Detecting leaks in evaporative emission control systems

Abstract

In a vehicle with an internal combustion engine, a system retains evaporative emissions of hydrocarbon vapor from fuel. A computer detects the occurrence of a leak in that system by receiving (410) a fuel tank pressure value (MEASURED_PRESSURE) from a tank pressure sensor. While the engine is running, the computer interacts with a controller to create a vacuum condition in the system. The computer operates (430) a pre-trained network, with receiving (431) further sensor data (FURTHER_DATA) from the vehicle, and obtaining (432) a vacuum estimation value (ESTIMATED_PRESSURE) that represents the vacuum. The computer determines the leak by comparing (440) the measured pressure value (MEASURED_PRESSURE) with the estimated value (ESTIMATED_PRESSURE).

Description

DETECTING LEAKS IN EVAPORATIVE EMISSION CONTROL SYSTEMSTechnical Field

[0001] In general, the disclosure relates to vehicles having evaporative emission control systems to minimize fuel emissions, and more in particular, the disclosure relates to a computer and to a computer-implemented method used to detect and to eventually quantify leaks in such control systems.Background

[0002] Vehicles with combustion engines may emit fuel vapors to the atmosphere. The vapors can comprise hydrocarbons (HCs). From an environmental protection perspective, such emissions are not desired. The amounts of permissible emissions are defined in technical standards, usually referred to by acronyms, such as US-CARB 1968.2, Euro 6.2, CN6, and others.

[0003] Much simplified, fuel such as gasoline is injected from a fuel tank to the engine, and the vapors often escape from the tank or from other vehicle components.

[0004] To reduce emissions, vehicles are generally equipped with so-called evaporative emission control systems (also EVAP systems). Such systems usually comprise a canister with lines (or "pipes"), such as a line to and from the fuel tank, a purge line to the engine, a vent line to a filter, in combination with a number of valves. The hydrocarbon vapors can be temporarily stored in the canister (that has adsorbent materials, such as activated carbon), and from time to time the vapors are forwarded to the engine (or "purged", via the purge line) to be consumed by the engine.

[0005] Although that principle appears simple, leakages may occur at the valves, the pumps, the pipes, as well as at the fuel tank and so on, whereby the emission control systems periodically perform leakage tests. In other words, the systems have diagnostics functions.

[0006] It is known in the art to implement tank leakage diagnosis modules or the like, andsome of the modules have electrical pumps to apply pressure, and have solenoid valves to conduct air, etc. Module components may fail, and consequently a failure in the module may incorrectly show up as leak.

[0007] US 2014 / 0069394 Al byJentz et al. describes an evaporative emission control system that employs vapor canisters. There is a focus on detecting unintended closing of mechanical valves during leak tests.

[0008] There is a need to implement leak detection in an alternative approach.

[0009] In a vehicle with an internal combustion engine, The invention relates to a system configured to retain evaporative emissions of hydrocarbon vapor from fuel such as .e.g gasoline. The system is typically associated to an internal combustion engine, in particular where the engine is on-board a vehicle. A computer detects the occurrence of a leak in that system by receiving a fuel tank pressure value from a tank pressure sensor. While the engine is running, the computer interacts with a controller to create a vacuum condition in the system. The computer operates a pre-trained network, with receiving further sensor data from the vehicle, and obtaining a vacuum estimation value that represents the vacuum. The computer determines the leak by comparing the measured pressure value with the estimated value.

[0010] More in detail, the disclosure presents a computer-implemented method to determine the occurrence of a leak in a system that is adapted to retain evaporative emissions of hydrocarbon vapor from fuel in an engine. The engine is an internal combustion engine of a vehicle. The system comprises a vapor canister and a plurality of connecting lines, wherein a first line connects the vapor canister to a fuel tank of the engine system, a second line connects the vapor canister to an air intake manifold of the engine system via a purge valve, and a third line connects the canister to a filter.

[0011] A computer receives a fuel tank pressure value from a tank pressure sensor that is associated with the fuel tank. The fuel tank pressure value represents the pressure inrelation to the surrounding atmosphere.

[0012] While the engine is running, the computer interacts with a controller to create a vacuum condition in the vapor canister, in the first line, in the second line and in the third line, as well as in the fuel tank. Interacting comprises to instruct the controller at a first time point, to close the vent valve and to open the purge valve, and at a second time point, when the fuel tank pressure data reach a pre-defined first threshold pressure value, to close the canister purge valve.

[0013] The computer operates a pre-trained network, with receiving further sensor data from the vehicle, and obtaining a vacuum estimation value that represents the vacuum in the fuel tank.

[0014] The computer compares the fuel tank pressure value and the vacuum estimation value. Thereby the computer determines the occurrence of the leak if the numerical difference between the fuel tank pressure value and the vacuum estimation value exceeds a second threshold pressure value.

[0015] Optionally, while the engine of the vehicle is still running and while the computer is interacting with the controller, the computer determines the time interval between the second time point and a third time point when the fuel tank pressure value has reached a third threshold pressure value, as decay time that becomes part of the further sensor data in the receiving step.

[0016] Optionally, the first threshold pressure value is - relative to the atmosphere - an under-pressure with a magnitude between 4 kPa and 0 kPa (zero corresponding to the atmosphere).

[0017] Optionally, the pre-defined first threshold pressure value has a magnitude of 3 kPa.

[0018] Optionally, the pre-defined third threshold pressure value is a fraction of the first threshold pressure value.

[0019] Optionally, the further sensor data in step receiving further comprise data that represent the temperature of the fuel tank. The computer obtains the dataselectively from any of the following: (i) from a temperature sensor that is associated with the fuel tank, or (ii) from a pre-trained auxiliary network that processes the temperature of the engine coolant, the temperature of the intake plenum, the time interval from engine start, the level of fuel in the fuel tank, or the pressure in the second line.

