A visual analysis method for multi-system progress data comparison in shipbuilding projects

Abstract

The present invention relates to the field of data analysis and management technology, and specifically to a visual analysis method for comparing multi-system progress data in a shipbuilding project. The method comprises the following steps: extracting progress-related feature data from the shipbuilding project's management system, human resources system, production system, financial system, and supply chain system to generate a standardized multidimensional progress feature matrix; performing a convolution operation on the multidimensional progress feature matrix with weight coefficients to generate a comprehensive progress deviation cloud map with temporal correlation; mapping the comprehensive progress deviation cloud map to a three-dimensional coordinate system; and updating the surface morphology in real time and outputting a critical path warning signal. The present invention, combining an adaptive sliding window with a dynamic weight convolution mechanism, significantly improves the real-time comparison accuracy and temporal traceability of cross-system progress deviations.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis and management, and in particular to a visual analysis method for comparing multi-system progress data of a shipbuilding project. Background Art

[0002] As the scale of shipbuilding projects becomes increasingly large, the management systems, human resources systems, production execution systems, financial management systems, and supply chain systems involved are interconnected and complex in their coordination, making it significantly more difficult to control the overall progress of projects. Existing project management methods are mostly based on single-system or linear-dimensional plan tracking, often ignoring time synchronization deviations between multiple systems, inconsistent data dimensions, and their combined impact on actual progress, making it difficult to promptly identify the root causes of progress imbalances and their evolution trends.

[0003] Furthermore, traditional progress analysis methods, which mostly rely on static reports or two-dimensional graphs, lack the ability to model and visualize the coupling relationships between progress deviations across different systems. This makes it difficult to meet the precise requirements of shipbuilding for dynamic changes, critical path control, and process conflict warnings. Existing methods, in particular, lack the ability to integrate multi-domain progress data, identify deviations at a granular level, and perform interactive analysis when faced with challenging conditions such as shared dock resources, complex lifting operations, and material delays. Summary of the Invention

[0004] The present invention provides a visual analysis method for comparing multi-system progress data of shipbuilding projects. This method integrates multi-system progress characteristics and has dynamic weight analysis and three-dimensional visualization capabilities to improve the progress transparency, control accuracy and abnormal response efficiency of shipbuilding projects.

[0005] A visual analysis method for comparing multi-system progress data of a shipbuilding project comprises the following steps:

[0006] S1, multi-system data standardization: Extract progress-related feature data from the shipbuilding project management system, human resources system, production system, financial system, and supply chain system, align the feature data with timestamps and unify the dimensions to generate a standardized multi-dimensional progress feature matrix;

[0007] S2, dynamic weight allocation and progress comparison: Based on the progress impact factors of each system in the multidimensional progress feature matrix, an adaptive sliding window algorithm is used to calculate the weight coefficient of each system in real time. The weight coefficient is convolved with the multidimensional progress feature matrix to generate a comprehensive progress deviation cloud map with time series correlation;

[0008] S3, 3D visualization model construction: Map the comprehensive progress deviation cloud map to a 3D coordinate system, where the X-axis represents the project timeline, the Y-axis represents the correlation between each system, and the Z-axis represents the deviation impact intensity. A visual comparison surface is generated through a spatial interpolation algorithm;

[0009] S4, interactive intelligent analysis: embeds adjustable threshold parameters in the visual comparison surface, adjusts the time slice range and weight distribution ratio through touch interaction, updates the surface morphology in real time and outputs critical path warning signals.

[0010] Optionally, the S1 specifically includes:

[0011] S11: Extract task node completion rate, approval process response time, and project milestone deviation from the management system; extract job type work hour saturation, skill matching, and staff turnover rate from the human resources system; extract segment construction completion rate, equipment overall efficiency, and quality inspection pass rate from the production system; extract budget execution rate, payment node achievement rate, and cost overrun coefficient from the financial system; and extract material completeness rate and logistics on-time rate from the supply chain system.

[0012] S12: Timestamp alignment uses the task node number as the association key to perform a bidirectional match between the production system's segmented construction timestamp and the management system's task node timestamp. When the deviation between the hoisting completion time recorded by the production system and the node deadline set by the management system exceeds the predetermined time, the actual completion time of the corresponding node in the management system is synchronously updated based on the production system timestamp.

