A marine engine test result intelligent evaluation method and system and a storage medium

Abstract

The application discloses a marine engine test result intelligent evaluation method and system and a storage medium, and the method comprises the following steps: collecting multi-dimensional test data of a target marine engine under a preset test scene, analyzing corresponding fuel mode identification and model characteristic parameters, distinguishing diesel and alternative fuel according to the fuel mode, and containing rated speed, stroke type, cylinder arrangement mode and the like in the model parameters. A joint index is constructed according to the fuel mode and the test scene, an evaluation strategy container is matched and positioned in a strategy knowledge graph, dynamic verification threshold values are generated by injecting the model parameters, and the graph stores differentiated calculation templates of multi-scene fuel combinations. The templates are called to solve the characteristics of the test data to obtain performance indicators, and the initial evaluation results are formed by comparing the dynamic threshold values. The final evaluation report containing a graded compliance level is generated by receiving manual interaction correction and confirmation through a visual interface. The application provides reliable technical support for the intelligent evaluation standard of ship power equipment.

Description

Technical Field

[0001] This invention belongs to the field of power equipment testing technology, and more specifically, relates to an intelligent evaluation method, system and storage medium for marine engine test results. Background Technology

[0002] Marine engines, as the core power equipment for ship navigation, directly determine the safety, economic efficiency, and maritime compliance of ships based on their operational performance, safety indicators, and environmental protection levels. With the ongoing low-carbon transformation of the global shipping industry, ship fuel systems are gradually expanding from the traditional single diesel fuel model to alternative fuels such as methanol and liquefied natural gas. Engine models are also showing a diversified development trend, exhibiting multiple speeds, multiple strokes, and multiple cylinder arrangements. Maritime regulations and classification society certification standards are becoming increasingly stringent, placing higher demands on the accuracy of performance evaluation and the standardization of compliance determination for various engine operating conditions.

[0003] Currently, the evaluation of marine engine test results in the industry still mainly relies on traditional manual judgment and fixed threshold comparison, which has many practical shortcomings. Most existing evaluation standards use uniform fixed thresholds, failing to differentiate combustion characteristics across different fuel modes and unable to adapt to the unique characteristics of various engine structures and rated operating conditions. This easily leads to problems such as insufficient adaptability of evaluation standards and significant deviations in judgment results. Furthermore, engine testing encompasses multi-dimensional data, including steady-state parameters and dynamic trend parameters. The data volume is enormous, and the operating conditions are complex. Manually sorting and calculating this data is labor-intensive and time-consuming, making it difficult to meet the needs of batch testing and certification.

[0004] Traditional assessment models lack systematic rule coordination capabilities. For specific tests such as shaft vibration, overspeed safety, and carbon emission calculations, there is no unified calculation and judgment logic tailored to different scenarios. The assessment process relies on the subjective judgment of experienced technicians, resulting in insufficient consistency and credibility of the results. Static fixed thresholds cannot adapt to dynamic changes in engine operating conditions, making it difficult to construct time-series evaluation benchmarks and easily overlooking potential performance hazards during dynamic response processes. The overall assessment process is fragmented; data processing, indicator comparison, and grade determination lack integrated intelligent support. Report preparation and compliance grading rely entirely on manual compilation, making it difficult to meet the standardized and intelligent testing and assessment needs of modern marine power equipment. A new intelligent assessment solution is urgently needed to address the industry's existing pain points. Summary of the Invention

[0005] This invention aims to overcome the industry pain points of existing marine engine testing and evaluation, which rely on human experience, have fixed and rigid evaluation standards, and cannot adapt to the differences between multiple fuel modes and engine models. It establishes differentiated calculation logic based on strategy knowledge graph, adaptively generates dynamic verification thresholds, realizes automatic feature calculation and intelligent initial evaluation of test data, and improves evaluation efficiency and result consistency with a visual manual correction mechanism, thus meeting the practical application needs of compliance classification testing and certification of marine power equipment.

[0006] To address the aforementioned deficiencies or improvement needs of existing technologies, as a first aspect of this invention, the present invention provides an intelligent evaluation method for marine engine test results, comprising: S1. Obtain multi-dimensional test data of the target marine engine under a preset test scenario, and parse the fuel mode identifier and engine characteristic parameters corresponding to the test data; the fuel mode identifier is used to distinguish between diesel mode and alternative fuel mode; the engine characteristic parameters include at least rated speed, stroke type and cylinder arrangement. S2. Based on the fuel mode identifier and the preset test scenario, construct a joint index key value, perform address matching in the preset strategy knowledge graph, locate the target evaluation strategy container, and inject the model feature parameters into the target evaluation strategy container to generate a dynamic verification threshold; wherein, the strategy knowledge graph stores differentiated calculation logic templates for different combinations of fuel modes and test scenarios; S3. Call the differentiated calculation logic template in the target evaluation strategy container to perform feature calculation on the multi-dimensional test data to obtain performance evaluation indicators, and compare the performance evaluation indicators with the dynamic verification threshold to generate preliminary evaluation results; S4. In response to the user's interactive instructions, the preliminary assessment results are corrected and confirmed in the visual assessment interface, and a final assessment report containing the graded compliance level is generated based on the corrected results.

