A method for online rapid health assessment of a diesel engine for power generation

Abstract

The application discloses a kind of diesel engine online fast health evaluation methods for power generation, first, determine the health state classification basis of diesel generator unit, the health state of diesel generator unit is divided into multiple levels;Second, the health state evaluation parameter selection principle of diesel generator unit is established, and the appropriate evaluation parameter is selected for different systems;Then, the health benchmark model of diesel generator unit is established, the model is trained using the monitoring data under health state, and the health degradation index is designed based on Mahalanobis distance, the index of different numerical range is corresponded with health state grade;Finally, for fault state, its fault type is determined based on association rule classification.This method is based on the health benchmark model of diesel generator unit, and the distance between the state characteristic vector of diesel generator unit in abnormal state and the standard state characteristic vector output by health benchmark model can be evaluated by measuring the health degree of running state, and the greater the distance, the farther the running state deviates from the health state.

Description

Technical Field

[0001] This invention relates to the field of ship power system automation and equipment condition monitoring technology, and more specifically, to a method for online rapid health assessment of diesel engines used for power generation. Background Technology

[0002] Marine diesel engines, as a crucial component of marine propulsion systems, provide power output for ships. Therefore, improving the performance of marine diesel engines is key to the healthy development of the shipbuilding industry. Once a marine diesel engine fails, the entire ship cannot operate normally. Many factors can cause marine diesel engine failures, such as excessive wear of a single component or overuse of the entire engine. The internal structure of marine diesel engines is complex, resulting in numerous unstable factors and making management and maintenance challenging. Currently, the operation and maintenance management of marine diesel engines is still primarily based on reactive repairs and periodic inspections. Typically, marine diesel engine failures have certain warning signs; the occurrence of a small fault inevitably leads to a complete engine failure. Therefore, it is necessary to introduce an online rapid health assessment method for diesel engines. Health status assessment uses sensors to acquire information reflecting the equipment's operating status, extracting effective features to assess the equipment's health condition. Diesel generator unit health assessment is the foundation for fault diagnosis and condition-based maintenance, effectively improving operational safety, reliability, and economy.

[0003] Among existing related patent technologies, patent document (CN113464274A) discloses a diesel generator health status assessment system and method. This system divides the generator lifecycle into multiple stages and matches operating data with a preset "fault type subset" using rules. If a match is successful, a preset score is deducted to calculate the health score. This technology has the following technical problems: it is essentially a "static rule matching + deduction" mechanism, which cannot reflect the continuous and gradual degradation process of the performance of complex diesel engine subsystems (such as combustion and turbocharging coupling); moreover, the scoring mechanism is crude and cannot accurately quantify the actual physical distance of the equipment from the health benchmark, exhibiting extremely poor ability to capture minor, gradual faults. For example, patent document (CN112380641B) discloses an emergency diesel engine health status evaluation method. This method uses a fault tree model to obtain the minimum cut set and importance weights, classifies characteristic parameters into performance degradation type and life loss type, calculates the degree of degradation, and then weights and sums the results to obtain the health score. This technology has the following technical problems: First, it relies heavily on expert experience and prior knowledge of fault trees, making it difficult to exhaust all fault modes for complex coupled systems; second, its degradation degree calculation is based on a linear formula for a single parameter (such as the ratio of normal value to alarm value), severing the covariance relationship between multidimensional parameters, making it susceptible to interference from fluctuations in a single parameter and resulting in low evaluation accuracy; finally, it does not solve the problem of sensitivity decay of health indicators in the severe fault range, causing the health value to change gradually when the equipment is extremely deteriorated, making it impossible to accurately trigger high-level alarms.