[0013] S13, dimensionality is unified and processed including outlier processing, and finally a standardized multi-dimensional progress feature matrix is generated. The row vectors of the matrix represent time series segments, and the column vectors are arranged according to domain division: management domain (3 dimensions), human resources domain (3 dimensions), production domain (3 dimensions), financial domain (3 dimensions), and supply chain domain (3 dimensions).

[0014] Optionally, the dimensional unification processing includes converting the working hour data into standard man-days (8 hours / man-day), normalizing the budget execution rate to a value in the interval [0,1], binarizing the equipment status data into a Boolean type (0-shutdown / 1-running), and mapping the number of days of supply chain material delay to a risk level value of [0,10] through a piecewise function; the outlier processing includes performing outlier cleaning on the processed data, setting a task node completion rate threshold range of [0,1.2], and interpolating data points outside the range using the mean of adjacent time periods.

[0015] Optionally, the S2 specifically includes:

[0016] S21, schedule impact factor extraction: Separate the core impact factors by domain from the multidimensional schedule feature matrix, including management domain factors, human domain factors, production domain factors, financial domain factors, and supply chain domain factors;

[0017] S22, Adaptive Sliding Window Algorithm Execution: Initialize the window length based on the total project cycle, with a step size equal to the inverse of the human resource turnover rate. Automatically shorten the window length when it detects that the overall efficiency of production equipment has dropped too high for multiple consecutive window periods. If the supply chain risk level exceeds the risk threshold, a supply chain risk penalty factor is added as the window slides. The supply chain risk level is mapped to a supply chain risk level value in the range [0,10] based on the number of days of supply chain material delay using a piecewise function.

[0018] S23, convolution operation design: Construct a three-dimensional convolution kernel, whose time dimension matches the adaptive sliding window and whose spatial dimension corresponds to the five-domain feature channels; perform depthwise separable convolution on the dynamic weight coefficient matrix and the multidimensional progress feature matrix, and initialize the convolution kernel weights using the inter-domain correlation strength matrix;

[0019] S24, comprehensive progress deviation cloud map generation: The convolution output is time-normalized, and the deviation and random fluctuation components are separated through feature decoupling technology; the deviation intensity is mapped to the RGB color space, where the red channel represents the dominant deviation in the production domain, the blue channel represents the dominant deviation in the supply chain domain, and the brightness of the green channel is positively correlated with the deviation intensity in the financial domain, forming a comprehensive progress deviation cloud map.

[0020] Optionally, the supply chain risk penalty factor is calculated as: ;in, is the supply chain risk penalty factor, The number of days the material is delayed. is the risk transmission coefficient, is the supply chain risk level value, and the risk transmission coefficient is determined based on the Pearson correlation coefficient between material delay and schedule deviation in historical data.

[0021] Optionally, the management domain factor is the product of the task node completion rate and the milestone deviation value; the human domain factor is the weighted harmonic mean of the work hour saturation and skill matching of the job type; the production domain factor is calculated by the geometric mean of the segmented construction completion rate, the equipment comprehensive efficiency and the quality inspection pass rate; the financial domain factor is the ratio of the budget execution rate to the cost overrun coefficient; the supply chain domain factor is based on the product of the material completeness rate and the logistics punctuality rate.

[0022] Optionally, the mapping rule for mapping the comprehensive progress deviation cloud map to the three-dimensional coordinate system includes:

[0023] The X-axis timeline uses a piecewise nonlinear scale, compressing the project plan timeline to a 1:1 ratio. The actual progress timeline dynamically expands and contracts according to the deviation intensity. The expansion factor is calculated as: ;in, is the scaling factor of the time axis, is the normalized deviation intensity of the Z axis from the comprehensive progress deviation cloud chart;

[0024] The Y-axis correlation calculation uses a cross-domain coupling algorithm: a five-domain correlation matrix is constructed, with basic correlation weights set between the management domain and the production domain, and a dynamic correlation weight set between the supply chain domain and the finance domain as the square root of the budget execution rate. Eigenvectors are extracted through cluster analysis, and the first principal component is projected onto the Y-axis to form a correlation scale for each system.