[0007] Further, in step S2, a joint index key-value pair is constructed based on the fuel mode identifier and the preset test scenario, and address matching is performed in a preset policy knowledge graph, including: The fuel mode identifier is used as the first-level node of the graph, and the preset test scenario is used as the second-level node. The third-level nodes of the knowledge graph store a set of parameterized rules associated with the model's feature parameters; The addressing and matching process specifically involves: traversing the knowledge graph based on the joint index key value, locating the corresponding third-level node, and extracting the basic calculation formula and benchmark threshold coefficient from the parameterized rule set.

[0008] Furthermore, the generation of the dynamic verification threshold in S2 includes: The target evaluation strategy container is parsed to identify the parameter type of the physical quantity to be evaluated, including steady-state scalar parameters and dynamic trend parameters. If the parameter type is a steady-state scalar parameter, then obtain the physical quantity calculation formula defined for the current fuel mode identifier. The physical quantity calculation formula includes basic physical parameter placeholders and correction coefficient placeholders related to the fuel mode. From the engine feature parameters, extract the rated operating condition parameters corresponding to the basic physical parameter placeholders, and extract the fuel characteristic factors corresponding to the correction coefficient placeholders. Use the rated operating condition parameters to determine the basic threshold benchmark, and perform nonlinear correction calculation on the basic threshold benchmark according to the fuel characteristic factors to obtain the dynamic verification threshold. If the parameter type is a dynamic trend parameter, then a preset standard characteristic curve model and tolerance band algorithm are obtained; the inertial time constant reflecting the dynamic response characteristics is extracted from the aircraft characteristic parameters; the standard characteristic curve model is subjected to time-domain scaling transformation to match the response speed of the current aircraft model; and the dynamic tolerance coefficient is calculated according to the fuel characteristic factor. Based on the transformed standard characteristic curve model and the dynamic tolerance coefficient, a threshold envelope that changes with time is constructed, and the threshold envelope is used as the dynamic verification threshold.

[0009] Furthermore, the differentiated calculation logic template in S2 includes: performing environmental protection calculations using a modified carbon emission factor under the alternative fuel mode, specifically including the following steps: Obtain the baseline carbon emission factor for current alternative fuels Combustion efficiency correction factor and fuel carbon content correction factor ; Using the combustion efficiency correction coefficient And fuel carbon content correction factor For the benchmark carbon emission factor The correction is made to obtain the corrected carbon emission factor. The calculation formula is: ; Wherein, the combustion efficiency correction coefficient The calculation formula is: In the formula, To measure the concentration of unburned hydrocarbons, This represents the concentration of products from theoretically complete combustion. The fuel carbon content correction factor The calculation formula is: In the formula, and These represent the mass of carbon and the total mass of the current alternative fuels, respectively. and These are the mass of carbon and the total mass of standard diesel fuel, respectively; based on the aforementioned modified carbon emission factor... Calculate the carbon emissions per unit power under the current operating conditions to complete the environmental impact assessment.

[0010] Furthermore, the differentiated calculation logic template in S2 also includes resonance determination logic for shaft vibration tests: The shaft vibration signal in the multi-dimensional test data is subjected to Fast Fourier Transform to extract frequency domain features; Identify the main harmonic components in the frequency domain features and calculate their corresponding torsional stress amplitudes; The torsional vibration stress amplitude is compared with the allowable value in the dynamic verification threshold. If the torsional vibration stress amplitude is less than the allowable value and there is no resonance peak in the spectrum, the inertial component is deemed to be qualified.

[0011] Furthermore, the differentiated calculation logic template in S2 also includes safety determination logic for overspeed tests: Monitor the speed response curve after the overspeed trigger moment to identify the starting point of the speed drop trend; Simultaneously monitor digital signals of fuel supply status to identify the trigger point of the fuel cut-off command; Calculate the time difference between the starting point and the trigger time. If the time difference is less than a preset response delay threshold and the peak rotation speed does not exceed the mechanical limit, then generate a qualified safety logic response indicator signal.

[0012] Furthermore, in step S3, the differentiated calculation logic template in the target evaluation strategy container is invoked to perform feature decomposition on the multi-dimensional test data to obtain performance evaluation metrics, specifically including: Based on the parameter types of the multi-dimensional test data, the corresponding data preprocessing logic is matched: if the parameter type is a steady-state scalar parameter, then the mean filtering algorithm is used to remove random noise and extract steady-state feature values; if the parameter type is a dynamic trend parameter, then the sliding window difference algorithm is used to extract the rate of change feature and obtain the dynamic trend feature value. The steady-state characteristic value or the dynamic trend characteristic value is input into the differentiated calculation logic template, and the corresponding calculation logic branch is called to perform feature calculation: if the input is a steady-state characteristic value, the steady-state performance evaluation index is calculated using the energy balance algorithm based on the first law of thermodynamics; if the input is a dynamic trend characteristic value, the transient performance evaluation index is calculated using the dynamic response algorithm based on time series analysis. The steady-state performance evaluation index or the transient performance evaluation index is output as the final performance evaluation index.