[0004] To address the technical problems in existing technologies, such as the inability to accurately reflect the degradation of multi-physics coupling, the low evaluation accuracy caused by independent parameter weighting, and the loss of sensitivity of health indicators under extreme faults, this invention proposes an online rapid health assessment method for diesel generators. Summary of the Invention

[0005] This invention proposes an online rapid health assessment method for diesel generator sets, a core component of marine power systems, to address the technical problems in existing diesel generator health assessments that cannot adapt to the continuous degradation of complex coupled systems and the decrease in the sensitivity of assessment indicators under extreme fault conditions.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A rapid online health assessment method for diesel generator sets includes a health status assessment method for the in-cylinder combustion system and a health status assessment method for the turbocharging system. First, the health status grading criteria for the generator set are determined, classifying the generator set's health status into multiple levels. Second, principles for selecting health status assessment parameters for the generator set are established, selecting appropriate assessment parameters for different systems. Then, a health baseline model for the generator set is established, trained using monitoring data under healthy conditions, and a health degradation index is designed based on Mahalanobis distance, mapping different numerical ranges of the index to health status levels. Finally, for fault states, the fault type is determined based on association rules.

[0008] Furthermore, the health status assessment parameters for the diesel engine's in-cylinder combustion system include: engine speed, injector drive current, injector flow rate, intake pressure, intake temperature, mean indicated pressure (IMEP), combustion start point (CA10), combustion center of gravity (CA50), maximum pressure rise rate, ignition delay period, and combustion duration.

[0009] Furthermore, the health status assessment parameters for the diesel engine turbocharger system include engine speed, torque, compressor inlet pressure, turbine outlet pressure, turbine inlet temperature, compressor inlet temperature, bypass valve opening, and turbocharger efficiency.

[0010] Furthermore, the health status of diesel generator units is divided into multiple levels: based on the actual operating conditions of the diesel generator units, their health status is divided into five levels from high to low: "healthy", "good", "average", "deteriorating", and "faulty".

[0011] Furthermore, the selection principles for the health status assessment parameters of diesel generator units are as follows: taking into account personnel, equipment, and environmental factors, the selection principles are based on the sensitivity, independence, variability, standardization, representativeness, and accessibility of the parameters.

[0012] Furthermore, a health decline index is designed based on Mahalanobis distance, and the index with different numerical ranges is mapped to the health status level. Specifically, based on the Mahalanobis distance algorithm, the Mahalanobis distance between the state feature vector composed of actual monitoring data and the standard state feature vector is calculated as the health decline index. The Mahalanobis distance normalization model is used to map the Mahalanobis distance to the health level HP. The Mahalanobis distance normalization model uses different normalization parameters for calculation in different intervals, with the Mahalanobis distance segmentation point as the boundary.

[0013] Furthermore, the formula for calculating Mahalanobis distance is: (1) In the formula, The Mahalanobis distance; This represents a vector, i.e., a sample to be evaluated; For sample set The mean, of which ; Let be the covariance matrix of the independent variables of the sample set.

[0014] Furthermore, Mahalanobis distance is normalized to health score. The calculation formula is as follows: (2) In the formula: For normalization parameters, This is the Mahalanobis distance.

[0015] Furthermore, to ensure that the assessed health status reflects reality, a two-segment method is used to calculate the health score for different Mahalanobis distance intervals. The calculation formula is as follows: (5) In the formula, These are the points where the Mahalanobis distance is segmented; The normalization parameter for the first segment; This is the normalization parameter for the second segment.

[0016] Furthermore, based on the numerical range of the health level HP, the current health status level of the diesel engine is determined, and the corresponding evaluation result is output; for data in a fault state, the fault type is determined based on association rules.

[0017] Compared with the prior art, the present invention has the following significant advantages: This invention establishes an online rapid health assessment method for marine diesel generator sets. First, it classifies the health status of the generator sets into levels, forming a unified standard and method. Second, it determines the health status parameters of the diesel engine's in-cylinder combustion and the turbocharging system. Finally, it uses a distance metric algorithm to measure the distance between the generator set's state feature vector under abnormal conditions and the standard state feature vector output by the health benchmark model, thereby assessing the degree of its operational health.