[0025] The Z-axis intensity mapping is logarithmic, and the baseline intensity value is set based on historical project progress deviations;

[0026] The spatial interpolation algorithm includes embedding dynamic constraints in the standard Kriging interpolation process;

[0027] The generation of the visual comparison surface includes converting the interpolated three-dimensional point cloud data into a NURBS surface, the number of surface control points is taken as the square root of the number of valid task nodes in the current time window, and special constraints are set during the surface parameterization process.

[0028] Optionally, the dynamic constraint conditions include:

[0029] ① When the overall equipment efficiency (OEE) of the production domain is lower than the efficiency threshold, a radial basis function penalty term is added to the interpolation equation, and the penalty coefficient is Calculated as: ;

[0030] ② For areas where the budget execution rate in the financial domain is lower than 70%, an anisotropic interpolation strategy is used to expand the search neighborhood to 2 times the standard deviation along the time axis.

[0031] Optionally, the interactive intelligent analysis in S4 includes embedding adjustable threshold panels in the four vertices of the visual comparison surface, setting the production cycle tolerance threshold, labor saturation threshold, supply chain risk diffusion coefficient and financial budget elasticity coefficient respectively, and driving the surface color temperature to change dynamically according to the HSV color circle when the threshold is adjusted.

[0032] Optionally, the interactive intelligent analysis also includes adjusting the time slice range and weight distribution ratio based on touch gesture execution, and using incremental recalculation to update the surface morphology in real time, while generating a three-level critical path warning and outputting an abnormality report based on the multi-domain deviation intensity and its correlation.

[0033] Beneficial effects of the present invention:

[0034] This invention extracts key progress characteristic indicators from five major systems: management, human resources, production, finance, and supply chain, and constructs a multi-dimensional progress characteristic matrix with unified dimensions, time alignment, and anomaly rejection mechanisms. This solves technical difficulties in the shipbuilding process, such as inconsistent multi-system data caliber, large dimensional span, and poor time correlation. Combining an adaptive sliding window with a dynamic weight convolution mechanism, this method significantly improves the real-time comparison accuracy and time series traceability of cross-system progress deviations.

[0035] The present invention uses a comprehensive schedule deviation cloud map and environmental variables such as tidal cycles to construct a three-dimensional comparison coordinate system. Spectral clustering is used to extract inter-domain coupling feature projections to form a Y-axis subsystem correlation scale. Industry-specific control mechanisms such as dock slide deformation compensation and welding stress disturbance mapping are introduced to ultimately generate an adjustable NURBS surface that preserves topology. This allows users to intuitively identify spatial concentration areas and key causes of schedule imbalances during dynamic interaction.

[0036] The present invention achieves real-time updating and feedback response of the progress surface by embedding four types of adjustment threshold parameter panels, supporting multi-dimensional touch gesture operations and a GPU-based incremental update mechanism; constructs three-level critical path warning rules, combines domain deviation fluctuations and coupling conflict conditions, and automatically generates abnormal time period reports and visual ripple alarms, thereby improving risk perception and scheduling decision-making efficiency in key process stages of shipbuilding tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the present invention or 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 for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 Schematic diagram of a method flow in an embodiment of the present invention;

[0039] Figure 2 Schematic diagram of three-dimensional coordinate mapping of deviation cloud map according to an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It is also noted that, to provide a more detailed description, the following embodiments are best and preferred embodiments, and those skilled in the art may employ alternative methods for implementing certain known technologies. Furthermore, the accompanying drawings are intended only to provide a more detailed description of the embodiments and are not intended to limit the present invention.

[0041] It should be noted that references in the specification to "one embodiment," "an embodiment," "exemplary embodiments," "some embodiments," etc. indicate that the described embodiments may include specific features, structures, or characteristics, but not necessarily every embodiment will include such specific features, structures, or characteristics. Furthermore, when specific features, structures, or characteristics are described in conjunction with an embodiment, it is within the knowledge of persons skilled in the relevant art to implement such features, structures, or characteristics in conjunction with other embodiments (whether or not explicitly described).

