the instruction offset recognition module of the present invention.

[0041] Figure 5 This is a flowchart illustrating the acquisition process of the working condition segment positioning module of the present invention.

[0042] Figure 6 This is a flowchart illustrating the acquisition process of the matching relationship screening module of the present invention. Detailed Implementation

[0043] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0044] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0045] In the embodiments of this invention, 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.

[0046] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0047] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0048] Please see Figure 1 This invention provides a technical solution: an intelligent design and simulation system for engineering technology research and development, the system comprising:

[0049] The evolution trend extraction module acquires stage design record data in the intelligent design and simulation process of engineering technology research and development, extracts the directional change amplitude, change rate and boundary span value of each design record in adjacent time periods, performs directional change value judgment based on the angular direction of the trend of three adjacent segments, performs lateral fluctuation intensity comparison and grouping of change rate and boundary span to generate trend label group, and generates stage behavior change combination table.

[0050] Based on the structure building module, according to the affiliation of trend label groups in the stage behavior change combination table, the continuous fluctuation trend between adjacent trend label groups is judged and the overlapping sections of change frequency are marked. For label groups whose trends have been judged to be continuous, a connection channel map is constructed, and the fluctuation duration interval value is statistically analyzed at the channel connection to eliminate short-cycle disturbances, generating a structural behavior progression path map.

[0051] The instruction offset recognition module extracts the set of design target behavior directions in the task book based on the direction data of each path segment in the structural behavior progression path map. It performs direction matching degree filtering based on the angle vector between the path direction and the target direction. It combines the matching degree filtering results with the number of distribution segments of the fluctuation segment in the path to perform overlap rate judgment and extract the deviation segment, generating a set of functional behavior offset segments.

[0052] The working condition segment positioning module extracts the start and end time point data from the functional behavior offset segment set, locates the load trigger point, boundary turning point and response reversal point in the corresponding time period in the simulation working condition sequence, performs jump interval judgment on the three types of trigger points on the time line and filters out high-density segments, merges and counts the trigger times of these segments and forms a segmented sequence, and generates a working condition disturbance response segment table.

[0053] The matching relationship screening module extracts the corresponding direction vector from the target set of design behavior based on the directional distribution results of each trigger sequence in the working condition disturbance response segment table. It performs start and end segment judgment and direction angle continuity comparison on the overlapping path segments between the response direction and the target direction, selects the path segments that satisfy both continuity and direction consistency as the matching path, and generates the simulation results of the engineering structure.

[0054] The stage behavior change combination table includes direction change angle labels, rate change level labels, and boundary span interval labels. The structural behavior progression path map includes trend channel links, fluctuation persistence statistics, and path segment sequence identifiers. The functional behavior offset segment set includes direction matching deviation segments, fluctuation distribution abnormal segments, and target behavior misalignment segments. The working condition disturbance response segment table includes high-frequency load triggering segments, boundary turning point concentrated segments, and response reversal dense segments. The engineering structure simulation results include direction continuous matching paths, target direction coverage segments, and start and end segment mapping relationships.

[0055] Please see Figure 2 The evolution trend extraction module includes a data change quantification submodule, a trend direction judgment submodule, and a fluctuation intensity classification submodule.

[0056] The data change quantification submodule acquires stage design record data in the intelligent design and simulation process of engineering technology research and development, detects various design record parameters in adjacent time periods, calculates the numerical differences of design parameters between adjacent time periods, calculates the change direction, change amplitude and change rate, monitors the range of design boundary parameters, quantifies the boundary span value, and establishes a parameter change feature set.

[0057] The acquired engineering technology R&D intelligent design and simulation process stage design record data is a design iteration log of an R&D project.

