A method for detecting cylinder bore scratching based on engine cold test

Abstract

The application discloses a kind of detection methods based on engine cold test machine's cylinder bore scratch, it is related to engine detection field, comprising: acquisition measured engine under constant speed condition Torque data, extract the maximum, minimum and average of the torque of each cylinder;When torque data is out of limit, determine target cylinder, and calculate the difference between its maximum and minimum, obtain single cylinder torque span value;If the span value exceeds the determination limit, then trigger exhaust back pressure test link, acquisition target cylinder single cylinder exhaust peak pressure;If the peak pressure is lower than the peak pressure of normal cylinder, and difference reaches abnormal drop threshold, then determine that there is cylinder bore scratch.The present application effectively amplifies weak fault signal by combining torque span feature and exhaust back pressure feature, without disassembly, cylinder bore scratch and pure mechanical bearing defect can be accurately distinguished, and the sensitivity and accuracy of online detection of mass production line are improved.

Description

Technical Field

[0001] This application relates to the field of engine testing, and more particularly to a method for detecting cylinder bore scratches based on an engine cold test machine. Background Technology

[0002] The engine cold test machine is an important quality inspection device before the automotive engine assembly line rolls off the production line. The cold test process does not involve fuel ignition. Instead, the engine is rotated by a drive motor, and various test data are collected to verify the assembly quality. This testing method meets the requirements of mass production.

[0003] During the towing test operation, the test system records a continuous torque sequence data covering a 720-degree crankshaft rotation angle. The alternating fluctuations of positive and negative torque reflect the changes in the frictional resistance of the internal mechanical components of the engine. In addition, the piston movement in the cylinder is accompanied by the compression and discharge of gas. The gas fluid dynamics characteristics determine the total exhaust volume and instantaneous exhaust pressure of the exhaust stroke, which can intuitively present the physical sealing state inside the cylinder. The torque data and exhaust pressure data constitute the basic support for evaluating the assembly quality of the engine assembly.

[0004] Current cold testing methods rely on single torque fluctuations for fault diagnosis. Both purely mechanical friction defects like main bearing scratches and cylinder bore scratches can cause abnormal torque. A single data point cannot distinguish between the two, necessitating engine assembly disassembly for defect localization. For minor cylinder bore scratches on shallow surfaces, the additional friction torque signal is extremely weak and easily masked by machining variations or background noise in the testing environment, leading to missed detections. Furthermore, existing testing procedures lack an independent testing step for exhaust back pressure and a mechanism to convert the minute cylinder leakage caused by cylinder bore scratches into instantaneous exhaust pressure decay. There is no related sealing test device to establish a sealed cavity and incorporate moving average filtering, failing to eliminate high-frequency airflow pulsation interference. The obtained sealing failure data lacks stability and cannot meet the operational requirements of accurately intercepting defective parts on the assembly line. Summary of the Invention

[0005] The purpose of this invention is to provide a method for detecting cylinder bore scratches based on an engine cold test machine, which solves the problem that existing cold test methods cannot distinguish between cylinder bore scratches and main bearing scratches without disassembly based on a single torque fluctuation, and that weak scratch features are easily masked, leading to missed detection.

[0006] This invention provides the following solution:

[0007] This invention provides a method for detecting cylinder bore scratches based on an engine cold test, comprising the following steps:

[0008] The torque data of the engine assembly under test is collected under constant speed conditions, and the maximum torque, minimum torque and average torque of each cylinder in the engine assembly under test are extracted from the torque data.

[0009] When the maximum torque, minimum torque, or average torque exceeds the standard limit, the cylinder that exceeds the limit will be designated as the target cylinder, and the single-cylinder torque span value of the target cylinder will be calculated based on the maximum torque and minimum torque.

[0010] When the single-cylinder torque span value exceeds the preset span judgment limit, the exhaust back pressure test is triggered, and the single-cylinder exhaust peak pressure of the target cylinder is collected in the exhaust back pressure test.

[0011] When the peak exhaust pressure of a single cylinder is lower than the preset peak exhaust pressure of a single cylinder of a normally sealed cylinder, and the difference between the peak exhaust pressure of a single cylinder of a normally sealed cylinder and the peak exhaust pressure of a single cylinder is greater than or equal to the threshold for judging abnormal exhaust pressure drop, it is determined that there is cylinder bore scratch in the engine assembly under test.

[0012] This invention achieves defect localization through a multi-physics cross-verification mechanism combining mechanical dynamics and fluid dynamics. When scratches exist on the inner wall of the cylinder, the oil lubricating film between the piston ring and the inner wall of the cylinder is damaged, causing an increase in frictional resistance. This results in an increase in the positive driving torque output by the servo motor during the compression stroke and a decrease in the reverse braking torque output during the expansion and exhaust stroke. The system amplifies this frictional anomaly signal through span difference calculation.

[0013] Simultaneously, the scratches create channels for gas leakage, causing compressed air to leak towards the crankcase, resulting in a reduction in the mass of fluid that can be discharged during the exhaust stroke and causing a drop in instantaneous exhaust pressure. The system utilizes the synchronous offset generated by the same physical defect in two different physical fields to form a cross-validation diagnosis, achieving cylinder bore scratch location without disassembly.

[0014] Preferably, the steps of collecting torque data of the engine assembly under constant speed and extracting the maximum torque, minimum torque and average torque of each cylinder in the engine assembly under test from the torque data specifically include: setting the data acquisition starting point to 90° before the top dead center of the first cylinder of the engine assembly under test, and continuously collecting dynamic torque data covering a 720° crankshaft rotation angle from the data acquisition starting point as torque data;

[0015] The torque data corresponding to a 720° crankshaft angle is divided into four independent single-cylinder data windows, each occupying a 180° crankshaft angle span. Within each single-cylinder data window, the maximum and minimum torque values ​​within the single-cylinder data window are filtered and locked by comparing the values ​​of each sampling point in the continuous torque sequence with the values ​​in the temporary register in real time, and the average torque value within the single-cylinder data window is calculated.

[0016] Preferably, the step of calculating the single-cylinder torque span value of the target cylinder based on the maximum and minimum torque values ​​specifically includes: determining the single-cylinder data window corresponding to the target cylinder; subtracting the maximum and minimum torque values ​​within the single-cylinder data window to obtain the absolute difference; and calculating and defining the absolute difference as the single-cylinder torque span value.

[0017] The signal principle of the single-cylinder torque span value lies in the fact that, under abnormal friction conditions, the positive driving torque required to maintain the target constant speed will be superimposed with an additional frictional resistance, while the reverse braking torque required to overcome rotational inertia will be subtracted from the aforementioned additional frictional resistance. Since the positive and negative offsets change in opposite directions, the difference between the maximum and minimum values ​​can be used to superimpose the originally singular additional frictional resistance, amplifying the fault characteristic signal to twice the numerical level and preventing it from being masked by machining variances and background noise.

[0018] Preferably, the steps for triggering the exhaust back pressure test include: controlling the sealing head of the exhaust back pressure test device to tightly press against the exhaust flange end face of the engine assembly under test, establishing a closed exhaust back pressure test chamber; maintaining the target test speed to continuously drive the engine assembly under test, limiting the duration of the exhaust back pressure test to 3 to 5 complete four-stroke working cycles; and using the exhaust pressure sensor of the exhaust back pressure test device to collect real-time exhaust instantaneous pressure data inside the exhaust back pressure test chamber for subsequent collection of the single-cylinder exhaust peak pressure of the target cylinder.