[0018] The method of this invention is based on a diesel generator unit health baseline model, where the deterioration or failure state of the diesel generator unit is considered an abnormal state deviating from the healthy background. The health level of the operating state can be assessed by measuring the distance between the state feature vector of the diesel generator unit under abnormal conditions and the standard state feature vector output by the health baseline model; the larger the distance, the further the operating state deviates from the healthy state. Mahalanobis distance is used to measure the difference between the actual operating state of the diesel generator unit and the health baseline model. Mahalanobis distance is a method for representing the covariance distance between data. Its advantages are that it is not affected by the dimensions of various parameters, the distance is independent of the measurement unit of the original data, and it can measure the distance between the sample to be evaluated and the known sample set, making it suitable for health assessment of equipment and industrial processes. Attached Figure Description

[0019] Figure 1 This is a schematic diagram illustrating equipment performance degradation. Figure 2 A graph showing the relationship between health status and Mahalanobis distance; Figure 3 This is a numerical model diagram of a diesel engine; Figure 4 This is a comparison chart of the model values ​​and actual values ​​of diesel engine parameters. Detailed Implementation

[0020] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0021] This invention proposes an online rapid health assessment method for a diesel engine used in power generation. This method assesses the health status of in-cylinder combustion in the diesel engine using parameters such as engine speed, injector drive current, injector flow rate, intake pressure, intake temperature, mean indicated pressure (IMEP), combustion start point (CA10), combustion center of gravity (CA50), maximum pressure rise rate, ignition delay period, and combustion duration. It also assesses the health status of the diesel engine's turbocharging system using parameters such as engine speed, torque, compressor inlet pressure, turbine outlet pressure, turbine inlet temperature, compressor inlet temperature, bypass valve opening, and boost efficiency.

[0022] First, the criteria for classifying the health status of diesel generator units were determined, and the health status of diesel generator units was divided into multiple levels. Second, the principle of parameter selection was established to select appropriate evaluation parameters for different systems. Then, a health benchmark model for diesel generator units was established, and the model was trained using monitoring data under the health status. Based on Mahalanobis distance, a health decline index was designed, and the indexes with different numerical ranges were mapped to the health status levels. Finally, for fault status, the fault type was determined based on association rules.

[0023] 1. The health status of diesel generator units is divided into multiple levels. The actual operating status of diesel generator units is categorized into five levels of health, from highest to lowest: "Healthy," "Good," "Average," "Deteriorating," and "Faulty." First, the concept of "health level" for diesel generator units is established. Health level is a dimensionless value between 0 and 1; the closer it is to 1, the better the health of the generator unit, and vice versa. Based on the health level value, the health status levels are divided, and the meaning of each level is as follows: (1) "Healthy" status. When the health score is less than or equal to 1 and greater than 0.8, the diesel generator unit is considered to be in a "healthy" status. This status indicates that the condition of each subsystem of the diesel generator unit is excellent and the unit is operating safely. Therefore, no additional maintenance measures are required. Maintenance work can be carried out according to the established maintenance plan and monitoring can be maintained.

[0024] (2) "Good" status. When the health level is less than or equal to 0.8 and greater than 0.6, the diesel generator unit is considered to be in a "good" status. This status indicates that the diesel generator unit is basically safe, but its performance has degraded to some extent. At this time, the inspection cycle should be shortened appropriately, maintenance measures should be taken to optimize equipment performance, common risk sources should be investigated, and monitoring of the diesel generator unit should be maintained.

[0025] (3) "Normal" state. When the health level is less than or equal to 0.6 and greater than 0.4, the diesel generator unit is considered to be in a "normal" state. This state indicates that the diesel generator unit is in a critical fault state. Although no equipment damage has occurred, it is on the verge of an accident. At this time, the maintenance plan should be adjusted, the maintenance cycle should be shortened, and monitoring of subsystems with poor operating conditions should be strengthened and their maintenance priority increased. The diesel generator unit must be monitored at all times.

