Fault detection in gas-fuelled engine systems

Abstract

An engine system 2 for a vehicle, comprising: a rail volume 28 for receiving gaseous fuel, a shut off valve 38 for controlling the supply of fuel to the rail volume and an injector assembly including

Description

The examples of the invention relate to systems, method and approaches for detecting leaks and emissions from gas-fuelled power plants, and particularly internal combustion engines fuelled at least partially with hydrogen or another gaseous fuel. Background Gaseous fuels such as hydrogen are promising alternative fuels to gasoline and diesel due to their potential for low or zero emissions. However gaseous fuels present some challenges relating to their containment and handling. Hydrogen, for example, has a very low density which makes it challenging to contain, transport and use in the context of a fuel system for an internal combustion engine. In general, following deactivation of a hydrogen-fueled engine, it is desirable to prevent any leakage of hydrogen as it presents a safety risk due to its combustibility and it is also a greenhouse gas. The injector becoming stuck open at the end of an injection event can result in unwanted leakage of fuel into the engine. In an injector for gasoline or diesel, for example, conventional methods for detecting stuck open injectors rely on a glitch detection where a voltage is measured as the valve needle hits the valve seat at the end of injection. Analysing the voltage can determine whether the valve needle has seated properly. However, in a gaseous fuel injector the valve seat may not be metallic and may instead be provided with a synthetic rubber seat, such as a neoprene seat. Such glitch detection methods cannot therefore be used for some gaseous fuel injectors. It is with these issues in mind that the embodiments of the invention have been devised. Summary of the Invention Against this background, examples of the invention provide, in a first aspect, an engine system for a vehicle, the engine system comprising: a rail volume for receiving gaseous fuel, a shut off valve for controlling the supply of fuel to the rail volume and an injector assembly including a plurality of fuel injectors. Each one of the injectors receives fuel from the rail volume and delivers fuel to an associated engine cylinder in an injection event. The system further comprises an air-fuel ratio sensor to measure the air-fuel ratio in an exhaust of the engine system, a rail sensor configured to measure the pressure of gas in the rail volume; and a controller configured to: receive air-fuel ratio data relating to the air-fuel ratio in the exhaust; determine, based on the air-fuel ratio data, that there is an air-fuel ratio abnormality associated with at least one of the cylinders, and perform a confirmatory action to confirm that at least one of the plurality of injectors associated with the respective cylinders has a leakage fault. Beneficially, therefore, the invention makes use of an air-fuel ratio sensor or lambda sensor’ or ‘oxygen sensor to derive initial information about a suspected fault with an injector. Then the system performs a further confirmatory action to confirm that at least one of the fuel injectors has a leakage fault. The invention extends to a corresponding method for determining a leak state in an engine system fora vehicle, comprising: receiving air-fuel ratio data relating to the air-fuel ratio in the exhaust; determining, based on the air-fuel ratio data, that there is an air-fuel ratio abnormality associated with at least one of the cylinders, performing a confirmatory action to confirm that at least one of the plurality of injectors associated with the respective cylinders has a leakage fault. The invention can also be expressed as a controller configured to execute the instructions defined by the method, and also a computer program product comprising computer readable instructions which, when performed by a suitable processor, implements the method as defined above. The air-fuel ratio data may be analysed to determine cylinder imbalance data which permits a specific one of the injectors to be identified as potentially faulty, which can then be diagnosed with more precision by the following confirmatory action. The confirmatory action can be performed in various ways. In some examples the cylinder imbalance data can be used to perform a confirmatory determination that a specific injector is faulty. The cylinder imbalance data may be used in combination with rail pressure data in order to confirm that the air-fuel ratio abnormality is indeed associated with a leaky fuel injector. In some examples, the gas leakage may be determined by stopping engine operation, isolating gas in the fuel rail or fuel volume, and then determining that the change in gas pressure or quantity or rate of change thereof exceeds acceptable levels. Such an approach therefore provides a two-stage leaky injector test which uses different data types: cylinder air fuel ratio data and fuel pressure data, which results in a reliable diagnostic approach. In an alternative approach, the confirmatory action may be performed by deactivating the fuel injector that is associated with a cylinder that has been identified as displaying an air-fuel ratio abnormality. Following deactivation of the suspected injector, a combustion activity can be monitored for that cylinder to identified if a leakage exists. The combustion signal can be provided by the air fuel ratio sensor, thereby providing an elegant solution because the same sensing equipment can be used for different purposes. Combustion activity can also be identified by other means, such as through the use of knock sensors or engine speed / torque sensors. The air-fuel ratio sensor may be a device that is operable to determine a wide range of air-fuel ratio in the exhaust gases of the engine system, such as a WRAP sensor. The invention is applicable to both port injection fuel systems in which the fuel injectors deliver fuel to the engine cylinders via the air inlet manifold, or to direct injection systems in which the fuel injectors deliver fuel to the engine cylinders directly. It will be appreciated that preferred and / or optional features of the first aspect of the invention may be incorporated alone or in appropriate combination within the second aspect of the invention also. Further optional and advantageous features are referenced in the detailed description and the appended claims. Brief Description of the Drawings Examples of the invention will now be described with reference to the following figures: Figure 1 is a schematic view of a gas-fuelled interna! combustion engine, being an example of an engine system to which the examples of the invention apply; Figure 2 is a flow diagram to illustrate the steps of a method which may be implemented by a controller of the engine system in Figure 1, to detect the presence of a faulty injector; Figures 3a-c illustrate a graph to show as a function of time (a) power state of the engine system (b) state of rail gas supply, and (c) rail pressure Figure 4 is a further flow diagram that illustrates steps in accordance with another example of the invention; Figure 5 is graph which illustrates a monitoring process of the method of Figure 4, showing the state of gas supply to the engine, rail pressure, and injector drive signals. Detailed Description In general, the examples of the invention provide an engine system for a vehicle in which a common rail delivers fuel to a plurality of injectors and in which a controller is configured to detect a leak of gas from the system, particularly a fuel injector teak where the shut off valve controls a supply of gaseous fuel to the common rail. The method is applicable to a port fuel injection system in which each of the fuel injectors of the engine system is arranged to inject fuel into an air inlet manifold at a dedicated channel which leads to a respective one of the engine cylinders. The method is also applicable to a direct injection system where the fuel is delivered directly to the cylinders of the engine. The method is performed while the engine is running in such a way as to minimise intrusion to the driver (i.e. the carrying out of the method should be imperceptible in an ideal case). To put the examples of the invention into technical context, a discussion of a typical internal combustion engine which is fuelled with gaseous hydrogen will now be described with reference to Figure 1. For the purpose of this description, and the main implementation of the invention, an engine system including a port fuel injection system will be described, but the invention may also be applicable to other configurations of fuel injection system. In overview, an internal combustion engine system 2 comprises an engine block 4, an air inlet system 6, a fuel delivery system 8 and an exhaust system 10. The engine system 2 further comprises a control unit 11, referred to as the engine control unit (ECU), which is adapted to receive data input 11.1 to sense operational parameters of the engine to provide suitable control output signals 11.2 to the engine system 2 to control its operation based on driver demands and