[0020] Optionally, the further sensor data in step receiving further comprises data that represents any of the following: the temperature of the coolant of the engine, the temperature of the intake manifold of the engine, the cycle time of the engine being the time interval elapsed from the last engine start, and the fuel level in the fuel tank.

[0021] Optionally, the further sensor data in step receiving further comprises data that represents: the pressure in the second line or data that indicates a change in that pressure, the lateral acceleration of the vehicle, or the longitudinal acceleration of the vehicle.

[0022] Optionally, the further sensor data in step receiving further comprises data that represents: the driving speed of the vehicle, the engine torque of the engine, the rotational speed of the engine, the ethanol concentration in the fuel, or the derivative of vehicle acceleration.

[0023] Optionally, in steps comparing, in case that the computer determines the occurrence of the leak because the numerical difference exceeds the first threshold pressure value, the computer also estimates the size of the leak according to the numerical difference, by pre-trained leak classes that correspond to numerical difference classes.

[0024] Optionally, the computer uses networks that are all-connected perceptron neural network with two layers.

[0025] A computer-implemented method to train a network is also disclosed. The training enables a computerto execute the computer-implemented method to determine the occurrence of a leak in a system. As explained, the system is adapted to retain evaporative emissions of hydrocarbon vapor from fuel in an engine, being an internalcombustion engine of a vehicle.

[0026] In the method to train the network, the computer performs the following: (i) for artificially created leaks with particular leak diameters, it performs the step receiving further sensor data of the method (to determine), by receiving the further sensor data from a vehicle in operation, and simultaneously also performs the step receiving fuel tank pressure value as measured pressure values from the tank pressure sensor from the same vehicle in operation, (ii) Separately for the particular leak diameters, the computer applies the further sensor data at the input of the network, and applies the measured pressure values at the output of the network as ground truth, and varies network weights to minimize the errors to the ground truth.

[0027] A computer program product that, when loaded into a memory of a computer system and executed by at least one processor of the computer system, causes the computer system to perform the steps of a computer-implemented method (to determine the leak) or the computer-implemented method (to train the network).

[0028] A computer is adapted to act as a diagnostics computer to determine the occurrence of a leak in a evaporative emission retaining system. It is adapted by comprising a plurality of modules to perform the steps of the computer-implemented method (to determine the leak).Brief Description of the Drawings

[0029] FIG. 1 illustrates a block diagram with a typical arrangement of vehicle components;

[0030] FIG. 2 illustrates a block diagram of a vapor canister with valves, with a controller and with a diagnostics computer;

[0031] FIG. 3 illustrates a block diagram of the computer, with pre-trained networks and with an evaluation module;

[0032] FIG. 4 is a flow-chart diagram of a computer-implemented method to determine the occurrence of a leak in a system that is adapted to retain evaporative emissions of hydrocarbon vapor from fuel of the vehicle;

[0033] FIG. 5 illustrates a line with an artificially added leak as well as a network under training; and

[0034] FIG. 6 illustrates a generic computer.Detailed DescriptionOverview to an engine system

[0035] FIG. 1 illustrates a block diagram with a typical arrangement of an engine system 100 ("engine system" in short), such as an internal combustion engine 120 ("engine in short") with its fuel delivery system including a fuel tank 130 for storing liquid fuel such as, for example, gasoline. Engine system 100 is typically part of a vehicle, such as a road vehicle.

[0036] As illustrated from right to left, fuel arrives via pipe 134 into fuel tank 130 when filling up. Pipe 134 is open only during tanking, otherwise pipe 134 remains closed, conventionally by fuel tank cap 136. The figure illustrates fuel tank 130 with tank pressure sensor 131 that measures the air pressure in the space above the surface of the fuel. Sensor 131 provides a fuel tank pressure value, here given as MEASURED_PRESSURE.

[0037] In operation, fuel is being injected to engine 120 via fuel injection circuit 135 that typically includes a fuel rail 121 to which fuel injectors 123 are connected to inject controlled fuel quantities into engine cylinders 125.

[0038] FIG. 1 further illustrates conventional parts of the engine system 100, such as air intake 101 (at the left side, with a filter) to provide air to the engine (via intake manifold 107 and throttle 103), and exhaust line 102 to discharge exhaust gases. The figure also illustrates an optional turbocharger 105 between intake 101 and exhaust line 102.

[0039] The fuel is for use by the engine, but a fraction of the fuel can evaporate. The resulting gases (such as hydrocarbon vapor) may leak to the atmosphere.

[0040] To keep the emission to the atmosphere at minimum, engine system 100 is further equipped with a system that is configured to retain the evaporative emissions, i.e., the hydrocarbon vapors emanating from the fuel delivery system. Such systems are known in the art, for example, as "evaporative emission control system" (cf. the Jentz reference), or simply "evaporative emissions systems". The figure illustrates the system by references 110, 111, 112, and 113. It comprises - at least - a vapor canister 110 (sometimes called "evaporative emission canister" or "vapor storage canister") and a plurality of pneumatic lines 111, 112, 113. The vapor canister 110 is generally configured to define an internal flow path for fuel vapors through one or more volumes of adsorbent material (e.g., activated carbon, in particular charcoal, or other appropriate adsorbent material). It typically comprises three ports: a vent port, a tank port and a purge port. The design of the vapor canister 110 being not the focus of the invention it will not be further detailed. In implementations, the system further comprises components to control its operations (i.e., to open or to close valves etc.), such as a controller (cf. FIG. 2).

[0041] First line 111 connects fuel tank 130 to vapor canister 110 (at the tank port). Second line 112 connects vapor canister 110 (at the purge port) to engine 120 via purge valve 115 (or "canister purge valve"). Sensor 116 measures the pressure in line 112. Third line 113 connects vapor canister 110 (at the vent port) to the atmosphere, preferably via a filter. The term "line" refers to any kind of piping adapted for conducting gas / fluid streams; in practice it may take the form of rigid or flexible ducts or hoses, e.g. made from metal, plastic or rubber.

