Description:Technical Field of invention
The present subject matter relates to a system and a method for monitoring the age of a sensor disposed in the exhaust channel of a power unit.
Background
Emission norms on vehicles with a power unit comprising an internal combustion engine, are put in place to control environmental pollution, make after treatment of exhaust gases a mandated industry standard for all vehicle manufacturers. One such technique employed for after treatment of exhaust gases is by passing the exhaust gases through a catalytic converter in a dedicated exhaust channel, before such gases are released in the environment. The catalytic convertor converts the toxic gases such as CO, NO and unburnt hydrocarbons present in the exhaust gases into nontoxic gases by using precious metals such as Rh, Pd, Ce etc. The catalytic converter can simultaneously oxidise carbon monoxide and hydrocarbons to carbon dioxide and water vapour, while reducing NOx to nitrogen dioxide (NO2). The ceramic substrate of the catalytic converter has the capability to store oxygen on the surface, thus it can supply oxygen into the reaction and store it back in the next phase of the reaction. Earlier, as open loop air fuel ratio control was permissible, fuel economy was improved by operating the engine in leaner air fuel ratios in low torque operating zones. As closed loop air fuel ratio control became mandatory, this necessitated better air fuel ratio control strategies. An air fuel ratio sensor (lambda sensor) is located in the exhaust manifold and provides the necessary feedback of the combustion information to the ECU (electronic control unit) of the vehicle. Typically, a binary type heated exhaust gas oxygen sensor is used that responds with a voltage signal of 200mV to 800mV. A lower voltage refers to a lean mixture, wherein more of oxygen is available for combustion than fuel, while a higher voltage refers a rich air fuel ratio in combustion.
Summary of the Invention
Current emission norms mandate that ageing of such oxygen sensors are monitored along with the health of the catalytic converter. Sensors and catalytic converters are disposed in various combinations and permutations in the exhaust channel to best suit the age detection mechanism designed by the manufacturer.
Such lambda sensors may suffer from aging due to various known or unknown reasons. In lambda sensors, majorly there are mainly three type of faults which may cause violations of emission norms. First, deviation in peak values of the sensor response, although it is very rare to see but deviation in extremes is also one type of sensor fault. In this condition the extreme values for rich or lean will be shifted upward or downward from the nominal value. Second is the delay in response of the sensor. This type of fault creates the time shift in the signal. The switching from lean to rich or rich to lean will be delayed from the nominal time. This delay values for the lean to rich and rich to lean transition will be different. Third is, a slow response of the sensor. The switch between two level (lean and rich) is analogous to first order system, the system response becomes sluggish for the old sensor. It can be interpreted as, the time constant of the system will increase over the time and the switching speed will be comparatively slower than new sensor. An aged sensor thus may show the air fuel mixture to be aged even though the mixture is lean. The signal from the aged sensor will be received by the ECU and it will try to reduce the fuel pulse width for the injector to make the mixture lean. Therefore, this kind of misinterpretation may cause the improper calculation of fuel correction. Therefore, incorrect / biased feedback can cause deviation from the stoichiometric ratio of the air fuel mixture. Thus, the old sensor causes the shift in the fuel pulse width command. Nitrogen Oxide (NOx) will be higher in lean mixture, and hydrocarbons (HC), carbon monoxide (CO) will be excess in a rich mixture. Also, light off temperature will be higher for an aged lambda sensor. It may cause poor fuel economy due to rich fuelling in cold start condition of the engine. Imbalance in fuelling damages the engine parts and causes unwanted behaviours like engine knocking, misfire, etc. Further, an aged sensor leads to higher conversion burden on the catalytic converter, which causes the converter to age more quickly. A faulty lambda sensor, therefore potentially leads to increased emissions, increased fuel consumption or engine damage.
The model-based system diagnosis methodologies are aimed to perform precise and accurate diagnostics but at the cost of computation throughput which may not be suitable for real time ECU implementation. Neural network-based methodologies have been tried for the system diagnosis. However, it requires huge data sets to train the network at every stage of sensor aging.
In an existing prior art on this subject matter, it is known to analyse the change in a fuel pulse width correlation with an output of an upstream lambda sensor in the exhaust channel. Thus, the lag of the fuel pulse width time duration serves as an indicator of the aging of the sensor. In another existing prior art on this subject matter, an imbalance parameter between the air fuel ratio in the cylinders of the engine are determined. Thus, both the known prior arts consider the lag of the fuel injector in its fuel pulse width as a measure of sensor aging. Moreover, these systems would require multiple detecting sensors and computing units, making the systems costlier. Furthermore, simply considering the lag in the fuel pulse width is not a very efficient system of determination of age of the sensor. In some implementations, physics-based modelling of the lambda sensor’s behaviour is done that is computationally expensive and requires a lot of time and efforts to achieve the right system parameters.
In view of the above, there is a need for a system and a method for monitoring the age of the lambda sensor disposed in the exhaust channel of a power unit, which addresses one or more of the limitations of the existing systems and methods stated above.