[0019] Preferably, the steps for collecting the peak exhaust pressure of a single cylinder in the exhaust back pressure test specifically include: establishing a data mapping relationship between the instantaneous exhaust pressure data and the crankshaft angle of the engine assembly under test; extracting continuous instantaneous exhaust pressure data within the single-cylinder data window of the corresponding exhaust stroke of the target cylinder; performing moving average filtering on the extracted instantaneous exhaust pressure data, and traversing the filtered data set to find the maximum value point, extracting the found maximum value point as the peak exhaust pressure of a single cylinder.

[0020] The principle of mass conservation of fluid inside the cylinder states that, assuming a constant intake charge, the change in the total amount of gas expelled during the exhaust stroke is entirely determined by the amount of blow-by caused by bore scratches. Under the constraint of a constant exhaust port cross-sectional area, a decrease in the total amount of gas expelled directly translates to a decrease in the peak exhaust pressure of a single cylinder.

[0021] Preferably, the method provided by the present invention further includes a step of determining defects in purely mechanical bearings, specifically including: when the single-cylinder torque span value exceeds the span determination limit and the single-cylinder exhaust peak pressure is lower than the single-cylinder exhaust peak pressure of a sealed normal cylinder, determining the difference between the single-cylinder exhaust peak pressure of the sealed normal cylinder and the single-cylinder exhaust peak pressure.

[0022] When the difference between the peak exhaust pressure of a normally sealed cylinder and the peak exhaust pressure of a single cylinder is less than the threshold for judging abnormal exhaust pressure drop, the tested engine assembly is determined to have a purely mechanical bearing defect. This indicates that the increased mechanical friction resistance shown in the torque data is not accompanied by the failure of the fluid sealing layer, thus pinpointing the fault source to purely mechanical moving parts such as the main bearing or connecting rod bearing that do not participate in cylinder sealing.

[0023] Preferably, the method provided by the present invention further includes a step of determining the span judgment limit, specifically including: extracting the normal single-cylinder torque span value corresponding to the qualified engine sample without scratch defects, and calculating the average span value of the normal single-cylinder torque span value; superimposing 3 times the statistical standard deviation on the average span value as the upper limit of the qualified envelope; and setting the upper limit of the qualified envelope as the span judgment limit.

[0024] Preferably, the method provided by the present invention further includes the step of determining the single-cylinder exhaust peak pressure of a properly sealed cylinder, specifically including: synchronously collecting the cold test exhaust pressure of qualified engine samples that have been verified to have no cylinder bore scratch defects, extracting the pressure peaks within the working cycle of each corresponding cylinder; calculating the statistical average value of the pressure peaks; and solidifying the statistical average value as a system comparison benchmark, which is used as the single-cylinder exhaust peak pressure of a properly sealed cylinder.

[0025] Preferably, the method provided by the present invention further includes a step of dynamically updating the judgment limit, specifically including: during the continuous collection of a large number of assembly cold test data of the verification engine, real-time monitoring of the data distribution trend of the single-cylinder torque span value and the extracted single-cylinder exhaust peak pressure; when it is found that the test values ​​of the tested engine assembly deviate from the normal concentrated distribution cluster and show an abnormal discrete state, the discrete samples are systematically marked.

[0026] The test extreme values ​​of qualified discrete samples that have been physically disassembled and confirmed to have no internal scratches are used as the normal envelope boundary. Based on the normal envelope boundary, the normal envelope boundary is shifted upward to set as a new span judgment limit, and the normal envelope boundary is shifted downward to set as a new judgment bottom line for the single-cylinder exhaust peak pressure of the sealed normal cylinder.

[0027] Preferably, before collecting torque data of the engine assembly under constant speed conditions, the process also includes setting the torque measurement environment, specifically including: setting the stable range of speed 150 rpm to 300 rpm below the maximum test speed as the target test speed, and dragging the engine assembly under test to perform low-speed lubrication operation and break-in operation.

[0028] Through low-speed lubrication and break-in operation, oil pressure is established inside the tested engine assembly, causing a continuous lubricating oil film to form on the contact surfaces of the piston, cylinder, and crankshaft. After the lubricating oil film has stably adhered and the operating conditions have stabilized, the torque measurement program is then executed. The steps for setting the torque measurement environment are used to eliminate torque noise interference caused by initial dry friction and oil pressure fluctuations, ensuring the complete representation of torque physical quantity acquisition density and defect characteristics.

[0029] The above solution achieves the following beneficial technical effects:

[0030] This invention employs a combined diagnostic approach that integrates torque data extraction and exhaust back pressure testing. Under cold test conditions, it first calculates the single-cylinder torque span of the target cylinder based on the extracted maximum and minimum torque values. When the span exceeds the limit, it triggers an exhaust back pressure test to collect the peak exhaust pressure of a single cylinder. By cross-comparing the abnormal characteristics of mechanical friction resistance with the characteristics of fluid seal leakage, an exclusive closed-loop diagnosis is formed. This solves the problem that existing technologies cannot accurately distinguish between cylinder bore scratches and main bearing scratches by relying solely on a single torque fluctuation. Without disassembling the engine assembly under test, it achieves accurate location and early interception of cylinder bore scratch defects.

[0031] This invention employs an extraction algorithm for single-cylinder torque span characteristics. It divides a continuous torque sequence covering a 720° crankshaft angle into independent single-cylinder data windows proportionally according to the firing order. The absolute difference between the maximum and minimum torque values ​​within each single-cylinder data window is calculated. Utilizing the physical characteristic that positive and negative torques shift synchronously in opposite directions under abnormal friction, the single additional friction torque is superimposed and amplified in the difference calculation. This amplifies weak fault characteristic signals, preventing minor scratches from being masked by machining basis dispersion or background noise, and improving the sensitivity of identifying shallow scratch defects.

[0032] This invention establishes a sealed cavity by setting up an exhaust back pressure test stage and a corresponding sealing test device, controlling the sealing head to press against the exhaust flange. Under constant speed and pure drag conditions, the peak exhaust pressure of a single cylinder is collected and extracted. Utilizing the characteristics of gas hydrodynamics, the small cylinder leakage caused by cylinder bore scratches is directly converted into a reduction in the total amount of gas discharged during the exhaust stroke and a significant attenuation of the instantaneous exhaust pressure. Combined with moving average filtering, high-frequency airflow pulsation interference is further eliminated, providing stable and reliable data support for confirming whether the target cylinder has physical seal failure. This fully meets the high-volume, high-cycle testing requirements of engine assembly lines. Attached Figure Description

[0033] Figure 1 This is a flowchart of the method of the present invention;

[0034] Figure 2 This is a graph showing the measured torque of the present invention.

[0035] Figure 3 This is a measured curve of the exhaust back pressure of the present invention;

[0036] Figure 4 This is a diagram showing the scratches on the cylinder bore of the present invention. Detailed Implementation

[0037] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] See attached document Figure 1 The present invention includes a cylinder bore scratch detection system based on an engine cold test machine. The detection system includes an AC servo drive motor, a dynamic torque sensor, a double-ended flexible coupling, a safety overload protection clutch, and a crankshaft adaptive docking fixture.

[0039] The AC servo drive motor, dynamic torque sensor, double-ended flexible coupling, safety overload protection clutch, and crankshaft adaptive docking fixture are coaxially aligned and arranged in series between the crankshaft of the drive system and the engine assembly under test.