[0042] As already mentioned, vapor canister 110 collects hydrocarbon vapors and from time to time, the vapors are purged to engine (via line 112, purge valve 115).Evaporative emission control system

[0043] FIG. 2 illustrates a block diagram of vapor canister 110 with purge valve 115 at line 112 and vent valve 114 at line 113, and well as a controller 300. The figure also illustrates a computer 200.

[0044] As symbolized by dashed lines, controller 300 controls valves 115 and 114 to open or to close. Controller 300 also instructs sensors to provide data, and so on, by sending control signals. The skilled person is able to implement such control signals without the need of further explanation herein.

[0045] While controller 300 is a component that is conventionally used in such systems, computer 200 implements a function to determine the occurrence of a leak in system 110 / 111 / 112 / 113. Information that indicates the presence or the absence of a leak is symbolized as LEAK_DATA. It is noted that a leak can also be detected if it would occur in fuel tank 130 or pipe 134. Likewise, the detection as a leak would also happen if fuel tank cap 136 is missing.

[0046] Computer 200 receives measurement values from sensors (e.g., MEASURED_PRESSURE) via controller 300 or otherwise, computer 200 performs a computer-implemented method (cf. method 400 in FIG. 4), and computer 200 instructs controller 300 (e.g., open or close valves). Computer 200 can also instruct a user interface (not illustrated) to inform a user 290 when a leak has been detected (e.g., to communicate LEAK_DATA in a user-friendly format). User 290 could be the driver of the vehicle, but data regarding leaks (i.e., LEAK_DATA) could be forwarded (via networks) to persons who are located remotely from the vehicle. It is also possible to store LEAK_DATA (for further consideration, when the vehicle is being inspected in a service station / workshop).

[0047] Computer 200 can be implemented by a computer that is part of controller, but the location of computer is not relevant.Computer with networks and evaluation module

[0048] FIG. 3 illustrates a block diagram of computer 200, with the following modules: pretrained network 210, according to its function also labelled "vacuum estimation network"; pre-trained network 220, according to its function also labelled "fuel tank temperature network"; as well as evaluation module 250. Networks 210 and 220 are pre-trained in the sense that they have been trained in advance. The description provides details regarding the training below in connection with FIG. 5.

[0049] Vacuum estimation network 210 and evaluation module 250 perform method (400 in FIG. 4) to determine the occurrence of a leak (in system 110 / 111 / 112 / 113, cf. FIG. 1). Network 220 is provided optionally (as auxiliary network), to provide data that represents the temperature in fuel tank 130 (cf. FIG. 1). This temperature can also (alternatively) be measured by a sensor that the skilled person can associate with tank 130.

[0050] Right-going arrows symbolize the data flow, "white" circles symbolize input nodes of networks 210 / 220, "black" circles symbolize the output nodes of networks 210 / 220. Numeric values (i.e., data) are presented here by UPPERCASE labels.

[0051] MEASURED_PRESSURE represents the pressure within fuel tank 130, determined by tank pressure sensor 131 that is associated with fuel tank 130 (cf. FIG. 1). In relation to the atmosphere that surrounds the engine, MEASURED_PRESSURE represents under-pressure. In other words, the gases above the surface of the fuel (in tank 130) have an absolute pressure that is smaller than the atmospheric pressure. This does not exclude the possibility of the tank pressure to be over-pressure, at least temporarily, for example when the method to detect leaks is not being performed.

[0052] Other data arrives from sensors that are physically distributed in the engine system 100. To perform the method in computer 200, there is no need to install additional sensors, because vehicles are equipped with sensors, and data can be re-used. In the following, the description gives an overview to such data, but it is noted that not all types of data have to be processed.

[0053] COOLANT_TEMPERATURE represents the temperature of the engine coolant. Appropriate sensors are known in the art, and in implementations, COOLANT_TEMPERATURE is the temperature shown to the driver on a display.

[0054] INTAKE_PLENUM_TEIVIPERATURE represents the temperature of intake manifold 121 (cf. FIG. 1).

[0055] ENGINE_CYCLE_TIME represents the time interval from engine start, measured, for example, in seconds. There is no need to have a dedicated sensor, the computer(engine control unit - ECU) that controls the operation of the engine has this data available.

[0056] FUEL_TANK_LEVEL represents the level of fuel in fuel tank 130. Again there is no need for a dedicated sensor because vehicles usually have that data available (such as for display to the driver). In the schematic block diagram of FIG. 1, FUEL_TANK_LEVEL would be about 50% of the overall tank volume.

[0057] PURGE_LINE_PRESSURE represents the pressure in second line 112 (cf. FIG. 1), and the data can be measured by line sensor 116 (cf. FIG. 1).

[0058] ATM_PRESSURE represents the atmospheric pressure. The data can be obtained from a dedicated sensor (i.e., a barometer), but it could also be obtained from an external source. Weather forecasts via public radio broadcast occasionally inform their listeners, and data can be transmitted by computer networks (for example, via the Internet, or in telematic applications). The skilled person can consider the geographic altitude to have ATM_PRESSURE as accurate as possible.

[0059] DECAY_TIME represents a time interval that is measured when computer 200 interacts with controller 300 (cf. FIG. 2). Computer 200 receives pressure data from tank pressure sensor 131 (i.e., MEASURED_PRESSURE). While engine 120 is running, computer 200 can interact (cf. method step 410, FIG. 4) with controller 300 to create a vacuum condition in vapor canister 110, in first line 111, in second line 112 and in third line 113, as well as in fuel tank 130. As no gas (i.e., air) should enter vapor canister 110, computer 200 instructs controller 300 to close vent valve 114. Subsequently, computer 200 instructs controller 300 to open purge valve 112 (at a first time point tl). Engine 120 is then creating a vacuum in the evaporative system (i.e., 110, 111, 112, 113). Engine 120 simply acts as a pump. Creating such a vacuum takes some time (until a second time point t2, until a threshold is reached, MEASURED_PRESSURE = TARGET_VACUUM). Threshold pressure value TARGET_VACUUM is - relative to the atmosphere - an under-pressure with a magnitude between 4 kPa and 0 kPa. The notation "0 kPa" stands for the atmospheric pressure. In implementations, TARGET_VACUUM has a magnitude of 3 kPa.

