GB2598785A

GB2598785A - Monitoring system

Info

Publication number: GB2598785A
Application number: GB2014443.2A
Authority: GB
Inventors: Pavey Philip; Matteo Bianchi Gian; Windsor Christopher; Mcarthur Adrian; Bean Alex
Original assignee: Jaguar Land Rover Ltd
Current assignee: Jaguar Land Rover Ltd
Priority date: 2020-09-14
Filing date: 2020-09-14
Publication date: 2022-03-16
Anticipated expiration: 2040-09-14
Also published as: GB2598785B; GB202014443D0

Abstract

A monitoring system (400) for a vehicle. The monitoring system comprises input means configured to receive vehicle sensor data (402) indicative of a state of at least one aspect of the vehicle; means configured to: estimate, in dependence on the vehicle sensor data, at least one wheel centre action (404) for at least one wheel of the vehicle, derive, from the at least one wheel centre action, a set of component-level forces (408) for at least one mechanical component of the vehicle, wherein the component-level forces are derived from a plurality of digital twins (406) associated with the at least one mechanical component, and calculate, from the component-level forces, a fatigue damage value (410) for the at least one mechanical component; and output means configured to output a signal indicative of the fatigue damage value (412).

Description

MONITORING SYSTEM

TECHNICAL FIELD

The present disclosure relates to monitoring components of a vehicle. Aspects of the invention relate to a monitoring system for a vehicle, to a method for monitoring components within a vehicle, to a non-transitory computer readable medium comprising computer readable instructions, and to a vehicle.

BACKGROUND

It is known that, when designing a vehicle, it is necessary to consider safety standards, performance to weight ratios, and design and validation processes. With regards to structural integrity, the components and systems of a car or motorcycle have to comply with two basic requirements. These are durability during normal customer operation and the ability of components to withstand special abuse events without damage and without causing the impairment of function or safety. Assumptions from customer driving habits, vehicle usage, variable road characteristics, and environmental conditions must be made to determine load and structural durability requirements. Special events are loads which occur infrequently during regular common normal usage of the vehicle. They must not lead to residual deformation or act as a precursor of fatigue failure. Misuse events, by contrast, are permitted to lead to a previously defined and readily identifiable damage -with customer safety remaining paramount.

Structural durability has always been a significant requirement in the process of designing and validating a vehicle, making an important contribution to the functional design and configuration of drive train, suspension, and body components. By fully utilising the material properties and employing geometric optimization, components can be exploited to their full capacity, ensuring a good performance to weight ratio is maintained. To achieve high degrees of efficiency in the use of a component, it is useful to examine and verify the loads and forces experienced. Therefore, the load spectrum for each component, encountered during ordinary customer usage, may be determined as comprehensively as possible. Besides these service loads, abuse event loads which occur infrequently during regular and appropriate use of the vehicle should also be considered. Such loads should not lead to any impairment of function or to safety relevant damage.

The development process for a vehicle consists of multiple design validation phases. Starting with load assumptions, the first component designs may be analysed by multi body simulations and finite element analysis calculations. The first strength and durability tests are conducted on test rigs based on load assumptions and simulated load data. Influences of the manufacturing process and component strength may also be evaluated. Finally, vehicle tests with close to pre-production vehicle models are conducted.

The initial load assumptions may be derived from past experience such as previous products, and expected vehicle performance for example category, mass, wheelbase, centre of gravity. Then the design systems are tested against specific test procedures. The test procedures are based on expected customer utilisation which in turn is currently based on the vehicle category. Each category is associated with the specific surface mix with the objective of gathering the system demand created by that surface mix. Establishing the surface mix that accurately represents current customer usage is a complex and time-consuming task. As such, it is not carried out on a regular basis. In this way this data to which the vehicles are designed and tested may be somewhat out of date. In summary, there is currently no link between the actual usage of vehicles by real customers and the test procedures for designing new vehicles. This may be referred to as a lack of up-to-date customer correlation.

It is an aim of the present invention to address one or more of the disadvantages associated with the

prior art.

SUMMARY OF THE INVENTION

Aspects and embodiments of the invention provide a monitoring system for a vehicle, a method for monitoring components within a vehicle, a non-transitory computer readable medium comprising computer readable instructions, and a vehicle as claimed in the appended claims.

According to an aspect of the present invention there is provided a monitoring system for a vehicle, the monitoring system comprising: means configured to: determine, in dependence on vehicle sensor data indicative of a state of at least one aspect of the vehicle, at least one wheel centre action for at least one wheel of the vehicle, estimate, from the at least one wheel centre action, a set of component-level forces for at least one mechanical component of the vehicle, wherein the component-level forces are estimated using a plurality of digital twins associated with the vehicle, and derive, from the component-level forces, a fatigue damage value for the at least one mechanical component.