[0042] In general, terms can be understood, at least in part, from their use in context. For example, depending at least in part on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending at least in part on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0043] like Figure 1-Figure 2 As shown, a visual analysis method for comparing multi-system progress data of a shipbuilding project includes the following steps:

[0044] S1, multi-system data standardization: Extract progress-related feature data from the shipbuilding project management system, human resources system, production system, financial system, and supply chain system, align the feature data with timestamps and unify the dimensions to generate a standardized multi-dimensional progress feature matrix;

[0045] S2, dynamic weight allocation and progress comparison: Based on the progress impact factors of each system in the multidimensional progress feature matrix, an adaptive sliding window algorithm is used to calculate the weight coefficient of each system in real time. The weight coefficient is convolved with the multidimensional progress feature matrix to generate a comprehensive progress deviation cloud map with time series correlation;

[0046] S3, 3D visualization model construction: Map the comprehensive progress deviation cloud map to a 3D coordinate system, where the X-axis represents the project timeline, the Y-axis represents the correlation between each system, and the Z-axis represents the deviation impact intensity. A visual comparison surface is generated through a spatial interpolation algorithm;

[0047] S4, interactive intelligent analysis: embeds adjustable threshold parameters in the visual comparison surface, adjusts the time slice range and weight distribution ratio through touch interaction, updates the surface morphology in real time and outputs critical path warning signals.

[0048] S1 specifically includes:

[0049] S11: Extract task node completion rate, approval process response time, and project milestone deviation from the management system; extract job type work hour saturation, skill matching, and staff turnover rate from the human resources system; extract segment construction completion rate, equipment overall efficiency, and quality inspection pass rate from the production system; extract budget execution rate, payment node achievement rate, and cost overrun coefficient from the financial system; and extract material completeness rate and logistics on-time rate from the supply chain system.

[0050] In the management system: task node completion rate , Approval process response time , Project Milestone Deviation ;

[0051] In the human resources system: job type and working hours saturation , skill matching , staff turnover rate ;

[0052] In the production system: staged construction completion , equipment overall efficiency OEE, quality inspection pass rate ;

[0053] In the financial system: budget execution rate , payment node achievement rate , cost overrun coefficient ;

[0054] In the supply chain system: Material completeness rate , logistics punctuality rate .

[0055] S12: Timestamp alignment uses the task node number as the association key to perform a bidirectional match between the production system's segmented construction timestamp and the management system's task node timestamp. When the deviation between the hoisting completion time recorded by the production system and the node deadline set by the management system exceeds the predetermined time, the actual completion time of the corresponding node in the management system is synchronously updated based on the production system timestamp.

[0056] S13, dimensionality is unified and processed including outlier processing, and finally a standardized multi-dimensional progress feature matrix is generated. The row vectors of the matrix represent time series segments, and the column vectors are arranged according to domain division: management domain (3 dimensions), human resources domain (3 dimensions), production domain (3 dimensions), financial domain (3 dimensions), and supply chain domain (3 dimensions).

[0057] Dimensional unification processing includes converting working hour data into standard man-days (8 hours / man-day), normalizing the budget execution rate to a value in the interval [0,1], binarizing equipment status data into a Boolean type (0-shutdown / 1-running), and mapping the number of days of supply chain material delay to a risk level value of [0,10] through a piecewise function; outlier processing includes performing outlier cleaning on the processed data, setting the task node completion rate threshold range [0,1.2], and interpolating data points outside the range using the mean of adjacent time periods.

[0058] Unified dimension processing: working hours are unified into standard man-days: ;in, Indicates the original number of working hours, Indicates standard man-day (8 hours / man-day);

[0059] Normalized budget execution rate: ;

[0060] Binarization of equipment operating status: ;

[0061] The multidimensional progress feature matrix is expressed as: ;

[0062] Each row : indicates the Time segments, each column is arranged in domain order as follows:

[0063] Management domain: ;

[0064] Human domain: ;

[0065] Production domain: ,OEE ;

[0066] Financial Domain: ;

[0067] Supply Chain Domain: ;

[0068] Finally, the Z-score method was used to standardize each column of data and retain 4 significant digits.

