Method and controller for measuring a temperature of a gas flow of an internal combustion engine

Abstract

A method of measuring a temperature of a gas flow of an internal combustion engine comprises obtaining a sensor temperature measurement of a gas flow of the internal combustion engine using a gas flow temperature sensor; obtaining an estimated temperature of the gas flow based on one or more operating conditions of the internal combustion engine; calculating a filtered estimated temperature based on one or more operating conditions of the internal combustion engine and a low pass filter; determining an adjusted temperature measurement of the gas flow based on the estimated temperature and a comparison of the filtered estimated temperature to the sensor temperature measurement, and outputting the adjusted temperature measurement of the gas flow. The estimated temperature may be based on an operating condition of one or more of: and exhaust gas recirculation (EGR) system, a turbocharger, an aftertreatment system, a fresh air intake, a charge air cooler.

Description

Field of the disclosure The present disclosure relates to internal combustion engines. In particular, the present disclosure relates to the sensing and measurement of a temperature of the internal combustion engine. Background The temperature of air which is used for combustion in an internal combustion engine may affect the performance / operation of the internal combustion engine. The temperature of the air at the intake to the internal combustion engine may vary considerably depending on e.g. ambient air temperature, exhaust gas recirculation, turbocharger operation and the like. In order to improve the performance of the internal combustion engine, and / or to have improved emissions performance, an engine control system of an internal combustion engine may control one or more setpoints of the internal combustion engine (e.g. airflow, fuel quantity etc) based on the temperature of the air at the intake. In some internal combustion engines, the temperature of the air at the intake is measured using an Intake Manifold Air Temperature (IMAT) sensor. This sensor is typically located at a point along the intake manifold which is downstream of any exhaust gas recirculation or turbocharger systems. Against this background, the present disclosure seeks to provide an improved, or at least commercially relevant alternative method and controller for measuring a temperature of a gas flow of an internal combustion engine. Summary According to a first aspect of the disclosure, a method of measuring a temperature of a gas flow of an internal combustion engine is provided. The method comprises: obtaining a sensor temperature measurement of a gas flow of the internal combustion engine using a gas flow temperature sensor of the internal combustion engine; obtaining an estimated temperature of the gas flow based on one or more operating conditions of the internal combustion engine; calculating a filtered estimated temperature based on one or more operating conditions of the internal combustion engine and a low pass filter; determining an adjusted temperature measurement of the gas flow based on the estimated temperature and a comparison of the filtered estimated temperature to the sensor temperature measurement; and outputting the adjusted temperature measurement of the gas flow. The present inventors have realised that when a temperature of a gas flow of an internal combustion engine changes, e.g. due to a relative rapid change in the output power of the internal combustion engine, there may be lag between the change in the gas flow temperature and a temperature measured by a sensor configured to measure the gas flow temperature (a gas flow temperature sensor). As gas flow temperature measurements may be used by an internal combustion engine controller in order to determine operating conditions for the internal combustion engine (e.g. fuel quantity injected, engine speed etc), it is important that the gas flow temperature measurements are accurate. By improving the accuracy of the gas flow temperature measurement, the emissions performance of the internal combustion engine may also be improved. Accordingly, the method of the first aspect aims to calculate an adjusted temperature measurement for a sensor temperature measurement of a gas flow of an internal combustion engine. The adjusted temperature measurement may be calculated, at least in part, based on an estimated temperature of the gas flow, wherein the estimated temperature is determined based on one or more operating conditions of the internal combustion engine. As such, the estimated temperature is determined without reference to the sensor temperature measurement of the gas flow. As the estimated temperature is based on one or more operating conditions of the internal combustion engine, the estimated temperature may respond more quickly to change in the gas flow temperature than the sensor temperature measurement. The present inventors have also realised that the estimated temperature of the internal combustion engine may not capture any differences between the actual performance of the internal combustion engine and the expected behaviour used to calculate the estimated temperature. To address this, the method of the first aspect calculates a filtered estimated temperature based on one or more operating conditions of the internal combustion engine and a low pass filter. As such, the filtered estimated temperature is smoothed, slower responding, version of the estimated temperature. The filtered estimated temperature can be compared to the sensor temperature measurement in order