[0060] At t2 the purge valve is closed, so that engine 120 does no longer act as a pump. Consequently, the vacuum decays. Much simplified, in a system with a leak, the decay is faster than in a system without a leak.

[0061] More in detail, computer 200 determines the time interval (DECAY_TIME) it takes to reach a decay threshold DECAY_TH RES HOLD. It is convenient to apply a pre-defined threshold, in the example, DECAY_THRESHOLD is a fraction of TARGET_VACUUM (e.g., 1 kPa decay from 3 kPa target).

[0062] VEHICLE_LATERAL_ACCELERATION and VEHICLE_LONGITUDINAL_ACCELERATION represent data that describe driving conditions of the vehicle. As the fuel in tank 130 is a liquid, it can shake inside the tank (so-called slosh effect) and therefore the pressure inside the tank (and inside the pipes and the canister) can change as well. Acceleration sensors are known in the art, and modern vehicles are equipped with them in many cases. Although the figure refers to acceleration in the direction of driving (LONGITUDINAL) and in orthogonal direction (LATERAL), there is no need to process both values.

[0063] FUEL_TANK_TEMPERATURE represents the temperature inside tank 130. FUEL_TANK_TEMPERATURE can be provided by (optional) network 220 as illustrated from the "black" circle, or can be provided by a sensor (not illustrated).

[0064] Much simplified, and with details to be explained next, network 210 receives FUEL_TANK_TEMPERATURE and receives FURTHER_DATA (e.g., ATM_PRESSURE, DECAY_TIME, ACCELERATION, FUEL_TANK_LEVEL, PURGE_LINE_PRESSURE) and provides an estimation of the pressure (inside the tank), here given by ESTIMATED_PRESSURE. Evaluation module 250 compares ESTIMATED_PRESSURE to MEASURED_PRESSURE. As illustrated, the result of the comparison can be processed as the difference between both values: A_PRESSURE.

[0065] The comparison indicates a value that would be expected for a leak-free system (i.e., system 110 / 111 / 112 / 113 without leaks). As a leak would influence MEASURED_PRESSURE, a difference between both values can indicate the presences of a leak.

[0066] Computer 200 determines the occurrence of the leak if the numerical difference (i.e., A_PRESSURE) between the fuel tank pressure value (MEASURED_PRESSURE) and the vacuum estimation value (ESTIMATED_PRESSURE) exceeds a first threshold pressure value. Using a threshold is dictated because ESTIMATED_PRESSURE is an estimation and not a measurement.

[0067] Optionally, computer 200 not only determines the occurrence of the leak, but also classifies the leak. As A_PRESSURE would depend on the size of the leak, a relation can be established.

[0068] For convenience of explanation, it is assumed that leaks can occur in the form of circular holes so that standardized leaks with particular hole diameters would lead to particular A_PRESSURE values.

[0069] Module 250 can implement a mapper function that maps A_PRESSURE values to particular standard leaks. The following example are provided for explanation only, and the skilled person can modify them: from A_PRESSURE = 0 to including A_PRESSURE = Al (Al being the first threshold value): LEAK_DATA = "no leak", from A_PRESSURE = Al to including A2: LEAK_DATA = leak with 0.3 mm diameter, from A_PRESSURE = A2 to including A3: LEAK_DATA = leak with 0.5 mm diameter, from A_PRESSURE = A3 to including A4: LEAK_DATA = leak with 0.7 mm diameter from A_PRESSURE = A4 to including A5: LEAK_DATA = leak with 1.0 mm diameter or larger.

[0070] The Al, A2, A3 etc. values stand for the threshold pressure values. In other words, particular leak sizes are mapped to particular pressure bands (such as "Al to including A2", etc.).

[0071] The diameters can be selected according to technical standards. Maps A_PRESSURE to standard leaks (or leak classes) can be determined by calibration and stored using look-up tables or the like. Alternatively, the mapping can be obtained by applying formulas.

[0072] It may happen that cap 136 is missing from pipe 134 (cf. FIG. 1). This is not necessarily a leak in the system, but would lead to a A_PRESSURE that would indicate such a "no cap" event.

[0073] The threshold values Al, A2, etc. can be obtained by training (details with FIG. 5).

[0074] The skilled person who is familiar with networks can select a suitable type of network. Networks 210 and 220 are artificial neural networks (ANN). Networks 210 and 220 can be implemented as two-layer feedforward perceptrons.

[0075] Further, data values can be normalized, for example, to normalized values between 0 and 1. For example, FUEL_TANK_LEVEL = 0.5 (half full, in FIG. 1), ATM_PRESSURE = 1 for about 1000 hPa, and so on.Method flow

[0076] FIG. 4 is a flow-chart diagram of computer-implemented method 400 to determine the occurrence of a leak in a system 110 / 111 / 112 / 113 that is adapted to retain evaporative emissions of hydrocarbon vapor from engine system 100.

[0077] As explained already, cf. FIG. 1, the system comprises first line 111 that connects vapor canister 110 to fuel tank 130 of engine system 100, second line 112 that connects vaporcanister 110 to engine 120 (of engine system 100) via purge valve 115, and third line 113 that connects vapor canister 110 to a filter.

[0078] In a method step receiving 410, computer 200 receives fuel tank pressure value MEASURED_PRESSURE from tank pressure sensor 131 (cf. FIG. 1, that is associated with the fuel tank 130). MEASURED_PRESSURE represents under-pressure in relation to the atmosphere that surrounds the engine system 100.