[0069] S2 specifically includes:

[0070] S21, Progress Impact Factor Extraction: From the Standardized Multidimensional Progress Feature Matrix The following core impact factors are extracted by dividing the five major domains:

[0071] Management Domain Factor : ;

[0072] Human Domain Factor : ;

[0073] Production Domain Factor : ;

[0074] Financial Domain Factors : ;

[0075] Supply Chain Domain Factors : ;

[0076] S22, adaptive sliding window algorithm:

[0077] Window initialization: ;

[0078] in, is the initial window length, is the total project period, is the sliding step length, is the staff turnover rate;

[0079] Window adjustment mechanism: If the equipment efficiency satisfies: ;

[0080] Then adjust: ;in, is the current weight coefficient of the financial domain;

[0081] Supply chain risk penalty factor: When , increase the weight penalty factor: ;in, is the risk penalty factor, The number of days the material is delayed. is the risk transmission coefficient (i.e. the Pearson correlation coefficient between material delay and schedule deviation), is the supply chain risk level value, which is mapped to a risk level value of [0,10] based on the number of days of supply chain material delay through a piecewise function: ; The number of days the material is delayed.

[0082] Special working condition detection: If the crane usage application volume in a certain time period satisfy: , then enable: window freezing mechanism (pause window sliding) and offline weight calculation mode (perform discrete optimization based on historical fragments).

[0083] S23, convolution operation design:

[0084] Convolution kernel settings: The convolution kernel shape is set to: ,in, is the time window length, 1 means spatial sliding (fixed), and 5 means five-domain feature channels;

[0085] Convolution kernel initialization: initial correlation strength matrix between domains The values are: Management and production domains: 0.7, finance and supply chain domains: 0.5;

[0086] Gating mechanism: If , then close the financial domain convolution channel: ;

[0087] Tidal cycle compensation parameters: According to the tidal cycle of the dock location Adjust the production domain weight time distribution function : ;in, Indicates at a point in time The original production domain weight coefficient on is used to measure the impact of the current production domain on the comprehensive schedule deviation. Indicates at a point in time The tidal cycle compensation factor is used to adjust the time weight distribution of production activities, reflecting the impact of tidal changes (such as high tide / low tide) on the feasible time window of production tasks (such as hull lifting, equipment installation, etc.), which is expressed as: , is the tidal cycle, is the amplitude of tidal fluctuation (experience value 0.1~0.3), is the phase adjustment constant used to synchronize the tide table with the production scheduling starting time.

[0088] S24, comprehensive progress deviation cloud diagram generation:

[0089] Time series normalization and smoothing: using double exponential smoothing method: ,in, represents the smoothed sequence, represents the original convolution output, Indicates the smoothing factor, with a value of 0.2-0.3;

[0090] Feature decoupling: convolution output signal Deconstructed into: ,in, Indicates systematic bias, represents the random fluctuation component;

[0091] Cloud color mapping: decouple the deviation component intensity Mapped to RGB channels respectively: .

[0092] S3 specifically includes:

[0093] S31, 3D coordinate mapping rules:

[0094] X-axis: Timeline mapping: uses a piecewise nonlinear coordinate scale. The project plan timeline uses a compression ratio of 1:1, and the actual progress timeline dynamically expands and contracts according to the deviation intensity. The expansion factor formula is as follows: ;in, is the time axis scaling factor, is the normalized deviation intensity from the Z-axis of the comprehensive progress deviation cloud chart;

[0095] Y-axis: System correlation mapping: Constructing a five-domain basic correlation matrix ,Example elements include: , use the spectral clustering method to perform eigendecomposition on the matrix and extract the first principal component vector ,map it to the Y-axis to form the subsystem correlation scale;

[0096] Z-axis: Deviation intensity mapping: Logarithmic transformation of cloud deviation intensity to enhance difference perception: ;in, is the absolute deviation value of the current time slice, is the 90th percentile of the historical project schedule deviation intensity (set to a logarithmic basis).

[0097] S32, spatial interpolation algorithm optimization: dynamic Kriging interpolation, when production domain equipment efficiency OEE , the interpolation function introduces the radial basis function (RBF) penalty term: , budget execution rate in the financial domain , enable anisotropic interpolation strategy and expand the search neighborhood in the time axis direction to: ; is the standard deviation of the progress indicator in the time series.