[0026] (4) "Deteriorating" State. When the health level is less than or equal to 0.4 and greater than 0.2, the diesel generator unit is considered to be in a "deteriorating" state. This state indicates that the health condition of the diesel generator unit is quite poor. Although no accident has occurred, the equipment has suffered a certain degree of damage, increasing maintenance costs. At this time, the standby diesel generator unit should be activated, and the continued use of the deteriorated equipment should be avoided as much as possible. It can only be used again after a comprehensive overhaul. If the actual situation does not allow the diesel generator unit to be shut down, it should be kept under constant monitoring and can be shut down at any time to prevent major accidents.

[0027] (5) "Fault" Status. When the health level is less than or equal to 0.2 and greater than 0, the diesel generator unit is considered to be in a "fault" status. This status indicates that the health condition of the diesel generator unit is extremely poor, and the unit is in an extremely dangerous state. At this time, the diesel generator unit must be unloaded and shut down, and the emergency maintenance plan must be activated immediately to organize the engine maintenance personnel to repair the destructive fault. Otherwise, it will not only cause significant economic losses, but may even threaten the personal safety of personnel.

[0028] The performance of mechanical equipment degrades over time, eventually leading to complete failure. The failure process consists of three stages: the normal operating stage, the performance degradation stage, and the functional failure stage. The equipment performance degradation process is as follows: Figure 1 As shown.

[0029] Figure 1The vertical axis represents equipment performance, which is not a fixed indicator but varies depending on the specific equipment. It needs to clearly reflect changes in equipment performance. For engines, this indicator may be output power, exhaust temperature margin, or other indicators, all of which decrease with the increase of diesel engine operating time. During the normal equipment phase, with equipment performance degradation point 1 as the boundary, a health level of 0.8 corresponds to this point. The phase before this point is the "healthy" state, and the phase from this point to equipment performance degradation point 2 is the "good" health state. A health level of 0.6 corresponds to equipment performance degradation point 2. The equipment performance decline phase lies between equipment performance degradation point 2 and the equipment failure point. Using the midpoint between these two points as the boundary, the health state of this phase is divided into "moderate" and "deteriorating". The interval from the midpoint to equipment performance degradation point 2 is the "moderate" health state, and the interval from the midpoint to the equipment failure point is the "deteriorating" health state. A health level of 0.4 corresponds to the midpoint, a health level of 0.2 corresponds to the equipment failure point, and the phase after the equipment failure point corresponds to the "failure" health state.

[0030] As the health status level of a diesel generator unit changes, corresponding measures need to be adjusted promptly. A decline in the health status level of a diesel generator unit from "healthy" to "faulty" indicates a continuous deterioration in its operational stability and safety, a rising failure rate, and a need for timely shortening of maintenance cycles.

[0031] 2. Principles for Selecting Health Status Assessment Parameters for Diesel Generator Units The parameters measured from the diesel generator unit directly reflect its health status. A diesel generator unit is a complex system composed of multiple subsystems, and its health assessment involves multiple levels. The overall health assessment requires the assessment of the health of each subsystem as a prerequisite. Selecting appropriate parameters as assessment indicators for each subsystem based on its characteristics is crucial for the health status assessment of the diesel generator unit. Taking into account factors such as personnel, equipment, and environment, the selection principles are the sensitivity, independence, variability, standardization, representativeness, and accessibility of the parameters. The meanings of each principle are as follows: (1) Parameter sensitivity principle The selected parameters must accurately reflect the health status of the diesel generator unit. Even if the health status changes slightly, the selected parameters should immediately show easily detectable changes. Through parameter fluctuations, maintenance personnel can perceive changes in the health status of the diesel generator unit.

[0032] (2) Principle of parameter independence Each parameter should be independent and not affect the others. If multiple parameters are coupled, the "parameter sensitivity principle" can be referred to, selecting the parameters that are more sensitive to changes in health status. Furthermore, parameters should be independent of the external environment and should not fluctuate significantly when affected by minor external factors.