[0013] Furthermore, in step S4, in response to user interaction commands, the preliminary evaluation results are corrected and confirmed in the visual evaluation interface, including: The performance evaluation index that fails the comparison results is highlighted in the visual evaluation interface, and the original test data fragment that caused the failure is displayed. Receive correction parameters or exemption reasons input by the user, re-execute the comparison step in S3 based on the correction parameters, or update the compliance level remarks information in the final assessment report based on the exemption reasons.

[0014] As a second aspect of the present invention, an intelligent evaluation system for marine engine test results is also provided, comprising: The data acquisition and parameter parsing unit is used to acquire multi-dimensional test data of the target marine engine under a preset test scenario, and parse the fuel mode identifier and engine characteristic parameters corresponding to the test data; the fuel mode identifier is used to distinguish between diesel mode and alternative fuel mode; the engine characteristic parameters include at least rated speed, stroke type and cylinder arrangement. The strategy matching and threshold generation unit is used to construct a joint index key value based on the fuel mode identifier and the preset test scenario, perform address matching in the preset strategy knowledge graph, locate the target evaluation strategy container, and inject the model feature parameters into the target evaluation strategy container to generate a dynamic verification threshold; wherein, the strategy knowledge graph stores differentiated calculation logic templates for different combinations of fuel modes and test scenarios; The feature calculation and comparison evaluation unit is used to call the differentiated calculation logic template in the target evaluation strategy container to perform feature calculation on the multi-dimensional test data, obtain performance evaluation indicators, and compare the performance evaluation indicators with the dynamic verification threshold to generate preliminary evaluation results. The report generation unit is used to respond to user interaction commands, correct and confirm the preliminary assessment results in the visual assessment interface, and generate a final assessment report containing the graded compliance level based on the corrected results.

[0015] As a third aspect of the invention, a computer-readable storage medium is also provided, on which a computer program is stored, which is executed by a processor as described in any one of the claims: an intelligent evaluation method for marine engine test results.

[0016] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects: 1. The intelligent evaluation method for marine engine test results of the present invention acquires multi-dimensional test data of the target marine engine under a preset test scenario, simultaneously analyzes the fuel mode identifier and engine model characteristic parameters corresponding to the test data, distinguishes between diesel mode and alternative fuel mode based on the fuel mode identifier, and characterizes the inherent attributes of the engine model based on parameters such as rated speed, stroke type, and cylinder arrangement. This method constructs a joint index key-value pair between the fuel mode identifier and the preset test scenario, completes addressing matching and locates the target evaluation strategy container in a preset strategy knowledge graph, and uses the knowledge graph to store differentiated calculation logic templates under different combinations. It can automatically match exclusive evaluation rules according to actual test conditions without manual setting of evaluation logic, achieving adaptation of evaluation rules under multiple fuel types, multiple test scenarios, and multiple engine models. This fundamentally solves the problems of arbitrary rule matching and poor adaptability in traditional evaluation methods, laying the foundation for subsequent threshold generation and index evaluation.

[0017] 2. The intelligent evaluation method for marine engine test results of the present invention injects engine model characteristic parameters into the target evaluation strategy container and generates dynamic verification thresholds according to the classification logic of steady-state scalar parameters and dynamic trend parameters. For steady-state scalar parameters, a basic threshold benchmark is determined based on rated operating condition parameters, and a fixed threshold value is obtained by performing nonlinear correction calculations in conjunction with fuel characteristic factors. For dynamic trend parameters, the time-domain scaling transformation of the standard characteristic curve is completed using the inertial time constant, and the dynamic tolerance coefficient is calculated in conjunction with the fuel characteristic factors to construct a time-series threshold envelope. This classification-based dynamic threshold generation method abandons the industry-standard unified fixed threshold mode, and can fit the structural characteristics, fuel combustion characteristics, and dynamic operating response laws of different engine models. It allows the evaluation criteria to adaptively adjust with the engine's own parameters and operating status, improving the fit and scientific nature of the test result threshold comparison, and avoiding misjudgments and omissions caused by fixed thresholds.

[0018] 3. The intelligent evaluation method for marine engine test results of this invention calls the differentiated calculation logic template within the target evaluation strategy container, selects the corresponding preprocessing method according to parameter type to complete the noise reduction and feature extraction of test data, and uses energy balance algorithm and time series analysis algorithm to solve steady-state and transient performance evaluation indicators respectively. The solved evaluation indicators are automatically compared with dynamic verification thresholds to generate preliminary evaluation results. At the same time, it relies on a visual evaluation interface to receive user interaction commands, highlights non-compliant indicators, and supports manual correction of parameters and supplementation of exemption reasons. Finally, an evaluation report with graded compliance level is generated based on the correction results. This method realizes the fully automated processing of test data calculation and indicator comparison, greatly reduces the workload of manual data verification and subjective judgment. The human-machine collaborative correction mode combines the efficiency of intelligent evaluation with the advantages of professional human verification, making the evaluation results more rigorous and standardized, and can also form standardized and traceable compliance evaluation documents. Attached Figure Description