sensor measurements, as is conventional. For example, a power-on or power-off request for the engine may be generated by a vehicle system in response to a user of the vehicle turning the ignition key to the ‘engine’ on or off position, respectively, or by pressing an ‘engine stop / start’ button on a vehicle screen interface or physical button, for example. This is often referred to as ‘key on / off’. The engine control unit 11 receives the key on / off signal in the data input 11.1. The engine control unit 11 includes a memory component 13. The memory component 13 stores data such as self-learnt control parameters and operating history data which can be retrieved by the ECU even after a power down cycle. The engine block 4 of the illustrated example comprises four combustion chambers 12, or cylinders, in an ‘in-line’ configuration. However, it should be noted that this is for illustrative purposes only and the engine block may comprise any suitable number of combustion chambers in any suitable configuration, as would be well understood by the skilled person. Common engine configurations are single cylinder engines, twin cylinders, triples, in-line sixes or V-sixes. and V8 engines. Herein, the term ‘combustion chamber’ will be considered synonymous with ‘engine cylinder’. The air inlet system 6 comprises an air inlet 14 which feeds fresh air into a network of air pipes 16 through an air filter 17, An air mass flow sensor 18 is provided to provide data to the control unit 11 (signals not shown) about the airflow entering the engine system 2. The network of air pipes 16 feeds incoming air through a compressor 19 and, subsequently, to an intercooler 20. The functionality of the compressor 19 and the intercooler 20 are known in the art so a further discussion will not be provided. The network of pipes 16 leads from the intercooler 20 through a throttle valve 22 to an air inlet duct or ‘manifold’ 24. As is known, the air inlet manifold 24 directs fresh air to each of the engine cylinders 12 of the engine block 4 via separate air channels (not identified). The fuel delivery system 8 comprises a set of one or more fuel delivery devices in the form of an injector assembly including a plurality of fuel injectors 26 (only one of which is labelled) that are arranged to inject combustible fuel, in this case hydrogen gas, into the fresh air flowing into the engine cylinders 12. In the illustrated example, there area plurality of fuel injectors 26, the number of which corresponds to the number of combustion chambers 12. The system is a port injection engine and so each of the fuel injectors 26 is arranged to inject fuel into the air inlet manifold 24 at a dedicated channel which leads to a respective one of the engine cylinders 12. The invention is also applicable to a direct fuel injection system in which the fuel injectors deliver fuel directly to the engine cylinder, and not via the air iniet manifold 24 (the direct injection system is not shown in the figures). The fuel injectors 26 are each connected to a fuel accumulator or ‘common rail’ 28. As is known, the common rail 28 provides a relatively large volume of fuel which is maintained at a predetermined, and controllable, pressure level which means that the fuel injectors 26 are connected to a source of fuel having a pressure level that is in essence static and is not affected by their operation. It should be noted, however, that the fuel pressure within the common rail 28 can be modified in use due to various requirements that are beyond the scope of this discussion. The pressure of fuel within the common rail 28 is determined by the control unit 11 or ‘controller’ and is monitored by means of a sensor in the form of a fuel or rail pressure sensor 30. The rail pressure sensor 30 is shown as being connected to the end of the common rail 28 which has an elongated shape, in this example. However, the shape of the common rail 28 and the relative position of the rail pressure sensor 30 are configurational aspects that are not central to the invention. The rail pressure sensor 30 provides an output signal (not shown), which is representative of the fuel pressure in the common rail 28, to the control unit 11. Fuel and air mixture in the engine cylinders 12 is ignited by respective spark plugs 31, in the usual manner. The common rail 28 is supplied with fuel by a fuel supply system 32. The fuel supply system 32 includes a pressurised fuel source or reservoir 34, a pressure regulating device 36, a shut-off valve 38 and a gas supply line 40 which connects the shut-off valve 38 to the common rail 28. In some examples, the shut-off valve 38 may be connected directly to the common rail 28 although it is usual fora length of gas supply line 40 to be present so that a desired separation distance may be achieved between the engine system 2 and the fuel supply system 32. The pressurised fuel source 34 or ‘fuel tank’ may suitably be configured to store hydrogen gas at an appropriate pressure level, which may be between 350 and 700 bar, whereas the pressure regulating device 36 is configured to reduce the gas pressure in the fuel tank 34 to a pressure suitable for injection, which may be between 5 bar and 10 bar but could be higher for some systems, for example direct injection systems. Together the gas supply line 40, the rail 28 and the injectors 26, and the various connections between these components, may be considered as the low pressure circuit LPC of the system. It should be noted that the configuration of the fuel supply system 32 is simplified for the purposes of this discussion and more components would be present in a practical system. However, the components shown here are the principal components relevant to the examples of the invention. Further sensing means may be provided for the control unit 11 in order for it to operate the engine system 2 effectively. In the illustrated example, the engine block 4 is equipped with a knock sensor 41. As is known in the art, a knock sensor provides a means to detect high frequency vibration of the engine block 4 from which a determination can be made about whether combustion has occurred within a particular combustion chamber 12 using associated software. The knock sensors and the associated software are able to discriminate between combustion occurring in different ones of the combustion chambers 12. A knock sensor is conventional technology and so further discussion will be omitted. The engine system 2 further includes a system sensor in the form of an air pressure sensor (or air inlet manifold pressure sensor) 42 which is configured to provide the control unit 11 with data relating to the pressure of air within the air inlet manifold 24. The pressure sensor 42 provides a pressure sensor output to the control unit 11 which is representative of the air pressure in the air inlet manifold 24. The engine system 2 further includes a crank position sensor 44 which is configured to provide the control unit 11 with data relating to the position and rotational speed of the crankshaft, ft should be noted that the crankshaft, pistons, intake and exhaust valves, and spark plugs are not shown on Figure 1, but their presence is implied. Data from the crankshaft position sensor 44 may be used by the control unit 11 to control fuel injection and ignition timing. Common mounting positions for the crankshaft position sensor 44 include on the engine flywheel (not shown), the camshaft (not shown) or the main crankshaft pulley (not shown). The crankshaft position sensor 44 is shown as being associated with the engine block 4 in Figure 4, for ease of illustration. A temperature sensor 45 may also be provided on or associated with the common rail 28 to provide the functionality of providing a measurement of the temperature of the gas within the common rail 28 as a data input 11.1 to the control unit 11. The temperature sensor 45 is shown here connected to the common rail 28 but other positions would be acceptable, for example attached to the fuel supply line 40 or the shut-off valve 38. The functionality of the temperature sensor 45 and the functionality of the rail pressure sensor 30 may also be combined into a single unit or package. Such temperature sensing functionality may also be determined by a suitable temperature sensing algorithm that predicts the gas temperature based on ambient temperature, engine loading, tank temperature and any other appropriate factors, as is known in the art. Note that the rail pressure sensor 30, the knock sensor 41, the air pressure sensor 42, the temperature sensor 45 and the crank position sensor 44 may communicate with the control unit 11 in a conventional manner to provide it suitable data input 11.1. This may be achieved by suitable wired connections, or through the connection of a CAN-bus (Controller Area Network) (not shown) which is conventional in automotive technology. Other sensors may be included in the system, such as a temperature sensor (not shown) for measuring the temperature of the fuel in the rail 28. The engine system 2 further comprises a starter motor 46 which is configured to turn the crankshaft (not shown) in order to initiate self-sustaining power-producing operation of the engine