[0040] The AC servo drive motor provides constant speed closed-loop control for uniformly driving the tested engine assembly to rotate. The dynamic torque sensor adopts a strain-type elastic shaft structure to sense the resistance torque micro-deformation during rotation and synchronously outputs torque numerical signals and speed pulse signals. The double-ended elastic coupling connects the output end of the AC servo drive motor, the two ends of the dynamic torque sensor, and the crankshaft adaptive docking fixture, respectively.

[0041] The double-ended flexible coupling is used to compensate for coaxiality deviation and absorb operating vibration. The safety overload protection clutch is arranged in series between the dynamic torque sensor and the engine assembly under test. It is used to disengage the transmission connection when the resistance suddenly increases. The crankshaft adaptive docking fixture is used to clamp the flywheel or the rear end of the crankshaft of the engine assembly under test to establish a backlash-free rigid transmission docking.

[0042] The testing system also includes an exhaust back pressure testing device, which includes an exhaust pressure sensor, a plug, a ball valve, a muffler, a drive cylinder, and a slide. The drive cylinder is mounted on the slide and is used to drive the plug to block the exhaust port of the engine assembly under test. The exhaust pressure sensor is connected in the air passage where the plug is located and is used to collect the instantaneous exhaust pressure data of the engine assembly under test during the exhaust stroke.

[0043] The machine tool automatically seals and clamps the engine assembly under test, the AC servo drive motor pulls the engine assembly under test to rotate, and the detection system controls the engine assembly under test to perform a starting torque test.

[0044] After the starting torque test is passed, the testing system enters the operating torque test phase. In the operating torque test phase, the testing system collects the torque data of the tested engine assembly under constant speed conditions and extracts the maximum torque, minimum torque and average torque of each cylinder.

[0045] When the maximum torque, minimum torque, and average torque meet the standard limits, the test result is normal, and the machine tool automatically retracts the sealing. When the maximum torque, minimum torque, or average torque exceeds the standard limits, the test system enters the torque span calculation stage.

[0046] The torque span calculation step calculates the single-cylinder torque span value of the target cylinder based on the extracted maximum and minimum torque values. When the single-cylinder torque span value exceeds the preset span judgment limit, the detection system determines that there is abnormal friction inside the engine assembly under test. At this time, since the torque data alone cannot accurately distinguish between tile scratches and cylinder bore scratches, the detection system then triggers the exhaust back pressure test step for joint diagnosis.

[0047] In the exhaust back pressure test, the exhaust pressure sensor collects the peak exhaust pressure of the target cylinder. When the peak exhaust pressure of the single cylinder is lower than the peak exhaust pressure of the single cylinder under normal conditions, the detection system determines that there is cylinder bore scratch in the engine assembly under test.

[0048] The setting for engine cold test torque measurement needs to be performed under conditions of constant speed and stable friction. It is used to identify machining and assembly defects in the internal components of the engine assembly under test. The setting of the torque measurement environment specifically includes setting the timing of torque measurement and the test speed.

[0049] The testing system sets the execution timing of the torque measurement program. The testing system arranges the execution node of the running torque measurement in the middle and rear section of the entire cold test process. The AC servo drive motor drives the engine assembly under test to perform low-speed lubrication and break-in operation in the early stage of the cold test process.

[0050] Low-speed lubrication and break-in operations establish the internal oil pressure of the engine assembly under test, allowing a continuous lubricating oil film to form on the contact surfaces of the piston, cylinder, and crankshaft. Once the lubricating oil film is stably adhered and the moving parts are operating stably, the testing system executes the torque measurement program. By scheduling the torque measurement program after the low-speed lubrication and break-in operations, the torque data measured by the dynamic torque sensor can reflect the actual friction and resistance changes of the engine assembly under stable operating conditions. This avoids interference from torque noise caused by dry friction or oil pressure fluctuations during the early cold test period on the determination of cylinder bore scratches, thus improving the consistency of test data.

[0051] The test speed of the torque measurement program in the detection system needs to be set by comprehensively considering the time resolution of data acquisition and the efficiency of mass production testing. If the speed of the AC servo drive motor driving the engine assembly under test is set too high, the crankshaft angle of the engine assembly under test will change faster. If the crankshaft angle changes too fast, the local friction peaks caused by cylinder bore scratches will be missed within the limited sampling period of the dynamic torque sensor, reducing the detection rate of scratch defects.

[0052] The detection system sets the target test speed of the torque measurement program to a stable range of 150 rpm to 300 rpm below the maximum test speed. This speed range limit ensures that the dynamic torque sensor can collect a sufficient density of torque physical quantities within a single crankshaft rotation angle, so that subtle defect characteristics are fully reflected in the test curve, while also taking into account the test cycle set by the assembly line.

[0053] For the constant speed closed-loop control strategy of AC servo drive motor to maintain the target test speed and the anti-interference filtering method of dynamic torque sensor, those skilled in the art can use conventional proportional-integral-derivative control algorithm and low-pass filter circuit to implement it.

[0054] After confirming that the engine assembly under test has reached the target test speed, the detection system performs synchronous acquisition and window division of torque data.

[0055] The detection system establishes the acquisition benchmark and acquisition cycle for torque data. The data acquisition starting point is set at 90° before the top dead center of the first cylinder of the engine assembly under test. The dynamic torque sensor continuously acquires dynamic torque data covering a 720° crankshaft rotation angle starting from the data acquisition starting point. The 720° crankshaft rotation angle corresponds to a complete four-stroke working cycle of the engine assembly under test. The complete four-stroke working cycle covers the torque changes during the compression process and the torque changes during the expansion and power stroke of all cylinders.

[0056] The testing system defines torque data characteristics based on the mechanical working state inside the engine assembly under test. When the cylinder piston is in the upward compression stroke, the gas compression reaction force inside the cylinder forms a resistance torque. The AC servo drive motor outputs positive torque to overcome the resistance torque and maintain a constant speed. The positive torque reaches its maximum value when the piston reaches the top dead center of the compression stroke. After the piston passes the top dead center of the compression stroke and enters the expansion and exhaust stroke, the compressed gas inside the cylinder releases energy, forming a positive pushing torque. The AC servo drive motor correspondingly reduces its output torque and outputs a reverse braking torque to counteract rotational inertia. The reverse braking torque reaches its minimum value during the expansion and exhaust stroke.

[0057] The testing system divides the continuous torque data within a 720° crankshaft rotation period into single-cylinder data windows. For the four-cylinder engine assembly under test, the system divides the 720° crankshaft rotation period into four independent single-cylinder data windows according to the firing order 1-3-4-2. These four independent single-cylinder data windows correspond sequentially to the complete working process of the first, third, fourth, and second cylinders of the engine assembly under test. Each single-cylinder data window occupies a 180° crankshaft rotation span.

[0058] The detection system independently extracts feature data within each defined single-cylinder data window. Key extracted feature parameters include the maximum, minimum, and average torque values ​​within each single-cylinder data window, as well as the crankshaft angle corresponding to the extreme torque values. The maximum torque corresponds to the peak torque at top dead center of the compression stroke of a single cylinder, while the minimum torque corresponds to the valley torque during the expansion and exhaust stroke. The average torque reflects the overall resistance load level of a single cylinder within one working stroke. These extracted feature parameters are used to compare the operational consistency between the cylinders of the tested engine assembly, providing fundamental data for subsequent judgment of abnormal fluctuations in single cylinders.

