Cold test process for diesel engine

Abstract

The invention discloses a diesel engine cold test process, which relates to the technical field of engine detection, and comprises the following steps of: acquiring working parameters fed back in an engine detection process, analyzing an engine comprehensive deviation influence index, calling a numerical value detected by each noise detection point, and acquiring a noise trend coefficient; and matching with a fault noise distribution diagram stored in a database, and analyzing a noise fault matching coefficient. According to the cold test process for the diesel engine, the working parameters obtained during detection are compared with the standard of the engine, so that the fault of the engine is quickly responded, then, the noise generation position is analyzed by utilizing the noise condition generated by the engine in the detection process, and the detection accuracy is improved. The fault type is determined by comparing the fault noise distribution diagram with a fault noise distribution diagram obtained by previous detection, so that the time required by operation is effectively reduced, and a basis is provided for subsequent fault search, so that the fault location can be quickly determined, and a corresponding repair scheme can be given.

Description

Technical Field

[0001] The invention relates to the technical field of engine detection, and in particular to a diesel engine cold test process. Background Art

[0002] Performance testing is a key process in diesel engine testing. The traditional testing method uses thermal testing technology.

[0003] The principle of thermal testing technology: When the engine is burning and working, the dynamometer is used to control the engine under different working conditions to test various parameters of the engine's thermal cycle. The overall power, torque and fuel consumption of the engine are detected.

[0004] There are problems with current thermal testing technology: low test accuracy, inability to accurately locate faults, long test time (about 30 minutes), and low efficiency. The test process not only consumes fuel, but also requires cooling water and test bench exhaust treatment facilities, and the cost of use and maintenance is relatively high.

[0005] Principle of cold test technology: Use an electric motor to reverse the engine to simulate the engine's operating state, use sensors to dynamically collect various parameters that reflect the engine's performance, and use system software to calculate and analyze, compare the differences between the measured parameters and the standard state, and make a quantitative judgment on the engine quality.

[0006] The Chinese patent with the existing patent application number "CN202010506419.X" discloses a method for diagnosing engine cold test faults, including the following steps: performing an engine cold test and obtaining m engine samples and dividing them into m1 normal engine samples and m2 faulty engine samples, then arranging the data of the m1 normal engine samples and the m2 faulty engine samples to form m3 normal engine samples and m4 faulty engine samples, so that m3 and m4 satisfy the following relationship: and m3+m4≧100, building an LSTM neural network, and importing the processed sample data into the LSTM neural network for training and self-testing, then performing a cold test on the engine to be diagnosed and extracting the test data, and inputting the test data into the self-tested model, and the model automatically outputs the diagnosis result. Although the LSTM neural network is built to perform intelligent fault diagnosis on the cold test of the engine to reduce the dependence of the test on the tester, since the model is obtained by analyzing, arranging and classifying the existing data, it takes a lot of time to train to ensure accuracy. In addition, since the neural network is a "black box nature", it is difficult to understand how they make predictions or decisions. This lack of transparency can cause problems in fields where trust and explainability are crucial, such as healthcare or finance.

[0007] However, during the implementation of the above technical solution, it was found that there were at least the following technical problems:

[0008] The amount of data that needs to be processed is large, and it is impossible to intuitively understand the engine's detection information. The existing engine cold test process mainly relies on neural models, but due to the accuracy of the neural model's execution, it relies on the amount of data obtained during its training. Secondly, as the neural network continues to learn, the data it needs to calculate is also increasing. Even in actual detection, it is also necessary to spend a lot of time and energy on processing the data obtained from engine feedback, resulting in the engine detection. The response cannot be fast and effective. In addition, because neural networks are "black box" (that is, all operations and calculation methods are calculated in the neural network itself), it is difficult to understand how they make predictions or decisions, resulting in the inability to utilize the various parameters obtained during the detection process, so that the subsequent improvement direction cannot be effectively determined. For this reason, we propose a diesel engine cold test process. Summary of the invention

[0009] 1. Technical issues to be resolved

[0010] In view of the deficiencies of the prior art, the present invention provides a diesel engine cold test process to solve the technical problems that when the prior engine cold test process is used, the data volume is large and the detection information during the engine detection process cannot be accurately understood.

[0011] (II) Technical solution

[0012] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0013] A diesel engine cold test process, the test process comprising the following steps:

[0014] Retrieving the corresponding detection strategy according to the engine model to be tested, executing the detection strategy and obtaining the working parameters fed back during the engine detection process, comparing the working parameters with the corresponding preset error allowable threshold range, and classifying the working parameters, wherein the working parameters include engine speed, engine noise decibel, rotational resistance, inlet and outlet air temperature and engine internal pressure;

[0015] The engine comprehensive deviation impact index is analyzed according to the classified working parameters, and the engine comprehensive deviation impact index is compared with the preset correctable threshold. When the engine comprehensive deviation impact index is greater than the correctable threshold, a risk verification instruction is issued, the detection strategy is adjusted, and the working parameters fed back by the engine during the detection process are recorded; otherwise, a factory instruction is issued;

[0016] When receiving the risk verification instruction, the engine speed is maintained, the values ​​detected at each noise detection point are retrieved, the noise trend coefficient is obtained, and then matched with the fault noise distribution map stored in the database, the noise fault matching coefficient is analyzed, and the noise fault matching coefficient is sorted from large to small, and the engine fault working parameters corresponding to the noise fault matching coefficient are retrieved in sequence, and combined with the engine working parameter information, and the fault matching degree is calculated, and then the fault matching degree is compared with the fault matching threshold. When the fault matching degree ≥ the fault matching threshold, a parts maintenance instruction is issued to repair the engine parts corresponding to the fault matching degree; otherwise, a risk maintenance instruction is issued to send the fault data with the largest fault matching degree to the maintenance department;

[0017] The repaired engine is retested and the working parameters obtained during the retesting process are obtained. The noise trend coefficients before and after the repair are compared, and then the fault repair evaluation value is calculated based on the engine comprehensive deviation impact index. Then, based on the comparison result between the fault repair evaluation value and the preset fault repair effect threshold, the adjustment strategy is executed to adjust the impact proportional coefficient when calculating the fault matching degree.