[0019] Figure 1 This is a flowchart of an intelligent evaluation method for marine engine test results according to an embodiment of the present invention; Figure 2 This is a visual interface illustration of an embodiment of the present invention. Figure 1 ; Figure 3 This is a visual interface illustration of an embodiment of the present invention. Figure 2 ; Figure 4 This is a schematic diagram of the system units in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0021] Example 1 Please refer to Figure 1 This embodiment 1 provides an intelligent evaluation method for marine engine test results, including: S1. Obtain multi-dimensional test data of the target marine engine under a preset test scenario, and parse the fuel mode identifier and engine characteristic parameters corresponding to the test data; the fuel mode identifier is used to distinguish between diesel mode and alternative fuel mode; the engine characteristic parameters include at least rated speed, stroke type and cylinder arrangement. S2. Based on the fuel mode identifier and the preset test scenario, construct a joint index key value, perform address matching in the preset strategy knowledge graph, locate the target evaluation strategy container, and inject the model feature parameters into the target evaluation strategy container to generate a dynamic verification threshold; wherein, the strategy knowledge graph stores differentiated calculation logic templates for different combinations of fuel modes and test scenarios; S3. Call the differentiated calculation logic template in the target evaluation strategy container to perform feature calculation on the multi-dimensional test data to obtain performance evaluation indicators, and compare the performance evaluation indicators with the dynamic verification threshold to generate preliminary evaluation results; S4. In response to the user's interactive instructions, the preliminary assessment results are corrected and confirmed in the visual assessment interface, and a final assessment report containing the graded compliance level is generated based on the corrected results.

[0022] This embodiment 1 further elaborates on the above steps.

[0023] (1) Data acquisition and parameter analysis To ensure that the evaluation results are consistent with actual working conditions, it is necessary to first set up a standardized test environment for the target marine engine to be tested, based on industry standards, to establish a foundation for subsequent data collection.

[0024] During the marine engine test data acquisition phase, the system first connects to the engine integrated test and control bench via an industrial Ethernet interface to acquire multi-dimensional test data of the target engine under preset test scenarios in real time. This data covers four main categories: combustion parameters, thermodynamic parameters, mechanical parameters, and emission parameters, specifically including key indicators such as cylinder combustion pressure, exhaust temperature, coolant temperature, lubricating oil pressure, fuel consumption, and NOx emission concentration. The data acquisition frequency is dynamically adjusted according to the parameter type; steady-state parameters are sampled at 1Hz, while dynamic parameters such as pressure fluctuations are sampled at a high frequency of 10kHz to ensure data integrity.

[0025] After acquiring the raw data, the fuel mode identifier parsing module is immediately started. This module determines the fuel type through a dual verification mechanism. The specific process is as follows: First, the fuel selection register status of the engine ECU is read. If the register value is 0x01, it is determined to be diesel mode; if it is 0x02, it is determined to be alternative fuel mode. At the same time, the results of intake air composition spectral analysis are cross-validated to ensure the accuracy of the identification. For example, in one test, the ECU register value was detected as 0x02 and the spectral analysis showed methane characteristic peaks, which ultimately confirmed that the fuel mode was LNG alternative fuel mode.

[0026] Subsequently, the engine characteristic parameter analysis module extracts core parameters such as rated speed, stroke type, and cylinder arrangement by parsing the XML metadata file in the test task sheet. For the cylinder arrangement, a geometric topology analysis algorithm is used to read the CAD data of the engine 3D model and automatically identify the cylinder arrangement by calculating the spatial coordinate relationship of the center lines of each cylinder, providing geometric information for subsequent vibration modal analysis.

[0027] Finally, all parsed parameters are encapsulated into a JSON-formatted "test metadata object," which contains key fields such as fuel mode identifier, aircraft characteristic parameters, test scenario identifier, and timestamp. This object is then injected into the evaluation process context, laying the foundation for subsequent policy matching.

[0028] (2) Strategy matching and threshold generation After determining the fuel mode identifier and the preset test scenario type, a joint index key value is constructed to carry out traversal addressing and matching in the pre-built strategy knowledge graph.

[0029] The strategy knowledge graph includes pre-defined templates for various differentiated computational logics, each tailored to the evaluation requirements of different test projects. For example, in the alternative fuel model, the environmental impact calculation uses a modified carbon emission factor, specifically including the following steps: Obtain the baseline carbon emission factor for current alternative fuels Combustion efficiency correction factor and fuel carbon content correction factor ; Using the combustion efficiency correction coefficient And fuel carbon content correction factor For the benchmark carbon emission factor The correction is made to obtain the corrected carbon emission factor. The calculation formula is: ; Wherein, the combustion efficiency correction coefficient The calculation formula is: In the formula, To measure the concentration of unburned hydrocarbons, This represents the concentration of products from theoretically complete combustion. The fuel carbon content correction factor The calculation formula is: In the formula, and These represent the mass of carbon and the total mass of the current alternative fuels, respectively. and These are the mass of carbon and the total mass of standard diesel fuel, respectively; based on the aforementioned modified carbon emission factor... Calculate the carbon emissions per unit power under the current operating conditions to complete the environmental impact assessment.