to provide an indication of any differences between the actual performance of the internal combustion engine and the expected behaviour used to calculate the estimated temperature (in particular in steady state conditions). Thus, the method of the first aspect determines an adjusted temperature measurement of the gas flow based on the estimated temperature (which can respond to short-term changes in gas flow temperature) and the comparison of the filtered estimated temperature to the sensor temperature measurement (which in particular can account for any differences in steady state behaviour). Brief discussion of the figures Embodiments of this disclosure will now be described with refence to the following figures in which: Fig. 1 is a block diagram of internal combustion engine controller according to an embodiment of the disclosure; Fig. 2 is a set of graphs showing the change in IMAP and I MAT on an internal combustion engine; Fig. 3 is a block diagram of a method according to this disclosure; Fig. 4 is a set of graphs showing the iterative calculation of the trim parameter; Fig. 5 is a Simulink (RTM) block model of a controller used to implement the method; Figs 6a, 6b and 6c are graphs showing the behaviour sensor temperature measurement, estimated temperature, filtered estimated temperature and trim parameter in response to changing operating conditions of the internal combustion engine; Fig. 7 is a block diagram implementation of a mixed gas equation according to this disclosure; and Figs 8a, 8b and 8c are further graphs showing the behaviour sensor temperature measurement, estimated temperature, filtered estimated temperature and trim parameter in response to changing operating conditions of the internal combustion engine. Detailed description The present disclosure relates to a method and system for measuring a temperature of a gas flow of an internal combustion engine. The following description discusses embodiments of the disclosure with reference to the measurement of a gas flow temperature in an intake manifold of an internal combustion engine. Of course, it will be appreciated that the present disclosure is not limited to the measurement of intake manifold gas flow temperatures. According to an embodiment of the disclosure, an internal combustion engine 1 is provided. The internal combustion engine 1 comprises an intake manifold 10 which is configured to distribute air to the cylinders (not shown) of the internal combustion engine 1. The intake manifold may receive gas flow from one or more sources. For example, in the embodiment of Fig. 1, the intake manifold may receive fresh air which may have been compressed by turbocharger (not shown). The intake manifold may receive exhaust gas which has been recirculated via an exhaust gas recirculation (EGR) circuit (not shown). In some embodiments, a charge air cooler (not shown) may also be provided between the compressor and intake manifold. As such, air in the intake manifold 10 may be composed of fresh air inducted through the compressor mixed with exhaust gas from the exhaust gas recirculation (EGR) circuit. It will be appreciated that a temperature of the gas output from the turbocharger compressor (Turbocharger compressor outlet air temperature) may be dependent on the amount of work done by the turbocharger and thus rises and falls with engine load. The presence of a charge air cooler between the compressor and the intake manifold 10 may dampen the impact of compressor outlet air temperature fluctuations on intake manifold air temperature (IMAT) to an extent, but fluctuation may still be significant. Internal combustion engines 1 without aftercooling may have a greater variation in IMAT than those with aftercooling. The rate of change of temperature seen by the physical IMAT sensor is limited by the transfer function of the sensor itself, the low pass filter for the ECM input and heat soaking of the metal around the sensor installation. A temperature sensor 20 may be located in, or proximate to, an outer wall of the intake manifold 10. The temperature sensor 20 may be configured to sense the IMAT and to output a reading indicative of the IMAT to a controller (not shown) of the internal combustion engine 1. Fig. 2 shows some example data from a block load on a C2.8 T DOC engine. Note that intake manifold pressure (IMAP) rises quickly but the IMAT sensor reading takes almost 5 minutes to settle at its final value. The IMAT measurement may be used by a controller (not shown) of the internal combustion engine 1 for the calculation of various set points for the internal combustion engine 1. In particular, the IMAT value may be used, at least in part, to determine one or more of desired IMAP, EGR, and the Fuel Air Ratio Control (FARC) fuel limit which dictates the maximum amount of fuel that can be injected into the combustion chamber based on the intake manifold air density (IMAD). As such, a controller using a delayed IMAT signal due to an IMAT measurement which is relatively slow to respond compared to real changes in air temperature may suffer from at least the issues listed in Table 1. Engine Conditions Measured IMAT Calculated IMAD FARC Fuel Limit Result Running with low load after significant period at high load > Actual < Actual < Required Poor transient response (e.g. stall during block load) Increasing load from cold initial condition < Actual > Actual > Required Excessive engine out soot Table 1 Embodiments of this disclosure aim to mitigates these issues by estimating the true IMAT more accurately than the raw data output by the temperature sensor 20. According to an embodiment of this disclosure, a method 100 of measuring