[0058] Table 1: Example of Design Iteration Log Data

[0059]

[0060] As shown in Table 1, the log includes the design iteration number, recording time, key design parameter P1 (e.g., main operating temperature), performance response index M1 (e.g., energy conversion efficiency), and constraint index C1 (e.g., structural integrity factor). First, the design recording parameters in adjacent time periods are examined. Taking design iteration 15 and design iteration 16 as an example, the value of parameter P1 in iteration 15 is... In iteration 16, this value is The difference in design parameter values ​​between adjacent time periods is calculated to obtain the magnitude of directional change. And calculate the rate of change value, that is, the amplitude value and the time interval ( The ratio of ) is Monitor the design boundary parameter range; the task specification requires that the binding indicator C1 must be maintained within... Within this range, this is the design boundary. If the value of C1 in iteration 15 is... In iteration 16, for Then its boundary utilization rate is given by the formula The utilization rate of iteration 15 is calculated. The utilization rate of iteration 16 is The quantized boundary span value is the absolute difference between the two utilization rates. The direction change magnitude of parameter P1 ( ), rate of change ( The boundary span value of index C1 () Integrate and establish a set of parameter change features.

[0061] The trend direction judgment submodule calls the direction change amplitude value in the parameter change feature set, selects the direction change data of three adjacent time periods, calculates the angle between the three change directions, judges the degree of deviation between the angle and the benchmark value, classifies the directional category according to the deviation range, performs directional classification operation on each group of three trend directions, and generates directional trend labels.

[0062] The direction change amplitude value of the key design parameter P1 in the parameter change feature set is called, and three adjacent change time periods are selected. Let their amplitude value sequence be... , , To calculate the angle between the three changing directions, and The change process can be viewed as a two-dimensional vector, where the time component is 1 unit and the numerical change component, i.e., the amplitude value, constitutes the vector. and Calculate the angle between two vectors. First, calculate its dot product. Then calculate their respective modulus. and Then the cosine of the included angle is Find the included angle The degree of deviation between the included angle and the reference value is used to determine the directional judgment reference value. The setting was achieved by statistically analyzing the angles of 500 continuously changing vectors in 100 similar completed R&D projects that were in a stable optimization phase, and obtaining their mean value. The standard deviation is ,Pick As a benchmark, directional categories are classified according to the range of deviation, where the angle is... The interval is defined as "consistent evolution". The range is for "corrective adjustments", while... The interval is considered a "reverse transition," due to the calculated... Within the first interval, a directional classification operation is performed on the trend of this group, classifying it as "consistent evolution" and generating a directional trend identifier.

[0063] The volatility intensity classification submodule calculates the mean and standard deviation of the rate of change for each time period based on the rate of change and boundary span values ​​in the parameter change feature set, compares the deviation between the single time period value and the mean, determines the volatility intensity level, performs horizontal intensity comparison in combination with directional trend indicators, groups and generates trend label combinations, and establishes a stage behavior change combination table.

[0064] Based on the rate of change and boundary span values ​​of all 200 design iterations in the parameter change feature set, firstly, for the rate of change values ​​(e.g., parameter P1)... Quantization is performed, and the arithmetic mean and standard deviation of these 200 rate values ​​are calculated to obtain the mean. Standard deviation Compare the rate values ​​of a single iteration. with the mean The degree of deviation is calculated using standard scores. Then, the fluctuation intensity level is determined, and the level is classified based on the absolute value of the standard score: It was determined to be "low-level fluctuation"; It was determined to be "moderate fluctuation"; It was determined to be "highly volatile" because The rate fluctuations during this period are classified as "low-degree fluctuations," and the boundary span values ​​(e.g., index C1) are defined as follows. Perform the same process, assuming the average boundary span of 200 iterations is... Standard deviation Then its standard score Similarly, it belongs to "low-degree fluctuation". Combined with the directional trend indicators (consistent evolution) generated in the previous module, a horizontal intensity comparison is performed. The directional indicators and intensity levels of the same iteration cycle are combined and grouped to generate trend label combinations. Finally, the labels of all iteration cycles are collected to establish a stage behavior change combination table.

[0065] Table 2: Examples of Combinations of Stage-Specific Behavioral Changes

[0066]

[0067] Table 2 provides an example structure for the stage behavior change combination table, which integrates the dynamic behavior characteristics of key parameters and indicators in each iteration.