[0030] Meanwhile, the resonance determination logic for shaft vibration test includes the following execution steps: performing fast Fourier transform on the shaft vibration signal in the multi-dimensional test data to extract frequency domain features; Identify the main harmonic components in the frequency domain features and calculate their corresponding torsional stress amplitudes; The torsional vibration stress amplitude is compared with the allowable value in the dynamic verification threshold. If the torsional vibration stress amplitude is less than the allowable value and there is no resonance peak in the spectrum, the inertial component is deemed to be qualified.

[0031] In addition, the safety determination logic for overspeed tests includes the following execution steps: Monitor the speed response curve after the overspeed trigger moment to identify the starting point of the speed drop trend; Simultaneously monitor digital signals of fuel supply status to identify the trigger point of the fuel cut-off command; Calculate the time difference between the starting point and the trigger time. If the time difference is less than a preset response delay threshold and the peak rotation speed does not exceed the mechanical limit, then generate a qualified safety logic response indicator signal.

[0032] The above three judgment logics are merely individual calculation logic templates based on these three scenarios. For other scenarios, corresponding calculation logic templates need to be introduced. In specific implementation, this will be set by technical personnel according to requirements. Meanwhile, the knowledge graph adopts a three-level node architecture design, with the fuel mode identifier as the first-level node and the preset test scenario as the second-level node. The third-level node of the knowledge graph stores a set of parameterized rules associated with the engine model's feature parameters. By traversing the knowledge graph according to the joint index key value, locating the corresponding third-level node, and extracting the basic calculation formulas and benchmark threshold coefficients from the parameterized rule set, a dedicated evaluation strategy container adapted to the current engine operating condition and engine model is locked.

[0033] After locating the target evaluation strategy container, the parsed engine characteristic parameters are imported into the container to construct the dynamic verification threshold. In the implementation of the intelligent evaluation method for marine engine test results, the steps for generating the dynamic verification threshold first include parameter type identification and logical branch selection. The identifier of the target physical quantity to be evaluated is obtained, such as exhaust temperature or turbocharger speed, and the target evaluation strategy container is accessed to read the corresponding metadata tag and determine whether its attribute is steady-state scalar or dynamic trend type. If it is determined to be steady-state scalar, numerical calculation logic based on nonlinear correction is executed; if it is determined to be dynamic trend type, trajectory generation logic based on characteristic curve mapping is executed.

[0034] When the target physical quantity is a steady-state scalar, numerical calculation logic based on nonlinear correction is executed. Placeholders containing fundamental physical parameters are retrieved from the target evaluation strategy container. and correction coefficient placeholders The formula for calculating the physical quantity. Extract the rated operating parameters of this parameter under standard conditions from the input set of machine model characteristic parameters. This serves as the baseline threshold. Simultaneously, based on the current fuel mode identifier, the corresponding fuel characteristic factors are retrieved and extracted. The fuel characteristic factor is used to perform a nonlinear correction calculation on the basic threshold benchmark to generate the final single numerical threshold. Its mathematical expression is: ; When the target physical quantity exhibits a dynamic trend, trajectory generation logic based on characteristic curve mapping is executed. The baseline characteristic curve function is obtained from the target evaluation strategy container. And a dynamic tolerance algorithm. The inertial time constant, reflecting the engine's dynamic response characteristics, is extracted from the engine's characteristic parameters. The constant is then used to perform a time-domain scaling transformation on the baseline characteristic curve function to obtain a transformed curve that matches the response speed of the current model. Its expression is: ; Subsequently, based on fuel characteristic factors and inertial time constant Calculate dynamic tolerance coefficient The dynamic tolerance coefficient is calculated based on a correlation model between fuel characteristic factors and dynamic response characteristics, specifically: ;in, This is the tolerance ratio, which is preset by the evaluation strategy container.

[0035] Finally, based on the transformed baseline curve and dynamic tolerance coefficient Construct a dynamic verification threshold envelope that varies over time, with its upper boundary... and lower boundary They are respectively: The envelope of this dynamic verification threshold is the final output dynamic verification threshold, which is used for real-time comparison and anomaly detection of the test data.

[0036] (3) Feature calculation and comparison evaluation After completing the matching of the evaluation strategy container and the generation of dynamic verification thresholds, all multi-dimensional test data are first classified and distinguished according to the inherent attributes of the parameters, dividing them into two major categories: steady-state scalar parameters and dynamic trend parameters. Then, appropriate data preprocessing methods are matched for different categories of parameters.

[0037] For the collected steady-state scalar parameters, such as exhaust temperature, coolant temperature, and lubricating oil pressure, a mean filtering algorithm is used for preprocessing. First, a time window is set. (For example, 60 seconds). Then, within this window, the sampling sequence... The arithmetic mean is calculated using the following formula: This removes random noise from the sensor and extracts steady-state characteristic values ​​that represent the current operating conditions.

[0038] For dynamic trend parameters, such as cylinder combustion pressure, crankshaft instantaneous speed, and load change rate, a sliding window difference algorithm is used. First, the sliding step size is set to... Next, the ratio of the difference in feature values ​​within adjacent windows to the time step size is calculated, i.e. This allows us to extract the rate characteristics of parameter changes over time, obtain dynamic trend characteristic values, and capture the transient response characteristics of the engine under varying operating conditions.