a temperature of a gas flow of an internal combustion engine 1 is provided. A block diagram of the method 100 is shown in Fig. 3. According to this embodiment, the temperature to be measured is the temperature of gas flowing through the intake manifold 10 (IMAT). A Simulink (RTM) block model of a controller used to implement the method 100 is shown in Fig. 4. As shown in Fig. 3, method 100 comprises a step 101 of obtaining a sensor temperature measurement of a gas flow of the internal combustion engine 1 using a gas flow temperature sensor of the internal combustion engine. According to the embodiment, the temperature sensor 20 may be used to obtain a raw IMAT measurement. Next, the method 100 comprises a step 102 of obtaining an estimated temperature of the gas flow based on one or more operating conditions of the internal combustion engine. As such, the method comprises obtaining an estimated IMAT based on one or more operating conditions of the internal combustion engine. For example, in some embodiments, IMAT may be estimated based on a lookup table with suitably chosen axes calibrated on a nominal engine. For some embodiments, in particular where the internal combustion engine comprises an EGR, such an approach may not fully account for variations in the performance of EGR system (due to EGR cooling fouling for example). As such, in some embodiments this may be addressed by instead storing the turbocharger compressor outlet gas flow temperature (TempCompOut) in a lookup table and using Equation 1 to calculate IMAT. EMAF x TempEGR + FMAF x TempCompOut Estimated IMAT =-------------------------------------- (1) TMAF k } According to Equation 1, EMAF is the EGR mass flow, TempEGR is the EGR temperature, FMAF is the fresh air mass flow, and TMAF is the total air mass flow. The parameter TempCompOut may be obtained from a lookup table which is calibrated for the internal combustion engine 1. As such, in some embodiments, the estimated IMAT may be calculated using a mixed gas equation including an estimate of the turbocharger compressor outlet gas flow temperature, a total air mass flow rate, an EGR gas flow temperature, and a fresh air intake mass flow rate of the internal combustion engine 1. The lookup table may use one or more operating conditions of the internal combustion engine in order to calculate TempCompOut. For example, according to the present embodiment, the compressor out air temperature (turbocharger compressor outlet gas flow temperature) may be calculated based on the engine speed, fuel quantity, ambient temperature, and the barometric temperature. In other embodiments, different combinations of parameters may be used to provide a suitable lookup table. The skilled person will appreciate that the lookup table / calculation used may depend on the internal combustion engine 1 and its application, as well as the available operating condition data. As such, one or more operating conditions of the internal combustion engine 1 may be used to obtain the estimated I MAT. While the method 100 is concerned with the measurement of IMAT, the present disclosure is not limited to the measurement of IMAT. As such, in some embodiments, the estimated temperature of the gas flow may be calculated based on an operating condition of one or more of: an exhaust gas recirculation (EGR) system, a turbocharger, an aftertreatment system, a fresh air intake, a charge air cooler. For example, the skilled person may utilise a suitable mixed gas equation, or other calculation to determine an estimated temperature. For example, the estimated temperature may be calculated using a mixed gas equation combining parameters indicative of two or more gas flows selected from the group comprising: a fresh air intake gas flow, an EGR gas flow, a turbocharger compressor gas flow, a turbocharger turbine gas flow, an aftertreatment gas flow, and a total internal combustion engine gas flow, and a charge air flow. In the mixed gas equation of Equation 1, a look-up table is used to calculate the compressor out air temperature. It will be appreciated that the present disclosure is not limited to the calculation of the compressor out air temperature using a look-up table. As such, in some embodiments one or more operating conditions of the internal combustion engine used to calculate the estimated temperature may be determined using a respective look up table. To account for additional system variability (both engine-to-engine and other factors not covered by compressor temperature map inputs) a trim may be applied based on the temperature sensor 20 reading. The trim should result in an adjusted IMAT (i.e. an adjusted temperature measurement) converging with the raw IMAT (output by the temperature sensor 20) in steady-state conditions. Thus, in some embodiments, the adjusted IMAT may be given by Equation 2: Adjusted IMAT = Estimated IMAT x Trim (2) As such, in some embodiments, determining the adjusted temperature IMAT based on the estimated IMAT and a comparison of the filtered estimated IMAT and the raw IMAT may comprise calculating a trim parameter based on a comparison of the filtered estimated IMAT to the raw IMAT. The adjusted IMAT may then be determined by adjusting the estimated IMAT based on the trim parameter. While the trim parameter could be calculated based on the difference between estimated IMAT and raw IMAT when the engine is running in steady-state conditions state (i.e. constant speed, torque, massflow) and raw IMAT is stabilised at its final steady-state value. However, in reality these conditions may be rarely fulfilled, resulting in a low update frequency and hence the possibility of a trim parameter