[0018] Preferably, during the engine detection, an electric motor is used as a power source to drive the engine crankshaft to rotate, and the detection strategy refers to controlling the output parameters of the motor, wherein the output parameters of the motor include the speed of the motor during output, the change in speed per unit time, and the output torque.

[0019] Preferably, the specific process of analyzing the engine comprehensive deviation impact index is as follows:

[0020] Set the duration of the detection cycle, the initial speed and the boost speed, which are denoted as T tjg 、r cs 、r es ;

[0021] Obtain the working parameters of the engine at the initial speed, and compare them with the corresponding preset error tolerance threshold interval. The endpoints of the error tolerance interval include the upper error limit value and the lower error limit value. When the working parameter is greater than the upper error limit or the working parameter is less than the lower error limit, a risk warning is issued, and the working parameter and the corresponding endpoint of the error tolerance interval are recorded; otherwise, a qualified instruction is issued;

[0022] When receiving a risk warning, we analyze the formula When the risk warning is issued, the sum of the deviation ratios of the engine operating parameters Gtoh ig , where Z represents the total number of operating parameters greater than the upper limit of the error, represents the i-th working parameter in the total number Z, Respectively represent the preset i-th working parameters The corresponding upper and lower error limits, X represents the total number of working parameters that are less than the lower error limit, represents the jth working parameter in the total number X, The preset values ​​represent the jth working parameters The corresponding upper and lower error limits;

[0023] When receiving a qualified instruction, by analyzing the formula When the qualified instruction is issued, the sum of the deviation ratios of the engine operating parameters Gtth ig , where C represents the total number of working parameters within the error tolerance range, represents the nth working parameter in the total number C, The preset values ​​represent the nth working parameters. The corresponding upper and lower error limits;

[0024] By analyzing the formula Get the deviation influence index Lern ​​of the engine at the initial speed cs , where tkgn represents the correction factor of the preset deviation influence index, e represents the natural constant, and ε 1 , ε 2 They represent the weights of the sum of the deviation ratios under the preset risk warning and the sum of the deviation ratios under the qualified instructions, ΔT ig Indicates the impact factor corresponding to the preset unit detection time;

[0025] Similarly, according to the analysis method of the deviation influence index of the engine at the initial speed, the deviation influence index of the engine at the increased speed is obtained, which is recorded as Lern es ;

[0026] By analyzing the formula Lern g =Lern es ×tog 1 +Lern cs ×tog 2 Get the engine comprehensive deviation influence index Lern g , where tog 1 、tog 2 They respectively represent the weights of the deviation influence index under the preset initial speed and the deviation influence index under the increased speed.

[0027] Preferably, during the analysis of the engine comprehensive deviation impact index, the values ​​of the engine operating parameters before and after the initial speed duration are recorded and recorded as When the parameter changes If the error is 0, a calibration instruction is issued and the correction strategy is executed; otherwise, no response is made;

[0028] Among them, when the correction strategy is received, the engine is controlled to stop rotating, and the values ​​of the engine's working parameters before and after the initial speed period are re-recorded. When the parameter change difference in the two comparison processes is greater than 1.2% or when the parameter change difference is ≤1.2% in the first comparison, and the parameter change difference is greater than 1.2% in the second comparison, a risk verification instruction is issued to adjust the detection strategy; otherwise, the result of the second test is used as the basis to analyze the engine's comprehensive deviation impact index.

[0029] Preferably, the process of obtaining the noise tendency coefficient is as follows:

[0030] Randomly select any detection point and obtain the standard noise decibel value and actual noise decibel value of the detection point, which are recorded as Ptyn dc , Qtyn dc ;

[0031] Set the screening distance, take the detection point being detected as the center of the circle and the screening distance as the radius as the screening area, and obtain the decibel values ​​detected by other detection points in the area;

[0032] By analyzing the formula Get the noise trend coefficient Yhomh of the current detection point t , where M represents the number of other detection points in the screening area corresponding to the current detection point, Qtyn i Ptyn represents the actual noise decibel value detected at the i-th detection point. i represents the standard noise decibel value of the i-th detection point, ypmh and mhk represent the weights of the change ratio of the current detection point and the average change of the detection points in the screening area respectively;

[0033] The noise trend coefficient Yhomh of the current detection point t It is compared with the preset trend threshold. When the noise trend coefficient is ≥ the trend threshold, a fault matching instruction is issued to match the fault noise distribution map of the detection point in the database and the detection points in the corresponding screening area, and the noise fault matching coefficient is analyzed; otherwise, no response is given.