system 2 Turning now to the exhaust system 10, combustion gases from the combustion chamber 12 feed into an exhaust duct or ‘manifold’ 50 which combines the gas out flow into a single pipe which leads to a turbine 52. As is known, the turbine 52 is connected to the compressor 19 and, together, the turbine 52 and the compressor 19 constitute a turbocharger of the engine system 2. Turbochargers provide a means to increase the density of the charge of air delivered to the combustion chambers 12, thereby providing more efficient and powerful combustion. However, their use is not essential to operation. Turbochargers are known in automotive technology so a full discussion will not be provided here for the sake of brevity. The exhaust manifold 50 is connected to an exhaust passage 51. In the illustrated example, the turbine 52 is located between the exhaust passage 51 and the exhaust manifold 50, although this is not essential as not all interna! combustion engine to which the invention applies include turbochargers. The exhaust passage 51 may include suitable mufflers or silencers to dissipate the sound energy carried by the exhaust gases, although such components are not shown in Figure 1. The exhaust passage 51 also includes a sensor 53 for detecting the air-fuel ratio in the exhaust gas. Such sensors are known in the art and will be referred to as the AFR sensor 53 from now on. The AFR sensor 53 may be any suitable air-fuel ratio sensor that is known in the art. Such sensors may be referred to as lambda sensors or oxygen sensors. A particularly suitable type of sensor is known as a wide-range air fuel ratio sensor, also known as a ‘WRAF’ sensor. WRAP sensors have a fast response time and, moreover, have the ability to measure air-fuel ratio across a wide range in a ‘linear’ manner which enables a fast and accurate detection of the air-fuel ratio (or the oxygen content). An air-fuel ratio sensor may form part of a wider system that is responsible for maintaining an appropriate fuelling ratio for the cylinders 12 of the engine. Such a system may be configured to adjust the fuel injection into the cylinders 12 to maintain the air-fuel ratio close to target values, as is known in the art. A known approach is to incorporate within the functionality of the control unit 11 an individual Cylinder Fuelling Control (ICFC) function 55 which is responsible for monitoring the output of the AFR sensor 53 and for applying appropriate control adjustments to the fuel injectors 26 to ensure that a correct fuel delivery to the cylinders. What is more, AFR sensors, and particularly WRAF sensors, may be used to control the fuelling in individual cylinders because of their fast response time. Such a control strategy is well known in the art, for example from US6382198B1. As shown in Figure 1, the ICFC function 55 is connected to the output of the AFR sensor 53 which may provide the ICFC with data relating to whether the air fuel ratio is rich, lean or at stoichiometry. Typically, the sensor output is linear between maximum and minimum values which provides a high resolution of the actual air fuel ratio of the cylinders. In particular, the AFR sensor 53 is configured to provide data relating to the air-fuel ratio of each of the cylinders 12 as its fast response time means that its data resolution can be timed with the travelling of exhaust gases through the exhaust passage 51. The control unit 11 may be configured to receive the sensor output directly and therefore calculate the air-fuel ratio of the cylinders 12 by internal algorithms or it may be configured to receive air-fuel ratio information from the ICFC function 55. By either process, the control unit 11 is able to determine the extent to which the air-fuel ratio is balanced between the numerous cylinders 12 of the engine, which may be referred to collectively herein as ‘cylinder imbalance data’. It should be noted that in the above discussion, the fuel delivery system 8 is configured into a port injection arrangement which means that the fuel injectors 26 are arranged to inject fuel into the air inlet manifold 24 so that the injected fuel, in this hydrogen gas, is mixed with fresh air in the inlet manifold before entering the cylinders 12 of the engine block. The invention, however, is equally applicable to a direct injection system. The skilled person would appreciate that the engine system 2 that is the focus of the above discussion has been simplified for present purposes and that in practice an engine system would be more complex. However, the illustrated engine system 2 is intended to demonstrate the principal components and subsystems that are relevant to the examples of the invention. As has been discussed above, the control unit 11 is operable to perform various engine monitoring and control objectives to manage the performance of the vehicle into which it is installed. The general operation of the control unit 11 would be well known to the skilled person and is outside of the scope of this discussion. It should be appreciated that the control unit 11 may be any suitable control environment provided by the engine system 2. The control unit 11 may be the “engine ECU” of the engine system or it may be another control unit which is configured to carry out other performance and monitoring tasks within the engine system 2 of the broader vehicle. In particular, the control unit 11 may be a control environment provided specifically for the purposes of performing the method. Irrespective of the functionality of the control unit 11, it will be appreciated that the control unit 11 has the necessary memory (e.g. NVM 13), processing environment (not shown) and communications interface (not shown) to be integrated into the engine system and the broader system of the associated vehicle. One challenge associated with hydrogen-fuelled engines is the potential for leaks to occur. In the examples of the invention, a method is presented in which a stuck open fuel injector can be detected and action taken to prevent further leaks. This may arise, for example, due to mechanical tolerances or as a result of debris or another foreign object being embedded in the valve seat. It is desirable to be able to detect leaks promptly so that action can be taken to reduce the impact of the leak. It is also desirable to detect leaks through a process that has minimal impacton the user of the vehicle. For example, it is undesirable to carry out a process through which fuelling is interrupted or stopped completely in order to infer that a leakage is present somewhere in the fuel injection system. An example algorithm or method can be implemented by the control unit 11 in order to determine when the valve needle of an injector 26 has become stuck open and there is a gas leak. The method is based on the principle that a leaking injector will result In a gas and air mixture excessively rich in gas being delivered into the associated cylinder of the engine. This means that the amount of gas in the cylinder will exceed what would be delivered under properly functioning conditions and so the exhaust gases would tend to be too rich as well. The presence of the AFR sensor 53 and / or the ICFC function 55, means that data relating to the combustion gases for each cylinder can be interrogated and a determination can be made about the potential for a leaking fuel injector. During normal engine operation the injectors of the engine system are open and closed in accordance with the driver commands to deliver the required fuel quantity to the engine cylinder during a plurality of injection events. An algorithm of the control unit 11 performs, intermittently, a leak detection method in accordance with one aspect of the invention. Referring to Figure 2, a method 100 in accordance with an example is initiated at step 102. Initiation of the method 100 may occur periodically at a selected frequency that is suitable to achieve robust monitoring for leakages without inducing excessive processor loading. It is envisaged that the method may be initiated at a frequency between 20ms and 1 second, for example around 100ms. This is merely exemplary. Once the process has initiated, the method 100 enters an air-fuel ratio monitoring loop at steps 104 and 106. At step 104 of the monitoring loop, a measurement of air-fuel ratio is performed. The, at step 106, it is checked whether the measured air-fuel ratio is within an acceptable range. More specifically, it is checked whether the air-fuel ratio within the cylinders 12 of the engine system 2 are indicating a ‘rich mixture’, which is significantly higher than a target air-fuel ratio. The precise levels of what would constitute an unacceptably rich air-fuel ratio may be specified during configuration of the system and would be within the capabilities of the skilled person. At this point, it should be noted that the monitoring of the air-fuel ratio may be conducted on the basis of a ‘average air fuel ratio which would apply to all of the cylinders 12 at the same time. Alternatively, the monitoring of the air-fuel ratio may be conducted on a cylinder-by-cylinder basis, as is possible with current WRAF sensor technology as mentioned above, and which is part of conventional ICFC