In one aspect, a system for monitoring an age of a lambda sensor disposed in an exhaust channel of a power unit is disclosed. The system comprises a lambda sensor. The lambda sensor is disposed in the exhaust channel of the power unit of the vehicle. The system also comprises one or more sensors coupled to the power unit. The one or more sensors are configured to generate one or more sensor parameters of the power unit. Further, the system comprises a control unit. The control unit is operably coupled to the lambda sensor and the one or more sensors. The control unit is configured to receive an exhaust status from the lambda sensor and the one or more generated parameters from the one or more sensors, and determine the age of the lambda sensor using the received exhaust status and the one or more generated sensor parameters based on a computation technique. The computation technique is adapted to estimate a gain factor of the lambda sensor based on the received exhaust status and the one or more generated sensor parameters.
In an embodiment, the control unit is configured to determine the age of the lambda sensor using the received exhaust status and the one or more generated sensor parameters based on a the computation technique by generating an estimated exhaust status of the lambda sensor using the one or more sensor parameters from a current engine cycle by a sensor model, and determining the gain factor using the computation technique to reduce an error between the estimated exhaust status and the received status of the lambda sensor. The sensor model is stored in the control unit and the sensor model is a mathematical model of the lambda sensor defined as rate of change of the exhaust status of the lambda sensor is equal to a product of the gain factor and an incremental change in at least of the one or more sensor parameters, wherein the at least one of the one or more sensor parameters is a fuel pulse width.
In an embodiment, the control unit is configured to classify the lambda sensor as one of a fresh condition, an intermediate condition, an aged condition, based on the gain factor using a classification module.
In an embodiment, the system comprises a diagnostics unit. The diagnostics unit is communicably coupled to the control unit. The diagnostics unit is configured to generate an audio-visual indication to alert a user of a vehicle of the age of the lambda sensor. The audio visual indication is generated upon at least one of determined aged condition of the lambda sensor and on reception of a user generated request signal.
In an embodiment, the control unit is configured to generate a fuel pulse width for injecting a fuel into a combustion chamber of the power unit, the fuel pulse width is determined based upon the exhaust status of the lambda sensor and one or more sensor parameters from a throttle position sensor and a temperature manifold absolute pressure sensor.
In an embodiment, the computation technique is a recursive least square estimation technique with a Kalman filter module. In an embodiment, the lambda sensor is disposed upstream of a catalytic convertor in the exhaust channel. In an embodiment, the control unit comprises an input-output module, a storage module, a lambda sensor output estimation module, a calculation module implementing the computation technique and a classification unit for determining the age of the lambda sensor using the received exhaust status and the one or more generated sensor parameters based on the computation technique.
In an embodiment, the control unit is configured to store the generated gain factor of the lambda sensor, and to classify the stored gain factor of the lambda sensor, after every predetermined number of engine cycles, to determine the age of the lambda sensor, the storage is local or remote.
In an embodiment, the control unit is configured to pre-process the received exhaust status from the lambda sensor and the one or more generated parameters from the one or more sensors to prevent high frequency gain switching.
The one or more sensor parameters comprising a throttle position, a temperature manifold absolute pressure, engine speed, and a fuel pulse width and the one or more sensors is a throttle position sensor (TPS), and a Temperature Manifold Absolute Pressure (TMAP) sensor.
In another aspect, a method for monitoring a lambda sensor disposed in an exhaust channel of a power unit is disclosed. The method comprises the steps of: generating, by a lambda sensor, an exhaust status of the power unit, generating, by one or more sensors, sensor parameters corresponding the power unit, receiving, by a control unit, the exhaust status and the one or more sensor parameters, generating, by the control unit, an estimated exhaust status using the received exhaust status and the one or more sensor parameters from a current engine cycle, determining, by the control unit, a gain factor by comparing the estimated exhaust status and the received exhaust status using a computation technique, and determining, by the control unit, an age of the lambda sensor based on the determined gain factor.
In an embodiment, the step of determining the age of the lambda sensor based on the determined gain factor comprising classifying the age of the lambda sensor as one of a fresh condition, an intermediate condition, an aged condition, based on the gain factor using a classification module. Classifying the age comprises classifying the determined gain factor, by the control unit, after every predetermined number of engine cycles, to determine the age of the lambda sensor.
In an embodiment, the method comprises pre-processing the received exhaust status from the lambda sensor and the one or more generated sensor parameters from the one or more sensors to prevent high frequency gain switching
In an embodiment, the step of determining the gain factor using the computation technique comprises reducing an error between the estimated exhaust status and the received status of the lambda sensor, where the estimated exhaust status of the lambda sensor is generated by a sensor model using the one or more sensor parameters from the current engine cycle.
In an embodiment, the method comprises generating an audio-visual indication, by a diagnostics unit, to alert a user of a vehicle of the age of the lambda sensor, the audio visual indication is generated upon at least one of a determined aged condition of the lambda sensor and on reception of a user generated request signal.
In an embodiment, the method comprises generating, by the control unit, a fuel pulse width for injecting a fuel into a combustion chamber of the power unit, the fuel pulse width is determined based upon the exhaust status of the lambda sensor and one or more sensor parameters from a throttle position sensor and a temperature manifold absolute pressure sensor.