[0034] Preferably, the analysis process of the noise fault matching coefficient is as follows:

[0035] The noise decibel value of each detection point is retrieved to generate a noise distribution curve diagram, wherein the noise distribution curve diagram has the detection point number as the X-axis and the noise decibel value as the Y-axis;

[0036] Compare the overlap with the fault noise distribution graph stored in the database. When the image overlap is ≥80%, a qualified match instruction is issued, and the fault data corresponding to the fault noise distribution graph is obtained to calculate the noise fault matching coefficient; otherwise, a failed match instruction is issued, and the images are sorted according to the overlap with the noise distribution curve, and the noise fault matching coefficient is calculated and sent to the maintenance department;

[0037] By analyzing the formula Get the noise fault matching coefficient Tgonu tob , where B represents the total number of detection points, represents the average fault noise decibel value of the kth detection point in the database, Qtyn k represents the actual noise decibel value detected at the kth detection point, represents the maximum actual noise decibel value and the minimum actual noise decibel value detected by the kth detection point, Yhomh k represents the noise trend coefficient of the kth detection point, σ 1 , σ 2 Respectively represent the preset noise trend coefficient and the weight of the deviation ratio. When the image overlap is ≥98%, Recorded as 0.

[0038] Preferably, before calculating the fault matching degree, the fault operating parameters and the current engine operating parameters corresponding to the noise fault matching coefficient are retrieved, and according to the comparison result between the ratio between the two and the preset standard deviation threshold ΔDtk, the current engine operating parameters are divided into matching parameters and non-matching parameters, wherein the fault operating parameters are the operating parameters fed back when the engine fails, and the formula based on which the fault matching degree is based is as follows:

[0039]

[0040] In the formula, Mjhg qx represents the sum of the deviation ratios of the matching parameters, X represents the total number of matching parameters, Indicates the xth working parameter in the matching parameters. Indicates that the fault working parameters are The corresponding parameter average value, Mjhg qy represents the sum of the deviation proportions of non-matching parameters, Y represents the total number of non-matching parameters, Indicates the yth working parameter among the non-matching parameters. Indicates that the fault working parameters are The corresponding parameter average value, Gtd represents the fault matching degree, ω 1 ,ω 2 They represent the weights of the preset matching parameter deviation ratio and non-matching parameter deviation ratio, respectively. Indicates the proportional coefficient of the current fault, a>1, and when (Mjhg qx ×ω 1 -Mjhg qy ×ω 2 )≤0, the fault matching degree Gtd is recorded as 0.

[0041] Preferably, the formula for the fault repair evaluation value is as follows:

[0042]

[0043] In the formula, Yhomh gt , Yhomh kt They represent the average noise trend coefficient of all detection points before and after fault repair, Lern gt , Lern kt They respectively represent the engine comprehensive deviation influence index of the engine operating parameters before and after the fault is repaired. The Ysng and Ltno sub-tables represent the proportional coefficient of the preset average noise trend coefficient and the engine comprehensive deviation influence index. G represents the constant correction coefficient.

[0044] Preferably, the adjustment strategy of the impact ratio coefficient is as follows: setting an adjustment ratio, denoted as θ, comparing the fault repair evaluation value with a preset fault repair effect threshold, and when the fault repair evaluation value is greater than the fault repair effect threshold, an upward adjustment instruction is issued to As the modified impact ratio coefficient; otherwise, a downward adjustment instruction is issued to As the corrected impact proportional coefficient.

[0045] Preferably, the testing process further comprises:

[0046] Get the previous n adjustment data, use the fault type as the X-axis and the number of adjustments as the Y-axis to generate a fault adjustment trend chart. If the error occurs, the system will issue an equipment repair instruction to the maintenance department; otherwise, no response will be given.

[0047] (III) Beneficial effects

[0048] 1. The working parameters obtained during the test are compared with the engine standards to determine the deviation between the current engine and the required ones, and an engine comprehensive deviation impact index that can reflect the engine status is generated to determine whether it meets the factory standards. Secondly, by using the noise emitted by the engine during the test, the location of the noise is analyzed, and then compared with the fault noise distribution map obtained by previous tests to determine the fault type, which effectively reduces the time required for calculations and provides a basis for subsequent fault finding, so as to quickly determine the fault location, provide corresponding repair plans, and improve maintenance efficiency.

[0049] 2. By retrieving the fault working parameters of the engine under the corresponding fault type at the same time as the fault feedback, and then comparing them with the working parameters fed back during the current engine detection, a fault matching degree that can be used to evaluate whether the fault search is accurate is obtained, so as to evaluate the fault type found through noise and ensure the accuracy of the fault type judgment. In addition, the impact proportional coefficients of different fault types are adjusted through the working parameters fed back by the engine after maintenance to determine the efficiency of fault assessment during engine detection. At the same time, the adjustment frequency of the impact proportional coefficient of the same type of fault within a period of time is used to determine the frequency of occurrence of the fault, so as to overhaul and adjust the equipment on the corresponding production line of the engine according to the situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings.

[0051] Figure 1 It is the overall flow chart of the present invention;

[0052] Figure 2 A noise distribution curve diagram in an embodiment of the present invention;

[0053] Figure 3 It is a structural diagram of an engine detection device in an embodiment of the present invention;

[0054] Figure 4 It is a schematic diagram of the structure of the connection between the engine crankshaft and the motor in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] The embodiment of the present application provides a diesel engine cold test process to solve the technical problems that when using the existing engine cold test process, the amount of data is large and the detection information in the engine detection process cannot be accurately understood. When using the existing engine cold test process, the working parameters obtained during the detection are compared with the engine standards to determine the deviation between the current engine and the required requirements, and generate an engine comprehensive deviation impact index that can reflect the engine status to determine whether it meets the factory standards. Secondly, by utilizing the noise emitted by the engine during the detection process, the location of the noise is analyzed, and then compared with the fault noise distribution map obtained by previous detection to determine the fault type, which effectively reduces the time required for calculation and also provides a basis for subsequent fault finding, so as to quickly determine the fault location, provide a corresponding repair plan, and improve the efficiency of maintenance.