functionality. The monitoring ioop steps 104,106 may continue for a predefined time period, it is envisaged that the monitoring loop may run continuously in order to monitor for an out of bounds air-fuel ratio, or it may be configured to run for a specific time period, for example in the order of one or more seconds. The control unit 11 may be operable to obtain data relating to the air-fuel ratio from the AFR sensor 53 or from the ICFC function 55. The obtained data, e.g. via a direct measurement, may indicate the ‘bulk’ air-fuel ratio or may indicate an air-fuel ratio specific to each one of the plurality of the cylinders 12. As an alternative, the control unit 11 may be operable to obtain an indication from the ICFC function about whether one or more of the cylinders is “running rich” such that a leakage may exist with the fuel injector associated with that cylinder 12. If the control unit 11 detects that one or more of the cylinders 12 is showing an incorrect air fuel ratio, then the process moves to step 108 at which point the control unit 11 issues an engine stop command. The engine stop command may be issued by the control unit 11 directly or a request may be made to another control system of the engine system or wider vehicle system. On a stationary power application such as a generator, the engine system 2 may shut down at this point (if safe), although it should be appreciated that in other applications an engine stop may be carried out when it is safe to do so. At this point, it should be noted that the leakage detection method 100 has identified a possible leaking fuel injector 26 through the monitoring of the air-fuel ratio of the engine system 2, for either all of the injectors or for individual injectors. Using the airfuel ratio is beneficial because monitoring can take place as a background task and it does not require an intervention in engine operation. In this case, step 106 is able to flag an initial indication of a possible injector leak. Following this, a confirmatory action will be performed in order to confirm that an injector ieak exists or whether the incorrect air-fuel ratio may be due to other factors beyond the scope of the ieak detection method 100. Once an engine stop has been initiated at step 108, the engine system 2 does not require fuel to be delivered and so the control unit 11 is configured to isolate the common rail 28 from the fuel tank 34 by operating the shut-off valve 38 to isolate fuel in the common rail 28 and the fuel supply line 40 from the fuel tank 34. This is indicated at method step 110. In this case, this means commanding the shut-off valve 38 into a closed position, in the event that the shut-off valve 38 is already closed at this point the term ‘operating’ covers checking the position status of the shut-off valve 38 and leaving it in the closed position. At this point, the control unit 11 also controls the fuel injectors 26 to configure them into a disabled state, so no fuel is injected into the manifold 24. Moreover, the spark plugs 31 may be deactivated at this point, although they may also be left active which may minimise the risk of leaked hydrogen escaping from the engine unburned. Once the shut-off valve 38 is closed, fuel injection is ceased, the method can then proceed to monitor the gas within the isolated fuel volume represented by the common rail 28 and the gas supply line 40. Closure of the shut off valve 38 defines the start of a test phase for the system during which the rail pressure is measured periodically. Rail pressure measurements are then monitored and sampled continuously, at step 112, through the test phase. At this point in the method 100, it should be appreciated that the pressure of gas, and also the quantity of gas, within the isolated fuel volume is, or can be assumed to be, the same as the pressure / quantity of gas within the isolated fuel volume at the point that the shut-off valve 38 was closed, at method step 110. From this point, therefore, the control unit 11 is configured to monitor for leaks within the isolated fuel volume before shutting down the engine system 2 fully, if appropriate to do so. In this way, the control unit 11 is able to confirm whether there is a leakage due to a faulty injector, as suggested by the identification of an incorrect air-fuel ratio at step 106, that would mean that it would be unsafe to allow the engine system 2 to be restarted at a later date. In such a circumstance, a suitable alert may be provided to the user, for example through a visual interface or attention-getter on the vehicle dashboard which would indicate that attention is required and that the engine system 2 is unable to be re-started. Optionally, during this test phase, a partial purge of the gas within the isolated fuel volume may be carried out. A gas purge may be performed in order to reduce the quantity of gas in the isolated fuel volume as a safety measure whilst still leaving a suitable quantity of gas in the isolated fuel volume at a suitable pressure so that leak testing can be performed. Such purging processes and hardware are outside the scope of this discussion. However, the process of purging the gas from the fuel volume may involve feeding the purged gas through a reactor chamber to reduce the concentration of hydrogen in the air, or to combust the hydrogen. Purging of some of the gas within the isolated fuel volume may also be achieved through leaving the engine system 2 running (fuel injectors operating and spark plugs activated) for a period of time whilst the shut off valve 38 is closed. Purging may also be carried out after the test phase is completed. For further illustration, reference will be made also to Figures 3a-3c which illustrates the quantity of gas within the isolated fuel volume during a period covered by the method of the invention. In Figure 3c, the gas quantity in the isolated fuel volume is shown on the Y-axis and time is represented on the X-axis. In Figures 3a-3c, it should be appreciated that the time prior to T1 corresponds to normal operation of the engine system 2, as is represented by step 102 in Figure 2. At T1, therefore, the operational state of the engine system 2 transitions from a powerproducing state to a non-power-producing state, in response to the control unit issuing the engine stop command at step 108, as discussed above. Figures 3a and 3b show, respectively, the engine system 2 transitioning between a power-on status and a power-off status (Figure 3a) and the shut-off valve 38 transitioning from an open position to a closed position. Although the transitions in the two signals are shown as happening simultaneously, it should be appreciated that precisely matched timing is not required. An optional purge operation (not shown in Figure 2) is shown at the time interval between 11 to T2, during which it can be observed that the quantity of gas within the isolated fuel volume is reduced from a high level Q0, to a second level, Q1, Q1 is a configurable threshold or target level which may be based on pressure or gas quantity. At this point, the control unit 11 has isolated a defined quantity (or pressure, as applicable) of gas within the fuel volume defined by the common rail 28 and the fuel supply line 40. The control unit 11 has therefore established initial conditions from which it is able to run a diagnostic subroutine. To this end, the control unit 11 is configured to monitor the quantity of gas within the fuel volume in the common rail 28 and the fuel supply line 40, herein after referred to simply as the Isolated fuel volume'. To do so, the control unit 11 is configured to determine the quantity of gas in the isolated fuel volume at two or more sample times that are spaced apart temporally. To this end, the control unit 11 uses data obtained by the rail pressure sensor 30 at first and second sample times, which are illustrated at T2 and T3 on Figure 3c. To cater for high leakages rates, a relatively fast sample time separation is envisaged to be most suitable, for example between 5ms and 15ms, for example 10ms. It will be noted that since the shut-off valve 38 is closed, the fuel pressure sensor 30 is measuring the pressure of gas not just in the common rail 28 but also in the fuel supply line 40. Therefore, it is the entire volume of fuel isolated between the shut-off valve 38 and the injectors 26 that is being determined. Since this volume may be considerable, and since the fuel supply line 40 may extend over a significant length (e.g. from the fuel tank 34 to the common rail 28 which may be a different ends of the vehicle), determining the quantity of gas present rather than simply the gas pressure is believed to provide a more accurate determination. To this end, the control unit 11 is configured to use the ideal gas law to determine the quantity of gas based on the measured gas pressure, since the volume of the common rail 28 and the fuel supply line 40 is known, and the temperature of the common rail 28 can also be determined using a suitable temperature sensor 45. It should be noted that the quantity of gas in the fuel volume may be determined by use of the ideal gas law, as mentioned above, because the volume of the common rail 28 and the fuel supply line 