According to another aspect of the present invention there is provided a monitoring system for a vehicle, the monitoring system comprising: input means configured to receive vehicle sensor data indicative of a state of at least one aspect of the vehicle; means configured to: determine, in dependence on the vehicle sensor data, at least one wheel centre action for at least one wheel of the vehicle, estimate, from the at least one wheel centre action, a set of component-level forces for at least one mechanical component of the vehicle, wherein the component-level forces are estimated using data from a plurality of digital twins associated with the vehicle, and derive, from the component-level forces, a fatigue damage value for the at least one mechanical component; and output means configured to output a signal indicative of the fatigue damage value.

In an embodiment, there is provided a monitoring system as described above, wherein the monitoring system comprises an electronic processor and an electronic memory device, wherein the input means comprise the electronic processor having an electrical input for receiving the vehicle sensor data, and the output means comprise the electronic processor having an electronic output configured to output the signal indicative of the at least one fatigue value; wherein the means configured to determine, in dependence on the vehicle sensor data, at least one wheel centre action for at least one wheel of the vehicle, estimate, from the at least one wheel centre action, a set of component-level forces for at least one mechanical component of the vehicle, wherein the component-level forces are estimated using data from a plurality of digital twins associated with the vehicle, and derive, from the component-level forces, a fatigue damage value for the at least one mechanical component comprise the processor being configured to access the memory device and execute the instructions stored therein such that it is operable to determine, in dependence on the vehicle sensor data, at least one wheel centre action for at least one wheel of the vehicle, estimate, from the at least one wheel centre action, a set of component-level forces for at least one mechanical component of the vehicle, wherein the component-level forces are estimated using data from a plurality of digital twins associated with the vehicle, and calculate, from the component-level forces, a fatigue damage value for the at least one mechanical component.

The monitoring system provides for estimating damage caused to vehicle components by external inputs such as terrain and driver behaviour and further allows the classification of the damage caused by these external inputs to the vehicle. The monitoring system provides measurements of product performance for the monitored components based on actual user behaviour and usage of the vehicles. These measurements may then be used in a variety of advantageous manners. For example, if the fatigue damage value calculated by the monitoring system indicates that a component may need maintenance or replacement, an alert may be generated for the driver. The measurement data may also be fed back to the vehicle manufacturer to inform the design of future vehicle models. Presently, information on the conditions vehicles actually experience while in use may not be available to vehicle designers. Furthermore, the measurements may be used to adapt the behaviour of active systems within the vehicle to improve vehicle operation, for example, to enhance the lifetime of the component in question. The invention provides the capability to learn from customer usage of vehicles after they have left the manufacturer's control and to change the vehicle behaviours to improve the like of chassis components. Advantageously, the monitoring system does not require the installation of any additional hardware, such as sensors, on the vehicle, as it uses the sensors data from sensors already present.

In an embodiment, the monitoring system comprises determining means configured to determine the wheel centre action for the at least one wheel of the vehicle; estimating means configured to estimate the component-level forces for the at least one mechanical component; and deriving means to derive the at least one fatigue damage value.

The set of component level-forces may comprise a component-level force for each of the plurality of digital twins. In this way, the use of multiple digital twins is used to capture the potential variation in the component-level forces due to the manufacturing variability of the components of the vehicle Optionally, each of the set of component-level forces is used to derive a provisional fatigue damage value, and a final fatigue damage value is obtained from the plurality of provisional fatigue damage values. In this way, the digital twins allow the manufacturing variability of the mechanical component to be included in monitoring the fatigue damage of the components of the vehicle.

The plurality of digital twins may be generated by sampling a digital twin template at a plurality of points.

This is an efficient way of providing a plurality of digital twins as so include the manufacturing variability in the fatigue monitoring. The plurality of points may be sampled randomly. Stochastic analysis techniques may be used to carry out the sampling. Such techniques allow for a broad selection of samples to be taken.

The sensor data may relate to a specific journey on a specific road portion such that the fatigue damage value is associated with the specific road portion. This facilitates classifying road portions according to the fatigue damage associated with them.

Optionally, the monitoring system is configured to calculate at least one combined fatigue damage value for a combination of road portions. In this way, fatigue damage may be monitored for journeys, and it not limited to single road portions.