Brief Description of Drawings
Reference will be made to embodiments of the invention, examples of which may be illustrated in accompanying figures. These figures are intended to be illustrative, not limiting. Although the invention is generally described in context of these embodiments, it should be understood that it is not intended to limit the scope of the invention to these particular embodiments.
Figure 1 is an exemplary illustration of a block diagram of the system for monitoring a lambda sensor of a vehicle.
Figure 2 is an exemplary in-depth illustration the system for monitoring the lambda sensor.
Figure 3 is an exemplary illustration of a flow chart of the method for monitoring the lambda.
Figure 4 is an exemplary graphical representation of one or more sensor parameters of the power unit,
Figure 5 is an exemplary graphical representation of an estimation error between the estimated exhaust status and the received exhaust status for a corresponding engine cycle, and
Figure 6 is an exemplary graphical representation of the lambda sensor gain factor determined over a period of time.
Detailed Description
Various features and embodiments of the present invention here will be discernible from the following description thereof, set out hereunder.
Figure 1 is an exemplary illustration of a block diagram of a system 100 for monitoring an age of a lambda sensor 109 of a vehicle. The system 100 comprises a power unit 100a typically disposed in a vehicle. Specifically, the block diagram represents the flow of air through the various parts of the power unit 100a. The power unit 100a broadly consists of three components which are in the path of the air flow through the power unit 100a, a throttle body 101, an engine 102, and an exhaust channel 103. Each of these three components consists of multiple sensors and actuators. These sensors and actuators are communicatively connected to the control unit 104. In the exemplary embodiment of the present invention, the control unit 104 is further communicatively connected to a diagnostic unit 110.
A throttle body 101 is a part of the power unit 100a which serves an inlet for air from the atmosphere, and essentially the air contains oxygen, which when mixed with a hydrocarbon based fuel, becomes combustible. This mixture of air and fuel is combusted in a combustion chamber of the engine 102, which drives a piston, transferring power to the moving parts of the vehicle. The combustion causes a residue of exhaust gases, which exit the power unit 100a through the exhaust channel 103 (also known as exhaust manifold) of the vehicle. A catalytic converter is placed in this exhaust channel 103 of the power unit 100a. The throttle body 101 consists of an accelerator (not shown) which is controlled by a user of the vehicle, to control the opening of the air intake (not shown) of the power unit 100a. One or more sensors are placed in the throttle body 101 of the power unit 100a. These include a throttle position sensor (TPS) 105, and a Temperature Manifold Absolute Pressure (TMAP) sensor 106. The TPS 105 is used to monitor the air intake of the engine 102. This sensor is usually located on a butterfly spindle / shaft (not shown), so that it can directly monitor the position of the throttle opening. The butterfly shaft regulates the airflow, and is actuated in response to the motion of the accelerator / throttle grip controlled by the user. The TMAP sensor 106 is used to measure the temperature and pressure of the intake air, which is then mixed with the fuel and combusted in the engine 102. These sensors TMAP sensor 106, the TPS 105, a fuel injector 108 generate sensor parameters in the system 100. The sensor parameters are a throttle position, a temperature manifold absolute pressure, a pulser coil position or engine speed, and a fuel pulse width
An engine 102 primarily consists of at least one combustion chamber (not shown), at least one piston (not shown), and at least one rotating shaft (not shown) coupled to the piston rod (not shown). The fuel is injected into the combustion chamber of the engine 102 by the fuel injector 108. The engine 102 may be a spark ignition internal combustion engine. In a single engine cycle, a mixture of air and fuel is combusted once, and the piston moves 4 times along the length of the combustion chamber, rotating a shaft connected to the piston by 720 degrees. A spark plug 107 is used to ignite the fuel and air in a compressed state during the combustion stroke of the engine cycle.
The gases resulting from the combustion are let out of the outlet valves of the engine, and into the exhaust channel 103. The exhaust channel 103 primarily serves the function of directing the exhaust gases to the rear of the vehicle so that the user of the vehicle is not affected by them. Further, the exhaust channel 103 cools the gases, reduces the noise from the emission of the gases through one or more mufflers / silencers, and holds one or more catalytic converters for converting toxic gases such as CO, NO and unburnt hydrocarbons present in the exhaust gases into nontoxic gases by using precious metals such as Rh, Pd, Ce etc. The lambda sensor 109 is placed in the exhaust channel 103 to monitor the oxygen content in the exhaust gases to give an indication on the type of air fuel mixture (lean or rich), and also monitor the age of the catalytic converter(s). In an embodiment, the lambda sensor 109 is disposed upstream of the catalytic convertor in the exhaust channel 103. The lambda sensor 109 detects an exhaust status, and further transmits this exhaust status to the control unit 104. The term “exhaust status” refers to a signal generated by the lambda sensor 109 that indicates of the type of air fuel mixture (lean or rich) being combusted in the engine. The lambda sensor 109 responds to the measured air fuel ratio (AFR) with respect to a stoichiometric ratio. If the measured AFR is lesser than the stoichiometric value, then the value of Lambda will be lesser than 1 and the output of the switch type lambda sensor will be exhaust status voltage value will be greater than 800mV. If the measured AFR is greater than the stoichiometric value, then the value of Lambda will be greater than 1 and the exhaust status voltage value will be lesser than 200mV.
Mathematically, Lambda =(A/F)/(A/F)stoichiometric
The control unit 104 as per the present invention is an electronic computing device. The control unit 104 may be an electronic control unit of the engine 102 that is involved in the closed loop air fuel ratio control. The control unit 104 may be the engine control unit (ECU) or may be a control unit in the vehicle apart from ECU. It is configured for receiving input signals from one or more sensors, that is, the TPS 105 and the TMAP sensor 106 and generating an output for an actuator, or another computing device. That is, the control unit 104 is operably coupled to the lambda sensor 109 and the one or more sensors 105, 106, 108 and determines the age of the lambda sensor 109 using the exhaust status and the sensor parameters based on a computation technique. In an embodiment, the control unit 104 may determine the fuel pulse width based on the ON time and OFF time of the fuel injector 108. In another embodiment, the control unit 104 is configured to generate a fuel pulse width for injecting a fuel and the fuel pulse width is determined based on the exhaust status of the lambda sensor 109 and one or more sensor parameters from the TPS 105 and the TMAP sensor 106. This determination of the fuel pulse width may be performed in the closed loop control of the AFR of the vehicle.
The diagnostics unit 110 as per the present invention is configured to generate an audio-visual indication to alert a user of a vehicle of the age of the lambda sensor 109. In an embodiment, the same diagnostics unit 110 may be used to alert the user of the vehicle regarding any failure or malfunction in the vehicle.
Figure 2 is an exemplary in-depth illustration of the system 100 for monitoring the lambda sensor 109. The control unit 104 is configured to determine age of the lambda sensor 109 using the received exhaust status and the one or more generated sensor parameters based on a computation technique by generating an estimated exhaust status of the lambda sensor 109 using said one or more sensor parameters from a current engine cycle by a sensor model, and determining the gain factor using the numerical computation technique to reduce an error between the estimated exhaust status and the received exhaust status of the lambda sensor 109. The sensor model is stored in the control unit 109 and the sensor model is a mathematical model of the lambda sensor 109 defined as rate of change of the exhaust status of the lambda sensor 109 being equal to a product of the gain factor and an incremental change in at least of the one or more sensor parameters, wherein the at least one of the one or more sensor parameters being a fuel pulse width.
As per the present embodiment, the control unit 104 comprises an input-output module 104a, a storage module 104b and a processing module 104c. The input-output module 104a, storage module 104b, and processing module 104c are communicatively connected to each other. The input-output module 104a is configured to receive the exhaust status from the lambda sensor 109 and sensor parameters from the one or more sensors 105, 106, 108. The processing module 104c comprises a lambda sensor output estimation module 104d, a calculation module 104e, and a classification module 104f. The processing module 104c, through its component modules as mentioned above, is configured to generate a gain factor of the lambda sensor 109, which is then stored in the storage module 104b of the control unit 104. In an embodiment, the sensor model is stored in the storage module 104b of the control unit 104. The lambda sensor output estimation module 104d is configured to generate a simulated output of the exhaust status in a subsequent engine cycle based on the received exhaust status and the sensor parameters received from the lambda sensor 109 and the one or more sensors 105, 106, 108 in a current engine cycle by applying these inputs in the current cycle on the sensor model. The calculation model 104e is configured to execute a numerical computation technique to calculate the gain factor of the lambda sensor 109. The numerical computation technique is based on a recursive least square estimation with a Kalman filter. The mathematical sensor model is given is equation (1):
V ̇_L=K_L.Δψ (1)
In discrete form:
V ̂_L (k+1)=V_L (k)+K_L.t_s.Δψ (2)
V ̂_L (k+1)=[K_L 1][■(V_L (k)@t_s.Δψ)] (3)
V ̂_L (k+1)=θ.ϕ (4)