40 are known, and the temperature of the common rail 28 can also be determined using the temperature sensor 45. This approach may be particularly suited to when the pressure is relatively low e.g. around 1300kPa and temperature is relatively high (e.g. above critical point for hydrogen). In another approach, a density lookup table may be used which may be suitable for any pressure and temperature combinations. A suitable density table may be stored in the control unit 11 with pressure and temperature as the two axes. The control unit may then determine the density of the gas from pressure and temperature measurements with standard linear interpolation. Density data can be sourced from experimental work or published sources. Gas quantity can thus be determined from the density data because volume is a known quantity. The time interval between T2 and T3, shown here as T_delta, between the two data samples Q1 and Q2, is selected in order to provide a suitable time period to detect an appreciable drop in gas quantity but to avoid false positives and increasing the length of the analysis time. It is envisaged that a suitable time period for T.delta is between 5ms and 15ms, as mentioned above. At this point, the control unit 11 is configured to carry out suitable calculations so that it can be determined whether a leak is present. As an example of such a determination, the control unit 11 is configured to evaluate the change in gas quantity in the isolated fuel volume. This is illustrated as Q__delta in Figure 3c If the change is gas quantity between T2 and T3 exceeds a predetermined leakage gas threshold, then the control unit 11 can infer that a leak is present. The threshold may be set to take into account a 'dead band’ to avoid false positives due to measurement noise on the signal. As a second determination, the control unit 11 is configured to determine the rate of change of gas quantity in the isolated fuel volume between T2 and T3. This can be achieved by dividing the change in gas quantity Q_delta by the time interval T_delta. The rate of change of gas quantity can provide an indication of the severity of the leak from the isolated fuel volume. If the rate of change of gas quantity exceeds a predetermined leakage rate threshold, then the control unit 11 can infer that a leakage is present The leakage rate of gas may be determined in units of standard cubic centimetres per minute, ‘seem’ which is also expressed as cm3STP / min which defines the volume of gas corrected to a standard pressure and temperature. The first and second determinations may be performed as alternatives or may be performed together. It is believed that the second determination can provide a more accurate determination that a leakage is present and, possibly, provides the option of determining where the leak is located based on the severity of the rate of change of gas quantity. In this connection, it should be noted that the leakage from the common rail 28 and associated components may be at a wide range of rates, in which case a slow leakage rate may mean that checking just the pressure of gas in the fuel volume might not indicate a problem. Checking the rate of change of gas quantity is considered to be more effective at identifying slow leakage rates, it is believed that monitoring the rate of change of gas quantity can provide a more accurate determination that a leakage is present and, possibly, provides the option of determining where the leak is located based on the severity of the rate of change of gas quantity. Following the monitoring of the gas quantify, at step 112, the method proceeds to decision step 114 where it is determined whether the change in gas quantity is within allowable bounds which would indicate whether or not a leak is present. If, as a result of the above determinations, the control unit 11 determines that a leak is not present, then the method proceeds to step 116. Here, the method has determined that air-fuel ratio abnormality is not attributabie to a leaking injector and so the engine can be allowed to restart. It should be noted here that the air-fuel ratio abnormality may be attributable to other causes but that this is outside the scope of the method 100. If, at step 114, it is determined that the gas quantity within the isolated fuel volume is outside of allowable bounds, then the method is able to infer that a fuel injector leakage is confirmed and proceeds to step 118. At step 118, the method 100 is operable to evaluate the data obtained from the AFR sensor 53 and / or the ICFC function 55 to assess whether there is an imbalance of air-fuel ratio between the cylinders 12 of the engine system 2. This data may be obtained at the point the method executes step 118, or it may be stored at the point the air-fuel ratio is evaluated at steps 104,106. If it is determined that there is an imbalance of air-fuel ratio within the cylinders 12 at step 118, then the method 100 proceeds to step 120 where it may be identified that a leakage fault exists with a specific one of the fuel injectors associated with the cylinder 12. Accordingly, a fault message may be issued at step 122, and logged in memory as appropriate, at step 122. Other suitable actions can be taken at this point, as would be appropriate for the identification of a leakage fault with a specific fuel injector. For example the engine start function may be disabled, or the faulty fuel injector may be disabled, if the engine system 2 is capable of running on fewer than a total complement of cylinders 12. It should be noted at this point that the evaluation o cylinder imbalance may be based on air-fuel ratio that is gathered at steps 106 and 106. Returning to decision step 118, if at this point it is determined that there is no appreciable imbalance of air-fuel ratio between the engine cylinders 12, then the method 100 proceeds to step 124. At this point, if there is no significance air-fuel ratio imbalance between the cylinders 12, then various scenarios may exist. In principle, it is possible for all fuel injectors to have a leakage fault, although such a condition is very unlikely. Another possibility is that one of the fuel injectors has a leakage fault and that the leaked gas has travelled to the other cylinders 12, which may depend on the design of the inlet manifold 24. It can, however, be inferred that at least one of the fuel injectors has a leakage fault, even if the specific injector cannot be identified. The method 100 proceeds to step 122 wherein an appropriate fault response is carried out as before. However, the precise form of fault response may depend on the leakage scenarios that have been identified. For example, at step 124 it has been determined that there is no identifiable air-fuel ratio imbalance between the cylinders 12 of the engine and so an appropriate action may be to allow the engine to continue to operate, particularly if the leakage quantity and / or leakage rate identified at step 114 is not excessive. Alternatively, the response may be taken to disable engine starting. A further example method 200 in accordance with the invention will now be explained with reference to Figure 4. As will be apparent from the discussion that follows, the leakage detection method 200 described here makes use of data from the AFR sensor 53 to identify that a leakage fault may be present with one or more of the fuel injectors 26 of the engine system 2 and is then operable to perform a confirmatory action to identify the leakage fault more specifically. It will also be appreciated that, compared to the method 100 described previously, the following leakage detection method 200 does not rely on the engine system 2 being stopped in order to carry out the confirmatory action. Referring to Figure 4, a method 200 in accordance with an example is initiated at step 202. Initiation of the method 200 may occur periodically at a selected frequency that is suitable to achieve robust monitoring for leakages without inducing excessive processor loading, in a similar manner to method 100 as described above. Once the process has initiated, the method 200 enters an air-fuel ratio monitoring loop at steps 204 and 206. At step 204 of the monitoring loop, a measurement of air-fuel ratio is performed. Then, at step 206, it is checked whether the measured air-fuel ratio is within an acceptable range, for example whether the air-fuel ratio within the cylinders 12 of the engine system 2 are indicating a ‘rich mixture’, which is significantly higher than stoichiometry. The precise levels of what would constitute an unacceptably rich air-fuel ratio may be specified during configuration of the system and would be within the capabilities of the skilled person. The monitoring loop steps 204,206 may continue fora predefined time period. It is envisaged that the monitoring loop may run continuously in order to monitor for an out of bounds air-fuel ratio, or it may be configured to run for a specific time period, for example in the order of one or more seconds. It should be noted here that in the example method 200 illustrated in Figure 4, the control unit 11 is operable to monitor the air fuel ratio in respect of each cylinder of the engine system 2, e.g. as