[0059] For peak-finding algorithms and data statistics algorithms that extract torque extrema within a data window and calculate the average torque value, those skilled in the art can use conventional extrema search programs and integral averaging programs.

[0060] Under the pure drag test condition without ignition, the total rotational torque collected in real time by the dynamic torque sensor outputting the AC servo drive motor is equal to the sum of the various mechanical resistance loads inside the engine assembly under test. The detection system establishes a dynamic decomposition model of the cold test rotational torque, decomposing the total rotational torque collected in real time by the dynamic torque sensor into three parts of resistance load. The formula of the above dynamic decomposition model is:

[0061] ;

[0062] In the formula, This indicates the total rotational torque collected in real time by the dynamic torque sensor; Indicates the compression resistance torque; This represents the frictional resistance torque; This indicates the torque loss of the mechanism.

[0063] Compression resistance torque refers to the load resistance torque generated when the piston compresses the air inside the cylinder during its upward movement, and the air acts in the opposite direction on the piston. Compression resistance torque is the source of torque fluctuations during cold testing. The better the compression sealing of each cylinder, the more stable the compression resistance torque. When there are bore scratches on the cylinder wall, the scratches will cause compressed air leakage inside the cylinder. This air leakage causes a local attenuation of the cylinder's compression resistance torque and distorts the waveform of the total rotational torque collected in real time by the dynamic torque sensor. The compression resistance torque reaches its peak value near the top dead center of the compression stroke. The detection system extracts the value at this peak value to evaluate the cylinder sealing performance.

[0064] Frictional resistance torque refers to the resistance torque generated by the contact friction between various moving parts inside the tested engine assembly. The moving parts specifically include the moving pairs formed by the piston and cylinder wall, piston rings and cylinder wall, and crankshaft and connecting rod. When there are scratches on the cylinder wall, the lubricating oil film on the cylinder wall surface is damaged. The damage to the lubricating oil film leads to an increase in the contact resistance between the piston rings and the cylinder wall. The increase in contact resistance will cause the frictional resistance torque to rise, which directly causes the total rotational torque collected by the dynamic torque sensor in real time to be too large in a local stroke and form an abnormal torque peak.

[0065] Mechanical loss torque refers to the inherent mechanical loss torque generated by the auxiliary structures of the tested engine assembly during operation. The auxiliary structures specifically include the valve train, oil seals, and timing system. The value of mechanical loss torque tends to stabilize under constant speed conditions. The testing system obtains the mechanical loss torque of the standard prototype by performing multi-cycle drag tests on a known defect-free standard prototype before testing, and then solidifies it as the background resistance benchmark value for subsequent comparison of the differences in the tested engine assembly.

[0066] The detection system collects the change data of total rotational torque in real time through the dynamic torque sensor, and based on the above-mentioned dynamic decomposition model, focuses on monitoring the local abnormal fluctuations caused by compression resistance torque and friction resistance torque, thereby providing a data basis for determining whether there are cylinder bore scratches in the tested engine assembly.

[0067] The threshold for determining whether there is an anomaly is obtained by statistical analysis of a large sample of qualified engines. The detection system calculates the average value of the total rotational torque collected in real time by the dynamic torque sensor of the qualified engine sample as the baseline, and sets the range of 10% to 20% above the baseline as the threshold for determining that the torque is too large.

[0068] When the engine assembly under test is in a non-ignition drag state, the detection system analyzes the abnormal state of the friction pair through a dynamic decomposition model. When there are cylinder bore scratches or bearing damage inside the engine assembly under test, the lubrication state between the piston ring and the cylinder wall or between the crankshaft and the bearing changes, causing the originally stable frictional resistance torque to deviate. Under the logic of the AC servo drive motor maintaining a constant speed closed-loop control, the change in friction loss will directly cause the total rotational torque collected in real time by the dynamic torque sensor to deviate.

[0069] The detection system analyzes the torque offset mechanism during the forward drag phase. When the cylinder piston is in the compression stroke, the AC servo drive motor needs to overcome the compression resistance torque and frictional resistance torque. When friction defects such as cylinder bore scratches occur, the frictional resistance torque increases. In order to counteract the additional resistance and maintain the target test speed, the AC servo drive motor must increase the torque output. The torque output relationship under abnormal working conditions is expressed as follows:

[0070] ;

[0071] In the formula, This indicates the positive driving torque required for an AC servo drive motor to maintain a constant speed under abnormal operating conditions. This indicates the standard positive driving torque under normal operating conditions; This refers to the additional torque caused by defects such as cylinder bore scratches or abnormal bearing plates, which result in an extra increase in frictional resistance.

[0072] The detection system analyzes the torque offset mechanism during the reverse braking phase. After the cylinder piston passes the top dead center of compression and enters the expansion stroke, the energy released by the compressed air inside the cylinder is converted into the power to drive the crankshaft to rotate. At this time, the AC servo drive motor is in a braking state, absorbing the expansion energy by outputting reverse torque to prevent the instantaneous speed of the tested engine assembly from exceeding the target test speed. If there is abnormal friction caused by cylinder bore scratches, the abnormal friction will consume some of the expansion energy, resulting in a reduction in the braking load that the AC servo drive motor needs to bear. The reverse braking torque relationship at this time is expressed as:

[0073] ;

[0074] In the formula, This indicates the reverse braking torque output by the AC servo drive motor to overcome the crankshaft rotational inertia and fallback resistance under abnormal operating conditions; This indicates the standard reverse braking torque under normal operating conditions. This refers to the additional torque caused by defects such as cylinder bore scratches or abnormal bearing plates, which result in an extra increase in frictional resistance.

[0075] The standard forward drag torque and standard reverse braking torque under normal operating conditions are determined as follows: The testing system pre-collects cold test data from 50 to 100 known qualified engine samples without cylinder bore scratches, calculates the average torque of these samples at each crankshaft angle within the corresponding cylinder working window in the target test speed range of 150 rpm to 300 rpm, and then solidifies the calculated average values ​​as the standard forward drag torque and standard reverse braking torque under normal operating conditions, respectively.

[0076] The method for determining the additional torque caused by defects such as cylinder bore scratches and abnormal bearings is as follows: By artificially setting scratches with a depth of 0.1mm to 0.3mm on the inner wall of the cylinder of the standard prototype, the deviation of the total rotational torque collected in real time by the dynamic torque sensor from the standard curve is measured, thereby establishing a quantitative correspondence between the physical size of the scratches and the additional torque caused by defects such as cylinder bore scratches and abnormal bearings.

[0077] Normally, the threshold for determining the additional torque caused by defects such as cylinder bore scratches and abnormal bearings is set to 5Nm to 15Nm. When the detection system detects that the absolute values ​​of both the positive and negative offsets exceed this threshold, it is preliminarily determined that there is abnormal friction inside the engine assembly being tested.

[0078] The detection system can identify abnormal friction inside the engine assembly by comparing the positive offset of the total rotational torque collected in real time by the dynamic torque sensor during the compression stroke and the negative offset during the expansion stroke. Based on the physical characteristics of synchronous offset of positive and negative torque, it provides a theoretical basis for subsequent extraction of torque span characteristics and amplification of defect signals, ensuring that torque fluctuations caused by changes in cylinder sealing and torque anomalies caused by changes in friction properties can be effectively distinguished.