[0068] Please see Figure 3 The dependency structure building module includes a trend continuity judgment submodule, a channel map construction submodule, and a disturbance filtering submodule;

[0069] The trend continuity judgment submodule detects the angle of change of fluctuation direction between adjacent trend label groups based on the trend label group affiliation relationship in the stage behavior change combination table, calculates the angle difference and compares it with the continuity judgment threshold, marks the start and end positions of the overlapping change frequency segment, judges the trend continuity status between label groups, performs a classification operation on the trend continuity label groups, and obtains the continuity label set.

[0070] Based on the trend label group affiliation relationship of each iteration cycle in the stage behavior change combination table, detect the fluctuation direction change angle between two adjacent label groups, such as Tag-16 of iteration 16 and Tag-17 of iteration 17. This angle is the included angle calculated in paragraph 2. If the included angle of Tag-16 is... The internal included angle of Tag-17 was calculated to be The difference in the direction angle is then calculated as follows: The difference is compared with the continuity determination threshold. The threshold is compared. By performing K-Means clustering (K=2) analysis on historical R&D data, two paths were distinguished: "continuous optimization" and "oscillating exploration." The average angle difference between the cluster centers of the two types was [missing information]. Therefore, it is set Due to the calculated value Less than The direction is determined to be continuous, and the start and end positions of the overlapping segments of the frequency change are marked. It is checked whether both Tag-16 and Tag-17 contain the analysis of the key parameter P1. If both contain it, the overlap is marked, and the direction continuity between the tag groups is determined to be "continuous". The classification operation is performed on all adjacent tag groups with continuous direction to obtain the continuous tag set.

[0071] The channel map construction submodule calls the data of the continuous label group from the continuous label set, establishes the coordinates of the connecting nodes between the label groups, constructs the connecting paths between the nodes, draws the spatial distribution of the connecting channels, determines the duration of the fluctuation interval at the channel connection, calculates the difference between the interval duration value and the set benchmark value, and generates the channel connection map.

[0072] Retrieve tag group data from the continuous tag set that has been determined to be continuous, such as the set containing consecutive Tags-16, Tags-17, and Tags-18. Establish the coordinates of the connecting nodes between these tag groups, treating each tag group as a node in the graph, and construct directed connection paths between continuous nodes, forming a structure like... The channels are plotted, and their spatial distribution is connected. Then, the duration of fluctuation intervals at the channel connections is measured. This is defined as the sum of the time covered by each tag group constituting the connection. If each iteration cycle is 5 hours, then the duration of the three-node channel is... Calculate the interval duration (15 hours) and the set baseline value. The difference, the benchmark value The settings were based on the average shortest time to complete one "design-simulation verification-data analysis" feedback loop in the R&D process. Statistical analysis of the time consumed in 30 such loops showed a 10th percentile time of 8.5 hours. The time frame was set to 8.5 hours, and the calculated difference was: The system integrates node, connection, and duration data to generate a channel connection graph.

[0073] The perturbation filtering submodule filters segments whose interval duration exceeds the short-period perturbation threshold based on the fluctuation duration interval value in the channel connection graph, removes path nodes with intervals below the threshold, retains connection channels that meet the continuity requirements, reconstructs the channel connection relationship after filtering, and establishes a structural behavior progressive path graph.

[0074] Based on the fluctuation duration interval values ​​of each connected channel in the channel connectivity map, for example, the duration of channel C1 is 15.0 hours, and another exploratory channel C2 has a duration of 6.0 hours, the screening interval duration exceeds the short-period perturbation threshold. The segment, the threshold The threshold was set by analyzing 20 historically abandoned R&D sub-paths and statistically analyzing the distribution of their continuous evolution time. It was found that the 95th percentile was 9.0 hours. This threshold was designed to effectively filter out short-lived explorations that failed to produce stable results, while ensuring that the threshold was not less than a benchmark value. (8.5 hours) will Set to 9.0 hours, and then perform the following filtering based on the duration of channel C1. The duration of channel C2 will be retained. If a channel with a duration of less than 9.0 hours is removed, its associated edges will be discarded. If the nodes connected to the channel are removed, they will be removed as well. All connected channels that meet the persistence requirements will be retained. A new graph composed of the retained channels and nodes will be reconstructed to establish a structural behavior progression path graph.