θ=[K_L 1] and ϕ=[■(V_L (k)@t_s.Δψ)]
The estimated exhaust status output V ̂_L at the (k+1)th instant can be expressed in the terms of a parameter vector θ and a measurement vector ϕ. The parameter vector θ contains the gain factor K_L and 1. The measurement vector is having V_L (k), that is measured exhaust status signal at the kth instance and second element is t_s sampling time of the exhaust status, and Δψ, which is the change in the fuel pulse width from the observed mean value (OMV) of the fuel pulse width as given by the fuel injector 108 or the control unit 104 of the vehicle. The OMV for fuel pulse width is calculated by the lambda sensor output estimation module 104d by considering the observed average value of fuel pulse width of past n engine cycles and an incremental change is calculated by subtracting the current fuel command from the OMV. In an embodiment, the lambda sensor output estimation module 104d computes the fuel pulse width for the fuel injector 108 depending on inputs from the lambda sensor 109 and the other one or more sensors 105, 106, 108.
The adaptive sensor gain factor KL captures the sensor aging effects such as speed of the sensor response, delay, etc. and it lumps all other sensor related parameters such as temperature and speed together. In equation (1), the rate of change of lambda sensor voltage signal or exhaust status is directly proportional to an incremental change in fuel pulse width with the proportionality coefficient KL termed as sensor gain factor. A fresh lambda sensor will be very sensitive to change in fuel command so the sensor gain factor KL will higher for new sensor then it will further reduce as the sensor 109 will become less responsive due to aging.
Incremental change in fuel pulse width will be input to the lambda sensor 109 placed in the exhaust channel 103 and the sensor model of the lambda sensor 109. The system output will be measured in the form of electrical voltage signal, that is, received exhaust status from the lambda sensor 109 and the estimated exhaust status from the sensor model will be calculated using equation (3). Difference of both outputs is calculated as an estimation error. This estimation error is fed back to a calculation module 104e to update the parameter vector. Updated system parameter gain factor KL is again supplied into the sensor model. The error will be calculated and manipulation in gain factor KL will be continued to minimize the estimation error. This formulation will estimate the output of the Lambda sensor 109 at the k+1th instance depending on the measured parameters at the kth instance. Thus, the value of the gain factor KL can be determined from the formulation. The output will be the value of sensor gain factor KL. Thus, the above formulae are the mathematical representation of the lambda sensor output estimation module 104d and the calculation module 104e.
The numerical computation technique used in the calculation module 104e to estimate the parameter KL is an online recursive least square estimation technique with a Kalman filter module. The recursive least square technique uses an iterative algebraic approach in conjunction with a matrix inversion lemma to define an algebraic recursion relationship between a Kalman gain vector and a Kalman filter (not shown) for processing the gain factor KL, wherein the recursive least square technique approaches the Kalman filter with reduced required throughput in a signal processor (not shown). The Kalman filter along with the recursive least square technique consider the total square errors rather than mean square errors during the numerical computation. Further, a new data sample of the measured input, that is, the measurement vector, and output will be received by the calculation module 104e and further estimation error will be calculated in the calculation module 104e by subtracting the estimated output from a measured output (received exhaust status). The estimation error will be fed to a Kalman gain update block (not shown) and then a pseudo inverse matrix will be updated jointly along with the gain factor KL. The gain factor KL will be recursively updated since the input data (received exhaust status and the OMV of the fuel pulse width) is continuously streaming. Thus, the calculation module 104e updates the gain factor KL.
In an embodiment, in the input-output module 104a, the measured sensor data, that is the received exhaust status and the fuel pulse width is passed through a low pass filter to avoid the case of high frequency gain switching. Filtering of the fuel pulse width signal is thus done using a moving average filter. For example, window size of the moving average filter is taken as 10 samples whereas measured data is logged at sampling rate of 10 milliseconds.
The gain factor KL as determined here will be stored in the storage module 104b through the communicative coupling between the calculation module 104e and the storage module 104b. In an embodiment, the values of the gain factor KL may be stored remotely in a cloud-based database. The classification module 104f of the control unit 104 classifies the stored lambda sensor gain factor repetitively after a predetermined number of engine cycles, or when a user generated request signal is received through the input-output module 104a. The sensor gain factor converts to different value for differently aged sensor 109. The estimated value of the sensor gain factor is used by the classification module 104f to quantify the lambda sensor age. The classification module 104f employs a trained classifier to classify the age of the lambda sensor as one of a fresh condition, an intermediate condition, an aged condition, based on the gain factor that reduces the estimation error. The input-output module 104a further communicates the determined condition of the lambda sensor 109 to the diagnostics unit 110. The diagnostics unit 110 generates an audio-visual indication to alert a user of a vehicle of the age of the lambda sensor 109. The audio-visual indication is generated if aged condition of the lambda sensor 109 is determined or on reception of a user generated request signal.