[0039] After data preprocessing, the extracted feature values ​​are input into the differential calculation logic template, and different calculation branches are automatically invoked based on the feature type. If the input is a steady-state feature value, the energy balance algorithm branch based on the first law of thermodynamics is activated. This algorithm constructs the energy conservation equation within the engine cylinder. By combining fuel consumption and combustion efficiency parameters, steady-state performance evaluation indicators such as indicated thermal efficiency and effective fuel consumption rate are calculated. For example, indicated work is calculated using the measured combustion pressure product, and mechanical efficiency is calculated by combining it with the output shaft torque, thereby quantitatively evaluating the engine's economy during stable operation.

[0040] If the input is a dynamic trend characteristic value, the algorithm switches to the dynamic response algorithm branch based on time series analysis. This branch uses an autoregressive moving average model to model the pressure fluctuation sequence and calculates indicators such as dynamic response delay time, overshoot, and settling time. For example, during the process of an engine suddenly increasing from 50% load to 90% load, transient performance evaluation indicators are calculated by analyzing the speed drop curve and recovery time to assess the sensitivity and combustion stability of the speed control system.

[0041] Finally, the calculated steady-state or transient performance evaluation indicators are standardized and encapsulated, and compared with the dynamic verification thresholds preset in the strategy container. If the indicator falls within the threshold range, a preliminary evaluation result of "qualified" is generated; if it exceeds the threshold, anomalies are marked and an early warning report is generated, completing the intelligent graded evaluation of the target engine performance.

[0042] (4) Report generation After the initial assessment is completed, the report generation stage begins. This stage responds to user interaction commands, allowing for the correction and confirmation of the initial assessment results within a visual assessment interface. Finally, an assessment report containing the tiered compliance levels is generated. The specific process is as follows: First, the preliminary comparison results of all performance evaluation indicators are displayed in a dashboard format on the front-end visual evaluation interface. For indicators that fail the comparison, such as "NOx emissions exceeding the standard" or "transient speed recovery time being too long," a red highlight border is used to warn the user, and the original test data segment that caused the indicator to fail is automatically displayed. For example, when "NOx emissions exceeding the standard" is highlighted, the exhaust temperature curve, air-fuel ratio fluctuation curve, and fuel injection quantity data for the corresponding time period simultaneously pop up at the bottom of the interface, allowing users to intuitively trace the root cause of the anomaly.

[0043] Subsequently, the system receives correction parameters or exemption reasons input by the user. If the user believes that the data anomaly is caused by a sensor malfunction, they can input correction parameters, such as "reduce the NOx sensor reading by 5%". Upon receiving this instruction, the system automatically re-executes the comparison and evaluation, that is, it recalculates the performance evaluation indicators based on the corrected data and compares them again with the dynamic verification threshold to update the evaluation results.

[0044] If a user confirms that the data is authentic and valid, but the indicator temporarily exceeds the limit due to special operating conditions (such as extreme sea conditions), they can enter an exemption reason, such as "Encountered level 8 winds and waves during the test, which constitutes force majeure." This reason will be recorded as a compliance level note, and the compliance level of the indicator will be marked as "conditionally compliant" in the final assessment report. At the same time, the reason for the exemption will be noted in the report's remarks column to ensure the transparency and traceability of the assessment results.

[0045] Finally, based on the revised or confirmed assessment results, a final assessment report is generated, which includes a tiered compliance level. The report adopts a three-tier compliance level system: "Fully Compliant," "Conditionally Compliant," and "Non-Compliant," and includes detailed data support and notes to provide a basis for performance certification and operation and maintenance decisions for marine engines.

[0046] Meanwhile, the visualization method provided in this embodiment presents complex background evaluation data intuitively to the user through a graphical user interface. For example... Figure 2 As shown, a global monitoring area is built at the top of the interface, displaying key engine performance indicators in real time in a dashboard format. These indicators cover power performance and thermodynamic parameters, specifically including engine speed (e.g., 53.26 rpm), engine power (2389.7 kW), fuel and alternative fuel consumption rates, exhaust temperature and pressure, and the temperature, pressure, and flow rate of boost air. This layout allows operators to clearly understand the engine's current operating status and quickly identify any abnormalities such as overheating or overpressure.

[0047] The lower middle section of the interface focuses on controlling the test scenario and displaying object information. The left-hand area provides a "Test Command" input interface, allowing users to set the engine speed and load, and issue test commands, enabling interactive control of the testing process. The right-hand area lists detailed metadata about the tested object, such as object name (Dalian Maritime_Two-Stroke Methanol Engine Model), specific model (6G50ME), rated power (6400kW), and cylinder bore, among other physical parameters. This area also displays a 3D model or structural diagram of the engine, clarifying the current "Test Scenario" configuration. This corresponds to the strategy knowledge graph matching logic mentioned earlier, ensuring users understand the engine type and operating condition background upon which the current evaluation is based.