[0079] After completing the division of the single-cylinder data window, the detection system extracts the extreme values ​​within the single-cylinder data window and constructs feature parameters that can amplify the fault signal.

[0080] The detection system performs numerical traversal and comparison procedures within each 180° crankshaft angle span of a single-cylinder data window to identify torque peaks and troughs. Due to the physical characteristics of the tested engine assembly during the compression and expansion strokes, the maximum torque within a single-cylinder data window typically occurs within ±10° of the piston's top dead center position during compression, while the minimum torque typically occurs in the middle to late stages of the expansion stroke. The detection system acquires a continuous torque sequence within the single-cylinder data window through high-frequency sampling. By comparing the value of each sampling point in the sequence with the value in the temporary register in real time, it filters and locks two feature points: the maximum torque within the single-cylinder data window and the minimum torque within the single-cylinder data window.

[0081] The detection system constructs a torque span feature using two extracted feature points. The calculated single-cylinder torque span value is defined as the absolute difference between the peak and valley torque values ​​within the same cylinder's working cycle. The formula for calculating the torque span feature is as follows:

[0082] ;

[0083] In the formula, This represents the calculated single-cylinder torque span value; This represents the maximum torque value within the single-cylinder data window, which corresponds to the positive driving torque required for the AC servo drive motor to maintain a constant speed under abnormal operating conditions. This represents the minimum torque value within the single-cylinder data window. Numerically, it corresponds to the reverse braking torque output by the AC servo drive motor under abnormal operating conditions, overcoming the crankshaft's rotational inertia and fallback resistance.

[0084] The detection system analyzes the amplification mechanism of the single-cylinder torque span value calculated by mathematical model on the defect signal. Based on the aforementioned torque offset mechanism under abnormal friction conditions, when the tested engine assembly has cylinder bore scratches, the positive torque will increase by an additional amount, and the negative torque will decrease by an additional amount. Substituting the offset relationship into the above torque span characteristic calculation formula, the following derivation formula is obtained:

[0085] ;

[0086] In the formula, This indicates the standard positive driving torque under normal operating conditions; This indicates the standard reverse braking torque under normal operating conditions. This refers to the additional torque caused by defects such as cylinder bore scratches or abnormal bearing plates, which result in an extra increase in frictional resistance.

[0087] The above derivation process clarifies that under normal operating conditions, the calculated single-cylinder torque span value is merely the difference between two standard values. However, under abnormal operating conditions with cylinder bore scratches, because the positive and negative offsets change in opposite directions, the additional torque resulting from the increased frictional resistance caused by defects such as cylinder bore scratches and bearing abnormalities is superimposed in the difference calculation, amplifying the fault characteristic signal to twice its original value. This signal amplification mechanism solves the technical problem that, in mass production environments, minute frictional anomalies are easily masked by the machining variations of the tested engine assembly and the background noise of the dynamic torque sensor, thus improving the sensitivity for identifying shallow scratch defects.

[0088] The detection system sets a span judgment limit for determining the calculated single-cylinder torque span value. The method for determining the above span judgment limit is as follows: statistically analyze the span distribution of a large sample of qualified engines, calculate the average span value of the qualified samples, and add 3 times the standard deviation to the average span value as the upper limit of the qualified envelope.

[0089] For four-cylinder engines in passenger vehicles, at target test speeds of 150 rpm to 300 rpm, the calculated single-cylinder torque range under normal operating conditions typically fluctuates between 250 Nm and 350 Nm. If the single-cylinder torque range exceeds the upper limit of the aforementioned acceptable envelope, the detection system can determine that the tested engine assembly has a suspected friction defect and trigger a subsequent exhaust back pressure joint detection procedure. Through this method, the detection system transforms the raw data collected by the dynamic torque sensor into a recognizable feature vector, providing data support for pinpointing the source of cylinder bore scratches.

[0090] For the specific code logic implementation of the numerical traversal and comparison program, those skilled in the art can use conventional loop comparison instructions and extreme value register access logic. The above-mentioned numerical traversal and comparison program is a well-known technology in this field and will not be described in detail here.

[0091] After the detection system identifies abnormal friction inside the engine assembly under test by the torque span characteristics, since the specific mechanical fault source cannot be located by relying solely on torque data, and cylinder bore scratches and bearing scratches cannot be directly distinguished, the detection system introduces exhaust back pressure test based on fluid dynamics characteristics for joint discrimination.

[0092] Under cold, un-ignited, pure compressed air operation conditions, the cylinder sealing performance of the tested engine assembly directly affects the fluid pressure in the exhaust manifold. Cylinder bore scratches disrupt the mechanical fit between the piston rings and the cylinder wall, leading to cylinder seal failure. The physical gaps created by cylinder bore scratches cause high-pressure compressed air inside the cylinder to leak towards the crankcase during the compression and power strokes. The formula relating the cylinder leakage state to the physical dimensions of the cylinder bore scratches and the internal cylinder pressure is as follows:

[0093] ;

[0094] In the formula, Indicates the amount of blow-by gas in the cylinder; Represents the function for mapping the amount of gas leakage; Indicates the clearance of cylinder bore scratches; This indicates the instantaneous pressure inside the cylinder.

[0095] Cylinder bore scratches typically manifest as irregular wear grooves with a depth of 0.1mm to 0.3mm in physical structure. The blow-by volume mapping function shows a positive correlation, meaning that the larger the cylinder bore scratches and the higher the instantaneous pressure inside the cylinder, the greater the blow-by volume. As the blow-by volume increases, the actual mass of gas that the tested engine assembly can expel through the exhaust valve during the exhaust stroke changes. Based on the law of conservation of mass of the fluid inside the cylinder, the formula for modeling the total amount of gas expelled during the exhaust stroke is:

[0096] ;

[0097] In the formula, Indicates the total amount of gas discharged during the exhaust stroke; Indicates the intake air volume; This indicates the amount of residual exhaust gas in the cylinder.

[0098] Under the premise of constant speed during cold testing and stable absolute pressure of the intake environment, the intake charge of the tested engine assembly in a single stroke remains constant. The amount of residual exhaust gas in the cylinder is determined by the inherent clearance volume of the combustion chamber of the tested engine assembly, which is also a fixed parameter. The change in the total amount of gas discharged during the exhaust stroke is entirely determined by the fluctuation in the amount of blow-by gas.

[0099] The total amount of gas discharged during the exhaust stroke directly determines the fluid pressure characteristics at the end of the exhaust port of the tested engine assembly. Under the physical boundary conditions that the exhaust valve of the tested engine assembly opens according to the camshaft profile and the exhaust passage flow area is fixed, the ratio between the instantaneous exhaust pressure and the total amount of gas discharged during the exhaust stroke can be simplified as follows:

[0100] ;

[0101] In the formula, Indicates the instantaneous pressure of the exhaust gas; Indicates a proportional relationship; This indicates the cross-sectional area of ​​the exhaust duct.

[0102] The detection system establishes a fault-locking logic for cylinder bore scratches using the aforementioned fluid dynamics derivation model. When cylinder bore scratches exist in the target cylinder of the tested engine assembly, the scratch gap increases, leading to an increase in cylinder blow-by. This increased blow-by consumes the air mass originally enclosed within the cylinder and participating in the complete compression cycle, resulting in a decrease in the total exhaust gas volume during the final exhaust stroke. Under the constraint of a constant exhaust passage cross-sectional area, this reduction in the total exhaust gas volume ultimately manifests as a significant decrease in instantaneous exhaust pressure. The detection system identifies the specific location of the seal failure by monitoring and comparing the pressure extremes during the exhaust stroke of each cylinder. The inequality for determining exhaust characteristics is expressed as:

[0103] ;

[0104] In the formula, This indicates the peak exhaust pressure of a single cylinder in a cylinder with a cylinder bore scratch defect. This indicates the peak exhaust pressure of a single cylinder in a normally sealed cylinder.