Figure 3 is an exemplary illustration of a flowchart of a method 300 for monitoring age of the lambda sensor 109. The method 300 is performed by the components of the system 100 illustrated in Figures. 1-2. As depicted, at step 301, an exhaust status is generated by the lambda sensor 109. At step 302, one or more sensor parameters are generated by the one or more sensors 105, 106, 108 disposed on the power unit 100a. The one or more sensors include the throttle position sensor 105, the TMAP sensor 106, and also the fuel pulse width as provided to actuate the fuel injector108. At step 303, the input-output module 104a of the control unit 104 receives the inputs from the lambda sensor 109 and the one or more sensors 105, 106, 108 generated at steps 301 and 302. At step 304, a lambda sensor output estimation module 104d of a processing module 104c of the control unit 104 is configured to estimate the exhaust status for a subsequent engine cycle (k+1th instance) based on the inputs received at the current engine cycle (kth instance) at steps 301 and 302. At step 305, a calculation module 104e of the processing module 104c of the control unit 104 is configured to determine the lambda sensor gain factor by a numerical computation technique by comparing the estimated exhaust status generated in step 304, and the received exhaust status, which is defined by the mathematical formulation described above. At step 306, the classification module 104f of the processing module 104c of the control unit 104 determines the age of the lambda sensor 109 based on the determined gain factor. In an embodiment, the processing module 104c stores the determined gain factor in a storage module 104b of the control unit 104. Further, the classification module 104f classifies the stored lambda sensor gain factor automatically after a predetermined number of engine cycles and determines the age of lambda sensor 109 as one of a fresh condition, an intermediate condition, and an aged condition, based on the gain factor. Further, in an embodiment, the received exhaust status from said lambda sensor 109 and the one or more generated sensor parameters from the one or more sensors 105, 106, 108 are pre-processed to prevent high frequency gain switching.
In an embodiment, the numerical computation technique is a recursive least square estimation technique. For determining the gain factor using the computation technique, the control unit 104 reduces an error between the estimated exhaust status and the received status of the lambda sensor 109, where the estimated exhaust status of the lambda sensor 109 being generated by a sensor model using said one or more sensor parameters from said current engine cycle.
The method 300 further comprises generating an audio-visual indication, by a diagnostics unit 110, to alert a user of a vehicle of the age of the lambda sensor 109. The said audio-visual indication being generated upon at least one of a determined aged condition of the lambda sensor 109 and on reception of a user generated request signal. In an embodiment, the diagnostics unit 110 is configured to indicate the age of the lambda sensor 109 based on this classification once the gain factor reaches a predetermined threshold. The diagnostics unit 110 is also configured to indicate the age of the lambda sensor 109 if it receives a user generated request signal for displaying the age of the lambda sensor 109. The age is displayed as one of new, intermediate and old. The lambda sensor gain factor decreases with respect to one or more parameters of the power unit 100a with aging. A newer lambda sensor 109 will therefore have a higher gain factor (KL). The user can have a designated diagnostics tool being used to send the request signal. This signal may be sent over wired or wireless connections to the diagnostics unit. Therefore, in an embodiment, the diagnostics unit 110 is configured to have a wireless communication module, which can connect with the diagnostics tool of the user. In an embodiment, this tool may be the user’s mobile phone configured with an application for communicating with the diagnostics unit 110 of the system 100. The wireless communication module of the diagnostics unit 110 may function on one of a Wireless Fidelity (Wi-Fi), Bluetooth, Infrared, or any other commonly used wireless architecture. In an embodiment, the audio-visual indication may be on the instrument cluster of the vehicle or on user device such as smart phone in possession of the user.
In embodiment, the method 300 comprises generating, by the control unit 104, a fuel pulse width for injecting a fuel into the combustion chamber 102 of the power unit 100a. The fuel pulse width being determined based upon the exhaust status of the lambda sensor 109 and one or more sensor parameters from the TPS 105 and the TMAP sensor 106.
Figure 4 is an exemplary graphical representation of one or more sensor parameters of the power unit 100a. In this particular embodiment, the characteristics of the power unit 100a are represented by the throttle opening with respect to engine speed. The throttle opening is dependent on user input through an accelerator of the vehicle, and is measured by sensors such as the throttle position sensor 105 and the TMAP sensor 106. Increasing the throttle input by the user results in more air intake by the power unit 100a, and thus the control unit 104 then determines the ideal fuel pulse width for the increased throttle input, which translates to the fuel injected by the fuel injection system 108. More air and fuel into the combustion chamber of the engine 102 means that more exhaust gases are produced as a result, and the piston travels a larger distance in a shorter duration of time, increasing the speed of rotation of the main shaft of the engine 102, thus increasing the vehicle speed. Figure 4 illustrates that with more throttle opening, the engine speed rises as a general characteristic of the power unit 100a. The air fuel mixture is usually classified as being a rich mixture or a lean mixture. It is an objective of the ECU or the control unit 104 to keep the air fuel ratio at a stoichiometric level. Generally, that would mean that the fuel pulse width is usually determined with respect to the air intake as determined by the throttle position sensor 105 and the TMAP sensor 106, so that the ratio of air and fuel in the engine 102 is as close to the stoichiometric ratio as possible. The lambda sensor 109 is configured to determine the exhaust status, that is classify the exhaust gases as one resulting from a rich mixture and one resulting from a lean mixture. If the throttle opening is not changed by the user, based on the input received from the lambda sensor 109, the ECU or the control unit 104 will correct the fuel pulse width of the fuel injector 108. For vehicles using a switch type lambda sensor 109, this results in a lag in the response to the fuel pulse width correction. Existing age determining systems are known to use this lag as the sole factor for determining the age of the lambda sensor 109. A switch type lambda sensor 109 will generate a value between a binary range, that is between one of a 200mV and 800mV. The sampling time is usually fixed at a duration of 10 milliseconds. Therefore, the lambda sensor 109 will generate a signal voltage at one of the two values mentioned before every 10 milliseconds (ms). The ECU or the control unit 104 thus is configured to react to the change in the fuel pulse width at this same frequency. The estimated value of the lambda sensor output for a next engine cycle, as generated by the lambda sensor output estimation module 104d, is also generated at this same rate throughout the current single engine cycle. Thus, Figure 4 shows different engine operating points scattered over all the vehicle operable zones and at all engine operating points, the exhaust status and the sensor parameters are determined for 3 differently aged sensors- a new, intermediate and old sensor. The engine operating points are defined by different combinations of throttle position and engine speed.