which becomes outdated as conditions change (e.g. temperatures and air flows). While the trim parameter could be mapped as a function of some other variable in order to mitigate for this, the engine would need to run in steady-state conditions at the extents of the range of the independent variable which may not be possible. Thus, according to the present embodiment, an approach is employed whereby the estimated IMAT is filtered to match the response of the IMAT sensor on the calibration engine and this filtered estimated IMAT (i.e. a ‘slow’ estimated IMAT) is compared to the raw IMAT reading from the temperature sensor. As such, the method 100 further comprises a step 103 of calculating a filtered estimated temperature (filtered estimated IMAT) based on one or more operating conditions of the internal combustion engine and a low pass filter. In some embodiments, the filtered estimated temperature may be calculated based on the estimated temperature and a low pass filter. In the embodiment of method 100, calculating the filtered estimated IMAT may comprise calculating a filtered estimate of the turbocharger compressor outlet gas flow temperature based on the estimate of the turbocharger compressor outlet gas flow temperature and a low pass filter. The filtered estimated IMAT may then be calculated using a mixed gas equation including the filtered estimate of the turbocharger compressor outlet gas flow temperature, a total air mass flow rate, an EGR gas flow temperature, and a fresh air intake mass flow rate of the internal combustion engine. As such, the filtered estimated compressor temperature may be fed into a second instance of the mixed gas model (Equation 1). In this embodiment, placing the low pass filter upstream of the mixed gas model prevents double-filtering of the EGR temperature signal (which is already subject to its sensor transfer function etc). Also, an alternate form of Equation 1 may be used to allow the FMAF:TMAF ratio to be additionally filtered, reflecting the fact these mass flow estimations change very rapidly. The calculation of the filtered estimated IMAT is discussed in more detail below. Step 104 subsequently comprises determining an adjusted temperature measurement (adjusted IMAT) of the gas flow based on the estimated temperature and a comparison of the filtered estimated temperature to the sensor temperature measurement. As such, the comparison of the filtered estimated IMAT to the raw IMAT accounts for any error between the filtered estimated IMAT and the sensed raw IMAT. The key advantage of this approach is that engine stability is not required. By performing a real-time comparison between the filtered estimated IMAT and the sensed raw IMAT, a real-time indication (e.g. a trim parameter) of any difference between the theoretical and actual operation of the internal combustion engine 1 may be obtained. This negates the need to use a map or save the trim value(s) to non-volatile memory. As such, the risk of an engine running for long periods with an outdated trim parameter is diminished. In some embodiments, the trim parameter may be calculated via a recursive least squares (RLS) algorithm. Fig. 5 shows a plurality of graphs depicting graphically the calculation of the trim parameter. The aim of the RLS algorithm is to find the trim parameter which minimizes the total squared error between raw IMAT and filtered estimated IMAT. This is analogous to drawing a line of best fit through the collected datapoints plotted on a x-y graph of filtered estimated IMAT (slow IMAT estimate cp) against raw IMAT (y), and then using the gradient as the trim parameter. In this embodiment, for simplicity there is no offset term in the trim. The recursive architecture of the algorithm means that the trim may be adjusted each time a new datapoint is collected rather than being recalculated from scratch. Only the previous trim and the current measurement are needed. In step 105, the method 100 comprises outputting the adjust temperature measurement of the gas flow. That is, the adjusted IMAT value is used by the internal combustion engine rather than the raw IMAT value output by the temperature sensor 20. Thus, a method 100 may be provided for measuring IMAT using a temperature sensor 20. To further illustrate the method 100, Fig. 6a shows a graph of raw IMAT (sensed temperature), estimated IMAT (Raw Estimated temperature) and adjusted IMAT (Final Estimated Temperature). It will be appreciated from Fig. 6a that the adjusted IMAT value responds more quickly to a change (increase) in temperature than the raw IMAT from the temperature sensor 20. Further, after a period of time, the adjusted IMAT value tends towards the raw IMAT value output by the temperature sensor. Further, Fig. 6b shows a graph of the raw IMAT value (sensed temperature) and the filtered estimated IMAT (slow estimated temperature). Fig. 6c shows a graph for the change in trim parameter over time. Similar to Figs. 6a, 6b, and 6c, Figs. 8a, 8b, and 8c respectively show the response of the system when undergoing a decrease in temperature. As discussed above, in some embodiments estimated IMAT may be obtained based on a compressor out temperature value which is obtained from a map (e.g. a lookup table). The map may comprise inputs for engine speed, fuel quantity and barometric pressure. In some embodiments, the map output may be filtered by a 1st order (low pass) filter to reflect the fact changes in temperature are not instantaneous. The look-up table may be calibrated based on testbed measurements of the internal combustion engine 1, or any other suitable calibration method known to the skilled person. In some