[0048] To provide in-depth analysis of dynamic performance, this visualization method also offers trend analysis capabilities, such as... Figure 3 As shown, the bottom of the interface displays graphs showing how parameters change over time or under different operating conditions, such as the trend of engine speed. These waveforms allow evaluators to visually observe parameter fluctuations in steady state or response speed under varying operating conditions, thus aiding in the assessment of engine dynamic stability. This visualization solution, combining static numerical monitoring with dynamic trend analysis, significantly improves the convenience and accuracy of marine engine testing and evaluation.

[0049] Example 2 Please refer to Figure 4 This embodiment 2 provides an intelligent evaluation system for marine engine test results, including: The data acquisition and parameter parsing unit is used to acquire multi-dimensional test data of the target marine engine under a preset test scenario, and parse the fuel mode identifier and engine characteristic parameters corresponding to the test data; the fuel mode identifier is used to distinguish between diesel mode and alternative fuel mode; the engine characteristic parameters include at least rated speed, stroke type and cylinder arrangement. The strategy matching and threshold generation unit is used to construct a joint index key value based on the fuel mode identifier and the preset test scenario, perform address matching in the preset strategy knowledge graph, locate the target evaluation strategy container, and inject the model feature parameters into the target evaluation strategy container to generate a dynamic verification threshold; wherein, the strategy knowledge graph stores differentiated calculation logic templates for different combinations of fuel modes and test scenarios; The feature calculation and comparison evaluation unit is used to call the differentiated calculation logic template in the target evaluation strategy container to perform feature calculation on the multi-dimensional test data, obtain performance evaluation indicators, and compare the performance evaluation indicators with the dynamic verification threshold to generate preliminary evaluation results. The report generation unit is used to respond to user interaction commands, correct and confirm the preliminary assessment results in the visual assessment interface, and generate a final assessment report containing the graded compliance level based on the corrected results.

[0050] Example 3 This embodiment 3 also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement any step of an intelligent evaluation method for marine engine test results.

[0051] The computer-readable storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0052] For a description of the computer-readable storage medium provided in this application, please refer to the above method embodiments; further details will not be repeated here.

[0053] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for intelligent evaluation of marine engine test results, characterized in that, include: S1. Obtain multi-dimensional test data of the target marine engine under a preset test scenario, and parse the fuel mode identifier and engine characteristic parameters corresponding to the test data; the fuel mode identifier is used to distinguish between diesel mode and alternative fuel mode; the engine characteristic parameters include at least rated speed, stroke type and cylinder arrangement. S2. Based on the fuel mode identifier and the preset test scenario, construct a joint index key value, perform address matching in the preset strategy knowledge graph, locate the target evaluation strategy container, and inject the model feature parameters into the target evaluation strategy container to generate a dynamic verification threshold; wherein, the strategy knowledge graph stores differentiated calculation logic templates for different combinations of fuel modes and test scenarios; S3. Call the differentiated calculation logic template in the target evaluation strategy container to perform feature calculation on the multi-dimensional test data to obtain performance evaluation indicators, and compare the performance evaluation indicators with the dynamic verification threshold to generate preliminary evaluation results; S4. In response to the user's interactive instructions, the preliminary assessment results are corrected and confirmed in the visual assessment interface, and a final assessment report containing the graded compliance level is generated based on the corrected results.

2. The intelligent evaluation method for marine engine test results according to claim 1, characterized in that, In step S2, a joint index key-value pair is constructed based on the fuel mode identifier and the preset test scenario, and address matching is performed in the preset policy knowledge graph, including: The fuel mode identifier is used as the first-level node of the graph, and the preset test scenario is used as the second-level node. The third-level nodes of the knowledge graph store a set of parameterized rules associated with the model's feature parameters; The addressing and matching process specifically involves: traversing the knowledge graph based on the joint index key value, locating the corresponding third-level node, and extracting the basic calculation formula and benchmark threshold coefficient from the parameterized rule set.

3. The intelligent evaluation method for marine engine test results according to claim 1, characterized in that, The dynamic verification threshold generated in S2 includes: The target evaluation strategy container is parsed to identify the parameter type of the physical quantity to be evaluated, including steady-state scalar parameters and dynamic trend parameters. If the parameter type is a steady-state scalar parameter, then obtain the physical quantity calculation formula defined for the current fuel mode identifier. The physical quantity calculation formula includes basic physical parameter placeholders and correction coefficient placeholders related to the fuel mode. From the engine feature parameters, extract the rated operating condition parameters corresponding to the basic physical parameter placeholders, and extract the fuel characteristic factors corresponding to the correction coefficient placeholders. Use the rated operating condition parameters to determine the basic threshold benchmark, and perform nonlinear correction calculation on the basic threshold benchmark according to the fuel characteristic factors to obtain the dynamic verification threshold. If the parameter type is a dynamic trend parameter, then a preset standard characteristic curve model and tolerance band algorithm are obtained; the inertial time constant reflecting the dynamic response characteristics is extracted from the aircraft characteristic parameters; the standard characteristic curve model is subjected to time-domain scaling transformation to match the response speed of the current aircraft model; and the dynamic tolerance coefficient is calculated according to the fuel characteristic factor. Based on the transformed standard characteristic curve model and the dynamic tolerance coefficient, a threshold envelope that changes with time is constructed, and the threshold envelope is used as the dynamic verification threshold.