[0056] Embodiment: The technical solution in the embodiment of the present application is to solve the technical problem that the amount of data is large and the detection information in the engine detection process cannot be accurately understood when the existing engine cold test process is used. The overall idea is as follows:

[0057] In view of the problems existing in the prior art, the present invention provides a diesel engine cold test process, which comprises the following steps:

[0058] Current issues:

[0059] The existing engine cold detection process mainly relies on the neural network model to analyze various data during the engine detection process. However, due to the accuracy requirements of engine detection, a large amount of detection data needs to be processed during engine detection to avoid missing any details that are easily overlooked. Therefore, a neural network model is required to process a large amount of data, such as test project design, test beat theoretical calculation, cold test process design, cold test bench plane layout design, cold test bench overall structure layout design, cold test bench rear end drive solution design, plugging solution design and NVH test solution design, etc. The detection equipment includes boost pressure sensor, camshaft phase sensor, EGR pressure sensor, EGR temperature sensor, EGR pressure difference sensor, water temperature sensor, crankshaft speed sensor, oil pressure sensor, intake temperature sensor, electronic throttle, EGR valve, injector, fuel flow control unit, preheating relay.

[0060] Although neural models, especially deep learning models, have achieved remarkable success in many fields, they also have some shortcomings, mainly including:

[0061] (1) Long training time: Neural network models require a large amount of training data and a long training time, especially when processing large-scale data.

[0062] (2) Overfitting problem: The model performs well on the training data but poorly on the test set, especially when the data size is small or the model is too complex.

[0063] (3) Poor interpretability: The decision-making process of a neural network is usually a black box, and it is difficult to explain the decision basis of the model, which limits its application in application scenarios that require interpretability.

[0064] Technology R&D:

[0065] In order to solve this problem and cooperate with the existing engine cold inspection process, we have studied a method to compare the working parameters obtained during the inspection with the engine standards, so as to determine the deviation between the current engine and the required requirements, and generate an engine comprehensive deviation impact index that can reflect the engine status, so as to determine whether it meets the factory standards. Secondly, by using the noise emitted by the engine during the inspection, the location of the noise is analyzed, and then compared with the fault noise distribution map obtained by previous inspections, the fault type is determined, which effectively reduces the time required for calculations. At the same time, it also provides a basis for subsequent fault finding, so as to quickly determine the location of the fault, provide a corresponding repair plan, and improve the efficiency of maintenance. The details are as follows:

[0066] Failure Analysis

[0067] Since various faults will affect the engine detection, for example, if the engine piston is too large, the friction between it and the inner wall of the engine cylinder will increase, which will eventually manifest as resistance when the engine rotates. Therefore, the various working parameters obtained during engine detection can reflect the changes in the engine the fastest. The engine is a precision instrument with very high requirements for various parameters. This also shows from the side that the various parameters of the engine play a vital role in the evaluation of the engine. In addition, since various parameters will affect each other and accumulate with each other, all working parameters need to be evaluated comprehensively.

[0068] According to the engine model to be tested, the corresponding detection standard strategy is retrieved. Since different engine detection standards are different, when testing different engines, the working state of the engine is controlled by the control finger. Since the cold detection of the engine uses the motor as the power source to drive the engine crankshaft to rotate, the detection strategy is to control the speed of the motor output, the change in speed per unit time and the output torque, execute the detection strategy and obtain the working parameters fed back during the engine detection process (including engine speed, engine noise decibels, rotational resistance, inlet and outlet air temperature and engine internal pressure), and compare the working parameters with the corresponding preset error allowable threshold range to classify the working parameters. The classification method is as follows:

[0069] Record the engine speed at the initial cs The working parameters under the condition are compared with the corresponding preset error allowable threshold range.

[0070] Among them, the endpoints of the error tolerance range include the upper error limit and the lower error limit. When the working parameter is greater than the upper error limit or the working parameter is less than the lower error limit, a risk warning is issued, and the working parameter and the corresponding endpoint of the error tolerance range are recorded; otherwise, a qualified instruction is issued, thereby dividing the working parameters into risk parameters with risks and safety parameters without risks.

[0071]

[0072] In formula (1.1), Gtth ig represents the sum of the deviation proportions of the safety parameters, Z represents the total number of working parameters greater than the upper limit of the error, represents the i-th working parameter in the total number Z, Respectively represent the preset i-th working parameters The corresponding upper and lower error limits, X represents the total number of working parameters that are less than the lower error limit, represents the jth working parameter in the total number X, The preset values ​​represent the jth working parameters The corresponding upper and lower error limits;

[0073] In formula (1.2), Gtoh ig represents the sum of the deviation proportions of the risk parameters, C represents the total number of working parameters within the error tolerance range, represents the nth working parameter in the total number C, The preset values ​​represent the nth working parameters. The corresponding upper and lower error limits;

[0074] In formula (1.3), Lern cs It represents the deviation influence index of the engine at the initial speed, tkgn represents the correction factor of the preset deviation influence index, e represents the natural constant, ε 1 , ε 2 They represent the weights of the sum of the deviation ratios under the preset risk warning and the sum of the deviation ratios under the qualified instructions, respectively. tjg The duration of the detection cycle, ΔT ig Indicates the impact factor corresponding to the preset unit detection time.