would be possible by the use of a WRAF sensor as has been discussed previously. The method 200 is therefore responsive to check the air fuel ratio for each cylinder 12 of the engine system 2 and thereby able to provide an initial indication that there may be a leakage fault with a specific fuel injector 26 associated with the cylinder 12 for which an air-fuel ratio error has been detected. As has been discussed above, the control unit 11 may be able to obtain data relating to the air fuel ratio of each cylinder directly by way of the AFR sensor 53 or through communication with the ICFC function 55. If the control unit 11 detects that a specific one of the cylinders 12 is showing an incorrect air fuel ratio, then the control unit 11 is able to infer that there is a gas leakage present at the fuel injector 26 associated with the identified cylinder 12. The method 200 then moves on to conduct one or more confirmatory actions to confirm that a fuel leakage is indeed present, such that a leakage event can be identified reliably and robustly and to reduce false positives. One a leaking fuel injector has been preliminarily identified, at step 206, the method moves to step 208 at which point the control unit 11 deactivates the fuel injector 26 that is associated with the cylinder 12 for which an incorrect air-fuel ratio has been identified. From here, reference will be made to ‘suspected faulty cylinder or, similarly, ‘suspected fault injector’, in this respect. The method 200 then moves to step 210 at which combustion is monitored in respect of the suspected faulty cylinder. Combustion may be detected in various ways, which can be summarised for present purposes as: monitoring the air fuel ratio for the suspected faulty cylinder; monitoring engine “knock” for the suspected faulty cylinder and monitoring engine torque production for the suspected faulty cylinder. Since the fuel injector 26 in respect of the suspected fault cylinder has been deactivated, it can be inferred that a combustion event occurring in that cylinder must be due to a leakage of gas through the injector from the inlet manifold 24 to the cylinder 12. Since the fuel injectors 26 are not injecting fuel into the combustion chambers 12, any vibration from the engine block 4 that is consistent with a combustion event can be attributed to a fuel leakage from the common rail 28 into the combustion chambers 12. This may be due to faulty injector components, for example. In order to infer combustion from air-fuel ratio data, the control unit 11 may be operable at step 220 to monitor combustion activities in relation to the suspected fault cylinder by detecting air fuel ratio changes that would indicate that combustion is occurring in the relevant time window relating to the compression cycle of the suspected fault injector. As would be understood by a skilled person, the time window for analysis would be as the outlet charge from a particular cylinder is exhausted and passes the sensor, which may be known as transport delay. It should be noted here that it would be expected that air fuel ratio would transition from a rich indication to a lean indication when deactivating an injector. This is because whilst the injector is turned on, it is injecting demanded fuel but also leakage fuel into the associated cylinder which results in a rich air fuel ratio. After the injector has been deactivated, it will only leak fuel into the cylinder as there will be no injection of demanded fuel. Therefore, an air-fuel ratio shift from rich to lean will be expected. The control unit 11 may also be operable to monitor for combustion activities by monitoring the output from the one or more knock sensors 41 associated with the engine block 4. As is known, the one or more knock sensors 41 monitor for structure- borne vibrations from the engine block 4 and convert these to electrical voltage signals which are suitably filtered and evaluated by the control unit 11. The control unit 11 is may be configured to apply a predetermined combustion threshold that is indicative of combustion occurring within the combustion chambers 12. The control unit 11 may also be operable to monitor the output from the crank position sensor 44. As is known, the crank position sensor 44 provides a signal indicative of the crank position, for example, by use of an optical encoder sensing techniques, inductive sensing techniques and Hall effect sensing techniques. The crank position signal is then processed to provide crank speed / acceleration. The control unit 11 may monitor the crank data relative to the time window associated for the power stroke of the suspected faulty injector and applies a suitable speed / acceleration threshold that is indicative of combustion occurring within the suspected faulty cylinder 12 thereby providing an engine torque indicator. The monitoring that is implemented during method steps 220 may be applied for a set time period, for example 1 to 2 seconds, the precise length of which would be a calibratable value. If combustion is detected for the suspected faulty cylinder at decision step 220, then the method moves to step 222, at which point confirmation is provided that the fuel injector 26 associated with the suspected faulty cylinder 12 has a leakage fault and appropriate action can be taken. This may mean that the engine is commanded to an off position or that a restart is not permitted. At step 224 a suitable fault message is issued and logged in the memory of the control unit 11 for later evaluation purposes. If combustion is not detected over the monitoring period of method step 220, then the method flows to step 226. At this point, it should be noted that if combustion is not detected in respect of the suspected faulty cylinder, then it cannot be ruled out that the associated fuel injector 26 is fault free. This is because the leakage from the injector may be at too low a rate to trigger combustion when the cylinder runs through the power stroke. Method step 226 and the steps that follow it therefore provide a further means to carry out a confirmatory action to determine whether the fuel injector 26 that is associated with the suspected fault cylinder has a leakage fault. Although step 226 is shown as following step 220 in Figure 4, it should be appreciated that method step 226 et seq. may also be performed in paraHei with steps 222 and 224, as is indicated by the dashed arrow marked 220’. At step 226, the control unit 11 operates to isolate the common rail 28 from the fuel tank 34 by operating the shut-off valve 38 to isolate fuel in the common rail 28 and the fuel supply line 40 from the fuel tank 34. In this case, this means commanding the shut-off valve 38 into a closed position. Once the shut off valve 38 is closed, the control unit 11 is operable to monitor the pressure (or gas quantity) within the isolated fuel volume by means of the rail pressure sensor 30. However, rather than simply evaluating the behaviour of the rail pressure as a whole over a predetermined time period, in this example the method 200 is configured to analyse a rail pressure characteristic in respect of the suspected faulty cylinder during a time period which an injection event is would be expected. In this connection, the term ‘rail pressure characteristic’ should be understood as meaning a parameter that is indicative of, or linked to, the pressure of gas in the common rail 28. This could be, for example, rail pressure as measured directly by the rail pressure sensor 28, or the quantify of gas in the common rail / fuel volume as is determined by the ideal gas law and as has been discussed above in relation to method 100 as set out in Figure 2. In general, this part of the algorithm can be considered as evaluating the rail pressure characteristic during a time period when the suspected faulty cylinder 12 is due to experience an injection event, but where that associated fuel injector 26 is deactivated. In such a scenario, it would be expected that the rail pressure characteristic would be substantially constant over that time period. The evaluation may take place over an injection sequence as illustrated in Figure 5 by way of example. The injection sequence includes a series of injection events (when the injector is injecting fuel) separated by between-injection periods (when the injector is not injecting fuel). When a specific injector is injecting, a valve needle of that injector is lifted away from a valve seat, and when that injector is not injecting the valve needle of the injector is seated against the valve seat. Figure 5 aiso illustrates the rail pressure characteristic for an injector assembly for which the valve needle for one of the injectors (INJ3 in this case) is stuck open, thereby creating a leak path from the injector into the air inlet manifold (or in the case of the direct injection system, into the engine cylinder). The upper line graph shows the state of the gas supply to the common rail 28. This is initially ON and is then switched OFF at time T1 when the shut off valve 38 is closed to define the start of the test phase. The shut off valve 38 remains closed until the end of the test phase at time T2. The lower line graph shows the injector drive pulses which