[0105] The method for the detection system to obtain the peak exhaust pressure of a single cylinder of a normally sealed cylinder is as follows: at the target test speed of 150 rpm to 300 rpm, the detection system performs cold test exhaust pressure synchronously on a large number of qualified engine samples that have been verified to be free of cylinder bore scratches. The system extracts the peak pressure within the single cylinder data window and calculates the statistical average value, which is then solidified as the system comparison benchmark.

[0106] For conventional passenger car four-cylinder engines with a displacement of 1.5L to 2.0L, the preset peak exhaust pressure of a normally sealed cylinder is usually distributed in the range of 200kPa to 250kPa. The detection system sets the threshold for judging abnormal exhaust pressure drop to a drop of 20kPa to 30kPa relative to the system comparison benchmark. When the detection system measures that the peak exhaust pressure of the target cylinder is lower than the system comparison benchmark and the difference exceeds the judgment threshold, the fluid leakage characteristic is established. Combined with the mechanical friction characteristic of abnormally widened torque span in the previous step, the detection system forms an exclusive closed-loop diagnosis and accurately judges that there is cylinder bore scratch in the tested engine assembly.

[0107] For the high-frequency signal acquisition of instantaneous fluid pressure in the exhaust pipe and the experimental fitting of the specific mathematical coefficients of the blow-by volume mapping function, those skilled in the art can use conventional high-frequency dynamic pressure transmitter hardware combined with fluid simulation software for fitting.

[0108] After determining that there is frictional abnormality in the target cylinder during the torque span calculation stage, the detection system triggers the extraction program of single-cylinder exhaust back pressure peak characteristics online, and the detection system controls the exhaust back pressure testing device to establish a sealed testing environment.

[0109] After receiving the control command, the drive cylinder pushes the slide table to move towards the engine assembly under test. The slide table drives the sealing head to press tightly against the exhaust flange end face of the engine assembly under test. The flexible sealing gasket equipped on the end face of the sealing head deforms under pressure, cutting off the fluid passage between the internal exhaust passage of the engine assembly under test and the external atmospheric environment. The ball valve on the exhaust back pressure test device pipeline is simultaneously closed, so that the exhaust passage of the engine assembly under test and the exhaust back pressure test device together form a closed exhaust back pressure test chamber.

[0110] After the closed exhaust back pressure test chamber is established, the AC servo drive motor maintains the target test speed of 150rpm to 300rpm, continuously driving the engine assembly under test to run. The piston inside the engine assembly under test reciprocates in the cylinder, performing purely mechanical intake and exhaust actions. External air is continuously drawn into the cylinder through the intake port of the engine assembly under test, and is forcibly pumped into the exhaust back pressure test chamber by the piston during the exhaust stroke.

[0111] Because the end of the exhaust back pressure test chamber is blocked by a plug, the air pumped into the engine assembly under test accumulates inside the exhaust flow channel, causing the instantaneous exhaust pressure inside the exhaust back pressure test chamber to fluctuate periodically as the crankshaft rotates. To prevent the pressure inside the exhaust back pressure test chamber from accumulating indefinitely and exceeding the equipment's tolerance limit, the testing system limits the duration of the exhaust back pressure test to 3 to 5 complete four-stroke working cycles.

[0112] The exhaust pressure sensor probe is directly connected to the exhaust back pressure test chamber, which collects the instantaneous exhaust pressure signal inside the chamber in real time. The detection system simultaneously receives the speed pulse signal output by the dynamic torque sensor and establishes a data mapping relationship between the instantaneous exhaust pressure signal and the crankshaft angle of the engine assembly under test. The detection system uses the same window division rule as the torque data processing to divide the complete 720° crankshaft angle of the engine assembly under test into single-cylinder data windows corresponding to each cylinder's working cycle. The detection system extracts continuous instantaneous exhaust pressure data within the single-cylinder data window corresponding to the exhaust stroke of the target cylinder.

[0113] The detection system performs extreme value extraction on the intercepted instantaneous exhaust pressure data sequence. To eliminate high-frequency transient interference caused by airflow pulsation, the system performs moving average filtering on the intercepted instantaneous exhaust pressure data. Then, it iterates through the filtered dataset to find the maximum value point. The system defines the found maximum value point as the single-cylinder exhaust peak pressure. The feature extraction formula for the single-cylinder exhaust peak pressure is expressed as:

[0114] ;

[0115] In the formula, This represents the extracted peak exhaust pressure of a single cylinder; Indicates the instantaneous pressure of the exhaust gas; This represents the crankshaft angle variable within the single-cylinder data window corresponding to the exhaust stroke of the target cylinder. The value range is the 180° crankshaft angle span from the bottom dead center to the top dead center of the target cylinder piston. This represents a mathematical function that searches for the maximum value within a specified interval.

[0116] After the detection system obtains the extracted peak exhaust pressure of a single cylinder, it calculates the difference between the extracted peak exhaust pressure of a single cylinder and the peak exhaust pressure of a normally sealed cylinder. The peak exhaust pressure of a normally sealed cylinder is usually distributed in the range of 200 kPa to 250 kPa. The detection system sets the threshold for judging abnormal drop in exhaust pressure to 20 kPa to 30 kPa.

[0117] If the extracted peak exhaust pressure of a single cylinder is lower than the peak exhaust pressure of a normally sealed cylinder, and the difference between the peak exhaust pressure of a normally sealed cylinder and the extracted peak exhaust pressure is greater than or equal to the threshold for abnormal exhaust pressure drop, the detection system determines that the target cylinder of the tested engine assembly has a sealing failure, resulting in physical leakage characteristics. Combined with the previous torque span determination results, the detection system correlates the abnormal characteristics of purely mechanical frictional resistance with fluid sealing leakage characteristics, completing the data preparation for fault source localization.

[0118] For the temperature compensation calibration of the exhaust pressure sensor and the analog-to-digital conversion processing of the analog voltage signal, those skilled in the art can use conventional thermistor compensation circuits and high-frequency analog-to-digital conversion chips.

[0119] After acquiring the single-cylinder torque span value of the target cylinder and the extracted single-cylinder exhaust peak pressure, the detection system combines mechanical friction characteristics and fluid sealing characteristics to execute joint diagnostic logic. A single torque span test can detect abnormal mechanical resistance inside the tested engine assembly, but it cannot distinguish the specific physical location where the resistance is generated. When scratches occur on the friction mating surfaces of the crankshaft and main bearing inside the tested engine assembly, it will also cause an increase in frictional resistance torque, causing the waveform output by the dynamic torque sensor to shift in a positive or negative direction similar to cylinder bore scratches. The detection system cross-compares the data from the parallel exhaust back pressure dimension to convert the friction anomaly markers into clear defect location results.