[0079] While engine 120 is running, computer 200 continues with step interacting 420. Computer 200 interacts with controller 300 to create a vacuum condition in vapor canister 110 (also within lines 111, 112, 113, as well as in fuel tank 130). Interacting 420 comprises that computer 200 instructs controller 300.

[0080] At a first time point tl, there is an instruction to close vent valve 114 and to open the purge valve 112, and at a second time point t2, when the fuel tank pressure data MEASURED_PRESSURE reaches a pre-defined threshold pressure value TARGET_VACUUIVI, there is an instruction to close purge valve 115.

[0081] In step operating 430, computer 200 operates pre-trained network 210, with receiving 431 further sensor data FURTHER_DATA from engine system 100, and obtaining 422 a vacuum estimation value ESTIMATED_PRESSURE that represents the vacuum in fuel tank 130.

[0082] As explained, in step comparing 440, computer 200 compares the fuel tank pressure value MEASURED_PRESSURE with the vacuum estimation value ESTIMATED_PRESSURE. Computer 200 thereby determines the occurrence of the leak if the numerical difference A_PRESSURE between the fuel tank pressure value MEASURED_PRESSURE and the vacuum estimation value ESTIMATED_PRESSURE exceeds a threshold pressure value Al (or A2 etc.), as discussed above with FIG. 3.Method details

[0083] The description now continues with explaining optional method details. While engine 120 is still running - and while computer 200 is interacting 420 with controller 300, computer 200 determines the time interval between the second time point t2 (i.e., when valve 115 has been closed) and a third time point t3. At t3, the fuel tank pressure value MEASURED_PRESSURE has reached a pressure threshold DECAY_TARGET, and the time between t2 and t3 is the decay time DECAY_TIME that becomes part of the further sensor data FURTHER_DATA in receiving step 431.

[0084] As the threshold pressure value TARGET_VACUUIVI is - relative to the atmosphere - an under-pressure, its magnitude can be between 4 kPa and 0 kPa. For example, predefined first threshold pressure value TARGET_VACUUIVI has a magnitude of 3 kPa.

[0085] The pre-defined pressure value DECAY_TARGET can be a fraction of the threshold pressure value TARGET_VACUUIVI. It is convenient to measure the time (DECAY_TIME) when the vacuum has decayed to one third of its original value (at t2). For a value 3 kPa (at t2), this target would be 1 kPa. The skilled person can usemodified values, the absolute values are less important. The relative change (such as by the factor 1 / 3) is the basis for measuring DECAY_TIME.Further data

[0086] Data has already been introduced with FIG. 3. Further sensor data FURTHER_DATA (in step receiving 441) can comprises data that represent the temperature of the fuel tank FUEL_TANK_TEMPERATURE. There are two options, so that computer 200 can obtain the data selectively from any of the following:

[0087] a) Computer 200 can receive the data from a temperature sensor that is associated with the fuel tank 130.

[0088] b) Computer 200 can use a pre-trained auxiliary network 220 (cf. FIG. 3) that processes- the temperature of the engine coolant (COOLANT_TEMP),- the temperature of the intake plenum (INTAKE_PLENUM_TEI\ / IP),- the time interval from engine start (ENGINE_CYCLE_TIME),- the level of fuel in the fuel tank (130, FUEL_TANK_LEVEL),- the pressure in the second line (112, PURGE_LINE_PRESSURE).

[0089] It is also possible to receive further sensor data FURTHER_DATA along the following: a) the pressure in second line 112 or data that indicates a change in that pressure, b) the lateral acceleration (VEHICLE_LATERAL_ACCELERATION) of the vehicle, c) the longitudinal acceleration (VEHICLE_LONGITUDINAL_ACCELERATION) of the vehicle.

[0090] In further embodiments, using the following is contemplated as well: the driving speed of vehicle 100, the engine torque of engine 120, the rotational speed of engine 120 (i.e., "turns per minute"), the ethanol concentration in the fuel, and the derivative of vehicle acceleration (i.e., so-called "jerk").

[0091] In step comparing 440, in case that the computer (200) determines the occurrence of the leak because the numerical difference (A_PRESSURE) exceeds the first threshold pressure value, the computer (200) also estimates the size of the leakaccording to the numerical difference, by pre-trained leak classes that correspond to numerical difference classes.Training the network

[0092] FIG. 5 illustrates line 112 with an artificially added leak as well as network 210 under training (with its input on the left side and its output on the right side).

[0093] Line 112 is shown near purge valve 115, and the picture above the line is a magnification of the line with a leak hole having diameter D. The skilled person does not have to create the leak inside line 112, but can alternatively create the leak in a by-pass line (connecting both ends of closed valve 115).

[0094] Using line 112 is convenient for illustration, but this leak can be made at other lines as well.

[0095] Leak diameter D is presented as a plurality of size variation, from D = 0 (no leak at all, leak-free system), to D = 1 (large leak) or even larger. The leak size variations correspond to the threshold mapping discussed above for FIG. 3.

[0096] To train network, data is being collected as training data, such as FURTHER_DATA (and in principle the data presented by UPPERCASE labels in FIG. 3). As the pressure can be measured (MEASURED_PRESSURE) by sensor 131 (cf. FIG. 1), measured pressure values serve as ground truth (being the training data being applied on the output of the network).

[0097] By varying diameter D (and by continuing to collect FURTHER_DATA), the network learns to detect diameter typical relations (between input and output). Over time it learns that, for example, a particular A_PRESSURE (within a plus / minus tolerance band) correspond to a particular standard leak (or corresponds to a leak-free system, with D = 0).

[0098] FIG. 5 does not illustrate how (auxiliary) network 220 can be trained to estimate the temperature, but the principle is the same: the measured temperature (from a dedicated sensor) would be the ground truth. Leaks do not have to be applied artificially.