[0075] Please see Figure 4 The instruction offset recognition module includes a target direction extraction submodule, a matching degree filtering submodule, and an offset fragment recognition submodule;

[0076] The target direction extraction submodule obtains the target behavior direction parameters designed in the task book based on the path segment direction data in the structural behavior progressive path map, detects the target direction angle value and direction vector coordinates, calculates the vector angle between the path segment direction angle and the target direction angle, determines the angle deviation value between each path segment and the target direction, and establishes a set of direction matching degree values.

[0077] Based on the directional data of each path segment (such as the retained path P1, which is composed of channel C1) in the structural behavior progression path map, this directional data constitutes an evolution vector in a multi-dimensional design space, formed by the average changes of core design parameters and performance indicators in each label group constituting the path. For simplicity, a two-dimensional example is used, assuming the average directional vector of path P1 is... =(Parameter P1 change: +4.0K, Performance M1 change: +0.8%), obtain the design target behavior direction parameters in the task description. The task description requires that "within the acceptable range of constraint index C1, parameter P1 decreases by 10K (X-axis), and performance M1 increases by 2% (Y-axis)". This objective constitutes the target vector. Calculate the path segment orientation vector Angle and target direction vector Angle (Located in the second quadrant, actually) ), calculate the angle between the two vectors, i.e. This is the angle deviation value. Repeat this calculation for all path segments in the map to establish a set of directional matching degree values.

[0078] The matching degree filtering submodule calls the matching degree values ​​of each path segment in the direction matching degree value set, filters the path segments whose matching degree values ​​exceed the set matching threshold, counts the number of fluctuating segments in the path, calculates the overlap rate between the number of fluctuating segments and the total number of path segments, judges the overlap rate and compares it with the overlap judgment benchmark value, and obtains the overlap rate judgment label.

[0079] Call the angular deviation values ​​of each path segment in the direction matching degree value set (e.g., path P1 deviation). Path P2 deviation The filter matching score exceeds the set matching threshold. The path segment, the threshold Based on the accuracy requirements of the R&D task, if the task specification stipulates that the allowable deviation angle of the design direction is... ,but Perform the filter: P1( )> P2( ) Therefore, P1 is a path segment with poor matching. For P1, we count the number of fluctuating segments in its path (assuming it consists of 8 nodes). A fluctuating segment is defined as a node marked as "moderately fluctuating" or "highly fluctuating" in segment 3. Assuming that 3 out of the 8 nodes in P1 are "moderately fluctuating", the number of fluctuating segments is 3. We then calculate the overlap rate between the number of fluctuating segments and the total number of segments in the path. The overlap rate is compared with the overlap determination benchmark value. The comparison results show that the benchmark value This study reviewed 15 known R&D failure cases that failed due to process oscillations. The average percentage of oscillating segments in these paths was 0.60, with a standard deviation of 0.15. To identify such unstable paths, a benchmark value was set. ,because The volatility of path P1 did not exceed the benchmark, and the overlap rate judgment indicator was obtained (P1: volatility is normal).

[0080] The offset segment identification submodule judges the identification and matching degree filtering results based on the overlap rate, selects path segments with an overlap rate lower than the benchmark value and a matching degree that does not meet the requirements, marks the start and end positions of the offset segments, extracts the behavioral feature data of the offset segments, classifies the type attributes of the offset segments, and generates a set of functional behavioral offset segments.

[0081] Based on the overlap rate judgment indicator (P1: normal volatility) and the matching degree screening result (P1 angle deviation) Not satisfied (Requirements), select those with an overlap rate lower than the benchmark value ( And the matching degree does not meet the requirements (deviation). The path segment P1 satisfies this binary condition. This combination of conditions aims to identify stable but directional exploratory behaviors during the R&D process. The start and end positions of this deviation segment are marked, i.e., the starting iteration number (e.g., No. 25) of the starting node and the ending iteration number (e.g., No. 32) of the ending node of path P1. The behavioral feature data of this deviation segment, i.e., the average directional vector of P1, is extracted. and The angle deviation is classified as "steady-state direction deviation" and P1 and its time range are treated as a record to generate a set of functional behavior offset segments.

[0082] Please see Figure 5 The working condition segment positioning module includes a trigger point positioning submodule, a density filtering submodule, and a statistical segmentation submodule.