Figure 5 is an exemplary graphical representation of an estimation error between the estimated exhaust status and the received exhaust status for a corresponding engine cycle. An estimation error is therefore measured, which is one of the parameters of the numerical computation technique used to the determine the lambda sensor gain factor at step 305 in figure 3. This is required as the aging of the lambda sensor 109 cannot usually be defined in terms of usage or actual duration of operation. There are various reasons why a lambda sensor 109 might age. The reactive metal on the sensor 109, usually zirconium (Zr), might go through oxidation and reduction reactions in the exhaust channel 103, or oil and grease in the exhaust channel 103 might get deposited on the surface of the reactive metal. It is also possible that the sensor may get physically damaged due to the vehicle usage. The failures of the sensor 109 generally result in various anomalies of the output of the sensor 109. The output may shift towards rich state, or the designated upper and lower limits of the output voltage of the sensor can change, and also a delay might get introduced in the response time of the sensor 109, due to these damages to the sensor 109 as mentioned above. All these factors contribute towards aging of the sensor 109, which might get aged when a vehicle has covered a distance of 50,000 kms, or it might well as good as a new sensor even when the vehicle has covered a distance of 150,000 kms. Thus, it is vital to determine the age of the sensor 109 in order to ensure that the air fuel ratio in the combustion chamber of the power unit 100a remains as close to the stoichiometric ratio as possible. The error in the estimation of the lambda sensor output, as generated by the lambda sensor output estimation module 104d is therefore vital to the determination of the gain factor of the lambda sensor 109. In the graph, the estimated and the received values converge to make the estimation error zero for the all three kinds of differently aged lambda sensors with different gain factors as determined by the numerical computation technique. For example, the gain factor is determined to be 0.3, 0.22 and 0.18 for fresh, intermediate aged, and old lambda sensor, respectively. An observation from this graph is that recursive least square estimation technique to determine the gain factor is an adaptive filter algorithm that minimizes the least square error and update the parameters recursively. This method is proved to be computational efficient for automotive applications with a computation time of for example, 6 microseconds per observation to publish the estimated gain factor.
Figure 6 is an exemplary graphical representation of the lambda sensor gain factor determined over a period of time and stored in the storage module 104b of the control unit 104. In this exemplary illustration, the lambda sensor gain factor is shown against corresponding values of normalised engine speed and normalised throttle opening. As shown in the figure, a new lambda sensor 109 will have a much higher gain factor compared to a lambda sensor 109 of an intermediate age. The gain factor of the aged (old) lambda sensor 109 as shown in the figure is much less than that of an intermediately aged lambda sensor 109. Sensor gain factor varies non-linearly over the age. As seen the figure, difference between new and intermediate sensor gain factor is much higher than the difference between intermediate and old sensor gain factor. Sensor gain factor values are distributed in engine operating point plan. It is evident that the gain is highly dependent on the sensor age only because, variation with respect to engine operating point is very less. In the recursive least square estimation technique, the root mean square error will further reduce as the number of observations will increase. The sensor gain will converge more distinctly as the number of samples increases. The classification module 104f will classify the lambda sensor as new, intermediate or old, after every predetermined number of engine cycles, depending on the value of the lambda sensor gain factor, depending on which range it falls into. In an embodiment, the classifying module can consider the stored values of the gain factor for a predetermined number of engine cycles.
The claimed invention as disclosed above is not routine, conventional or well understood in the art, as the claimed aspects enable the following solutions to the existing problems in conventional technologies. Specifically, this is an onboard control unit implementable method for real time estimation of sensor gain factor which lead us to objective determination of degradation in the lambda sensor. The determined value of gain factor is validated with three differently aged lambda sensors and mapped to objectively diagnose the health of the sensor with less computationally intensive computations by the control unit of the vehicle. Also, the system recursively takes the measured data and updates the gain factor. Thus, a real-time diagnosis is possible through the system. Additionally, the system takes into account the exhaust status, and the one or more sensor parameters, making the system more robust. As a result, the accuracy of monitoring of the age of the lambda sensor using the system is enhanced vis-à-vis the conventional systems. Moreover, the system is capable of quantifying age or state of health of the lambda sensor and alerting the user of the vehicle about a need for replacement or servicing. Suitable classification or regression method can be utilized for real time monitoring of the lambda sensor health. It is possible on-board vehicle ECU or the algorithm can be deployed over the cloud server also. The input data for the algorithm can be posted continuously from the vehicle via any communication channel. A computationally less intensive simplified reduced order mathematical sensor model is employed to replicate the behaviour of the lambda sensor to estimate the exhaust status. The method is estimating the gain factor online the vehicle with good convergence and accuracy.