4. The intelligent evaluation method for marine engine test results according to claim 1, characterized in that, The differentiated calculation logic template in S2 includes: using a modified carbon emission factor for environmental protection calculations under the alternative fuel mode, specifically including the following steps: Obtain the baseline carbon emission factor for current alternative fuels Combustion efficiency correction factor and fuel carbon content correction factor ; Using the combustion efficiency correction coefficient And fuel carbon content correction factor For the benchmark carbon emission factor The correction is made to obtain the corrected carbon emission factor. The calculation formula is: ; Wherein, the combustion efficiency correction coefficient The calculation formula is: In the formula, To measure the concentration of unburned hydrocarbons, This represents the concentration of products from theoretically complete combustion. The fuel carbon content correction factor The calculation formula is: In the formula, and These represent the mass of carbon and the total mass of the current alternative fuels, respectively. and These are the mass of carbon and the total mass of standard diesel fuel, respectively; based on the aforementioned modified carbon emission factor... Calculate the carbon emissions per unit power under the current operating conditions to complete the environmental impact assessment.

5. The intelligent evaluation method for marine engine test results according to claim 1, characterized in that, The differentiated calculation logic template in S2 also includes resonance determination logic for shaft vibration tests: The shaft vibration signal in the multi-dimensional test data is subjected to Fast Fourier Transform to extract frequency domain features; Identify the main harmonic components in the frequency domain features and calculate their corresponding torsional stress amplitudes; The torsional vibration stress amplitude is compared with the allowable value in the dynamic verification threshold. If the torsional vibration stress amplitude is less than the allowable value and there is no resonance peak in the spectrum, the inertial component is deemed to be qualified.

6. The intelligent evaluation method for marine engine test results according to claim 1, characterized in that, The differentiated calculation logic template in S2 also includes safety determination logic for overspeed tests: Monitor the speed response curve after the overspeed trigger moment to identify the starting point of the speed drop trend; Simultaneously monitor digital signals of fuel supply status to identify the trigger point of the fuel cut-off command; Calculate the time difference between the starting point and the trigger time. If the time difference is less than a preset response delay threshold and the peak rotation speed does not exceed the mechanical limit, then generate a qualified safety logic response indicator signal.

7. The intelligent evaluation method for marine engine test results according to claim 1, characterized in that, In step S3, the differential calculation logic template in the target evaluation strategy container is called to perform feature decomposition on the multi-dimensional test data to obtain performance evaluation indicators, specifically including: Based on the parameter types of the multi-dimensional test data, the corresponding data preprocessing logic is matched: if the parameter type is a steady-state scalar parameter, then the mean filtering algorithm is used to remove random noise and extract steady-state feature values; if the parameter type is a dynamic trend parameter, then the sliding window difference algorithm is used to extract the rate of change feature and obtain the dynamic trend feature value. The steady-state characteristic value or the dynamic trend characteristic value is input into the differentiated calculation logic template, and the corresponding calculation logic branch is called to perform feature calculation: if the input is a steady-state characteristic value, the steady-state performance evaluation index is calculated using the energy balance algorithm based on the first law of thermodynamics; if the input is a dynamic trend characteristic value, the transient performance evaluation index is calculated using the dynamic response algorithm based on time series analysis. The steady-state performance evaluation index or the transient performance evaluation index is output as the final performance evaluation index.

8. The intelligent evaluation method for marine engine test results according to claim 1, characterized in that, In step S4, in response to user interaction commands, the preliminary evaluation results are corrected and confirmed in the visual evaluation interface, including: The performance evaluation index that fails the comparison results is highlighted in the visual evaluation interface, and the original test data fragment that caused the failure is displayed. Receive correction parameters or exemption reasons input by the user, re-execute the comparison step in S3 based on the correction parameters, or update the compliance level remarks information in the final assessment report based on the exemption reasons.

9. An intelligent evaluation system for marine engine test results, characterized in that, include: The data acquisition and parameter parsing unit is used to acquire multi-dimensional test data of the target marine engine under a preset test scenario, and parse the fuel mode identifier and engine characteristic parameters corresponding to the test data; the fuel mode identifier is used to distinguish between diesel mode and alternative fuel mode; the engine characteristic parameters include at least rated speed, stroke type and cylinder arrangement. The strategy matching and threshold generation unit is used to construct a joint index key value based on the fuel mode identifier and the preset test scenario, perform address matching in the preset strategy knowledge graph, locate the target evaluation strategy container, and inject the model feature parameters into the target evaluation strategy container to generate a dynamic verification threshold; wherein, the strategy knowledge graph stores differentiated calculation logic templates for different combinations of fuel modes and test scenarios; The feature calculation and comparison evaluation unit is used to call the differentiated calculation logic template in the target evaluation strategy container to perform feature calculation on the multi-dimensional test data, obtain performance evaluation indicators, and compare the performance evaluation indicators with the dynamic verification threshold to generate preliminary evaluation results. The report generation unit is used to respond to user interaction commands, correct and confirm the preliminary assessment results in the visual assessment interface, and generate a final assessment report containing the graded compliance level based on the corrected results.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program is executed by a processor as described in any one of claims 1-8: a method for intelligent evaluation of marine engine test results.

Images