[0083] The trigger point location submodule extracts the start and end time point data from the functional behavior offset segment set, locates the load trigger point time in the corresponding time period in the simulation working condition sequence, detects the location of the boundary inflection point, monitors the change time of the response reversal point, calculates the time interval of the three types of trigger points on the time line, records the time coordinates and type identifiers of each trigger point, and establishes the trigger point time series.

[0084] Extract the start and end time data of the "steady-state direction deviation" segment P1 from the functional behavior offset segment set. This covers the simulation time series from hour 125 to hour 160. Locate various simulation event points within this time period [125h, 160h] in the simulation condition sequence. Searching the simulation log reveals: at 130.2h, a "cyclic load of condition set A" (load trigger point L1) was applied; at 138.5h, the "boundary constraint condition #2 of the solution model" (boundary inflection point B1) was modified; at 147.8h, the key performance indicator M1 "showed a nonlinear inflection point followed by a trend reversal" (response reversal point R1); and at 155.1h, a "pulse load of condition set B" (load trigger point L2) was applied. Calculate the interval between these trigger points on the timeline. The interval between L1 and B1 is... The interval between B1 and R1 The interval between R1 and L2 Record the time coordinates and type identifiers of each trigger point to form an array S=[(130.2h,'L'),(138.5h,'B'),(147.8h,'R'),(155.1h,'L')], and establish the trigger point time series.

[0085] The density filtering submodule calls the time interval data of each trigger point in the trigger point time series, judges the comparison result of the interval length of adjacent trigger points with the density filtering threshold, filters the trigger point group whose interval length is lower than the threshold, marks the start and end boundary positions of the trigger point group, calculates the number density value of trigger points in the group, and obtains the trigger point density distribution.

[0086] The time interval data of each trigger point in the trigger point time series S, i.e., the interval array I={8.3h,9.3h,7.3h}, is retrieved to determine the interval length between adjacent trigger points and the density filtering threshold. The comparison results, the threshold The settings were based on the typical execution time of a standard set of operating conditions in simulation analysis. If the average time from loading to stable response for a standard set of operating conditions is 8.0 hours, then the settings are as follows: Hours are used to capture consecutive event chains that are faster than the response cycle of a single standard operating condition, performing comparisons: intervals. ;interval ;interval By filtering out trigger point groups with intervals shorter than a threshold, the interval between R1 (147.8h) and L2 (155.1h) was found to be... The conditions are met, therefore {R1, L2} constitutes a trigger point group G1. The start and end boundary positions of this group are marked as [147.8h, 155.1h]. Calculate the number density value of the trigger points within the group. The density of G1 is Repeat this process for all offset segments per hour to obtain the distribution of all high-density trigger point groups.

[0087] The statistical segmentation submodule, based on the trigger point group data in the trigger point density distribution, counts the number of load triggers, the number of boundary turnings, and the number of response reversals in each density segment, merges the counts of the same type of triggers, arranges the density segment sequence in chronological order, constructs the connection relationship between segments, and generates a table of operating condition disturbance response segments.

[0088] Based on the trigger point group data in the trigger point density distribution (e.g., group G1, time [147.8h, 155.1h], containing {R1, L2}), the occurrence frequency of each type of simulation event in this density segment is counted. For G1: the load triggering frequency is 1 (L2), the boundary turning frequency is 0, and the response reversal frequency is 1 (R1). The combined statistical results yield the feature vector of G1 as (load: 1, boundary: 0, response: 1). Assuming there is another high-density segment G2, its feature vector is (load: 3, boundary: 0, response: 0). These density segment sequences are arranged in chronological order to construct the temporal relationship between segments, and the statistical results are compiled into a table.

[0089] Table 3: Statistics of Disturbance Segments under Simulation Conditions

[0090]

[0091] As shown in Table 3, this table quantifies the high-incidence sections of potential simulation condition disturbances that lead to design deviations and their event composition, generating a table of condition disturbance response sections.