[0075] This gives a feedback initial speed r cs The deviation influence index of the engine operating parameters is Lern cs , so as to feedback whether the working state of the engine is stable, to ensure that the data obtained is the working parameter that can reflect the real state of the engine, rather than the influence caused by the engine running-in problem or the adaptability between parts during assembly. Provide a verification plan.

[0076] Verification plan: record the engine at the initial speed r cs The values ​​of the working parameters before and after the duration are recorded as When the parameter changes When , it means that the working parameters obtained by the engine have undergone a very large change, so a calibration instruction is issued at this time to execute the correction strategy; otherwise, it means that the working state of the engine is constant, and no response is given. The working parameters obtained can reflect the working state of the engine.

[0077] Among them, the correction strategy is: control the engine to stop rotating, and re-record the values ​​of the engine's working parameters before and after the initial speed period. When the parameter change difference is greater than 1.2% during the two comparisons, it means that the engine has adaptation, running-in or other problems and cannot participate in subsequent inspections; when the parameter change difference is ≤1.2% during the first comparison, and the parameter change difference is greater than 1.2% during the second comparison, it means that the engine has a fault and needs to be inspected. Therefore, when the above two situations occur, a risk verification instruction must be issued and the detection strategy must be adjusted.

[0078] On the contrary, when the parameter change difference is ≤1.2% during the two comparisons or the parameter change difference is >1.2% during the first comparison, and ≤1.2% during the second comparison, it indicates that the engine tends to be stable during detection. Therefore, the result of the second detection is used as the basis to analyze the engine deviation impact index.

[0079] During the test, an important problem was found. Since the engine is used to provide speed for the car, and the speed of the car depends on the coordination between the gearbox and the engine speed, the engine has an initial speed of r. cs The deviation influence index Lern cs , can only reflect the state of the engine at a constant speed, but cannot reflect the characteristics of the engine speed change. Therefore, the above method is used to obtain the engine speed when increasing the speed r es The deviation influence index under es , thus obtaining a deviation impact index Lern ​​that can reflect the speed of change es Combined with the deviation influence index Lern ​​under uniform speed cs Calculate the engine comprehensive deviation impact index Lern g , the formula is as follows:

[0080] Lern g =Lern es ×tog 1 +Lern cs ×tog 2

[0081] In the formula, tog 1 、tog 2 They represent the weights of the deviation impact index at the preset initial speed and the deviation impact index at the increased speed. In this way, an engine comprehensive deviation impact index that can reflect the engine status can be obtained to determine whether it meets the factory standards, and the detection effect is more comprehensive. That is, when the engine comprehensive deviation impact index is greater than the correctable threshold, a risk verification instruction is issued, the detection strategy is adjusted, and the working parameters fed back by the engine during the detection process are recorded; otherwise, a factory instruction is issued.

[0082] Troubleshooting

[0083] When the system issues a risk verification instruction, the engine speed is maintained first, and the values ​​detected at each noise detection point are retrieved. The noise detection point is set on the machine tool before the engine is tested. When the engine fails, it is mainly manifested by friction, and noise is generated along with friction. Therefore, the distribution of noise can be detected to detect the change of noise, and the location of the fault can be inferred. For example, if one of the pistons is worn, the piston will rub against the cylinder body and generate noise, and the energy of noise propagation will continue to decay, so the closer to the piston, the greater the noise, and vice versa, the farther away, the smaller the noise. In order to facilitate intuitive observation, it is necessary to calculate the noise trend coefficient Yhomh based on the distribution of noise. t , randomly select any detection point and calculate the noise trend coefficient Yhomh of the detection point t The formula is as follows:

[0084]

[0085] Where M represents the number of other detection points in the screening area corresponding to the current detection point, Ptyn dc Indicates the standard noise decibel value of the current detection point, Qtyn dc Indicates the actual noise decibel value at the current detection point, Qtyn i Ptyn represents the actual noise decibel value detected at the i-th detection point. i It represents the standard noise decibel value of the ith detection point, ypmh and mhk represent the weights of the change ratio of the current detection point and the average change of the detection points in the screening area, respectively.

[0086] The noise trend coefficient Yhomh of the current detection point tCompare with the preset trend threshold. When the noise trend coefficient ≥ trend threshold, it means that the noise may be emitted by the location of the detection point. Therefore, the fault may be located around the detection point. Then it is necessary to match what type of fault is caused. Therefore, a fault matching instruction is issued to match the fault noise distribution map of the detection point in the database and the detection points in the corresponding screening area, and analyze the noise fault matching coefficient; otherwise, no response is made.

[0087] Screening area: Take the detection point being detected as the center of the circle and the screening distance (which can be set as needed, generally 30cm) as the radius as the screening area, and obtain the decibel values ​​detected by other detection points in the area. It can be simply understood as taking the detection point being detected as the origin and calculating the distance between other detection points and it. As long as the distance is less than the set screening distance, they are all detection points in the screening area.

[0088] Fault Matching

[0089] Obtain historical fault data, that is, the fault conditions that have occurred in the database, and then compare them with the noise trend coefficient Yhomh of the current detection point t The feedback data is matched to analyze the noise fault matching coefficient.

[0090] The principle is to generate a noise distribution curve by calling the noise decibel value of each detection point. The noise distribution curve uses the detection point number as the X-axis and the noise decibel value as the Y-axis. Then, the curve is compared with the fault noise distribution map stored in the database for coincidence (image comparison).

