Engine crankcase ventilation line leak detection system

Abstract

TITLE: Crankcase Gas Recirculation Line Leak Detection System Crankcase gas recirculation line (5) leak detection system of an internal combustion engine (Mx) connecting the outlet of the oil separator (4) to the intake line (6) between the air filter (7) and the turbocharger (8), comprising: a pressure sensor (C1, C2, C3) respectively downstream of the turbocharger (8), downstream of the throttle body (9), at the outlet of the oil separator (4), and - a learning machine (LMx) receiving the pressure signals (P1-3) from the sensors and the engine field signals from the engine (Mx) to generate a fault signal (SDj) by a combination of the pressure signals (P1-3), interpreted by the machine (LMx) as corresponding to a leak in the recirculation line (5). The learning machine (LMx) is prepared from a model learning machine (LMa) of a model detection system (100a), applied to a test engine (Ma).Figure 1.

Description

Title of the invention: Engine crankcase gas recovery line leak detection system FIELD OF INVENTION

[0001] The present invention relates to a leak detection system for the crankcase gas recovery line connecting the oil separator outlet to the engine intake line downstream of the air filter and upstream of the turbocharger.

[0002] Such detection systems are known. They require complicated implementation methods and their reliability, and especially compliance with the very strict regulations of certain countries, is not guaranteed.

[0003] PURPOSE OF THE INVENTION

[0004] The present invention aims to develop a leak detection system for the crankcase gas recovery line guaranteeing the detection of even a small leak other than a simple rupture of the intake line and adaptable to any type of internal combustion engine of a vehicle.

[0005] DESCRIPTION AND ADVANTAGES OF THE INVENTION

[0006] To this end, the invention relates to a leak detection system of the type defined above comprising a pressure sensor downstream of the turbocharger, a pressure sensor downstream of the throttle body, a pressure sensor at the outlet of the oil separator, a learning machine receiving the pressure signals from the sensors and the speed and load signals (engine field) of the engine to generate a fault signal by the pressures, interpreted by the learning machine as corresponding to a leak equal to a limit leak imposed by the regulations in the return line, the learning machine being obtained from a model learning machine of a model detection system, applied to a test engine, identical to the engines which will be equipped with the detection system and having operated under simulation conditions of leaks for the operating engine field (speed, load) of the engines. Brief description of the drawings

[0007] The present invention will be described in more detail below with reference to an embodiment of a crankcase gas leak detection system at the outlet of an oil separator shown in the accompanying drawing, in which:

[0008] [Fig. 1] Diagram of the detection system according to the invention,

[0009] [Fig.2] Schematic of the model detection system applied to a vehicle engine.

[0010] DESCRIPTION OF A METHOD OF EMBODIMENT OF THE INVENTION

[0011] Figure [1] shows the diagram of a 100x pipe leak detection system 5. Crankcase gas return of an Mx engine.

[0012] The Mx engine is schematically represented by a cylinder 1 and its crankcase 2 connected by a crankcase gas line 3 to the oil separator 4. The oil separator 4 separates the oil from the crankcase gases which arrive in the return line 5 via the PCV valve 51. The return line 5 opens into the air supply line 6 downstream of the air filter 7. The line 6 is connected to the turbocharger 8. The turbocharger supplies the throttle body 9 at the inlet of the intake manifold 10 connected to cylinder 1.

[0013] The return line 5 must be monitored as it may be the cause of crankcase gas leaks.

[0014] The operation of the internal combustion engine Ma is managed by a central unit 100a.

[0015] The system includes Cl-3 pressure sensors feeding an LMx learning machine also receiving operating signals: R speed and CH load of the Mx engine to generate an output signal confirming the tightness of the crankcase gas return line 5 or a fault signal SD indicating a leak in the return line 5 reaching a leak-limit SE imposed by regulation.

[0016] The imposed leakage limit is the smallest leak at any point on the return line 5, regardless of its location on the return line 5. This regulation imposes a very low threshold or leakage limit that must not be exceeded. This means that it is sufficient to detect the smallest leak, and not a series of small leaks whose cumulative effect could reach the leakage limit threshold.

[0017] The imposed threshold also disregards the condition of the actual engine equipped with the lOOx system, whether this engine is new or old, whether its air filter is clear or clogged, or even blocked.

[0018] According to [Fig.1], in the lOOx system the first sensor Cl is installed at the outlet of the turbocharger 8 to provide the pressure signal PI at the outlet of the turbocharger 8.

[0019] The second sensor C2 installed at the outlet of the throttle body 9 measures the air pressure at that location and provides a corresponding pressure signal P2.

[0020] The third sensor C3 is installed at valve 51 at the outlet of the oil separator 4 to measure the pressure and provide the corresponding signal P3. The sensor C3 associated with valve 51 can be installed upstream or downstream of it.