[0092] Please see Figure 6 The matching relationship filtering module includes a direction vector extraction submodule, a segment correspondence judgment submodule, and a path matching selection submodule;

[0093] The direction vector extraction submodule extracts the corresponding direction vector coordinates from the target set of design behavior based on the direction distribution results of each trigger sequence in the working condition disturbance response segment table, calculates the response direction angle value and the target direction angle value, detects the start and end time nodes of overlapping path segments, measures the direction change amplitude of path segments, and establishes a direction matching dataset.

[0094] Based on the trigger sequence (R1, L2) and its directional distribution results of segment G1 in the disturbance response segment table (see Table 3), the combined impact vector of the R1 event (response reversal) and the L2 event (pulse load) on the design objectives (parameter P1, performance M1) is extracted from the detailed data of the simulation output. Assuming that the R1 event causes P1 to increase by 1.5K and M1 to decrease by 0.2%, the vector is constructed. The L2 event causes P1 to increase by 2.0K and M1 to increase by 0.1%, forming a vector. Then the overall response direction vector of segment G1 is Extract the design behavior target vector Calculate the response direction Angle and the angle with the target direction By comparison, the start and end time nodes of the overlapping path segments were detected, i.e., the time period [147.8h, 155.1h] in which G1 is located. The directional change amplitude of path segment G1 was measured, i.e., the angle between its response direction and the target direction, which is approximately... Establish a direction matching dataset {ID:G1,Time:[147.8,155.1],Angle_Diff:170.33°}.

[0095] The segment correspondence judgment submodule calls the path segment time node data in the direction matching dataset to determine the time overlap range between the start and end segments of the response direction and the target direction segment, calculates the continuous change rate of the direction angle between adjacent path segments, compares the difference between the continuous change rate and the continuity judgment threshold, filters the path segments that meet the continuity requirements, and obtains the continuity evaluation result.

[0096] The data from each segment in the direction matching dataset is called, such as {ID:G1,Angle_Diff:170.33°} and another segment {ID:G2,Angle_Diff:15.2°}. The overlap between the time range [147.8h,155.1h] of the response direction start and end segment G1 and the time segment of the target direction is determined. Since the target direction is continuously valid throughout the entire R&D cycle, the G1 time period is completely covered. The rate of change of the directional angle between adjacent path segments G1 and G2 is calculated, and the response vector of G1 is determined. Assume that the response vector of G2 is G1's angle G2's angle The angle changes to The time interval between G1 and G2 is Then the continuous rate of change This rate of change is compared with the continuity threshold. Comparison, this threshold Based on the statistical analysis of the angle change rate during the smooth transition phase of design direction in historical successful cases, the 75th percentile (P75) of its distribution is taken as the threshold. If P75 = ,because This indicates that the direction change process from G1 to G2 is smooth in terms of rate, and the path segment (G1,G2) satisfies the continuity requirement, resulting in a continuity evaluation result (Pair(G1,G2): Pass).

[0097] The path matching and selection submodule selects path segments that meet both continuity and direction consistency conditions based on the continuity evaluation results and direction consistency comparison data, marks the matching path segment identifiers, constructs the time series relationship of the matching paths, summarizes the simulation parameters and response characteristics of the path segments, and generates the simulation results of the engineering structure.

[0098] Based on the continuity assessment results (Pair(G1,G2): Pass) and the directional consistency comparison data (G1 deviation: 170.33°, G2 deviation: 15.2°), path segments that simultaneously meet both continuity and directional consistency conditions are selected. The directional consistency criterion adopts the matching threshold set in paragraph 8. G1 deviation The directional consistency is not satisfied; the deviation of G2 is... To ensure directional consistency, although the transition from G1 to G2 is smooth, the response direction of G1 itself is severely contrary to the design objective. Therefore, G1 is identified as a disturbance source causing the deviation and is eliminated. G2 is selected as the effective simulation response segment consistent with the objective, and a matching path segment identifier (ID: G2) is marked. A matching path time series relationship is constructed. If a subsequent G3 that meets the conditions exists, then a... The matching paths are identified, and the simulation conditions and detailed response characteristics corresponding to these matching path segments (such as G2) are summarized to form a set of key data that has been filtered and reveals how the design objectives are achieved under specific operating conditions. This data is then used to generate simulation results for the engineering structure.