The monitoring system may be configured to define a road severity index for each road portion for which sensor data is received, wherein the road severity index is based on the fatigue damage value for the at least one component. The road severity index may be vehicle specific. Having data on road severity may be useful to manage fatigue damage within a vehicle.

Optionally, the monitoring system is configured to generate a warning if the combined fatigue value is above a predefined threshold. In this way, components suffering excessive fatigue damage may be monitored by the user, and potentially received maintenance or be replace if necessary.

The monitoring system may be configured to maintain a component database comprising fatigue damage values for a plurality of mechanical components of the vehicle. In this way, the mechanical components may be monitored, and their fatigue damage data is available for review, and updating as appropriate.

The monitoring system may be configured to maintain a road portion database comprising road severity index information for each road portion. In this way, the vehicle may adapt active control systems to suit conditions of a previously travelled road portion.

Optionally, the monitoring system is configured to derive a durability control signal from the fatigue damage value, and the output is configured to output the durability control signal to at least one control system of the vehicle. In this way, the control systems of the vehicle may adapt the operation of the vehicle in response to the fatigue damage. The control systems may include active control systems.

The durability control signal may be based on the road severity index. In this way, the vehicle may be controlled based on the vehicle's previous experience on that road portion.

The durability control signal may comprise an adaption enable component to control whether the at least one control system should apply an adaption, and a control level component to control the level of the adaptation. This is an efficient manner to manage adaption of control systems within a vehicle.

At least a part of the monitoring system may be located on the vehicle. At least a part of the monitoring system is located remote from the vehicle. In this way, the monitoring system may take advantage of cloud storage and cloud processing.

Optionally, the monitoring system is configured to consolidate the sensor data. In this way, the sensor data may be adapted for more efficient processing.

The wheel centre actions may comprise at least one of: wheel centre forces, wheel moments, and kinematic position of the suspension. The values may be used to accurately characterise the wheel centre.

The aspects of the vehicle whose states may be indicated by the sensor data may include engine torque; wheel hub vertical acceleration; intelligent tyre; brake pressure; steering angle, speed and torque; body Inertial Measurement Unit (IMU); wheel speed and odometer; ride height; Global Navigation Satellite Systems (GNSS); Ambient temperature; and odometer. These aspects provide thorough data on the operation of the vehicle, and so allow accurate estimation of the fatigue damage.

According to yet another aspect of the present invention there is provided a method for monitoring components within a vehicle, the method comprising determining, in dependence on vehicle sensor data indicative of a state of at least one aspect of the vehicle, at least one wheel centre action for at least one wheel of the vehicle, estimating from the at least one wheel centre action, a set of component-level forces for at least one mechanical component of the vehicle, wherein the component-level forces are estimated using data from a plurality of digital twins associated with the vehicle, and calculating, from the component-level forces, a fatigue damage value for the at least one mechanical component; and outputting a signal indicative of the fatigue damage value.

The method may further comprise defining a road severity index for each road portion for which sensor data is received, wherein the road severity index is based on the fatigue damage value for the at least one component.

Optionally, the method comprises deriving a durability control signal from the fatigue damage value, and the output is configured to output the durability control signal to at least one control system of the vehicle.

According to a further aspect of the invention, there is provided a non-transitory computer readable medium comprising computer readable instructions that, when executed by a processor, cause performance of the method.

According to a still further aspect of the invention, there is provided a vehicle comprising the monitoring system discussed above.

Within the scope of this application it is expressly intended that the various aspects, embodiments, examples and alternatives set out in the preceding paragraphs, in the claims and/or in the following description and drawings, and in particular the individual features thereof, may be taken independently or in any combination. That is, all embodiments and/or features of any embodiment can be combined in any way and/or combination, unless such features are incompatible. The applicant reserves the right to change any originally filed claim or file any new claim accordingly, including the right to amend any originally filed claim to depend from and/or incorporate any feature of any other claim although not originally claimed in that manner.

BRIEF DESCRIPTION OF THE DRAWINGS

One or more embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings, in which: Figure 1 shows a block diagram of a monitoring system in accordance with an embodiment of the invention; Figure 2 shows a block diagram of a system including the monitoring system in accordance with an embodiment of the invention; Figure 3 shows an example chart illustrating a rainflow analysis; Figures 4(a) and 4(b) are diagrams illustrating an example of Latin hypercube sampling; Figure 5 is a block diagram of a further monitoring system according to an embodiment of the invention; Figure 6 is a block diagram of a data repository forming part of a monitoring system according to an embodiment of the invention; Figure 7 is a block diagram of a monitoring system according to an embodiment of the invention including a durability control module; Figure 8 is a flow diagram of a method according to an embodiment of the invention; Figure 9 shows a vehicle in accordance with an embodiment of the invention.