embodiments, the estimated IMAT value may be low pass filtered. In some embodiments, the estimated compressor temperature out value may be low pass filtered, wherein the filtered estimated compressor temperature out it in turn used to in the mixed gas equation to calculate the filtered estimated IMAT. The estimated compressor temperature out value output from the associated map may be filtered by a 1st order (low pass) filter to reflect the fact changes in temperature are not instantaneous. The look-up table may be calibrated based on testbed measurements of the internal combustion engine 1, or any other suitable calibration method known to the skilled person. In some embodiments, the low-pass filter used may be a first order filter having a constant filter factor. Preferably, the filtered estimated IMAT value calculated is a good match to data output by an ideal temperature sensor 20 on a calibration engine. As such, the performance of the low pass filter for the compressor out temperature is therefore important for calculating the trim parameter. In some embodiments, a first-order filter may be used since its behaviour is well-understood. However, known non-linear behaviour may be further accounted for with a mapped filter factor (i.e. a control map for the filter factor). As such, in some embodiments the low pass filter used to calculate the filtered estimated temperature comprises a filter factor, wherein the filter factor is determined from a low pass filter control map based on one or more operating conditions of the internal combustion engine. In order to determine a filter factor control map, empirical analysis of testbed block load data (with raw IMAT values and estimated IMAT included) for the calibration engine may be used. It will be appreciated that the precise filter factor control map values may vary for different internal combustion engine designed. For example, the filter factor may be determined from the low pass filter control map based on one or more of: the estimate of the turbocharger compressor outlet gas flow temperature, the engine speed, the rate of change of the estimate of the turbocharger compressor outlet gas flow temperature, and the sensor temperature measurement. It will be appreciated that the value of the filter factor in the low pass filter control map may depend on the design of the internal combustion engine. In general, the low pass filter control map included the following general data trends. For changes in the direction temperature change (i.e. where the sensor cools more slowly than it heats), it was found that a larger filter factor required (less filtering) is required for rising temperatures. It was found that a smaller filter factor may be required (more filtering) for falling temperatures. For engine speeds it was found that the temperature sensor 20 may respond quicker for higher engine speeds. Accordingly, a larger filter factor required (less filtering) may be applied for higher engine speeds, and a smaller filter factor (more filtering) may be applied for lower engine speeds. When considering the rate of change of the estimated compressor temperature out, it was found that the initial impulse response may be faster than a subsequent convergence to the target temperature. Accordingly, a larger filter factor may be required (less filtering) for relatively high rates of change, and a smaller filter factor (more filtering) may be used to slower rates of change. In some embodiments, the absolute rate of change (ROC) of the estimated compressor temperature out may be calculated via a filtered derivative and this may be placed in series with another derivative block such that the double derivative is also calculated to give a signal which can be thought of as the acceleration of temperature. The final metric may be the speed of temperature change plus 2x the acceleration of the temperature change. This effectively may be used as a basic method for manipulating the shape of the sensor transfer function to match the real sensor. The present design may be implemented in a straightforward manner with standard Simulink blocks. The skilled person will appreciate that alternative filter designs which may have a suitable shape in their base format may also be used. Thus, the filtered estimated IMAT may be calculated according to embodiments of this disclosure. In some embodiments, one or more of the other parameters indicative of an operating condition of the internal combustion engine 1 may also be subject to low pass filtering. For example, upfront low pass filtering of the TMAF and FMAF input signals may be performed as these are signals may have a relatively fast response, while real air temperature change may be expected to be much slower. To incorporate the low pass filtering of the TMAF and FMAF input signals, in some embodiments the mixed gas equation (e.g. Equation 1) may be rearranged to Equation 3 as shown below. This formulation of the equation allows the FMAF:TMAF ratio can be filtered as shown in Fig. 7. As shown in Fig. 7, a slow Mass Air Flow (MAF) filter factor may be applied to the low pass filter for the ratio FMAF:TMAF to tune the amount of filtering provided when calculating estimated IMAT (vlMAT_SlowEstlMAT). Estimated!MAT = FMAF x TempCompOut + fl — FMAF\ x TempEGR (3) As shown in Fig. 4, the trim parameter may be calculated based on a comparison of the filtered estimated temperature to the sensor temperature measurement using a recursive least squares function. The intent of the trim parameter is to mitigate for errors in the estimated IMAT due to noise factors which result in a difference between the engine being controlled and the nominal