[0099] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An intelligent design and simulation system for engineering technology research and development, characterized in that, The system includes: The evolution trend extraction module acquires stage data in the engineering technology research and development process, extracts the directional change amplitude, change rate and boundary span of continuous stages, judges the consistency of direction and the degree of fluctuation, groups them into trend segment labels, and generates a stage behavior change combination table. Based on the trend label combination in the stage behavior change combination table, the module determines the directional consistency and fluctuation frequency segment overlap of adjacent label groups, establishes connectable channels, filters out the span of channel discontinuities and removes short jump segments, and generates a structural behavior progression path map. The instruction offset recognition module extracts the design target direction based on the distribution of path segments in the structural behavior progression path map, determines the degree of angular offset between the path and the target direction, filters path segments with low overlap rate, and generates a set of functional behavior offset segments. The working condition segment positioning module extracts the start and end time point data from the functional behavior offset segment set, performs density judgment on the load jump points, boundary switching points and direction reversal points in the simulation sequence during that period, filters the continuous jump areas and groups them according to the number of triggers, and generates a working condition disturbance response segment table.

2. The intelligent design and simulation system for engineering technology research and development according to claim 1, characterized in that: The stage behavior change combination table includes direction change angle labels, rate change level labels, and boundary span interval labels. The structural behavior progression path map includes trend channel links, fluctuation persistence statistics, and path segment sequence identifiers. The functional behavior offset segment set includes direction matching deviation segments, fluctuation distribution abnormal segments, and target behavior misalignment segments. The working condition disturbance response segment table includes high-frequency load triggering segments, boundary turning point concentrated segments, and response reversal dense segments.

3. The intelligent design and simulation system for engineering technology research and development according to claim 1, characterized in that, The evolution trend extraction module includes a data change quantification submodule, a trend direction judgment submodule, and a fluctuation intensity classification submodule; The data change quantification submodule acquires stage design record data in the intelligent design and simulation process of engineering technology research and development, detects various design record parameters in adjacent time periods, calculates the numerical differences of design parameters between adjacent time periods, calculates the change direction, change amplitude and change rate, monitors the range of design boundary parameters, quantifies the boundary span value, and establishes a parameter change feature set. The trend direction judgment submodule calls the direction change amplitude value in the parameter change feature set, selects the direction change data of three adjacent time periods, calculates the angle between the three change directions, judges the degree of deviation between the angle and the benchmark value, divides the directional category according to the deviation range, performs directional classification operation on each group of three-segment trend, and generates directional trend identifier. The fluctuation intensity classification submodule calculates the mean and standard deviation of the change rate for each time period based on the change rate value and boundary span value in the parameter change feature set, compares the deviation between the single time period value and the mean, determines the fluctuation intensity level, performs horizontal intensity comparison in combination with directional trend indicators, groups and generates trend label combinations, and establishes a stage behavior change combination table.

4. The intelligent design and simulation system for engineering technology research and development according to claim 1, characterized in that, The dependency structure construction module includes a trend continuity judgment submodule, a channel map construction submodule, and a disturbance filtering submodule. The trend continuity judgment submodule detects the angle of change of fluctuation direction between adjacent trend label groups according to the trend label group affiliation relationship in the stage behavior change combination table, calculates the angle difference and compares it with the continuity judgment threshold, marks the start and end positions of the overlapping change frequency segment, judges the trend continuity status between label groups, performs a classification operation on the trend continuity label groups, and obtains a continuous label set. The channel map construction submodule calls the data of the continuous label group in the continuous label set, establishes the coordinates of the connecting nodes between the label groups, constructs the connecting paths between the nodes, draws the spatial distribution of the connecting channels, determines the duration of the fluctuation at the channel connection, calculates the difference between the interval duration value and the set benchmark value, and generates the channel connection map. The disturbance filtering submodule, based on the fluctuation duration interval values ​​in the channel connection graph, filters segments whose interval duration exceeds the short-period disturbance threshold, removes path nodes with intervals below the threshold, retains connection channels that meet the duration requirements, reconstructs the channel connection relationships after filtering, and establishes a structural behavior progressive path graph.