DETAILED DESCRIPTION

A monitoring system for a vehicle in accordance with an embodiment of the present invention is described herein with reference to the accompanying Figure 1. The monitoring system, indicated generally by the reference numeral 100, comprises an input means 102 and an output means 104. The input means 102 is configured to receive vehicle sensor data from the vehicle. The vehicle sensor data is indicative of a state of at least one aspect of the vehicle. The output means 104 is configured to output a signal indicative of fatigue damage value for a mechanical component of the vehicle. The monitoring system 100 further comprises determining means 106, wherein the determining means are configured to determine a wheel centre action for a wheel of the vehicle, in dependence on the vehicle sensor data. The monitoring system further comprises estimating means 108. The estimating means 108 is configured to estimate a set of component-level forces for the mechanical components of the vehicle from the wheel centre action. The component-level forces are estimated using data from a plurality of digital twins associated with the vehicle. The monitoring system 100 comprises deriving means 110. The deriving means is configured to derive the fatigue damage value for the mechanical component. The fatigue damage value is derived from the set component-level forces.

Referring to Figure 2, there is shown a block diagram of the monitoring system 100 according to Figure 1 showing the input thereto and output therefrom. The monitoring system 100 receives the vehicle sensor data 112 from the vehicle and outputs a signal 114 indicative of the fatigue damage value. The vehicle sensor data may be associated with a specific journey on a specific road portion.

Road conditions and driver behaviour can be considered as the main contributors to vehicle durability. These elements are outside the control of the vehicle manufacturers. A typical vehicle is equipped with a number of on-board sensors (not shown). The input means 102 of the monitoring system 100 is configured to receive sensor data from the vehicle with which the monitoring system 100 is associated.

The on-board sensors which may provide data to the monitoring system include but are not limited to engine torque; wheel hub vertical acceleration; low frequency wheel vertical force via intelligent tyre technologies; brake pressure; steering angle, speed and torque; body Inertial Measurement Unit (IMU); wheel speed and odometer; ride height; Global Navigation Satellite Systems (GNSS); ambient temperature; and vehicle odometer. Many vehicles may further comprise active systems that process sensor data to control aspects of the vehicles operation and improve passenger safety and comfort.

Such active systems include but are not limited to active damper, Electric Active Roll Control (eARC), electronic air suspension, and active differential coupling. These active systems may generate control parameters. The measured values of certain control parameters may be included in the vehicle sensor data received by the input means 102 of the monitoring systems. These may include active damper current, Electric Active Roll Control (eARC) actuator current, electronic air suspension pressures, and active differential coupling torque. It will be understood that the sensor data is not limited to the tpes listed above. Any sensor fitted to the body or suspension of the vehicle may provide useful data for the monitoring system. In some cases, certain sensor outputs may be used to provide a redundancy check.

Sensor signals, Sn, can be assumed to be a function of the driver (D) and road (R) inputs. The sensor signals may be monitored on CAN, Flexray, Ethernet busses or other suitable busses.

(SI, Sp, S3, S4, = f(D,R) After the vehicle sensor data has been received, it may be consolidated. The consolidation operation may comprise cleaning of the sensor data, alignment of the data, removal of non-value-added signals, consistency checks, and other techniques. The consolidation operation may comprise evaluating the data, saving portions evaluated to be useful, and discarding portions evaluated not to be of interest. For example, when monitoring fatigue, sensor data indicating a period where the vehicle was cruising on a motorway may not be considered relevant, as motorway cruising does not contribute significantly to fatigue of the vehicle components. As such, the consolidated sensor data may be continuous, or may be associated with discrete portions of journeys. The vehicle location is recorded as part of the vehicle sensor data and will be included in the consolidated sensor data. The distance travelled is also included in the consolidated sensor data, and in the case of discrete journey portions, the distance covered in that journey portion is covered. The consolidated sensor data is represented as tn.

Once the monitoring system 100 has received the appropriate vehicle sensor data, it determines a wheel centre action, based on the vehicle sensor data. The determining may be carried out by the determining means 106. Wheel centre actions comprise wheel centre forces and kinematic positions for the wheel in question. The wheel centre forces may comprise longitudinal force, lateral force, and vertical force. The kinematic positions may include driving torque, braking torque, vertical displacement and wheel direction.