[0098] S33, Visual Surface Generation: Generate NURBS surface (Non-Uniform Rational B-Spline) using 3D interpolation point cloud. The number of control points is set to: , Indicates the number of valid task nodes in the current time window;

[0099] Special surface constraints: During the peak lifting hours (09:00–11:00 and 14:00–16:00), a maximum curvature threshold is set to prevent excessive visual deformation; when multiple docks are operating in parallel, the subdivision surface structure is automatically generated based on the topology preservation conditions.

[0100] Also includes dynamic rendering enhancements:

[0101] Create a color-bias type mapping relationship:

[0102] Red channel intensity = production domain deviation intensity × 0.8 + supply chain deviation intensity × 0.2;

[0103] Blue channel intensity = financial domain bias intensity × 0.6 + management domain bias intensity × 0.4;

[0104] Green channel intensity = human domain bias intensity × 0.7 + cross-domain coupling bias × 0.3;

[0105] Equidistration lines are superimposed on the surface of the curved surface, and the line width changes dynamically with the tide height. The line width is magnified to 1.5 times during high tide.

[0106] S4 specifically includes:

[0107] S41, adjustable threshold parameter embedding, sets dynamic adjustment panels at the four vertices of the visual comparison surface, corresponding to:

[0108] ① Production cycle tolerance threshold (range [0.8, 1.2] times the benchmark cycle);

[0109] ② Human resource saturation threshold (set in the range of [70%, 130%] according to job classification);

[0110] ③ Supply chain risk diffusion coefficient (graded adjustment level 1-10);

[0111] ④ Financial budget elasticity coefficient (continuous adjustment range 0.5-2.0);

[0112] The threshold parameter is dynamically bound to the color temperature of the surface vertices. When the threshold is adjusted, the vertex color rotates clockwise by the corresponding angle along the HSV color wheel.

[0113] S42, touch interaction logic design:

[0114] Time slice adjustment uses a two-finger pinch gesture:

[0115] ① Horizontal sliding controls the start and end points of the time window, with an accuracy of 15 minutes;

[0116] ② Pinch vertically to adjust window density, supporting six levels of zoom from hourly to weekly;

[0117] The weight distribution ratio is adjusted by the surface concave gesture:

[0118] ① Press and hold the surface area with one finger to trigger weight editing mode;

[0119] ② The depth of pressure along the surface normal is proportional to the reduction in weight (pressing down 1 cm corresponds to a 10% reduction in weight);

[0120] ③ Sliding friction coefficient feedback mechanism: When the adjustment amplitude exceeds 3 times the standard deviation of historical data, a vibration alarm is triggered;

[0121] S43, real-time update mechanism:

[0122] Establish an incremental recomputation pipeline:

[0123] ① After capturing the interaction instruction, the feature matrix subset of the affected domain is updated first (the update granularity is controlled within 10% of the data);

[0124] ② Use lightweight convolution kernel (3×3×3) for local weight convolution, skipping the aforementioned depth-wise separable convolution;

[0125] ③ The interpolation operation reuses the historical surface control point topology and only updates the Kriging interpolation patch for the changed area;

[0126] S44, critical path warning generation:

[0127] Set the trigger conditions for the third-level warning:

[0128] ① Yellow warning: The deviation intensity of a single domain is greater than 2 standard deviations of the baseline value for three consecutive windows;

[0129] ② Orange alert: The product of the correlation deviation strengths of the two domains is greater than 80% of the historical maximum;

[0130] ③ Red alert: Production domain deviation intensity × supply chain risk value > 10 and manpower saturation < 65%;

[0131] Warning signal outputs include:

[0132] ① Generate pulsed annular ripple diffusion effects in abnormal areas of the surface;

[0133] ② Automatically capture abnormal time slices to generate PDF reports, including:

[0134] Topological diagram of affected task nodes (reference the feature data in S1);

[0135] Weight adjustment trend curve (related to the window parameters in S1).