initiate the injection sequence for injectors I NJ 1 -INJ4. The injector drive pulses are applied at regular intervals to injectors INJI, INJ2, INJ3 and INJ4. Each injector drive pulse initiates an injection event in the associated injector whereby the valve needle of the injector is moved away from its valve seat to commence injection and, after a predetermined time interval, closes again to terminate injection when the drive pulse is removed. The middle line graph shows the rail pressure characteristic for an injector assembly including a partially stuck open injector (INJ3), identified as line A, and the rail pressure characteristic for an injector assembly in which all injectors are functioning normally, identified as Line B, albeit that the third fuel injector in the sequence, labelled INJ3, is deactivated. It will be noted here that the injector drive signal trace shows fuel injectors INJ1, INJ2, INJ4 as having an active drive signal, whereas fuel injector INJ3, does not have an active drive signal, as it is shown in ghosted lines. Since it has been determined that the injector associated with suspected faulty cylinder 12 is INJ3, the method 200 is operable to evaluate the pressure of fuel within the common rail 28 at an intermediate time period T3 during the injector sequence. With the shut off valve 38 closed, and with no leaks at any of the other injectors 26 or within the iow pressure system downstream of the shut off valve 38, it would be expected that the rail pressure in the common rail 28 stays constant during time period T3 since injector INJ3 is deactivated. However, as can be seen in Figure 5 by observing Line A, the rail pressure has a negative gradient highlighted at G3. This is to be compared with the rail pressure gradient observable in the same time period, T3, in plot A. Once all injectors have carried out an injection event in the test phase (four in the example shown), the shut off valve 38 is opened at time T2 to terminate the test phase, allowing fuel to once again fill the common rail 28 and normal operation to continue (assuming no leak is detected) until the next test phase. For each injector drive pulse, the rail pressure is sampled and recorded continuously through the test phase between T1 and T2, both during injection events and between injection events when no injector is injecting. The rail pressure is sampled at least at the rising edge of at least the injector drive pulse for the suspected faulty injector, in this case INJ3, and at the falling edge of the injector drive pulse, at least. Further samples may be taken in respect of the other injector drive pulses and between adjacent injector drive pulses for the derivation of further injector behaviour data. This ensures that the required gradient measures for the rail pressure are obtained to determine any fault or leak state, as described further below. It is worth noting that the timing between injections, and the frequency of sampling the rail pressure measurements, will be dependent on the engine speed and the number of injectors / cylinders. In the case of an injector assembly that is functioning with INJI, INJ2 and INJ4 in an active state, and INJ3 in a deactivated state, where there is no injector leak associated with IN J3, the rail pressure characteristic through the test phase would follow Une A, where the rail pressure starts at an initial rail pressure level RO and decreases after each injection event to a new, lower level between adjacent injection events. Between injection events the rail pressure remains stable at a constant level. After the final injector INJ4 has delivered an injection of fuel, the shut off valve 38 is opened again at T2 and so the rail pressure increases back to the initial rail pressure level RO. Notably, the rail pressure remains stable during T3, wherein the deactivated INJ3 would have ordinarily performed an injection event. Comparing Line A to Line B, it will be noted that there is a continuous decay in the rail pressure characteristic between injection events, rather than a stable constant rail pressure value as in Une B. This is because there is a constant leak through INJ3 into the air inlet manifold 24, even if the injector is nominally closed. As a result, the pressure in the common rail is reduced to a much lower level at the end of the test phase than would otherwise have been the case when there is no leak, and, moreover, it is apparent that there is a reducing rail pressure gradient during time period T3 when an injection event in respect of INJ3 would have been initiated. In addition, looking at the rail pressure characteristic for INJ3, there is no change in the gradient of the rail pressure characteristic between the end of injection for IN J2, and the start of injection for IN J3. This is because the injector INJ3 is not responsive with the valve needle stuck partially open, so that opening and closing the leaky injector has no effect on the rate of decrease of rail pressure through this period, including for the period of the injection event for the subsequent injection events (corresponding to INJ1, INJ2 and INJ4), the gradient of the rail pressure characteristic is greater than the gradient for a normal functioning injector in which there is no leak, as identified at G1, G2 and G4, respectively. This is because when the other injectors I NJ 1, INJ2 and IN J4 are injecting, this is accompanied by an additional leak through INJ3 (as it is always stuck open). The rail pressure characteristic through the test phase (with the shut off valve 38 closed) can therefore be analysed to determine whether the fuel injector in respect of the suspected faulty cylinder is partially stuck open and non-responsive to injector drive commands. The defective injector (INJ3) can in this case be identified by looking for a non-zero rail pressure gradient (decreasing rail pressure) for an injection event by the injector when it has been deactivated, instead of a constant rail pressure (i.e. zero gradient), as would be expected for a normal functioning injector that has been deactivated. As mentioned above, the control unit 11 may be operable to assess the pressure gradients in respect of all of the injectors, in some examples. However, it is considered acceptable to just evaluate the pressure gradient in respect of the injection time period T3 relating to the suspected fault injector (INJ3). In the event that a leak is identified, an output fault signal can be generated to indicate a fault and / or engine operation can be ceased altogether (e.g. by shutting down the gas supply to the engine). This can be observed in Figure 4 where the process flow moves to step 222 whereat a leakage of the specific injector INJ3 is confirmed and appropriate action is taken, such as disabling engine restart or commanding an engine shut down, together with issuance of an injector fault message at step 224. if, as a result of the above determinations, the control unit 11 determines that a leak is not present with the identified suspected faulty injector INJ3, then the method 200 proceeds to step 232. Here, the method 200 has determined that the gas pressure evaluation in respect of the injector INJ3 is not determinative so it cannot be confirmed that the suspected injector INJ3 is leaking. Therefore, it must be assumed that the airfuel ratio abnormality is not attributable to a leaking injector and so the engine can be allowed to restart, and the shut off valve 38 is opened. It should be noted here that the air-fuel ratio abnormality may be attributable to other causes but that this is outside the scope of the method 200. In embodiments of the invention, the control unit 11 may generate an output to indicate a fault state within the engine system, for example in one of the injectors, and this output may be used to terminate operation of the engine system at the point at which a fault is detected (e.g. a partially stuck open injector, a fully stuck open injector), provide a fault signal to the driver to indicate that there is a leak or a fault and / or shut down the faulty injector. By performing the test under low speed / low load conditions, the test sensitivity is increased because the quantity of any leak is more prominent relative to the quality of fuel being injected through the test phase. This makes it easier to detect even a relatively small leak. Also, performing the test under low speed and load conditions (e.g. engine idle) allows the test phase duration to be prolonged before there is a requirement to recharge the common rail 28 by opening the shut off valve 38. It also ensures that the test is unobtrusive and the possibility for the driver to detect the test is minimised. it is important that, depending on which type of fault is being looked for, the rail pressure measurements are sampled frequently enough to be able the required gradient measurements to be determined, for comparison with the ideal scenario. The skilled person would understand that various modifications may be made to the specific examples of the invention discussed above without departing from the scope of the invention as defined by the ciaims. Some variants have been discussed above. Other will now be discussed below.