5. The intelligent design and simulation system for engineering technology research and development according to claim 1, characterized in that, The instruction offset recognition module includes a target direction extraction submodule, a matching degree filtering submodule, and an offset fragment recognition submodule; The target direction extraction submodule obtains the target behavior direction parameters designed in the task book based on the path segment direction data in the structural behavior progressive path map, detects the target direction angle value and direction vector coordinates, calculates the vector angle between the path segment direction angle and the target direction angle, determines the angle deviation value between each path segment and the target direction, and establishes a set of direction matching degree values. The matching degree filtering submodule calls the matching degree values ​​of each path segment in the direction matching degree value set, filters the path segments whose matching degree values ​​exceed the set matching threshold, counts the number of fluctuating segments in the path, calculates the overlap rate between the number of fluctuating segments and the total number of path segments, judges the overlap rate and compares it with the overlap judgment benchmark value, and obtains the overlap rate judgment identifier. The offset segment identification submodule, based on the overlap rate judgment identifier and matching degree filtering results, selects path segments with an overlap rate lower than the benchmark value and a matching degree that does not meet the requirements, marks the start and end positions of the offset segments, extracts the behavioral feature data of the offset segments, classifies the type attributes of the offset segments, and generates a set of functional behavioral offset segments.

6. The intelligent design and simulation system for engineering technology research and development according to claim 1, characterized in that, The working condition segment positioning module includes a trigger point positioning submodule, a density filtering submodule, and a statistical segmentation submodule; The trigger point positioning submodule extracts the start and end time point data from the set of functional behavior offset segments, locates the load trigger point time in the corresponding time period in the simulation working condition sequence, detects the location of the boundary turning point, monitors the change time of the response reversal point, calculates the time interval of the three types of trigger points on the time line, records the time coordinates and type identifiers of each trigger point, and establishes the trigger point time series. The density filtering submodule calls the time interval data of each trigger point in the trigger point time series, judges the comparison result of the interval duration of adjacent trigger points with the density filtering threshold, filters the trigger point group with an interval duration lower than the threshold, marks the start and end boundary positions of the trigger point group, calculates the number density value of trigger points in the group, and obtains the trigger point density distribution. The statistical segmentation submodule, based on the trigger point group data in the trigger point density distribution, counts the number of load triggers, boundary turning times, and response reversals in each density segment, merges the counts of the same type of triggers, arranges the density segment sequence in chronological order, constructs the connection relationship between segments, and generates a table of operating condition disturbance response segments.

7. The intelligent design and simulation system for engineering technology research and development according to claim 1, characterized in that, The system also includes: The matching relationship screening module extracts the design target direction vector based on the directional distribution results of each trigger sequence in the working condition disturbance response segment table, performs a judgment on the coincidence of the start and end times of the response direction and the consistency comparison of the direction angle, screens out continuous directional consistent segments, and generates engineering structure simulation results. The simulation results of the engineering structure include the directional continuous matching path, the target direction coverage segment, and the mapping relationship of the start and end segments.

8. The intelligent design and simulation system for engineering technology research and development according to claim 7, characterized in that, The matching relationship screening module includes a direction vector extraction submodule, a segment correspondence judgment submodule, and a path matching selection submodule; The direction vector extraction submodule extracts the corresponding direction vector coordinates from the design behavior target set based on the direction distribution results of each trigger sequence in the working condition disturbance response segment table, calculates the response direction angle value and the target direction angle value, detects the start and end time nodes of overlapping path segments, measures the direction change amplitude value of the path segment, and establishes a direction matching dataset. The segment correspondence judgment submodule calls the path segment time node data in the direction matching dataset to determine the time overlap range between the start and end segments of the response direction and the target direction segment, calculates the continuous change rate of the direction angle between adjacent path segments, compares the difference between the continuous change rate and the continuity judgment threshold, filters the path segments that meet the continuity requirements, and obtains the continuity evaluation result. The path matching and selection submodule, based on the continuity evaluation results and directional consistency comparison data, selects path segments that satisfy both continuity and directional consistency conditions, marks matching path segment identifiers, constructs time series relationships of matching paths, summarizes path segment simulation parameters and response characteristics, and generates engineering structure simulation results.

Images