[0033] (3) Principle of parameter variability Parameters should change with the operating time and performance of the diesel generator unit. Parameters such as rated power, rated speed, and rated voltage are fixed at the time of manufacture, are not affected by fluctuations in the operating status of the equipment, and do not change over time. Therefore, they are not suitable for assessing the health status of diesel generator units.

[0034] (4) Parameter standardization principle For parameters that are specified in existing national or industry standards or equipment manuals, the corresponding specifications must be strictly followed. For parameters whose value ranges are not clearly defined, the operating characteristics of the equipment should be fully considered, and the value range should be reasonably set.

[0035] (5) Principle of parameter representativeness The selected parameters should best reflect the system's operating status; they should be the most representative and typical parameters. Diesel generator units have complex structures and numerous parameters. By selecting representative parameters (such as those strongly related to certain important functions or those that directly describe performance), not only can measurement costs be reduced, but the complexity of parameter analysis can also be decreased.

[0036] (6) Parameter availability principle Not all parameters are easily measured, and the equipment for measuring some parameters is expensive; therefore, the difficulty of obtaining these parameters must be considered. While meeting the requirements for the health status assessment of the diesel generator unit, parameters that are relatively easy to measure should be selected whenever possible.

[0037] When actually assessing the health status of diesel generator units, if limited by budget costs and hardware equipment, it is not necessary to strictly follow the above six principles. The requirements can be appropriately relaxed, but the sensitivity, variability, and standardization of parameters should still be met.

[0038] 2.1 Parameters for assessing the health status of in-cylinder combustion in diesel engines In-cylinder combustion in diesel engines has a decisive impact on the performance of diesel generator sets. Abnormal combustion can lead to misfires, knocking, and other phenomena, causing unstable operation of the diesel engine and resulting in deterioration of power, fuel economy, and emissions performance. Therefore, the health status assessment of in-cylinder combustion in diesel engines should be an important part of the health status assessment of diesel generator sets. The in-cylinder working process of a diesel engine is very complex, involving a series of physicochemical processes such as spraying, combustion, flow, mass transfer, and heat transfer.

[0039] The fuel injection pattern of a diesel engine directly affects in-cylinder combustion. Control parameters such as injection start point, injection quantity, and injection pressure influence combustion parameters such as combustion start point (CA10), combustion center of gravity (CA50), and combustion duration. The fuel injection pattern is directly obtained through the injector drive current signal and the injector flow characteristic curve. Intake characteristics are another important factor affecting in-cylinder combustion, determined by both intake pressure and intake temperature. A good in-cylinder combustion process is controlled by coordinated fuel-air interaction. Machine learning is used to construct the coupling relationship between engine speed, fuel injection pattern, intake characteristics, and Weber combustion model parameters. The machine learning model is trained using measured data to build a machine learning-based benchmark model for in-cylinder combustion health, providing an accurate comparative standard for assessing the health of diesel engine in-cylinder combustion. The health of in-cylinder combustion is assessed by comparing the output parameters of the benchmark model with actual monitored parameters, including Mean Indicated Pressure (IMEP), CA10, and maximum pressure rise rate. Comparing the measured IMEP with the output of the health benchmark model (normal combustion) can determine whether a misfire has occurred. Comparing the measured CA10 with the output of the health baseline model can determine whether abnormal combustion, such as premature or delayed combustion, has occurred. Comparing the measured maximum pressure rise rate with the output of the health baseline model can determine whether the in-cylinder air-fuel mixture is good and whether rough combustion has occurred.

[0040] Taking into account the speed-based control method, injection characteristics, and intake characteristics of diesel engines, the parameters required for assessing the in-cylinder combustion health status of diesel engines should include speed, injector drive current, injector flow rate, intake pressure, intake temperature, IMEP, CA10, CA50, maximum pressure rise rate, ignition delay period, and combustion duration. In actual assessments, some of these parameters can be selected based on test conditions and operational requirements.