In an example, these wheel centre actions may be derived in the following ways, although it will be appreciated that other methods of determining these forces may be used. Vertical force may be derived from low frequency tyre vertical force, as measured by an intelligent tyre system, combined with higher frequency wheel hub vertical acceleration, using appropriate signal processing techniques. The corner wheel lateral force can be determined from the vertical force, wheel speed and steering angle. The sum of the lateral forces of the four corner wheels can be checked against the body IMU lateral acceleration, from the Inertial Measurement Unit (IMU). The longitudinal force per corner can be determined by vertical force, brake pressure, engine torque, wheel speed and differential coupling torque. The moments at each wheel corner can be determined knowing the suspension architecture and the forces in action there.

The wheel forces are a function of the sensor data from onboard vehicle sensors WFt = f(Si, Sz, Sz, S. Sr) where WEt is each wheel force vector as a function of the vehicle onboard sensors and i is the corner in question Wheel forces are assumed to consist of six components WEI =(F,c, Fy, Fz, Mx, my, Anz) where Fx, Fy, Fz are the forces in the x, y, and z direction respectively, and Mx, my, Mare the moments in the x, y, and z direction respectively.

Once the wheel centre forces and moments have been calculated providing the wheel centre actions, this information can be combined with vehicle ride height sensor data and used to derive component-level forces, using data from a set of digital twins for that vehicle model, according to the operation al= f(WFt, Pd where 'c indicates the individual chassis component and 't' indicates one of the set of digital twins for that vehicle model. The component-level forces can be derived from a combination of wheel forces and the kinematic position (toe, camber and vertical position) of the wheel (R) and determined from ride height sensors and/or the steering wheel angle. The digital twins provide data indicating the rating or parameter values for the mechanical components to facilitate the component-level forces calculations.

The set of digital twins may be referred to as a variability enhanced digital twin. The variability enhanced digital twin aims to represent the manufacturing variability of certain components within the vehicle. The individual component rating or parameter of each chassis component of interest fitted on a specific vehicle may not be known, however its nominal value and tolerance is known. These values are typically defined during the design and validation phases for the vehicle. The nominal values and tolerances may be used to define a digital twin template, from which one or more digital twins may be generated.

A digital twin template may be defined for each vehicle configuration of interest. Vehicle configuration may refer to the vehicle model and any further variations within that model such as engine type and size and other feature configurations. For each vehicle configuration of interest, a set of mechanical components of interest is chosen for inclusion in the digital twin template. The set of mechanical components of interest may include components whose manufacturing variability affects the load propagation through the system, such as main chassis suspension components. Each mechanical component has at least one component rating or parameter that is relevant for use in estimating component-level forces from wheel centre actions. Typical component parameters may include a bush stiffness, a link stiffness, damping, or the like. That component rating or parameter has a nominal value and tolerance, such that the range of acceptable values is included in the digital twin template. The digital twin template may thus define a multi-dimensional domain that captures the manufacturing variability of certain components within the vehicle. For a vehicle configuration with x components of interest having y ratings or parameters of interest, the digital twin template may be considered a y-dimensional domain. Selecting a point within that y-dimensional domain provides one specific digital twin, with a set of rating or parameter values that fall within their acceptable ranges. Each digital twin represents a unique combination of selected components and their associated parameters or ratings. Digital twins may be identified from the digital twin template by using stochastic analysis techniques. A digital twin template may specify approximately ten to twelve component ratings or parameters, but it will be understood that there may be a greater or lesser number.

Considering an example component such as a bush, it may have a bush stiffness characterised as a nominal value ±10%. Using stochastic analysis tools, for example Monte Carlo, Latin hypercube sampling, or other suitable techniques, it is possible to identify multiple digital twins, as desired. It will be understood that for each digital twin identified, one value for the force the component experiences is derived. Thus the use of a plurality of digital twins facilitates determining an range in which the real behaviour of the component in the vehicle will lie. An example of Latin hypercube sampling is described in relation to Figures 4(a) and 4(b).

Machine learning and/or semi-analytical multibody models may be used to perform the analysis and calculation to derive the component-level forces. In an example, a trained neural network is used to predict the component-level forces. The suspension model could also be implemented using a semi-analytical multibody model.