 
List of Reference Signs:

100 – system
100a – power unit
101 – throttle body
102 – engine with combustion chamber
103 – exhaust channel
104 – Control unit
104a – Input output module of the control unit
104b – Storage module of the control unit
104c – processing module of the control Unit
104d – lambda sensor output estimation module of the processing unit
104e – calculation module of the processing unit
104f – classification module of the processing unit
105 – Throttle position sensor
106 – temperature manifold absolute pressure sensor
107 – spark plug disposed on the engine
108 – fuel injection system for the engine with a fuel pulse monitor
109 – lambda sensor disposed in the exhaust channel
110 – diagnostics unit
, C , Claims:We Claim:
1. A system (100) for monitoring an age of a lambda sensor (109) disposed in an exhaust channel (103) of a power unit (100a), said system (100) comprising:
said lambda sensor (109) disposed in said exhaust channel (103);
one or more sensors (105, 106, 108) coupled to said power unit (100a), said one or more sensors (105, 106, 108) being configured to generate one or more sensor parameters of said power unit (100a); and
a control unit (104), said control unit (104) being operably coupled to said lambda sensor (109) and said one or more sensors (105, 106, 108), said control unit (104) being configured to:
receive an exhaust status from said lambda sensor (109) and said one or more generated parameters from said one or more sensors (105, 106, 108), and
determine said age of said lambda sensor (109) using the received exhaust status and the one or more generated sensor parameters based on a computation technique, said computation technique being adapted to estimate a gain factor of said lambda sensor (109) based on said received exhaust status and said one or more generated sensor parameters.