[0041] 2.2 Health Status Assessment Parameters of Diesel Engine Turbocharging System As a key component of diesel engines, the turbocharger enhances engine power. Assessing the turbocharger's health can effectively reduce the probability of turbocharger failure, extend its service life, and thus improve the operational stability and safety of the diesel generator set. Therefore, turbocharger health assessment is necessary. To ensure the normal operation of the turbocharger, its operating status needs to be monitored and operating parameters acquired, and a health baseline model established based on these parameters.

[0042] A turbocharger consists of a compressor, a turbine, and a connecting shaft. When exhaust gas from the diesel engine cylinders flows through the turbine, the residual heat and kinetic energy in the exhaust gas drives the turbine to rotate, which in turn drives the coaxial compressor to compress fresh air. This increases the amount of air entering the cylinder, ensuring uniform mixing of the working fluid and complete combustion. The efficiency of the turbocharging system determines the pressure increase and power gain before and after turbocharging; therefore, turbocharging efficiency is an important indicator for assessing the health of the turbocharging system.

[0043] For a properly matched turbocharged system, the turbocharger efficiency is mainly related to the diesel engine's operating conditions, the operating altitude, and the turbocharger system's operating parameters. Diesel engine operating parameters include engine speed and torque. Turbocharger system operating parameters include turbine inlet and outlet temperatures and pressures, compressor outlet temperatures and pressures, and bypass valve opening. A health benchmark model of the turbocharger system is constructed based on a neural network. The parameters required for evaluation include diesel engine speed, torque, compressor inlet pressure, turbine outlet pressure, turbine inlet temperature, compressor inlet temperature, bypass valve opening, and turbocharger efficiency. In actual evaluations, a subset of these parameters can be selected based on test conditions and operational requirements.

[0044] 3. Health Status Assessment Methods Mahalanois distance, proposed by statistician Mahalanois PC, is a method for representing the covariance distance between data. It can calculate the distance between a sample and a sample set, or measure the distance between two unknown sample sets. Compared to other distance algorithms, Mahalanois distance is unaffected by the dimensions of the parameters; the distance is independent of the measurement unit of the original data; and it can measure the distance between the sample to be evaluated and a known sample set, making it suitable for health assessments of equipment and industrial processes. The formula for calculating Mahalanois distance is shown in equation (1).

[0045] (1) In the formula, The Mahalanobis distance; This represents a vector, i.e., a sample to be evaluated; For sample set The mean, of which ; Let be the covariance matrix of the independent variables of the sample set.

[0046] A larger Mahalanobis distance indicates a greater deviation of the diesel generator unit from its normal operating state and a more severe performance degradation. A health degradation index is formed by measuring the deviation between the actual values ​​of performance parameters such as CA50, IMEP, and burst pressure of the diesel generator unit and the health benchmark model based on the Mahalanobis distance. Multi-stage alarm thresholds are set for the health degradation index to distinguish different health status levels of the diesel generator unit. To more intuitively describe the health status of the diesel generator unit, the Mahalanobis distance is normalized to a health degree. The calculation formula is shown in equation (2).

[0047] (2) In the formula: For normalization parameters, This is the Mahalanobis distance.

[0048] Health The value ranges between 0 and 1; a larger value indicates a better health condition for the diesel generator unit, while a smaller value indicates a more severe decline in health. Adjusting parameters... Value, health The sensitivity to changes in the health status of a diesel generator unit at different stages will change accordingly. Parameters Health Distance from Mahalanobis The influence of relationships, such as Figure 2 As shown. Parameters The larger, the healthier With Mahal distance The flatter the curve of change, the less sensitive health is to changes in health status.

[0049] By comparing measured parameters such as CA50, IMEP, and burst pressure with the output of the corresponding health benchmark model, the system health is calculated, and the health assessment of the key subsystems of the diesel generator unit is finally completed.