engine (on which the compressor temp map was calibrated). Some of these noise factors will be slow-moving or even constant and others will change quickly as engine running conditions change. To account for these different noise sources, the recursive least squares function may include a forgetting factor and a random walk component (discussed in more detail below). Example sources of noise / variation in the expected operation of the internal combustion engine are compressor efficiency variation (e.g. due to compressor fouling), IMAT sensor accuracy variation (e.g. part to part), IMAT sensor coolant temperature sensitivity and changes in ambient temperature, and the like. To account for these noise sources, the trim parameter may be calculated via a recursive least squares (RLS) algorithm. The aim of this algorithm is to find the trim parameter, 0, which minimizes the total squared error, VN(0) - that is the sum of the square of the difference between the raw IMAT (y(t)) and the filtered estimated IMAT with trim correction applied (y(t|0)) averaged across the total number of timesteps, N. W)= ^?-i(y(t|0)-y(t))2   (4) The trim parameter is defined as a gain such that the raw IMAT from temperature sensor 20 y(t) is assumed to be equal to filtered estimated IMAT, <p(t) multiplied by a trim parameter 0(t). The trim parameter is an estimate of the real gain and so may be referred to as 0(t). As such, it follow that the raw IMAT is: y(0 = 0(t) x <p(t) (5) With each timestep the total number of datapoints, N, increments by 1. Therefore, with each timestep the optimal trim parameter 0(t) which gives the smallest total squared error is likely to change, however 0(f) needn’t be recalculated from scratch. The recursive algorithm allows the previous trim parameter to be adjusted in each timestep given the newest y(t) and y(t|0). Fig. 5 is a simplified visualisation showing how the line of best fit is iterated as new data is processed. With each timestep there is a new filtered IMAT estimate, <|>(t), and a new raw IMAT measurement, y(t), allowing a new datapoint to be plotted on the <|)-y plane, thus changing the equation for the best fit line, y, by adjustment of the estimated trim parameter, The trim parameter may be updated recursively by way of a Kalman filter. As such, the trim parameter for 0(f) may be calculated based on the previous trim parameter, 0(t - 1) , a Kalman gain, Kt, and the squared prediction error (y(t|0) - Calculation of the Kalman gain, Kt, is not detailed here but can be found under the mask of the RLS library in the model. In some embodiments, the Kalman filter implementation may also include a forgetting factor and random walk inputs. The base RLS algorithm may assume the system being modelled is a linear time-invariant (LTI) system, meaning the output of the system at any given time is a linear combination of its input and past outputs. However, in reality the trim parameter is accounting for a group of noise factors which are neither linear nor timeinvariant. The forgetting factor and random walk features may mitigate for this. In some embodiments, the forgetting factor, A, may be a scalar value between 0 and 1, where a value closer to 1 gives more weight to recent measurements, and a value closer to 0 gives more weight to older measurements. When A is set to 1, the RLS algorithm gives equal weight to all past measurements, resulting in a slower adaptation to changes in the system. By contrast, when A is set to a smaller value, the RLS algorithm gives less weight to older measurements, resulting in a faster adaptation to changes in the system. However, setting A too small can cause the algorithm to be sensitive to measurement noise and result in an unstable estimate. The forgetting factor may be used to prevent the RLS algorithm from becoming over-fitted to past data and to improve its ability to track changes in the system over time. It enables the algorithm to forget past measurements that are no longer relevant to the current estimate of the parameter and adapt to changes in the system more quickly. Calibration of the forgetting factor is a trade-off between convergence speed and stability. The random walk input may be added to the co-variance matrix which is used in the calculation of the Kalman gain, Kt. The intention of the random walk is to introduce random noise into the Kalman gain such that the RLS system is forced to keep refining its solution as new measurement data comes in. A larger value for the random walk disturbs the RLS system more and thus makes it more sensitive to new data. A smaller value for the random walk disturbs the RLS system less and so results in a more stable output. Calibration of the random walk involves a similar trade-off to calibration of the forgetting factor. As such, in some embodiments, the random walk may be defined by a singular tuneable value. In some embodiments, the random walk may value may be based on a change detection logic such that the co-variance matrix is perturbed more when a global shift in conditions is detected based on the theory that the optimal trim value should be relatively stable in steady-state conditions but could shift significantly for a large shift in speed / load. A change detection strategy could be achieved as part of the RLS sub-system itself (factorisation of the RLS function may be required) or alternatively the distance moved in the compressor temperature map space (e.g. Euclidean) could offer an indication that the RLS model may need shifting. In some embodiments, the trim parameter may be subject to some additional system constraints in order to improve the operation of the internal combustion