[0091] When the image overlap is ≥80% (can be adjusted as needed, 80% is commonly set as the dividing line), it means that the fault corresponding to the fault noise distribution diagram is very likely to be a fault of the current engine, so the obtained fault data is used as the fault output of the current engine, that is, a qualified match instruction is issued, and the fault data corresponding to the fault noise distribution diagram is obtained, and the noise fault matching coefficient is calculated; otherwise, a failed match instruction is issued, but in order to provide a reference for the maintenance department, it is necessary to first sort according to the size of the overlap with the noise distribution curve diagram, and also calculate the noise fault matching coefficient and send it to the maintenance department.

[0092] Therefore, the noise fault matching coefficient Tgonu tob The calculation is divided into two categories, one is the image with a coincidence degree ≥ 80% and the other is the image with a coincidence degree ≥ 80%. The two types are as follows:

[0093]

[0094] In the formula, B represents the total number of detection points, represents the average fault noise decibel value of the kth detection point in the database, Qtyn k represents the actual noise decibel value detected at the kth detection point, represents the maximum actual noise decibel value and the minimum actual noise decibel value detected by the kth detection point, Yhomh k represents the noise trend coefficient of the kth detection point, σ 1 , σ 2 Respectively represent the preset noise trend coefficient and the weight of the deviation ratio. When the image overlap is ≥98%, Recorded as 0.

[0095] Matching check

[0096] Get the noise fault matching coefficient Tgonu tob After that, the engine fault working parameters corresponding to the noise fault matching coefficient are retrieved in order from large to small, and the parameters are combined with the engine working parameter information to calculate the fault matching degree to determine whether the fault is found wrong, so as to check the found data and improve the accuracy of fault query. Therefore, in this way, it is not only possible to respond quickly, but also to perform fine processing. Since the fine detection is based on the noise fault matching coefficient Tgonu tob Therefore, the data required for acceleration can be greatly reduced to maintain fast response while also achieving accurate detection.

[0097] Therefore, by using the noise analysis method, all the working parameters of the engine detection are coordinated, so that the comprehensive detection is not limited to the evaluation of passing noise. Therefore, before the fault matching degree is calculated, the fault working parameters corresponding to the noise fault matching coefficient and the current engine working parameters are retrieved, and according to the comparison result between the ratio between the two and the preset standard deviation threshold ΔDtk, the current engine working parameters are divided into matching parameters and non-matching parameters, among which the fault working parameters are the working parameters fed back when the engine fails. The formula for calculating the fault matching degree is as follows:

[0098]

[0099] In the formula, Mjhg qx represents the sum of the deviation ratios of the matching parameters, X represents the total number of matching parameters, Indicates the xth working parameter in the matching parameters. Indicates that the fault working parameters are The corresponding parameter average value, Mjhg qy represents the sum of the deviation proportions of non-matching parameters, Y represents the total number of non-matching parameters, Indicates the yth working parameter among the non-matching parameters. Indicates that the fault working parameters are The corresponding parameter average value. The sum of the deviation ratios of the matching parameters obtained is Mjhg qx The sum of the deviation ratios from the non-matching parameters Mjhg qy , calculate the fault matching degree.

[0100]

[0101] Where Gtd represents the fault matching degree, ω 1 ,ω 2 They represent the weights of the preset matching parameter deviation ratio and non-matching parameter deviation ratio, respectively. Indicates the proportional coefficient of the current fault, a>1, and when (Mjhg qx ×ω 1 -Mjhg qy ×ω 2 )≤0, the fault matching degree Gtd is recorded as 0.

[0102] The fault matching degree is compared with the fault matching threshold. When the fault matching degree is ≥ the fault matching threshold, a parts maintenance instruction is issued to repair the engine parts corresponding to the fault matching degree. Otherwise, a risk maintenance instruction is issued to send the fault data with the highest fault matching degree to the maintenance department to provide a basis for maintenance.

[0103] Error Correction

[0104] After the repaired engine is retested and the working parameters obtained during the retesting process are obtained, the noise trend coefficients before and after the repair are compared, and then combined with the engine comprehensive deviation impact index, the fault repair evaluation value is calculated. The formula based on the fault repair evaluation value Gghng is as follows:

[0105]

[0106] In the formula, Yhomh gt , Yhomh kt They represent the average noise trend coefficient of all detection points before and after fault repair, Lern gt , Lern ktThey respectively represent the engine comprehensive deviation impact index of the engine operating parameters before and after the fault repair, the Ysng and Ltno tables represent the preset average noise trend coefficient and the proportional coefficient of the engine comprehensive deviation impact index, and G represents the constant correction coefficient. Thus, the fault repair evaluation value Gghng is used to evaluate whether the problem has been solved before and after the parts are repaired. If the problem has been solved, it means that the problem has been found correctly; on the contrary, if the change is not significant, it means that the problem has not been solved and the problem does not correspond to the fault. Therefore, it is necessary to implement an adjustment strategy based on the comparison result between the fault repair evaluation value and the preset fault repair effect threshold to adjust the impact proportional coefficient when calculating the fault matching degree.

[0107] Adjustment strategy: Set the adjustment ratio, denoted as θ, and compare the fault repair evaluation value with the preset fault repair effect threshold. When the fault repair evaluation value is greater than the fault repair effect threshold, an upward adjustment instruction is issued. As the modified impact ratio coefficient; otherwise, a downward adjustment instruction is issued to As the corrected impact proportional coefficient, in order to avoid other faults not being able to respond in the first time due to the impact proportional coefficient being too high or too low, the adjustment times are set. When the same fault is adjusted up to the specified times, it will not be adjusted up again. Generally, the adjustment times are set to 5-7 times. For example, the number of increases or decreases is set to 5 times. When the impact proportional coefficient of a certain fault has been adjusted up five times, that is, After that, it will no longer be adjusted upward, but it can be adjusted downward normally.