Claims

1. An engine system for a vehicle, the engine system comprising;a rail volume (28) for receiving gaseous fuel;a shut off valve (38) for controlling the supply of fuel to the rail volume (28);an injector assembly including a plurality of fuel injectors (26), each one of the injectors (26) for receiving fuel from the rail volume (28) and for delivering fuel to an associated engine cylinder in an injection event;an air-fuel ratio sensor (53) configured to measure the air-fuel ratio in an exhaust of the engine system,a rail sensor (30) configured to measure the pressure of gas in the rail volume; anda controller (11) configured to:receive air-fuel ratio data relating to the air-fuel ratio in the exhaust;determine, based on the air-fuel ratio data, that there is an air-fuel ratio abnormality associated with at least one of the cylinders,perform a confirmatory action to confirm that at least one of the plurality of injectors associated with the respective cylinders has a leakage fault.

2. The engine system of Claim 1, wherein the controller (11) is further configured:to analyse the air-fuel ratio data to determine cylinder imbalance data, thereby to identify one of the cylinders that has an air-fuel ratio abnormality.

3. The engine system of Claim 2, wherein, in performing the confirmatory action, the controller is configured to identify a specific one of the plurality of injectors that has a leakage fault based at least in part on the cylinder imbalance data.

4. The engine system of Ciaims 2 or 3, wherein the controller is configured, in response to determining the presence of an air-fuel ratio abnormality, to perform the confirmatory action by:determining the presence of gas leakage from the fuel rail based on a signa! received from the fuel rail sensor.

5. The engine system of Claim 4, wherein, in performing the confirmatory action that a leakage fault is present, the controller is configured to identify that a specific one of the plurality of fuel injectors has a leakage fault based on the determined presence of gas leakage from the fuei rail and based on the determined cylinder imbalance data.

6. The engine system of Claims 4 or 5, wherein in determining the presence of gas leakage from the fuel rail, the controller is configured to:command the engine system to stop operation:operate the shut-off valve (38) to isolate gas in the fuel volume;determine a change in gas quantity or pressure and / or a rate of change in gas quantity or pressure over a predetermined time period,confirm that a leakage fault is present if the change and / or the rate of change of gas quantity or pressure exceeds a predetermined threshold.

7. The engine system of Claim 2, wherein the controller is configured, following the identification that one of the cylinders has an air-fuel ratio abnormality, to perform the confirmatory action by:deactivating the fuel injector associated with the identified cylinder; andanalysing a combustion signal associated with the identified cylinder to identify the presence of a combustion event,and confirming that the identified fuel injector has a leakage if a combustion event is identified.

8. The engine system of Claim 7, wherein the combustion signal is provided by the air-fuel ratio sensor.

9. The engine system of Claim 7 or 8, wherein the controller is further configured, where a combustion event is not detected, to perform the confirmatory action by:operating the shut-off valve to isolate gas in the fuel volume;determining a change in gas quantity or pressure and / or a rate of change in gas quantity or pressure over a predetermined time period,confirming that a leakage fault is present with the identified injector if the change and / or the rate of change of gas quantity or pressure exceeds a predetermined threshold and based on the cylinder imbalance data.

10. The engine system of Claim 7 or 8, wherein, wherein the controller is further configured, where a combustion event is not detected, to perform the confirmatory action by:evaluating a rail pressure characteristic in respect of the deactivated fuel injector at a fuei injection time period;comparing the rail pressure characteristic with an expected rail pressure characteristic for a non-fault injector during the fuel injection time period;determine that the deactivated injector is faulty based on the result of the comparison.

11. A method of determining a leak state in an engine system (2) for a vehicle, the engine system (2) comprising: a rail volume (28) for receiving gaseous fuel; a shut off valve (38) for controlling the supply of fuel to the rail volume (28); an injector assemblyincluding a plurality of fuel injectors (26), each one of the injectors (26) for receiving fuel from the rail volume (28) and for delivering fuel to an associated engine cylinder in an injection event; an air-fuel ratio sensor (53) configured to measure the air-fuel ratio in an exhaust of the engine system, and a rail sensor (30) configured to measure 5 the pressure of gas in the rail volume; the method comprising:receiving air-fuel ratio data relating to the air-fuel ratio in the exhaust;determining, based on the air-fuel ratio data, that there is an air-fuel ratio 10 abnormality associated with at least one of the cylinders,performing a confirmatory action to confirm that at least one of the plurality of injectors associated with the respective cylinders has a leakage fault.33

Images