[0120] The detection system analyzes the first data combination state to identify pure mechanical bearing defects. When the single-cylinder torque span value of the target cylinder exceeds the preset span judgment limit, but the instantaneous exhaust pressure characteristics of the same cylinder remain stable, the detection system determines that the tested engine assembly has tile scratches or other pure friction defects related to non-sealing pairs. The joint judgment logic formula for the state is:

[0121] ;

[0122] ;

[0123] In the formula, This indicates the torque span value of a single cylinder; Indicates the span determination limit; This indicates the peak exhaust pressure of a single cylinder in a normally sealed cylinder. This represents the extracted peak exhaust pressure of a single cylinder; This indicates the threshold for determining an abnormal drop in exhaust pressure.

[0124] The method for determining the span judgment limit is as follows: the detection system tests qualified engine samples without scratch defects, extracts the normal single-cylinder torque span value corresponding to the qualified engine sample, and adds 3 to 5 times the statistical standard deviation to the normal single-cylinder torque span value as the span judgment limit for judging abnormal friction.

[0125] Under the above data combination, the abnormal expansion of the single-cylinder torque span value proves that there is dry friction or micro-metal cutting resistance that blocks normal lubrication inside the tested engine assembly. The extracted single-cylinder exhaust peak pressure did not show a significant decrease and the pressure drop was less than the threshold for judging abnormal exhaust pressure drop, proving that the fluid sealing layer between the piston ring and the cylinder wall inside the target cylinder is intact and there is no leakage of compressed air towards the crankcase.

[0126] The abnormal friction location is on a purely mechanical moving part that does not participate in cylinder sealing. The detection system thus rules out the possibility of cylinder bore scratches and pinpoints the source of the fault to scratches or wear on the main bearing, connecting rod bearing, or camshaft journal.

[0127] The detection system analyzes the second data combination to pinpoint cylinder bore scratches. When the single-cylinder torque span of the target cylinder exceeds the span judgment limit, and the exhaust pressure of the same cylinder experiences synchronous decay, the detection system determines that the tested engine assembly has cylinder bore scratches. The joint judgment logic for cylinder bore scratches is expressed as follows:

[0128] ;

[0129] ;

[0130] In the formula, This indicates the torque span value of a single cylinder; Indicates the span determination limit; This indicates the peak exhaust pressure of a single cylinder in a normally sealed cylinder. This represents the extracted peak exhaust pressure of a single cylinder; This indicates the threshold for determining an abnormal drop in exhaust pressure.

[0131] Cylinder bore scratches have both mechanical friction and fluid leakage effects. The metal scoring grooves on the inner wall of the cylinder bore disrupt the oil lubrication film on the textured structure, increasing the dry friction resistance when the piston rings slide up and down. Simultaneously, the metal scoring grooves create a physical channel for gas leakage. These two physical effects are directly reflected in the test data as an increase in the single-cylinder torque span and a decrease in the extracted single-cylinder exhaust peak pressure. Within the tested engine assembly, only the piston and cylinder wall pair can simultaneously affect both mechanical resistance torque and cylinder volume exhaust pressure. The detection system utilizes the influence of the same physical defect in two different physical fields. Through the aforementioned logical cross-verification, without disassembling the tested engine assembly, interference from other friction components is eliminated, thus completing the determination of the cylinder bore scratch defect inside the target cylinder.

[0132] For the program algorithm that performs the above-mentioned conditional branch judgment and multivariate logic comparison within the detection system, those skilled in the art can use the comparison instructions and logic AND instructions inside the conventional programmable logic controller to implement it.

[0133] Specific application examples:

[0134] See attached document Figure 2 To be continued Figure 4 , Figure 2 This is a measured torque curve according to an embodiment of the present invention. Figure 3 This is a measured curve of exhaust back pressure according to an embodiment of the present invention. Figure 4 This is a photograph of a cylinder bore scratch according to an embodiment of the present invention.

[0135] After completing the aforementioned joint diagnostic logic deployment, the detection system performs data closed-loop verification by disassembling a physical engine sample, and calibrates specific judgment limits based on the closed-loop verification results.

[0136] The testing system acquires complete cold-run data of the engine assembly under test. Following the 1-3-4-2 firing order of a four-cylinder engine, the system divides the 720° crankshaft rotation angle into four independent evaluation intervals. For the attached... Figure 2For the test samples shown, the detection system calculated and extracted the single-cylinder torque span value for each evaluation interval. The single-cylinder torque span value of the first cylinder in the crankshaft angle interval of 0° to 180° is 313.93 Nm; the single-cylinder torque span value of the third cylinder in the crankshaft angle interval of 180° to 360° is 312.35 Nm; the single-cylinder torque span value of the fourth cylinder in the crankshaft angle interval of 360° to 540° is 312.55 Nm; and the single-cylinder torque span value of the second cylinder in the crankshaft angle interval of 540° to 720° is 312.54 Nm. By comparing the above extraction results horizontally, the single-cylinder torque span value of the first cylinder shows an obvious abnormally large feature, which meets the preliminary judgment condition of exceeding the span judgment limit.

[0137] The detection system simultaneously retrieves transient data collected from the above test samples during the exhaust back pressure test. (See attached...) Figure 3 As shown, the detection system extracted the peak values ​​of the exhaust pressure curves of four single cylinders. The extracted peak exhaust pressure of the first cylinder was 82.12 kPa; the extracted peak exhaust pressure of the second cylinder was 200.45 kPa; the extracted peak exhaust pressure of the third cylinder was 205.32 kPa; and the extracted peak exhaust pressure of the fourth cylinder was 230.36 kPa. Through numerical comparison, the extracted peak exhaust pressure of the first cylinder showed a significant decrease compared to the other three cylinders, and the decrease was much greater than the threshold for judging abnormal exhaust pressure decrease. Combined with the abnormal characteristics of the torque dimension, the detection system determined that the first cylinder of the test sample had a dual composite characteristic of cylinder seal failure and sudden increase in mechanical friction, and locked that the first cylinder had cylinder bore scratches.

[0138] To verify the accuracy of the testing system's judgment, technicians physically disassembled the test samples. (See attached image.) Figure 4 As shown, after endoscopic detection and physical dissection, it was confirmed that there were obvious mechanical scratches and grooves on the inner wall of the first cylinder of the test sample. The location of the physical damage was completely consistent with the data judgment result output by the detection system, proving that the multi-physics cross-validation logic can accurately locate internal defects without disassembling the engine structure.

[0139] After establishing the aforementioned verification closed loop, the testing system executes mass production calibration procedures for the span judgment limit and the abnormal exhaust pressure drop judgment threshold. The testing system continuously collects assembly cold test data from a large batch of verification engines on the mass production assembly line and monitors the data distribution trends of the single-cylinder torque span value and the extracted single-cylinder exhaust peak pressure in real time. When the testing system detects that the test values ​​of a specific engine deviate from the normal concentrated distribution cluster and exhibit an abnormally discrete state, the testing system systematically marks the discrete samples. Technicians then intercept and physically disassemble the marked discrete samples in sequence.

[0140] The inspection system uses the extreme values ​​of the tests on qualified discrete samples, confirmed to be free of internal scratches, as the normal envelope boundary. Based on this normal envelope boundary, the system sets the upward offset of the normal envelope boundary as the span judgment limit and the downward offset of the normal envelope boundary as the judgment baseline for the single-cylinder exhaust peak pressure of a sealed, normal cylinder. Through mapping analysis between the physical state of the entity and the discrete trend of the data, the inspection system dynamically updates the judgment limits, enabling the system to adapt to the processing discreteness deviations of different batches of parts.