[0108] Equipment failure assessment

[0109] Get the previous n adjustment data, use the fault type as the X-axis and the number of adjustments as the Y-axis to generate a fault adjustment trend chart. (The ratio can be adjusted), it means that the latest batch of engines have this kind of fault and need to be taken seriously, then send equipment adjustment instructions to the maintenance department; otherwise, no response will be made, and the number of adjustments n≤50; for example, the adjustment data of the first 20 times is retrieved, among which the piston fault occurs 3 times, then This indicates that the set ratio has been exceeded, so the piston production line needs to be inspected and checked.

[0110] Finally, it should be noted that: Obviously, the above embodiments are only examples for clearly explaining the present invention, and are not intended to limit the implementation methods. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from this are still within the scope of protection of the present invention.

Claims

1. A diesel engine cold test process, characterized in that: The testing process includes the following steps: Retrieving the corresponding detection strategy according to the engine model to be tested, executing the detection strategy and obtaining the working parameters fed back during the engine detection process, comparing the working parameters with the corresponding preset error allowable threshold range, and classifying the working parameters, wherein the working parameters include engine speed, engine noise decibel, rotational resistance, inlet and outlet air temperature and engine internal pressure; The engine comprehensive deviation impact index is analyzed according to the classified working parameters, and the engine comprehensive deviation impact index is compared with the preset correctable threshold. When the engine comprehensive deviation impact index is greater than the correctable threshold, a risk verification instruction is issued, the detection strategy is adjusted, and the working parameters fed back by the engine during the detection process are recorded; otherwise, a factory instruction is issued; When receiving the risk verification instruction, the engine speed is maintained, the values ​​detected at each noise detection point are retrieved, the noise trend coefficient is obtained, and then matched with the fault noise distribution map stored in the database, the noise fault matching coefficient is analyzed, and the noise fault matching coefficient is sorted from large to small, and the engine fault working parameters corresponding to the noise fault matching coefficient are retrieved in sequence, and combined with the engine working parameter information, and the fault matching degree is calculated, and then the fault matching degree is compared with the fault matching threshold. When the fault matching degree ≥ the fault matching threshold, a parts maintenance instruction is issued to repair the engine parts corresponding to the fault matching degree; otherwise, a risk maintenance instruction is issued to send the fault data with the largest fault matching degree to the maintenance department; The repaired engine is retested and the working parameters obtained during the retesting process are obtained. The noise trend coefficients before and after the repair are compared, and then the fault repair evaluation value is calculated based on the engine comprehensive deviation impact index. Then, based on the comparison result between the fault repair evaluation value and the preset fault repair effect threshold, the adjustment strategy is executed to adjust the impact proportional coefficient when calculating the fault matching degree.

2. A diesel engine cold test process as claimed in claim 1, characterized in that: During the engine detection, the motor is used as a power source to drive the engine crankshaft to rotate, and the detection strategy refers to controlling the output parameters of the motor, wherein the output parameters of the motor include the speed of the motor during output, the change in speed per unit time, and the output torque.

3. A diesel engine cold test process as claimed in claim 1, characterized in that: The specific process of analyzing the engine comprehensive deviation impact index is as follows: Set the duration of the detection cycle, the initial speed and the boost speed, which are denoted as T tjg 、r cs 、r es ; Obtain the working parameters of the engine at the initial speed, and compare them with the corresponding preset error tolerance threshold interval. The endpoints of the error tolerance interval include the upper error limit value and the lower error limit value. When the working parameter is greater than the upper error limit or the working parameter is less than the lower error limit, a risk warning is issued, and the working parameter and the corresponding endpoint of the error tolerance interval are recorded; otherwise, a qualified instruction is issued; When receiving a risk warning, we analyze the formula When the risk warning is issued, the sum of the deviation ratios of the engine operating parameters Gtoh ig , where Z represents the total number of operating parameters greater than the upper limit of the error, represents the i-th working parameter in the total number Z, Respectively represent the preset i-th working parameters The corresponding upper and lower error limits, X represents the total number of working parameters that are less than the lower error limit, represents the jth working parameter in the total number X, The preset values ​​represent the jth working parameters The corresponding upper and lower error limits; When receiving a qualified instruction, by analyzing the formula When the qualified instruction is issued, the sum of the deviation ratios of the engine operating parameters Gtth ig , where C represents the total number of working parameters within the error tolerance range, represents the nth working parameter in the total number C, The preset values ​​represent the nth working parameters. The corresponding upper and lower error limits; By analyzing the formula Get the deviation influence index Lern ​​of the engine at the initial speed cs , where tkgn represents the correction factor of the preset deviation impact index, e represents the natural constant, ε1 and ε2 represent the weights of the sum of the deviation proportions under the preset risk warning and the sum of the deviation proportions under the qualified instruction, respectively, ΔT ig Indicates the impact factor corresponding to the preset unit detection time; Similarly, according to the analysis method of the deviation influence index of the engine at the initial speed, the deviation influence index of the engine at the increased speed is obtained, which is recorded as Lern es ; By analyzing the formula Lern g =Lern es ×tog1+Lern cs ×tog2 to get the engine comprehensive deviation influence index Lern g , where tog1 and tog2 represent the weights of the deviation influence index under the preset initial speed and the deviation influence index under the increased speed, respectively.