[0099] It is possible to obtain training data for leak-like situations, for example, to detect a missing cap 136.Testing the system

[0100] The proper operation of purge valve 115 and vent valve 114 is a condition for performing method 400, but a test (in FIG. 4 as optional step 405) can be performed. Engine 120 is started, and both valves 115 and 114 are being closed. MEASURED_PRESSURE (from sensor 131) is monitored.

[0101] If MEASURED_PRESSURE reaches a calibrated threshold (obtained by laboratory measurements) or if the time to reach this calibrated threshold is too short (e.g., DECAY_TIME with a factor), then purge valve 115 is considered faulty.

[0102] In that case vent valve diagnostics are activated. The vacuum pressure inside fuel tank 130 is being monitored as vent valve 114 remains open. If the vacuum decay is too slow (or too fast) compared to the calibrated time then the canister vent valve is considered as faulty. FUEL_TANK_LEVEL is taken into account.

[0103] Further, if it would be impossible to reach 3 kPa (i.e., an example for the TARGET_VACUUM threshold), the presence of a relatively large "leak" can be assumed. But it does not have to be a leak in the system: if refueling event have been detected as well, a missing fuel cap is the most probable cause. LEAK_DATA can be presented to user 290 with an instruction to check of the cap is attached.Discussion

[0104] Using the neural network based approach, the number of sensors remains minimal, and data from existing sensors in the vehicle can be re-used. There is no need to install a vacuum pump or the like (the engine acts as a pump). It is not only possible to differentiate presence and absences of leaks, but also estimate the leak size (due to the calibration during training with leak holes with standardized diameters). Detecting relatively small leaks down to 0.3 mm diameter is possible.

[0105] Although the leaks are quantified in terms of hole diameters, actual leaks do not have to be circular holes. The actual shape of the leak is not identified, but that is notrequired.

[0106] Traditionally, vehicles with combustion engines only move when the engine is running. For so-called plugin hybrid vehicles that pictures changes. Such vehicles still have combustion engines but they also move when the (combustion) engine is not running. They may even drive in full electric mode, but fuel remains in the tank (and therefore emissions can still occur).Generic computer

[0107] FIG. 6 illustrates an example of a generic computer device which may be used with the techniques described here. FIG. 6 is a diagram that shows an example of a generic computer device 900 and a generic mobile computer device 950, which may be used with the techniques described here. Computing device 900 is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Generic computer device may 900 correspond to computer 200. Computing device 950 is intended to represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smart phones, driving assistance systems or board computers of vehicles and other similar computing devices. For example, computing device 950 may be used as a frontend by a user (e.g., an operator of a blast furnace) to interact with the computing device 900. The components shown here, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and / or claimed in this document.

[0108] Computing device 900 includes a processor 902, memory 904, a storage device 906, a high-speed interface 908 connecting to memory 904 and high-speed expansion ports 910, and a low speed interface 912 connecting to low speed bus 914 and storage device 906. Each of the components 902, 904, 906, 908, 910, and 912, are interconnected using various busses, and may be mounted on a common motherboard or in other manners as appropriate. The processor 902 can processinstructions for execution within the computing device 900, including instructions stored in the memory 904 or on the storage device 906 to display graphical information for a GUI on an external input / output device, such as display 916 coupled to high speed interface 908. In other implementations, multiple processors and / or multiple buses may be used, as appropriate, along with multiple memories and types of memory. Also, multiple computing devices 900 may be connected, with each device providing portions of the necessary operations (e.g., as a server bank, a group of blade servers, or a multi-processor system).

[0109] The memory 904 stores information within the computing device 900. In one implementation, the memory 904 is a volatile memory unit or units. In another implementation, the memory 904 is a non-volatile memory unit or units. The memory 904 may also be another form of computer-readable medium, such as a magnetic or optical disk.

[0110] The storage device 906 is capable of providing mass storage for the computing device 900. In one implementation, the storage device 906 may be or contain a computer- readable medium, such as a floppy disk device, a hard disk device, an optical disk device, or a tape device, a flash memory or other similar solid state memory device, or an array of devices, including devices in a storage area network or other configurations. A computer program product can be tangibly embodied in an information carrier. The computer program product may also contain instructions that, when executed, perform one or more methods, such as those described above. The information carrier is a computer- or machine-readable medium, such as the memory 904, the storage device 906, or memory on processor 902.

[0111] The high speed controller 908 manages bandwidth-intensive operations for the computing device 900, while the low speed controller 912 manages lower bandwidth-intensive operations. Such allocation of functions is exemplary only. In one implementation, the high-speed controller 908 is coupled to memory 904, display 916 (e.g., through a graphics processor or accelerator), and to high-speed expansion ports 910, which may accept various expansion cards (not shown). In the implementation, low-speed controller 912 is coupled to storage device 906 and low-speed expansion port 914. The low-speed expansion port, which may include various communication ports (e.g., USB, Bluetooth, Ethernet, wireless Ethernet) may be coupled to one or more input / output devices, such as a keyboard, a pointing device, a scanner, or a networking device such as a switch or router, e.g., through a network adapter.

[0112] The computing device 900 may be implemented in a number of different forms, as shown in the figure. For example, it may be implemented as a standard server 920, or multiple times in a group of such servers. It may also be implemented as part of a rack server system 924. In addition, it may be implemented in a personal computer such as a laptop computer 922. Alternatively, components from computing device 900 may be combined with other components in a mobile device (not shown), such as device 950. Each of such devices may contain one or more of computing device 900, 950, and an entire system may be made up of multiple computing devices 900, 950 communicating with each other.

[0113] Computing device 950 includes a processor 952, memory 964, an input / output device such as a display 954, a communication interface 966, and a transceiver 968, among other components. The device 950 may also be provided with a storage device, such as a microdrive or other device, to provide additional storage. Each of the components 950, 952, 964, 954, 966, and 968, are interconnected using various buses, and several of the components may be mounted on a common motherboard or in other manners as appropriate.