[0141] For the host computer software module of the detection system that performs distribution statistics and discrete trend plotting on massive test data, those skilled in the art can use conventional industrial big data analysis platforms and graphical report components to achieve this. Industrial big data statistics and discrete trend visualization are well-known technologies in this field.

Claims

1. A method for detecting cylinder bore scratches based on an engine cold test, characterized in that, Includes the following steps: The torque data of the engine assembly under test is collected under constant speed conditions, and the maximum torque, minimum torque and average torque of each cylinder in the engine assembly under test are extracted from the torque data. When the maximum torque, minimum torque, or average torque exceeds the standard limit, the cylinder that exceeds the limit will be designated as the target cylinder, and the single-cylinder torque span value of the target cylinder will be calculated based on the maximum torque and minimum torque. When the single-cylinder torque span value exceeds the preset span judgment limit, the exhaust back pressure test is triggered, and the single-cylinder exhaust peak pressure of the target cylinder is collected in the exhaust back pressure test. When the peak exhaust pressure of a single cylinder is lower than the preset peak exhaust pressure of a single cylinder of a normally sealed cylinder, and the difference between the peak exhaust pressure of the single cylinder of the normally sealed cylinder and the peak exhaust pressure of the single cylinder is greater than or equal to the threshold for judging abnormal exhaust pressure drop, it is determined that the tested engine assembly has cylinder bore scratches.

2. The method for detecting cylinder bore scratches based on an engine cold test machine according to claim 1, characterized in that, The steps of collecting torque data of the tested engine assembly under constant speed conditions and extracting the maximum, minimum, and average torque values ​​of each cylinder in the tested engine assembly from the torque data specifically include: The data acquisition starting point is set at 90° before the top dead center of the first cylinder of the engine assembly under test. Dynamic torque data covering 720° crankshaft rotation angle is continuously acquired from the data acquisition starting point as the torque data. The torque data corresponding to the 720° crankshaft angle is divided into four independent single-cylinder data windows, each single-cylinder data window occupies a 180° crankshaft angle span; Within each of the divided single-cylinder data windows, the maximum and minimum torque values ​​within the single-cylinder data window are filtered and locked by comparing the values ​​of each sampling point in the continuous torque sequence with the values ​​in the temporary register in real time, and the average torque value within the single-cylinder data window is calculated.

3. The method for detecting cylinder bore scratches based on an engine cold test machine according to claim 2, characterized in that, The steps for calculating the single-cylinder torque span value of the target cylinder based on the maximum torque value and the minimum torque value specifically include: Determine the single-cylinder data window corresponding to the target cylinder; The difference between the maximum torque value and the minimum torque value within the single-cylinder data window is calculated to obtain the absolute difference value. The absolute difference is calculated and defined as the single-cylinder torque span value.

4. The method for detecting cylinder bore scratches based on an engine cold test as described in claim 1, characterized in that, The steps for triggering the exhaust back pressure test include: The sealing head of the exhaust back pressure test device is tightly pressed against the exhaust flange end face of the engine assembly under test to establish a closed exhaust back pressure test cavity. The engine assembly under test is continuously driven to maintain the target test speed, and the duration of the exhaust back pressure test is limited to 3 to 5 complete four-stroke working cycles. The exhaust pressure sensor of the exhaust back pressure testing device is used to collect the instantaneous exhaust pressure data inside the exhaust back pressure testing chamber in real time, which is then used to collect the single-cylinder exhaust peak pressure of the target cylinder.

5. The method for detecting cylinder bore scratches based on an engine cold test as described in claim 4, characterized in that, The steps for collecting the peak exhaust pressure of a single cylinder of the target cylinder in the exhaust back pressure test specifically include: Establish a data mapping relationship between the instantaneous exhaust pressure data and the crankshaft angle of the engine assembly under test; Extract continuous instantaneous exhaust pressure data within the single-cylinder data window corresponding to the exhaust stroke of the target cylinder; The extracted instantaneous exhaust pressure data is subjected to moving average filtering, and the maximum value point is found by traversing the filtered data set. The maximum value point is then extracted as the single-cylinder exhaust peak pressure.

6. The method for detecting cylinder bore scratches based on an engine cold test as described in claim 1, characterized in that, The method also includes a step for determining defects in purely mechanical bearings, specifically including: When the single-cylinder torque span value exceeds the span judgment limit, and the single-cylinder exhaust peak pressure is lower than the single-cylinder exhaust peak pressure of the sealed normal cylinder, the difference between the single-cylinder exhaust peak pressure of the sealed normal cylinder and the single-cylinder exhaust peak pressure is determined. When the difference between the peak exhaust pressure of a single cylinder of the normally sealed cylinder and the peak exhaust pressure of the single cylinder is less than the threshold for judging abnormal exhaust pressure drop, it is determined that the engine assembly under test has a pure mechanical bearing defect.

7. The method for detecting cylinder bore scratches based on an engine cold test machine according to claim 2, characterized in that, The method further includes a step of determining the span limit, specifically including: Extract the normal single-cylinder torque span value corresponding to the qualified engine sample without scratch defects, and calculate the average span value of the normal single-cylinder torque span value; The average span value is superimposed with 3 times the statistical standard deviation as the upper limit of the qualified envelope; The upper limit of the qualified envelope is set as the span determination limit.

8. The method for detecting cylinder bore scratches based on an engine cold test machine according to claim 7, characterized in that, The method further includes the step of determining the peak exhaust pressure of a single cylinder of the sealed normal cylinder, specifically including: The qualified engine sample that has been verified to be free of cylinder bore scratches was subjected to cold test exhaust pressure synchronously, and the pressure peak value during the working cycle of each cylinder was extracted. Calculate the statistical average of the pressure peak values; The statistical average value is solidified as the system comparison benchmark and used as the single-cylinder exhaust peak pressure of the sealed normal cylinder.

9. The method for detecting cylinder bore scratches based on an engine cold test as described in claim 8, characterized in that, The method also includes a step of dynamically updating the judgment limit, specifically including: During the continuous collection of large-scale cold test data of the engine assembly, the distribution trend of the single-cylinder torque span value and the extracted single-cylinder exhaust peak pressure is monitored in real time. When the test values ​​of the engine assembly under test are found to deviate from the normal centralized distribution cluster and exhibit an abnormal discrete state, the discrete samples are systematically marked. The test extreme values ​​of qualified discrete samples that have been physically disassembled and confirmed to have no internal scratches are used as the normal envelope boundary. Based on the normal envelope boundary, the normal envelope boundary is shifted upward to set as the new span determination limit, and the normal envelope boundary is shifted downward to set as the new determination baseline for the single-cylinder exhaust peak pressure of the sealed normal cylinder.

10. The method for detecting cylinder bore scratches based on an engine cold test as described in claim 1, characterized in that, Before collecting torque data from the engine assembly under constant speed conditions, the process also includes setting up the torque measurement environment, which specifically includes: Set the stable range of speed from 150 rpm to 300 rpm below the maximum test speed as the target test speed, and drag the engine assembly under test to perform low-speed lubrication operation and break-in operation. Through the low-speed lubrication operation and break-in operation, oil pressure is established inside the tested engine assembly, causing a continuous lubricating oil film to form on the contact surfaces of the piston, cylinder and crankshaft. After the lubricating oil film has stably adhered and the operating condition has stabilized, the torque measurement program is then executed.

Images