4. A diesel engine cold test process as claimed in claim 3, characterized in that: During the analysis of the engine comprehensive deviation impact index, the values ​​of the engine working parameters before and after the initial speed duration are recorded and recorded as When the parameter changes If the error is 0, a calibration instruction is issued and the correction strategy is executed; otherwise, no response is made; Among them, when the correction strategy is received, the engine is controlled to stop rotating, and the values ​​of the engine's working parameters before and after the initial speed period are re-recorded. When the parameter change difference in the two comparison processes is greater than 1.2% or when the parameter change difference is ≤1.2% in the first comparison, and the parameter change difference is greater than 1.2% in the second comparison, a risk verification instruction is issued to adjust the detection strategy; otherwise, the result of the second test is used as the basis to analyze the engine's comprehensive deviation impact index.

5. A diesel engine cold test process as claimed in claim 4, characterized in that: The process of obtaining the noise tendency coefficient is as follows: Randomly select any detection point and obtain the standard noise decibel value and actual noise decibel value of the detection point, which are recorded as Ptyn dc 、Qtyn dc ; Set the screening distance, take the detection point being detected as the center of the circle and the screening distance as the radius as the screening area, and obtain the decibel values ​​detected by other detection points in the area; By analyzing the formula Get the noise trend coefficient Yhomh of the current detection point t , where M represents the number of other detection points in the screening area corresponding to the current detection point, Qtyn i Ptyn represents the actual noise decibel value detected at the i-th detection point. i represents the standard noise decibel value of the i-th detection point, ypmh and mhk represent the weights of the change ratio of the current detection point and the average change of the detection points in the screening area respectively; The noise trend coefficient Yhomh of the current detection point t Compare with the preset trend threshold. When the noise trend coefficient is ≥ the trend threshold, a fault matching instruction is issued to match the fault noise distribution map of the detection point in the database and the detection point in the corresponding screening area, and analyze the noise fault matching coefficient; Otherwise, no response will be given.

6. A diesel engine cold test process as claimed in claim 5, characterized in that: The analysis process of the noise fault matching coefficient is as follows: The noise decibel value of each detection point is retrieved to generate a noise distribution curve diagram, wherein the noise distribution curve diagram has the detection point number as the X-axis and the noise decibel value as the Y-axis; Compare the overlap with the fault noise distribution graph stored in the database. When the image overlap is ≥80%, a qualified match instruction is issued, and the fault data corresponding to the fault noise distribution graph is obtained to calculate the noise fault matching coefficient; otherwise, a failed match instruction is issued, and the images are sorted according to the overlap with the noise distribution curve, and the noise fault matching coefficient is calculated and sent to the maintenance department; By analyzing the formula Get the noise fault matching coefficient Tgonu tob , where B represents the total number of detection points, represents the average fault noise decibel value of the kth detection point in the database, Qtyn k represents the actual noise decibel value detected at the kth detection point, represents the maximum actual noise decibel value and the minimum actual noise decibel value detected by the kth detection point, Yhomh k represents the noise trend coefficient of the kth detection point, σ1 and σ2 represent the preset noise trend coefficient and the weight of the deviation ratio respectively. When the image overlap is ≥98%, Recorded as 0.

7. A diesel engine cold test process as claimed in claim 6, characterized in that: Before calculating the fault matching degree, the fault operating parameters and the current engine operating parameters corresponding to the noise fault matching coefficient are retrieved, and the current engine operating parameters are divided into matching parameters and non-matching parameters according to the comparison result between the ratio between the two and the preset standard deviation threshold ΔDtk. Among them, the fault operating parameters are the operating parameters fed back when the engine fails. The formula based on the fault matching degree is as follows: In the formula, Mjhg qx represents the sum of the deviation ratios of the matching parameters, X represents the total number of matching parameters, Indicates the xth working parameter in the matching parameters. Indicates the fault working parameters The corresponding parameter average value, Mjhg qy represents the sum of the deviation proportions of non-matching parameters, Y represents the total number of non-matching parameters, Indicates the yth working parameter among the non-matching parameters. Indicates the fault working parameters The corresponding parameter average value, Gtd represents the fault matching degree, ω1 and ω2 represent the weights of the preset matching parameter deviation ratio and non-matching parameter deviation ratio, respectively. Indicates the proportional coefficient of the current fault, a>1, and when (Mjhg qx ×ω1-Mjhg qy ×ω2)≤0, the fault matching degree Gtd is recorded as 0.

8. A diesel engine cold test process as claimed in claim 6, characterized in that: The formula for the fault repair evaluation value is as follows: In the formula, Yhomh gt , Yhomh kt They represent the average noise trend coefficient of all detection points before and after fault repair, Lern gt , Lern kt They respectively represent the engine comprehensive deviation influence index of the engine operating parameters before and after the fault is repaired. The Ysng and Ltno sub-tables represent the proportional coefficient of the preset average noise trend coefficient and the engine comprehensive deviation influence index. G represents the constant correction coefficient.

9. A diesel engine cold test process as claimed in claim 8, characterized in that: The adjustment strategy of the impact ratio coefficient is as follows: set the adjustment ratio, denoted as θ, compare the fault repair evaluation value with the preset fault repair effect threshold, and when the fault repair evaluation value is greater than the fault repair effect threshold, issue an upward adjustment instruction to θ) as the corrected impact proportional coefficient; otherwise, a downward adjustment instruction is issued to As the corrected impact proportional coefficient.

10. A diesel engine cold test process as claimed in claim 1, characterized in that: The testing process also includes: Get the previous n adjustment data, use the fault type as the X-axis and the number of adjustments as the Y-axis to generate a fault adjustment trend chart. If the error occurs, the system will issue an equipment repair instruction to the maintenance department; otherwise, no response will be given.

Images