In an example, the Wheel Forces and Position (fl/Ft. Pi) are used as input to a Non-Linear Auto-Regressive with External Input (NARX) model. This machine learning model uses the last estimated output vector (Y) as input to the next time step (W F P, Y), creating a time dependence of the output to the previous iteration. The output vector represents the component level forces. The machine learning model estimates how the load is cascaded in the suspension system. Such a model may be trained using actual data or using Computer-Aided Engineering (CAE) simulations, such as a semi-analytical multibody simulation. Each digital twin may be considered to represent a semi-analytical multi body model of that vehicle configuration. The process may comprise using a sampling technique to extract a set number of Digital Twins (t), representing a unique combination of components/parameters; then t N "variability enhanced Digital Twins" based on the previously extracted sets (each set makes up a DT). Each digital twin may be considered to represent a semi-analytical multi body model of that vehicle configuration. This semi-analytical digital twin can directly accept the Wheel Forces and position data and cascade to component-level forces.

As alternative, a semi-analytical model could be used to train a machine learning NARX model. A set of known inputs may be provided as an input to the semi-analytical model and the cascaded loads may then be used to train the ML NARX model. In this way, there would be a ML NARX model for each digital twin. This would allow the use of neural networks for identifying the component-level forces, which has a lower processing overhead than that of above-described multi-body model process.

Once the component-level forces have been derived, a fatigue damage value for the components of interest is derived by the monitoring system 100. The derivation may be carried out by the deriving means 110. A fatigue damage value may be calculated for each component for which a force has been calculated. The fatigue damage value represents the accumulated fatigue damage for the component in question, for the road portion from which the vehicle sensor data was recorded. A provisional fatigue damage value may be calculated based on each digital twin. A final fatigue damage value to be output may be obtained from the set of provisional fatigue damage values in an appropriate manner. For example, the fatigue damage value may be an average of the provisional fatigue damage values. Alternatively, in a scenario where a higher safety factor is preferred, the highest accumulated fatigue damage value may be used.

Methods for the calculation of accumulated fatigue damage are well known in the context of mechanical systems. For each component that is to have a fatigue damage value calculated, the minimum component-level fatigue damage capacity is used. Depending on the fatigue damage method employed, this could take the form of a simple stress-life curve based on measured material test data, for that specific component. It is also possible to use more extensive parameters describing the stress life S-N curve or to define a strain-life curve and the Ultimate Tensile Strength. Other parameters which could be defined include the Young modulus or the yield strength. These additional parameters would support the application of different fatigue damage calculation methods with varying levels of complexity.

A simple Wohler curve may be used to provide estimated damage or severity indicator. However, it will be apparent that a variety of other known techniques may be used to estimate the damage once the component-level forces are available.

The fatigue damage value may be calculated to represent the damage that occurred over a period of time. As such it may be derived from a time series of the derived component-level forces for the component in question. The time series of forces represents a number of load cycles experienced by the component. Each load cycle will have a certain level of damage associated therewith. The time series of forces may be processed to identify the load cycles. Such processing may use a rainf low cycle counting algorithm. The algorithm comprises the steps of extracting turning points from the signal; removing small amplitude cycles from the signal; and then applying a cycle extraction algorithm. Extracting the turning points from the signal reduces the number of data points in the time series so as to allow faster processing and reduce data storage requirements. This may be achieved using an algorithm known as a Peak-Valley filtering algorithm, which uses 1st and 2nd differentials of the signal to identify turning points and the nature of the turning point. Once the turning points have been extracted the next step is to remove the small amplitude signals cycles from the signal. This may also be referred to as hysteresis filtering. Removing the small amplitude cycles reduces the complexity of the fatigue calculations by reducing the number of input cycles. The threshold below which cycles are ignored may be defined for each component of interest. Next the remaining signal values are processed further to apply an algorithm to extract each cycle. One suitable algorithm is the four-point rainflow cycle counting algorithm.

An example to illustrate the four-point rainflow cycle counting algorithm is shown in Figure 3. There is shown a sequence of adjacent turning points labelled A, B, C and D. If 1B-CI is greater than or equal to IA-DI the algorithm moves to evaluate the next turning point. If 1B-CI is less than IA-DI, the points BC are extracted from the time series data, and stored in a list of rainflow cycles. While the four-point rainflow cycle counting algorithm is illustrated here, it will be apparent that other rainflow cycle counting algorithms may be used. When all of the relevant time series data has been analysed according to the rainflow algorithm, there will be a list of extracted rainflow cycles. These extracted rainflow cycles may be used to calculate the provisional fatigue damage value for the component in question. Each cycle contained in the list of extracted rainflow cycles has a specific damage contribution for each component on the vehicle for which it has been derived.