[0114] The processor 952 can execute instructions within the computing device 950, including instructions stored in the memory 964. The processor may be implemented as a chipset of chips that include separate and multiple analog and digital processors. The processor may provide, for example, for coordination of the other components of the device 950, such as control of user interfaces, applications run by device 950, and wireless communication by device 950.

[0115] Processor 952 may communicate with a user through control interface 958 and display interface 956 coupled to a display 954. The display 954 may be, for example,a TFT LCD (Thin-Film-Transistor Liquid Crystal Display) or an OLED (Organic Light Emitting Diode) display, or other appropriate display technology. The display interface 956 may comprise appropriate circuitry for driving the display 954 to present graphical and other information to a user. The control interface 958 may receive commands from a user and convert them for submission to the processor 952. In addition, an external interface 962 may be provide in communication with processor 952, so as to enable near area communication of device 950 with other devices. External interface 962 may provide, for example, for wired communication in some implementations, or for wireless communication in other implementations, and multiple interfaces may also be used.

[0116] The memory 964 stores information within the computing device 950. The memory 964 can be implemented as one or more of a computer-readable medium or media, a volatile memory unit or units, or a non-volatile memory unit or units. Expansion memory 984 may also be provided and connected to device 950 through expansion interface 982, which may include, for example, a SIMM (Single In Line Memory Module) card interface. Such expansion memory 984 may provide extra storage space for device 950, or may also store applications or other information for device 950. Specifically, expansion memory 984 may include instructions to carry out or supplement the processes described above, and may include secure information also. Thus, for example, expansion memory 984 may act as a security module for device 950, and may be programmed with instructions that permit secure use of device 950. In addition, secure applications may be provided via the SIMM cards, along with additional information, such as placing the identifying information on the SIMM card in a non-hackable manner.

[0117] The memory may include, for example, flash memory and / or NVRAM memory, as discussed below. In one implementation, a computer program product is tangibly embodied in an information carrier. The computer program product contains instructions that, when executed, perform one or more methods, such as those described above. The information carrier is a computer- or machine-readable medium, such as the memory 964, expansion memory 984, or memory on processor952 that may be received, for example, over transceiver 968 or external interface 962.

[0118] Device 950 may communicate wirelessly through communication interface 966, which may include digital signal processing circuitry where necessary. Communication interface 966 may provide for communications under various modes or protocols, such as GSM voice calls, SMS, EMS, or MMS messaging, CDMA, TDMA, PDC, WCDMA, CDMA2000, or GPRS, among others. Such communication may occur, for example, through radio-frequency transceiver 968. In addition, short-range communication may occur, such as using a Bluetooth, WiFi, or other such transceiver (not shown). In addition, GPS (Global Positioning System) receiver module 980 may provide additional navigation- and location-related wireless data to device 950, which may be used as appropriate by applications running on device 950.

[0119] Device 950 may also communicate audibly using audio codec 960, which may receive spoken information from a user and convert it to usable digital information. Audio codec 960 may likewise generate audible sound for a user, such as through a speaker, e.g., in a handset of device 950. Such sound may include sound from voice telephone calls, may include recorded sound (e.g., voice messages, music files, etc.) and may also include sound generated by applications operating on device 950.

[0120] The computing device 950 may be implemented in a number of different forms, as shown in the figure. For example, it may be implemented as a cellular telephone 980. It may also be implemented as part of a smart phone 982, personal digital assistant, or other similar mobile device.

[0121] Various implementations of the systems and techniques described here can be realized in digital electronic circuitry, integrated circuitry, specially designed ASICs (application specific integrated circuits), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at leastone input device, and at least one output device.

[0122] These computer programs (also known as programs, software, software applications or code) include machine instructions for a programmable processor, and can be implemented in a high-level procedural and / or object-oriented programming language, and / or in assembly / machine language. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus and / or device (e.g., magnetic discs, optical disks, memory, Programmable Logic Devices (PLDs)) used to provide machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and / or data to a programmable processor.

[0123] To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0124] The systems and techniques described here can be implemented in a computing device that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a client computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a widearea network ("WAN"), and the Internet.

[0125] The computing device can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

[0126] A number of embodiments have been described. Nevertheless, it will be understood that various modifications may be made without departing from the spirit and scope of the invention.

[0127] In addition, the logic flows depicted in the figures do not require the particular order shown, or sequential order, to achieve desirable results. In addition, other steps may be provided, or steps may be eliminated, from the described flows, and other components may be added to, or removed from, the described systems. Accordingly, other embodiments are within the scope of the following claims.

Claims

Claims1. Computer-implemented method (400) to determine the occurrence of a leak in a system (110, 111, 112, 113) that is adapted to retain evaporative emissions of hydrocarbon vapor from fuel in an engine (120), being an internal combustion engine of a vehicle, wherein the system (110, 111, 112, 113) comprises a vapor canister (110) and a plurality of connecting lines (111, 112, 113), wherein a first line (111) connects the vapor canister (110) to a fuel tank (130) of the engine system (100), a second line (112) connects the vapor canister (110) to an air intake manifold (121) of the engine system (100) via a purge valve (115), and a third line (113) connects the canister (110) to a filter, the method (400) comprising: a computer (200) receiving (410) a fuel tank pressure value (MEASURED_PRESSURE) from a tank pressure sensor (131) that is associated with the fuel tank (130), wherein the fuel tank pressure value (MEASURED_PRESSURE) represents the pressure in relation to the surrounding atmosphere; while the engine (120) is running, the computer (200) interacting (420) with a controller (300) to create a vacuum condition in the vapor canister (110), in the first line (111), in the second line (112) and in the third line (113), as well as in the fuel tank (130), wherein interacting (420) comprises to instruct the controller (300) at a first time point (tl), to close the vent valve (114) and to open the purge valve (112), and at a second time point (t2), when the fuel tank pressure data(MEASURED_PRESSURE) reach a pre-defined first threshold pressure value(TARGET_VACUUIVI), to close the canister purge valve (112);the computer (200) operating (430) a pre-trained network (210), with receiving (431) further sensor data (FURTHER_DATA) from the vehicle, and obtaining (432) a vacuum estimation value (ESTIMATED_PRESSURE) that represents the vacuum in the fuel tank (130), the computer (200) comparing (440) the fuel tank pressure value (MEASURED_PRESSURE) and the vacuum estimation value (ESTIMATED_PRESSURE), thereby determining the occurrence of the leak if the numerical difference (A_PRESSURE) between the fuel tank pressure value (MEASURED_PRESSURE) and the vacuum estimation value (ESTIMATED_PRESSURE) exceeds a second threshold pressure value (Al).