engine, for example, in some embodiments, the trim parameter may be not exceed a maximum threshold value and / or may not go below a minimum threshold value. As such, the system may be prevented from deviating too far from an expected operational value for the adjusted temperature. In some embodiments, a rate of change limit may also be imposed on the trim parameter. This may provide an additional way to slow the trimming behaviour down. This rate of change limit should be calibrated considering the trade-offs described above in the discussion of the RLS random walk and forgetting factor. In some embodiments, the method 100 of measuring a temperature according to embodiments of this disclosure may be implemented by a controller. The controller may decide when to implement the method 100. That is, in a first operational mode, the controller may choose to rely on the raw IMAT value from the temperature sensor 20, while in a second operational mode, the controller may implement method 100 and rely on the adjusted IMAT value calculated accordingly. For example, the controller may determine the internal combustion engine is operating in a first operational mode when the internal combustion engine is undergoing a start-up routine. Once the internal combustion engine 1 has been operating for a predetermined duration, the controller may determine that the internal combustion engine is operating in a second operation mode and switch to using the adjusted IMAT value. The controller may perform a plurality of checks to distinguish between a first operational mode and a second operational mode. The operational checks may include comparing one or more of: sensor operational status, engine run time, engine speed, IMAT temperature sensor absolute value, coolant temperature, start-up idle hold duration and the like against a threshold or expected value. In general, where the internal combustion engine is operating outside of expected operational parameters, the controller may elect to rely on the raw IMAT value from the temperature sensor 20 (i.e. the first operational mode). Industrial applicability Embodiments of this disclosure may be used to measure a temperature of a gas flow of an internal combustion engine with improved accuracy and / or responsiveness. While the above examples have discussed the determination of an adjusted temperature for an intake manifold temperature sensor, it will be appreciated that the present disclosure is not limited to such examples. As such, embodiments of the disclosure may be used to determine an adjusted temperature for any temperature sensor of an internal combustion engine configured to measure a temperature of a gas flow. For example, embodiments of this disclosure may comprise methods of measuring a temperature of a gas flow in one or more of: an exhaust gas recirculation (EGR) system, a turbocharger, and an aftertreatment system. Embodiments according to this disclosure aim to calculate an adjusted temperature measurement for a sensor temperature measurement of a gas flow of an internal combustion engine. The adjusted temperature measurement may be calculated, at least in part, based on an estimated temperature of the gas flow, wherein the estimated temperature is determined based on one or more operating conditions of the internal combustion engine. As such, the estimated temperature is determined without reference to the sensor temperature measurement of the gas flow. As the estimated temperature is based on one or more operating conditions of the internal combustion engine, the estimated temperature may respond more quickly to change in the gas flow temperature than the sensor temperature measurement. To account for any difference between the estimated temperature of the internal combustion engine and the actual performance of the internal combustion engine (e.g. operational noise) embodiments according to this disclosure calculate a filtered estimated temperature based on one or more operating conditions of the internal combustion engine and a low pass filter. As such, the filtered estimated temperature is smoothed, slower responding, version of the estimated temperature. The filtered estimated temperature can be compared to the sensor temperature measurement in order to provide an indication of any differences between the actual performance of the internal combustion engine and the expected behaviour used to calculate the estimated temperature (in particular in steady state conditions). Thus, the embodiments of this disclosure determine an adjusted temperature measurement 5 of the gas flow based on the estimated temperature (which can respond to short-term changes in gas flow temperature) and the comparison of the filtered estimated temperature to the sensor temperature measurement (which in particular can account for any differences in steady state behaviour). 10

Claims

2522CLAIMS:

1. A method of measuring a temperature of a gas flow of an internal combustion engine comprising:5 obtaining a sensor temperature measurement of a gas flow of the internalcombustion engine using a gas flow temperature sensor of the internal combustion engine;obtaining an estimated temperature of the gas flow based on one or more operating conditions of the internal combustion engine;calculating a filtered estimated temperature based on one or more operating10 conditions of the internal combustion engine and a low pass filter;determining an adjusted temperature measurement of the gas flow based on the estimated temperature and a comparison of the filtered estimated temperature to the sensor temperature measurement; andoutputting the adjusted temperature measurement of the gas flow.