The damage contribution per exact extracted rainflow cycle maybe calculated according to any of a number of known methods. A simple fatigue damage method using a Wohler curve and the PalmgrenMiner rule is described herein as an example. This method sums the damage contribution per extracted rainflow cycle. The Wohler curve used is based on the Basquin equation which and can be expressed as: = aRr.

where the parameters and a and p describe the Wohler curve and are defined by the user of the method. The damage contribution of each cycle i with range R in the list of rainflow cycle rainflow cycles is therefore expressed as: di3O,1= 1 /N1,1 where di is the fraction of component life this cycle represents, and N is the number of load cycles of range the component will survive before failure is identified. As each cycle is processed, the fraction of component life for each is summed together using the Palmgren-Miner rule: Doi= i=1 Uict This defines the total accumulated fatigue damage as a fraction of the total damage the component can survive before failure is predicted. Where D is greater than or equal to 1, it is predicted the component has reached the usable fatigue life and failure could occur.

This example shows how the provisional fatigue damage value contribution for a road portion consisting of a number of cycles may be calculated, based on a component level force derived using one digital twin from vehicle sensor data 112 from that road portion. Further provisional fatigue damage values are derived from at least some of the remaining multiple digital twins. It will be apparent that there are numerous alternative methods to calculate the fatigue damage contribution per cycle. For example, a Manson-Coffin approach, fatigue damage spectrum approach, or continuous non-linear damage mechanics approach may also be used. As discussed above, the provisional fatigue damage values are combined to form a fatigue damage value, which may be referred to as a final fatigue damage value or an overall fatigue damage value. The fatigue damage value may also be combined with the distance travelled to generate that fatigue damage. In this way, a normalized fatigue damage value is obtained. This normalised value may be referred to as a fatigue damage value per unit distance, for example per kilometre.

It will be understood that the monitoring system 100 calculates a fatigue damage value for specific road portions from which vehicle sensor data has been received. Fatigue damage values may be combined to generate a combined fatigue damage value representing a combination of road portions. A combined fatigue damage value may represent a single journey, a set of journeys defined by distance or time period. In one example, the combined fatigue damage value represents the total fatigue for the lifetime of the vehicle, according to the processed sensor data. Such a total fatigue may not correspond to the actual total fatigue experienced by the vehicle components, as vehicle sensor data may not be recorded at all times.

The fatigue damage value per unit distance may be used to calculate an indicator referred to as the road severity index. The road severity index may be calculated using a look-up table. The predefined look-up table represents the relationship between the fatigue damage value per unit distance and the road severity index. As the road severity index is based on the vehicle sensor data, it may be specific to the vehicle that provided the vehicle sensor data in question. An example look-up table is provided at Table 1 below. In Table 1, the unit distance used is one kilometre.

Fatigue Damage Value per km Road Severity Index Less than 0.01 0 Between 0.01 and 0.1 1 Between 0.1 and 0.5 2 Greater than 0.5 3

Table 1

The vehicle location data is processed so that each road portion for which a fatigue damage value has been calculated has an associated road severity index. A road portion may be an entity known as a road segment in a Geographical Information System (GIS) used in navigation and mapping systems. The road network topology is mapped as a set of links and nodes for which attributes may be associated. A road segment may be defined as a combination of topology nodes and links. The road severity indexes may be stored in a road portion database, with identifying information for the road portions such as road segments.

For each length of road surface for which fatigue damage values are calculated provided to the monitoring system 100, the monitoring system 100 will monitor the road severity index and will continue to accumulate the length of concurrent road surfaces until the road severity index changes to a different category. When a change in the road severity index category is observed, the monitoring system may at that point save, to memory, the road severity index for the length of data processed, the fatigue damage value for each component, and the GIS segment data. The monitoring system 100 may continue to monitor subsequent road portions.

While the present disclosure teaches the derivation of a road severity index based on vehicle sensor data, it will be understood that other methods of obtaining a useful road severity index are available. For example, it would be possible to use other parameters to determine the severity of a length of road.

These parameters may be used instead of the road severity index derived from the vehicle sensor data, or may be combined therewith to create a modified road severity index. Suitable parameters may include using the highest force measured per length and comparing this against a predefined value related to the engineering capacity of the component; the shock response spectrum or fatigue damage spectrum which is compared against a known spectrum and duration to which the component has been engineered.