[0136] The present invention encompasses any alternatives, modifications, equivalents, and solutions that fall within the spirit and scope of the present invention. To provide a thorough understanding of the present invention, specific details are described in detail below in connection with the preferred embodiments of the present invention, but those skilled in the art will be able to fully understand the present invention without these detailed descriptions. Furthermore, to avoid unnecessary confusion regarding the essence of the present invention, well-known methods, processes, procedures, components, and circuits have not been described in detail.

[0137] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A visual analysis method for comparing multi-system progress data of shipbuilding projects, characterized in that: The following steps are involved: S1, multi-system data standardization: Extract progress-related feature data from the shipbuilding project management system, human resources system, production system, financial system, and supply chain system, align the feature data with timestamps and unify the dimensions to generate a standardized multi-dimensional progress feature matrix; S2, dynamic weight allocation and progress comparison: Based on the progress impact factors of each system in the multidimensional progress feature matrix, an adaptive sliding window algorithm is used to calculate the weight coefficient of each system in real time. The weight coefficient is convolved with the multidimensional progress feature matrix to generate a comprehensive progress deviation cloud map with time series correlation; S3, 3D visualization model construction: Map the comprehensive progress deviation cloud map to a 3D coordinate system, where the X-axis represents the project timeline, the Y-axis represents the correlation between each system, and the Z-axis represents the deviation impact intensity. A visual comparison surface is generated through a spatial interpolation algorithm; S4, interactive intelligent analysis: embeds adjustable threshold parameters in the visual comparison surface, adjusts the time slice range and weight distribution ratio through touch interaction, updates the surface morphology in real time, and outputs critical path warning signals; The S2 specifically includes: S21, schedule impact factor extraction: Separate the core impact factors by domain from the multidimensional schedule feature matrix, including management domain factors, human domain factors, production domain factors, financial domain factors, and supply chain domain factors; S22, Adaptive Sliding Window Algorithm Execution: Initialize the window length based on the total project cycle, with a step size equal to the inverse of the human resource turnover rate. Automatically shorten the window length when it detects that the overall efficiency of production equipment has dropped too high for multiple consecutive window periods. If the supply chain risk level exceeds the risk threshold, a supply chain risk penalty factor is added as the window slides. The supply chain risk level is mapped to a supply chain risk level value in the range [0,10] based on the number of days of supply chain material delay using a piecewise function. S23, convolution operation design: Construct a three-dimensional convolution kernel, whose time dimension matches the adaptive sliding window and whose spatial dimension corresponds to the five-domain feature channels; perform depthwise separable convolution on the dynamic weight coefficient matrix and the multidimensional progress feature matrix, and initialize the convolution kernel weights using the inter-domain correlation strength matrix; S24, comprehensive schedule deviation cloud map generation: The convolution output is time-normalized and feature decoupling techniques are used to separate the deviation and random fluctuation components. The deviation intensity is mapped to the RGB color space, where the red channel represents the dominant deviation in the production domain, the blue channel represents the dominant deviation in the supply chain domain, and the brightness of the green channel is positively correlated with the deviation intensity in the finance domain, to form a comprehensive schedule deviation cloud map. The mapping rules for mapping the comprehensive progress deviation cloud map to the three-dimensional coordinate system include: The X-axis timeline uses a piecewise nonlinear scale, compressing the project plan timeline to a 1:1 ratio. The actual progress timeline dynamically expands and contracts according to the deviation intensity. The expansion factor is calculated as: ;in, is the scaling factor of the time axis, is the normalized deviation intensity of the Z axis from the comprehensive progress deviation cloud chart; The Y-axis correlation calculation uses a cross-domain coupling algorithm: a five-domain correlation matrix is constructed, with basic correlation weights set between the management domain and the production domain, and a dynamic correlation weight set between the supply chain domain and the finance domain as the square root of the budget execution rate. Eigenvectors are extracted through cluster analysis, and the first principal component is projected onto the Y-axis to form a correlation scale for each system. The Z-axis intensity mapping is logarithmic, and the baseline intensity value is set based on historical project progress deviations; The spatial interpolation algorithm includes embedding dynamic constraints in the standard Kriging interpolation process; The generation of the visual comparison surface includes converting the interpolated three-dimensional point cloud data into a NURBS surface, the number of surface control points is taken as the square root of the number of valid task nodes in the current time window, and special constraints are set during the surface parameterization process.