[0021] The pressure signals P1-3 from the Cl-3 sensors are transmitted to the LMx learning machine, which also receives signals representing the engine speed R and the load CH of the internal combustion engine Mx. By definition, the engine speed R is the rotational speed of the engine and the load CH is the amount of fuel introduced during each engine cycle.

[0022] The LMx learning machine monitors the feedback control 5 and emits a signal SD fault as soon as it detects a triplet of pressures T (Pl-3) associated with the leak-limit SE and meaning that a leak equal to the leak-limit SE is reached to warn the driver of the need to remedy this leak or possibly the programmed shutdown of the engine.

[0023] The learning of the LMx machine equipping the Mx series internal combustion engine of a vehicle is done with a model 100a detection system combined with a Ma test engine identical to the Mx series engines that equip a range of vehicles.

[0024] The 100X detection system, also called the "current" detection system, having the same structure as the model 100a detection system, the components of the two systems bear the same references except that the learning machine, called the current learning machine, bears the reference LMx and that of the model 100a system bears the reference LMa.

[0025] According to [Fig. 2], the model system 100a is applied to an engine Ma represented by a single cylinder 1 with an engine block 2 connected by a crankcase gas line 3 to an oil separator 4, itself connected by a valve 51 (PCV valve according to common terminology) to a return line 5 for the de-oiled crankcase gases. This line 5 is connected to the intake line 6 between the air filter 7 and the turbocharger 8. The outlet of the compressor 8 is connected to a throttle body 9, the outlet of which is connected by the manifold 10 to the combustion chamber of the engine 1.

[0026] The turbocharger outlet 8 is equipped with a pressure sensor Cl which measures the pressure PI and the manifold 10 downstream of the throttle body 9 is equipped with a pressure sensor C2 which measures the pressure P2 in the manifold.

[0027] Upstream or downstream of valve 51 the return line 5 is equipped with a pressure sensor C3 which provides the pressure P3.

[0028] The inlet of the pipe 5 at the inlet or outlet of the valve 51 is a leakage point bearing the reference El; the body of the pipe 5 bears the reference E2 and the junction of the pipe 5 with the supply pipe bears the reference E3.

[0029] The leakage points El, E2, E3 represent potential leakage points of the conduit 5, namely the junction at points El and E3 and the structure of the conduit 5 between these points El and E3.

[0030] At points El and E3, the leak may result from a failure of the junction going as far as the tearing of the return pipe 5. Between its two ends, the pipe 5 may have sealing defects such as perforations, or even a total rupture; these possibilities are schematically represented by the leak point E2.

[0031] The leakage points El, E2, E3 are each equipped with a leak simulator SF1, SF2, SF3 which are devices creating an artificial leak at the ends El, E3 of the return line 5 and an artificial leak from the body E3 of the return line 5 to be used for machine learning in LMa.

[0032] The SF1-3 leak simulators are connections on the return line 5 at points E1-E3, each having a calibrated hole (MF1-3) chosen by the engine manufacturer according to the limit leakage flow rate imposed by the standard.

[0033] For the learning of the LMa machine, a simulated leak is successively carried out at each leak point El-3 in any order, by the calibrated orifice MF1-3 associated with each leak point El-3.

[0034] A complete motor field (engine speed R, load CH) is then applied to this calibrated leak MF1 or MF2 or MF3 and the pressures PI, P2, P3 measured by sensors Cl, C2, C3 are recorded and which correspond at any instant to a triplet of pressures T(P1, P2, P3) associated with this limit leak MF1 or MF2 or MF3.

[0035] This learning phase will be applied to each SF1-3 leak simulator for its calibrated hole MF1 or MF2 or MF3 or the complete motor field.

[0036] The Pl-3 pressures are transmitted to the LMa model learning machine.

[0037] This yields a set of pressure triplets T(P1, P2, P3) representing the limiting leakage at one of the points El or E2 or E3. This leakage simulation is also applied to a motor Ma on which an absence of leakage is simulated, giving pressure triplets To (PI, P2, P3) for a complete motor field.

[0038] These two series of learning are carried out with the data of a normal engine also called nominal engine Mao whose air filter 7 works normally and also with the data of an engine Mal whose air filter 7 is clogged with limited porosity, but also with other engines MaX whose aging influences the flow of gas recirculation in the pipe 5, and therefore the triplet of pressures T(P1, P2, P3).

[0039] In all cases, whether the nominal Mao engine without a blocked air filter or an engine with a blocked air filter, the leak-free and leak-limit tests are applied to locations E1-E3 with the complete engine field.

[0040] Leaks and non-leaks are thus represented by the pressure triplets TF(P1, P2, P3) and To (PI, P2, P3) associated with the complete driving field (CH, R).