[0050] 4. In-cylinder combustion health assessment based on simulation data A model of the SC7H diesel engine was built using GT-POWER. This engine was modified from a single-stage turbocharger to a two-stage turbocharger. The model is shown below. Figure 3 As shown.

[0051] Six operating conditions were selected to verify the model, and the details of the operating conditions are shown in Table 1. The numerical model of the diesel engine was compared with the measured in-cylinder combustion pressure, crankshaft angle corresponding to the combustion pressure, high-pressure stage intercooler outlet pressure, high-pressure stage intercooler outlet temperature, and torque to verify the model. The results are shown in [Table 1]. Figure 4 .

[0052] Table 1. Model Validation Conditions

[0053] The average relative error of each parameter under all operating conditions was controlled within 6.25%, which indicates that the constructed diesel engine numerical model has high predictive reliability.

[0054] The effectiveness of the health status assessment method is verified using an uneven fuel supply fault as an example, specifically setting the fault as insufficient fuel supply from the injector in cylinder 1. A fault coefficient is introduced to simulate faults of varying severity. Insufficient fuel supply indicates that the cyclic fuel injection quantity is lower than normal. Rear diesel engine single-cylinder cycle fuel injection quantity It can be represented by equation (3).

[0055] (3) In the formula, It is the amount of fuel injected into a single cylinder of a diesel engine under fault-free conditions.

[0056] The fault coefficient ranges from [0, 1]. Using the fault coefficient under fault-free conditions as a baseline, the fault severity is gradually increased by setting a fault coefficient variation rate. For a certain type of fault, the fault coefficient variation rate... The calculation method is shown in equation (4).

[0057] (4) In the formula, The fault coefficient; This represents the fault coefficient value under fault-free conditions. The maximum failure factor; This represents the minimum failure factor.

[0058] As the failure coefficient variation rate increases, the fuel injection quantity of cylinder 1 gradually decreases from its original value, while the fuel injection quantity of the other cylinders remains unchanged. The correspondence between the failure coefficient variation rate and the diesel engine health status level is shown in Table 2.

[0059] Table 2. Change rate of diesel engine failure coefficient and health status level

[0060] The failure coefficient variation rate increased from 0 to 98.99% in increments of 1.01%, generating 99 data points per engine speed. Health was calculated separately for different engine speeds, with each speed having corresponding health status data and data to be evaluated. Data with a failure coefficient variation rate less than 2.5% were considered healthy data. The Mahalanobis distance between the samples to be evaluated and the healthy samples was calculated. The data used to calculate health included cylinder IMEP, cylinder CA50, cylinder combustion duration (10%-90%), and cylinder combustion pressure. Partial data for 2300 rpm is shown in Table 3.

[0061] Table 3. Partial evaluation data for diesel engines at 2300 rpm.

[0062] In Table 3, the Mahalanobis distance of the data samples in the healthy state is relatively small, not exceeding 2, but the Mahalanobis distance of the samples in other states is very large and increases sharply with the deterioration of health. The health status assessed based solely on a single health calculation formula is difficult to reflect the actual situation. Here, a two-segment method is used to calculate the health status in different Mahalanobis distance intervals, and the calculation formula is shown in Equation (5).

[0063] (5) In the formula, These are the points where the Mahalanobis distance is segmented; The normalization parameter for the first segment; This is the normalization parameter for the second segment.

[0064] The parameters and accuracy of the health status calculation formula at each rotational speed are shown in Table 4 after fitting.

[0065] Table 4. Health Status Calculation Formulas, Parameters, and Accuracy of Status Assessment

[0066] The accuracy of the condition assessment at all speeds was above 95%, with an average accuracy of 97.64%, meeting the task requirements and successfully verifying the effectiveness of the condition assessment method.

[0067] To verify the timeliness of the evaluation model, it is necessary to calculate the evaluation time at different speeds. The number of evaluation samples, total evaluation time, and average evaluation time per sample for each speed are shown in Table 5.