2. Method (400) according to claim 1, wherein - while the engine (120) of the vehicle is still running - and while the computer (200) is interacting (420) with the controller (300), the computer (200) determines the time interval between the second time point (t2) and a third time point (t3) when the fuel tank pressure value (MEASURED_PRESSURE) has reached a third threshold pressure value (DECAY_TARGET), as decay time (DECAY_TIME) that becomes part of the further sensor data (FURTHER_DATA) in the receiving step (431).

3. Method (400) according to claim 1, wherein the first threshold pressure value (TARGET_VACUUM) is - relative to the atmosphere - an under-pressure with a magnitude between 4 kPa and 0 kPa.

4. Method (400) according to claim 3, wherein the pre-defined first threshold pressure value (TARGET_VACUUM) has a magnitude of 3 kPa.

5. Method (400) according to any of claims 2 and 3, wherein the pre-defined third threshold pressure value (DECAY_TARGET) is a fraction of the first threshold pressure value (TARGET_VACUUM).

6. Method (400) according to any of claims 2 to 5, wherein the further sensor data (FURTHER_DATA) in step receiving (431) further comprise data that represent the temperature of the fuel tank (FUEL_TANK_TEMPERATURE), wherein the computer (200) obtains the data selectively from any of the following: from a temperature sensor that is associated with the fuel tank (130), or from a pre-trained auxiliary network (220) that processes the temperature of the engine coolant (COOLANT_TEMP), the temperature of the intake plenum (INTAKE_PLENUM_TEMP), the time interval from engine start (ENGINE_CYCLE_TIME), the level of fuel in the fuel tank (130, FUEL_TANK_LEVEL), the pressure in the second line (112, PURGE_LINE_PRESSURE).

7. Method (400) according to any of claims 2 to 6, wherein the further sensor data (FURTHER_DATA) in step receiving (431) further comprises data that represents any of the following: the temperature (COOLANT_TEMP) of the coolant of the engine (120), the temperature (INTAKE_PLENUM_TEMP) of the intake manifold (121) of the engine (120), the cycle time of the engine (ENGINE_CYCLE_TIME) being the time interval elapsed from the last engine start, and the fuel level in the fuel tank (FUEL_TANK_LEVEL).

8. Method (400) according to claim 7, wherein the further sensor data (FURTHER_DATA) in step receiving (431) further comprises data that represents: the pressure in the second line (112) or data that indicates a change in that pressure, the lateral acceleration (VEHICLE_LATERAL_ACCELERATION) of the vehicle, andthe longitudinal acceleration (VEHICLE_LONGITUDINAL_ACCELERATION) of the vehicle.

9. Method (400) according to claim 8, wherein the further sensor data (FURTHER_DATA) in step receiving (431) further comprises data that represents: the driving speed of the vehicle, the engine torque of the engine (120), the rotational speed of the engine (120), the ethanol concentration in the fuel, and the derivative of vehicle acceleration.

10. Method (400) according to any of claims 1 to 8, wherein in steps comparing (440), in case that the computer (200) determines the occurrence of the leak because the numerical difference (A_PRESSURE) exceeds the first threshold pressure value, the computer (200) also estimates the size of the leak according to the numerical difference, by pre-trained leak classes that correspond to numerical difference classes.

11. Method (400) according to any of claim 1 to 10, wherein the computer uses networks that are all-connected perceptron neural network with two layers.

12. Computer-implemented method to train a network (210) to enable a computer (200) to execute a further computer-implemented method (400) to determine the occurrence of a leak in a system (110, 111, 112, 113), wherein the system is adapted to retain evaporative emissions of hydrocarbon vapor from fuel in an engine (120), being an internal combustion engine of a vehicle, wherein the system (110, 111, 112, 113) comprises a vapor canister (110) and a plurality of connecting lines (111, 112, 113), with a first line (111) connecting the vapor canister (110) to a fuel tank (130) of the engine system (100), a second line (112) connecting the vapor canister (110) to an air intake manifold (121) of the engine system (100) via a purge valve (115), and a third line (113) connecting the canister (110) to a filter,wherein the further computer-implemented method (400) is defined by any of claims 1 to 11, wherein the method to train the network comprises: for artificially created leaks with particular leak diameters (D), performing the step receiving (431) further sensor data (FURTHER_DATA) of the further method (400), by receiving the further sensor data from a vehicle in operation, and simultaneously also performing the step receiving (410) fuel tank pressure value as measured pressure values from the tank pressure sensor (131) from the same vehicle in operation; and separately for the particular leak diameters (D), applying the further sensor data (FURTHER_DATA) at the input of the network, and applying the measured pressure values at the output of the network as ground truth, and varying network weights to minimize the errors to the ground truth.

13. A computer program product that, when loaded into a memory of a computer system and executed by at least one processor of the computer system, causes the computer system to perform the steps of a computer-implemented method according to any of the claims 1-11 or the computer-implemented method according to claim 12.

14. A computer adapted to act as a diagnostics computer to determine the occurrence of a leak in a evaporative emission retaining system, being adapted by comprising a plurality of modules to perform the steps of the computer-implemented method according to any of the claims 1 to 11.

Images