2. A visual analysis method for comparing multi-system progress data of a shipbuilding project according to claim 1, characterized in that: Said S1 specifically includes: S11: Extract task node completion rate, approval process response time, and project milestone deviation from the management system; extract job type work hour saturation, skill matching, and staff turnover rate from the human resources system; extract segment construction completion rate, equipment overall efficiency, and quality inspection pass rate from the production system; extract budget execution rate, payment node achievement rate, and cost overrun coefficient from the financial system; and extract material completeness rate and logistics on-time rate from the supply chain system. S12: Timestamp alignment uses the task node number as the association key to perform a bidirectional match between the production system's segmented construction timestamp and the management system's task node timestamp. When the deviation between the hoisting completion time recorded by the production system and the node deadline set by the management system exceeds the predetermined time, the actual completion time of the corresponding node in the management system is synchronously updated based on the production system timestamp. S13, dimensionality is unified and processed including outlier processing, and finally a standardized multi-dimensional progress feature matrix is generated. The row vectors of the matrix represent time series segments, and the column vectors are arranged according to domain division: management domain, human resources domain, production domain, financial domain, and supply chain domain.

3. A visual analysis method for comparing multi-system progress data of a shipbuilding project according to claim 1, characterized in that: The dimensional unification processing includes converting working hour data into standard man-days, normalizing the budget execution rate to a value in the interval [0, 1], binarizing the equipment status data into a Boolean type, and mapping the number of days of supply chain material delay to a risk level value of [0, 10] through a piecewise function; outlier processing includes performing outlier cleaning on the processed data, setting a task node completion rate threshold range of [0, 1.2], and interpolating data points outside the range using the mean of adjacent time periods.

4. A visual analysis method for comparing multi-system progress data of a shipbuilding project according to claim 1, characterized in that: The supply chain risk penalty factor is calculated as: ;in, is the supply chain risk penalty factor, The number of days the material is delayed. is the risk transmission coefficient, which is determined based on the Pearson correlation coefficient between material delay and schedule deviation in historical data.

5. A visual analysis method for comparing multi-system progress data of a shipbuilding project according to claim 1, characterized in that: The management domain factor is the product of the task node completion rate and the milestone deviation value; the human resources domain factor is the weighted harmonic mean of the work hour saturation and skill matching degree of the job type; the production domain factor is calculated by the geometric mean of the segmented construction completion degree, the equipment comprehensive efficiency and the quality inspection pass rate; the financial domain factor is the ratio of the budget execution rate to the cost overrun coefficient; the supply chain domain factor is based on the product of the material completeness rate and the logistics punctuality rate.

6. A visual analysis method for comparing multi-system progress data of a shipbuilding project according to claim 1, characterized in that: The dynamic constraints include: ① When the comprehensive efficiency of the production domain equipment is lower than the efficiency threshold, add the radial basis function penalty term in the interpolation equation, and the penalty coefficient Calculated as: , represents the overall equipment efficiency of the production domain; ② For areas where the budget execution rate in the financial domain is lower than 70%, an anisotropic interpolation strategy is used to expand the search neighborhood to 2 times the standard deviation along the time axis.

7. A visual analysis method for comparing multi-system progress data of a shipbuilding project according to claim 1, characterized in that: The interactive intelligent analysis in S4 includes embedding adjustable threshold panels at the four vertices of the visual comparison surface, respectively setting the production cycle tolerance threshold, labor saturation threshold, supply chain risk diffusion coefficient and financial budget elasticity coefficient. When the threshold is adjusted, the surface color temperature is driven to change dynamically according to the HSV color circle.

8. A visual analysis method for comparing multi-system progress data of a shipbuilding project according to claim 7, characterized in that: The interactive intelligent analysis also includes adjusting the time slice range and weight distribution ratio based on touch gesture execution, and using incremental recalculation to update the surface morphology in real time. At the same time, it generates a three-level critical path warning and outputs an abnormality report based on the multi-domain deviation intensity and its correlation relationship.

Images