[0068] Table 5 Model Evaluation Time

[0069] The total evaluation time at each rotational speed is less than 0.06 seconds, and the average evaluation time required for a single sample is less than 0.61 milliseconds, which meets the task requirements and proves that the state evaluation method has good timeliness.

[0070] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the invention. Those skilled in the art can make various changes and modifications without departing from the concept and scope of the present invention. Therefore, all technical solutions obtained by equivalent substitution or equivalent transformation fall within the protection scope of the present invention.

Claims

1. A method for online rapid health assessment of a diesel engine for power generation, characterized in that: This includes health status assessment methods for diesel engine in-cylinder combustion systems and diesel engine turbocharging systems. First, the criteria for classifying the health status of diesel generator sets are determined, and the health status of diesel generator sets is divided into multiple levels. Second, principles for selecting health status assessment parameters for diesel generator sets are established, and appropriate assessment parameters are selected for different systems. Then, a health benchmark model for the diesel generator unit is established, the model is trained using monitoring data under healthy conditions, a health decline index is designed based on Mahalanobis distance, and the indexes with different numerical ranges are mapped to health status levels; finally, for fault states, the fault type is determined based on association rules classification.

2. The method according to claim 1, characterized in that: The health status assessment parameters of the diesel engine in-cylinder combustion system include: engine speed, injector drive current, injector flow rate, intake pressure, intake temperature, mean indicated pressure (IMEP), combustion start point (CA10), combustion center of gravity (CA50), maximum pressure rise rate, ignition delay period, and combustion duration.

3. The method according to claim 1, characterized in that: The health status assessment parameters for a diesel engine turbocharger system include engine speed, torque, compressor inlet pressure, turbine outlet pressure, turbine inlet temperature, compressor inlet temperature, bypass valve opening, and turbocharger efficiency.

4. The method according to claim 1, characterized in that: The health status of diesel generator units is divided into multiple levels, specifically: based on the actual operating conditions of the diesel generator units, their health status is divided into five levels from high to low: "healthy" status, "good" status, "average" status, "deteriorating" status, and "faulty" status.

5. The method according to claim 1, characterized in that: The selection principle for parameters for the health status assessment of diesel generator units is as follows: taking into account personnel, equipment and environmental factors, the selection principle is based on the sensitivity, independence, variability, standardization, representativeness and accessibility of the parameters.

6. The method according to claim 1, characterized in that: The health decline index is designed based on Mahalanobis distance, and the index with different numerical ranges is mapped to the health status level. Specifically, the Mahalanobis distance is calculated between the state feature vector composed of actual monitoring data and the standard state feature vector based on the Mahalanobis distance algorithm, and used as the health decline index. The Mahalanobis distance normalization model is used to map the Mahalanobis distance to the health level HP. The Mahalanobis distance normalization model uses different normalization parameters for calculation in different intervals with the Mahalanobis distance segment points as the boundary.

7. The method according to claim 6, characterized in that: The formula for calculating Mahalanobis distance is: (1) In the formula, The Mahalanobis distance; This represents a vector, i.e., a sample to be evaluated; For sample set The mean, of which ; Let be the covariance matrix of the independent variables of the sample set.

8. The method according to claim 7, characterized in that: Mahalanobis distance normalized to health level The calculation formula is as follows: (2) In the formula: For normalization parameters, This is the Mahalanobis distance.

9. The method according to claim 8, characterized in that: To ensure that the assessed health status reflects reality, a two-segment method is used to calculate health scores for different Mahalanobis distance intervals. The calculation formula is as follows: (5) In the formula, These are the points where the Mahalanobis distance is segmented; The normalization parameter for the first segment; This is the normalization parameter for the second segment.

10. The method according to claim 9, characterized in that: Based on the numerical range of the health level HP, the current health status level of the diesel engine is determined, and the corresponding evaluation result is output; for data in a fault state, the fault type is determined based on association rules.

Images