2. The system (100) as claimed in claim 1, wherein said control unit (104) being configured to determine said age of said lambda sensor (109) using said received exhaust status and said one or more generated sensor parameters based on said computation technique by:
generating an estimated exhaust status of said lambda sensor (109) using said one or more sensor parameters from a current engine cycle by a sensor model, and
determining said gain factor using said computation technique to reduce an error between said estimated exhaust status and said received status of said lambda sensor (109).

3. The system (100) as claimed in claim 2, wherein said sensor model being stored in the control unit (104) and said sensor model being a mathematical model of said lambda sensor (109) defined as rate of change of said exhaust status of said lambda sensor (109) being equal to a product of said gain factor and an incremental change in at least of said one or more sensor parameters, wherein said at least one of said one or more sensor parameters being a fuel pulse width.

4. The system (100) as claimed in claim 1, wherein said control unit (104) being configured to classify said lambda sensor (109) as one of a fresh condition, an intermediate condition, and an aged condition, based on said gain factor using a classification module (104f).

5. The system (100) as claimed in claim 4, comprising a diagnostics unit (110) communicably coupled to said control unit (104), said diagnostics unit (110) being configured to generate an audio-visual indication to alert a user of a vehicle of said age of said lambda sensor (109), said audio visual indication being generated upon at least one of determined aged condition of said lambda sensor (109) and on reception of a user generated request signal.

6. The system (100) as claimed in claim 1, wherein said control unit (104) being configured to generate a fuel pulse width for injecting a fuel into a combustion chamber (102) of said power unit (100a), said fuel pulse width being determined based upon said exhaust status of said lambda sensor (109) and one or more sensor parameters from a throttle position sensor (105) and a temperature manifold absolute pressure sensor (106).

7. The system (100) as claimed in claim 1, wherein said computation technique being a recursive least square estimation technique with a Kalman filter module.

8. The system (100) as claimed in claim 1, wherein said lambda sensor (109) being disposed upstream of a catalytic convertor in said exhaust channel (103).

9. The system (100) as claimed in claim 1, wherein said control unit (104) being configured to store said generated gain factor of said lambda sensor (109) and to classify said stored gain factor of said lambda sensor (109), after every predetermined number of engine cycles, to determine said age of said lambda sensor (109), said storage being local or remote.

10. The system (100) as claimed in claim 1, wherein said control unit (104) being configured to pre-process said received exhaust status from said lambda sensor (109) and said one or more generated parameters from said one or more sensors (105, 106, 108) to prevent high frequency gain switching.

11. The system (100) as claimed in claim 1, wherein said one or more sensor parameters comprising a throttle position, a temperature manifold absolute pressure, engine speed, and a fuel pulse width and said one or more sensors (105, 106, 108) being a throttle position sensor (TPS) (105), and a Temperature Manifold Absolute Pressure (TMAP) sensor (106).

12. The system (100) as claimed in claim 1, wherein said control unit (104) comprising an input-output module (104a), a storage module (104b), a lambda sensor output estimation module (104d), a calculation module (104e) implementing the computation technique, and a classification module (104f) for determining said age of said lambda sensor (109) using said received exhaust status and said one or more generated sensor parameters based on said computation technique

13. A method (300) for monitoring an age of a lambda sensor (109) disposed in an exhaust channel (103) of a power unit (100a), said method (300) comprising the steps of:
generating, by a lambda sensor (109), an exhaust status of said power unit (100a);
generating, by one or more sensors (105, 106, 108), sensor parameters corresponding to said power unit (100a);
receiving, by a control unit (104), said exhaust status and said one or more sensor parameters;
generating, by said control unit (104), an estimated exhaust status using said received exhaust status and said one or more sensor parameters from a current engine cycle;
determining, by said control unit (104), a gain factor by comparing said estimated exhaust status and said received exhaust status using a computation technique; and
determining, by said control unit (104), an age of said lambda sensor (109) based on said determined gain factor.

14. The method (300) as claimed in claim 13, wherein determining said age of said lambda sensor (109) based on said determined gain factor comprising classifying said age of said lambda sensor (109) as one of a fresh condition, an intermediate condition, an aged condition, based on said gain factor using a classification module (104f), wherein classifying said age comprises classifying said determined gain factor, by said control unit (104), after every predetermined number of engine cycles, to determine said age of said lambda sensor (109).

15. The method (300) as claimed in claim 13, comprising pre-processing said received exhaust status from said lambda sensor (109) and said one or more generated sensor parameters from said one or more sensors (105, 106, 108) to prevent high frequency gain switching.

16. The method (300) as claimed in claim 13, wherein determining said gain factor using said computation technique comprises reducing an error between said estimated exhaust status and said received status of said lambda sensor (109), where said estimated exhaust status of said lambda sensor (109) being generated by a sensor model using said one or more sensor parameters from said current engine cycle.

17. The method (300) as claimed in claim 14, comprising generating an audio-visual indication, by a diagnostics unit (110), to alert a user of a vehicle of said age of said lambda sensor (109), said audio visual indication being generated upon at least one of a determined aged condition of said lambda sensor (109) and on reception of a user generated request signal.

18. The method (300) as claimed in claim 13, comprising generating, by said control unit (104), a fuel pulse width for injecting a fuel into a combustion chamber of said power unit (100a), said fuel pulse width being determined based upon said exhaust status of said lambda sensor (109) and one or more sensor parameters from a throttle position sensor (105) and a temperature manifold absolute pressure sensor (106).