152. A method according to claim 1, whereindetermining an adjusted temperature measurement of the gas flow based on the estimated temperature and a comparison of the filtered estimated temperature and the sensor temperature measurement comprises:20 calculating a trim parameter based on a comparison of the filtered estimatedtemperature to the sensor temperature measurement, whereinthe adjusted gas flow temperature is determined by adjusting the estimated temperature based on the trim parameter.25 3. A method according to claim 1 or claim 2, whereinthe estimated temperature is calculated based on an operating condition of one or more of: an exhaust gas recirculation (EGR) system, a turbocharger, an aftertreatment system, a fresh air intake, a charge air cooler.30 4. A method according to any preceding claim, whereinthe estimated temperature is calculated using a mixed gas equation combining parameters indicative of two or more gas flows selected from the group comprising: a fresh air intake gas flow, an EGR gas flow, a turbocharger compressor gas flow, a turbocharger turbine gas flow, an aftertreatment gas flow, and a total internal combustion engine gas35 flow.10 03 255. A method according to any preceding claim, whereinone or more operating conditions of the internal combustion engine used to calculate the estimated temperature are determined using a respective look up table.

56. A method according to any preceding claim, whereinthe gas flow temperature sensor is an intake manifold air temperature sensor such that the sensor temperature measurement is an intake manifold temperature of the internal combustion engine.

107. A method according to claim 6, whereinthe estimated temperature is calculated based on one more operating conditions of the internal combustion engine comprising an estimate of a turbocharger compressor outlet gas flow temperature.

158. A method according to claim 6 or claim 7, whereinthe estimated temperature is calculated using a mixed gas equation including an estimate of a turbocharger compressor outlet gas flow temperature, a total air mass flow rate, an EGR gas flow temperature, and a fresh air intake mass flow rate of the internal 20 combustion engine.

9. A method according to claim 8, whereinthe filtered estimated temperature is calculated based on the estimated temperature and a low pass filter.2510. A method according to claim 8, whereina filtered estimate of the turbocharger compressor outlet gas flow temperature is calculated based on the estimate of the turbocharger compressor outlet gas flow temperature and a low pass filter, and30 the filtered estimated temperature is calculated using a mixed gas equationincluding the filtered estimate of the turbocharger compressor outlet gas flow temperature, a total air mass flow rate, an EGR gas flow temperature, and a fresh air intake mass flow rate of the internal combustion engine.10 03 25wherein low pass filter used to calculate the filtered estimated temperature comprises a filter factor, wherein the filter factor is determined from a low pass filter control map based on one or more operating conditions of the internal combustion engine.5 12 A method according to claim 11, whereinthe filter factor is determined from the low pass filter control map based on one or more of: the estimate of the turbocharger compressor outlet gas flow temperature, the engine speed, the rate of change of the estimate of the turbocharger compressor outlet gas flow temperature, and the sensor temperature measurement.1013 A method according to any of claims 8 to 12, whereinthe fresh air intake mass flow rate and / or the total air mass flow rate are low pass filtered before calculating the filtered estimated temperature and / or the estimated temperature.1514. A method according to any of claims 2 to 13 when dependent on claim 2, wherein the trim parameter is calculated using a recursive least squares algorithm to minimise the squared difference between the filtered estimated temperature and the sensor temperature measurement over N time steps, where N is a positive integer.2015. An internal combustion engine controller configured to measure a temperature of a gas flow of an internal combustion engine, the internal combustion engine controller configured to:receive a sensor temperature measurement of a gas flow of an internal combustion 25 engine from a gas flow temperature sensor of the internal combustion engine;receive one or more signals indicative of one or more operating conditions of the internal combustion engine;calculate an estimated temperature of the gas flow based on the one or more operating conditions of the internal combustion engine;30 calculate a filtered estimated temperature based on the one or more operatingconditions of the internal combustion engine and a low pass filter;determine an adjusted temperature measurement of the gas flow based on the estimated temperature and a comparison of the filtered estimated temperature to the sensor temperature measurement; and35 output the adjusted temperature measurement of the gas flow.

16. An internal combustion engine comprising:a gas flow temperature sensor configured to measure a temperature of a gas flow of the internal combustion engine; andan internal combustion engine controller according to claim 15.LDCM

Images