[0041] This learning from the LMa machine is then transferred to the current learning machine LMx for vehicles equipped with an internal combustion engine Mx corresponding to the nominal engine Mao and the Mal engine with blocked air filter 7.

[0042] For safety reasons, the additional learning, namely that corresponding to the leak-free Mao and Mal motors, is also applied to the LMa machine to avoid possible detection errors since, in principle, the leak-free TF (PI, P2, P3) triplet sets and the leak-free To (PI, P2, P3) triplet sets are disjoint sets.

[0043] The analysis of the fault signals SD for a limit leak SE and the pressures PI, P2, P3 associated with this threshold makes it possible to evaluate the importance of each pressure PI or P2 or P3 in this detection and to simplify the detection system 100a and its learning machine LMa by taking into account only such or such pressure and to eliminate a non-essential sensor for the detection of the leak with the accuracy required by the regulations.

[0044] After the realization of the model machine LMa, adapted to the Ma motor and the SE leakage threshold, the LMa machine can be copied into the current machine LMx of the lOOx detection system intended for the Mx motors.

[0045] NOMENCLATURE OF MAIN ELEMENTS

[0046] 100a Model detection system

[0047] lOOx Current detection system

[0048] 1 Internal combustion engine cylinder

[0049] 2 Crankcase

[0050] 3 Crankcase gas passage

[0051] 4 Degreaser

[0052] 5 Oil-free gas return line

[0053] 51 Valve

[0054] 6 Air supply line

[0055] 7 Air filter

[0056] 8 Turbocharger

[0057] 9 Throttle body

[0058] 10 Inlet manifold

[0059] CH Engine load

[0060] R Engine speed

[0061] Cl, C2, C3 Pressure sensors

[0062] El-3 Potential Escape Points

[0063] LMa Model Learning Machine

[0064] LMx Current Learning Machine

[0065] Mao Rated Motor

[0066] Max Motor current

[0067] PI Turbocharger outlet pressure

[0068] P2 Pressure downstream of the throttle body

[0069] P3 Pressure at valve outlet 51

[0070] T (PI, P2, P3) Triplet of pressures

[0071] TF (PI, P2, P3) Triplet of pressures corresponding to the leak-limit

[0072] To (PI, P2, P3) Leak-free pressure triplets

[0073] SE Leak threshold / leak-limit

[0074]

[0075]

[0076]

[0077]

[0078]

[0079]

[0080] SD Fault signal SF1 Leak simulator at leak point El SF2 Leak simulator at leak point E2 SF3 Leak simulator at leak point E3 MF1 Calibrated leak rate imposed at point El MF2 Calibrated leak rate imposed at point E2 MF3 Calibrated leak rate imposed at point E3

Claims

Demands

1. A leak detection system for the crankcase gas return line (5) (2) of an internal combustion engine (Mx) connecting the oil separator outlet (4) to the intake line (6) between the air filter (7) and the turbocharger (8), a system comprising: - a pressure sensor (Cl) downstream of the turbocharger (8), - a pressure sensor (C2) downstream of the throttle body (9), - a pressure sensor (C3) at the outlet of the oil separator (4), - a learning machine (LMx) receiving the pressure signals (Pl-3) from the sensors (Cl-3) and the engine speed (R) and load (CH) signals (engine field) from the engine (Mx) to generate a fault signal (SD) by the pressures (Pl-3), interpreted by the learning machine (LMx) as corresponding to a leak equal to a limit leak imposed by regulations in the return line (5), - the learning machine (LMx) being obtained from a model learning machine (LMa) of a model detection system (100a), applied to a test engine (Ma), identical to the engines (Mx) which will be equipped with the detection system and having operated under simulated leak conditions for the operating engine field (engine speed R, (CH) load of the motors (Mx),- The learning from the LMa machine is transferred to the current learning machine LMx for vehicles equipped with an Mx internal combustion engine corresponding to the nominal Mao engine and the Mal engine with a blocked air filter 7, - The additional learning corresponding to the leak-free Mao and Mal motors is also applied to the LMa machine to avoid possible detection errors.

2. Leak detection system according to claim 1, characterized in that the model system (100a) comprises: - potential leakage points (El, E2, E3) at the inlet, outlet and at a point on the return line (5) of the motor (Ma), - The leakage points (El-3) are equipped with leak simulators (SF El-3) simulating imposed limit leaks, - pressure sensors (Cl-3) measuring pressures (PI, P2, P3) downstream of the turbocharger (8), downstream of the throttle body (9) and at the outlet of the oil separator (4), - a model learning machine (LMa), * receiving pressure signals (Pl-3) associated with limit leakage imposed at points (El-3) of the return line (5) and the complete motor field of the model motor (Ma) with an unblocked air filter (Mao or blocked Mal), * comparing pressure signals to emit a fault signal SD if the leak threshold (SE) is exceeded.

Images