Referring now to Figures 4(a) and 4(b), there is illustrated an example of Latin hypercube sampling for generating digital twins from a digital twin template. This may be used in deriving the estimated component level forces using the multiple digital twins. Figures 4(a) and 4(b) relate to an example comprising a pair of bushes for which the nominal stiffness K for their main axis is known, as well as their manufacturing tolerance. As such, this example digital twin template domain is 2-dimensional. Firstly, the Cumulative Function Distribution (CFD) for each bush is calculated and divided into N equal ranges (CFD / N), as shown in Figure 4A. In this case N is equal to five. For each of the equally partitioned range regions (j), a point within that region is randomly sampled j * CFD / N. In this way, the random points will not all be taken from the same local region of the CFD. Each sample corresponds to one digital twin. Each bush will be represented in N digital twins, and each digital twin will have a different value for the stiffness of each bush. In this way, the manufacturing variability of the stiffness value is represented. The two distributions are independent, and they can be randomly combined as 2-dimensional pairs as shown in Figure 4(b). Sampling is random for each element of the grid and only one sample is picked from each row and column as shown in Figure 4(b).

Referring now to Figure 5, there is shown a block diagram of an embodiment of the monitoring system indicated generally by the reference numeral 200. The monitoring system 200 comprises a component forces estimator 202. The component forces estimator 202 uses the multiple digital twins to take into account the manufacturing variability of a component so as to provide a reliable estimate of the force experienced by that component over a period of time. The component forces estimator 202 may be considered to correspond to the determining means 106 of Figure I. The outputs of the component forces estimator 202 are provided to a damage estimation module 204. The damage estimation module 204 calculates the damage that the component would have incurred if it had experienced the forces estimated by the component forces estimator. The damage estimation module 204 carries out a four-step process of extracting turning points from the time series data of the component forces; removing smaller amplitude cycles from the data; and then performing rainf low cycle extraction to arrive at a fatigue damage calculation. The damage estimation module 204 may be considered to correspond to the calculating means 110 of Figure I. The damage estimation module 204 produces two outputs, firstly a damage value 206 for each monitored component of the vehicle and secondly a road severity index 208 for each portion of road for which vehicle sensor data has been processed. The damage value 206 corresponds to the fatigue damage value discussed previously. The monitoring system 200 further comprises a data repository 210. The data repository 210 may be remote from the vehicle and may be implemented as a cloud storage solution. The road severity indexes 208 and the damage values 206 are stored in the data repository 210. The data repository 201 also stores road segment definitions 212 derived from GNSS location data 214 obtained from the vehicle. The monitoring system may include communication means for communication with remote components thereof, such as a remote data repository. The monitoring system may avail of communication means within the vehicle for communication with remote components thereof.

Within the data repository 210, the accumulated damage values may be stored in a damage table. The damage table comprises a list of the monitored chassis components with the integration over time of their accumulated damage per road segment and per component. The accumulated damage may thus be represented as follows: where L, is the accumulated damage per component c, and id refers to the road portion or road segment.

A value referred to as the residual component life may be calculated from the accumulated damage as follows: R, = I -L, where is Re is the residual component life of the component c.

Referring now to Figure 6, there is shown a block diagram of a data repository 210 for use by a monitoring system according to an embodiment of the invention. The data repository 210 comprises a component database 220 and a road database 220. The component database 220 may comprise fatigue damage values for the mechanical components of the vehicle that are being monitored by the monitoring system. The road database 220 may comprise the road severity indexes that have been derived by the monitoring system. The data repository may be located on the vehicle or stored remotely from the vehicle. In an example where the data repository 210 is stored remotely from the vehicle company, a local copy may be maintained at the vehicle. In a case where the road severity indexes that have been calculated exceed the space available to store them in the data repository, the oldest road severity index may be overwritten The damage values and the road severity indexes are derived from the vehicle sensor data of a specific vehicle, and as such the values obtained relate to that vehicle. The data repository of the monitoring system 200, if remote from the vehicle, may store data from more than one vehicle having the monitoring system. In such cases, the data will be associated with a unique identifier, such as the Vehicle Identification Number (VIN), for the vehicle from which it originates.

Referring now to Figure 7, there is shown a block diagram of an alternative embodiment of a monitoring system, indicated generally by the reference numeral 300, according to the invention. The monitoring system 300 of Figure 7 comprises the features of the monitoring system 100 as described herein in relation to Figure 1. It will be understood that the features of the monitoring system 200 as described herein in relation to Figure 5 may also be included in the monitoring system 300. The monitoring system 300 further comprises a durability control module 302. The durability control module 302 receives the outputs of the monitoring system 100 which include the road segment indexes and the fatigue damage values. The durability control module 302 also receives current vehicle status information 304. This current vehicle status information 304 includes location information for the vehicle and odometer information.

The durability control module 302 generates a durability control signal 306. The durability control signal 306 is derived based on the road severity index and the accumulated damage over time. The durability control module 302 uses the current vehicle location to identify the applicable road severity index for that location from the data repository. If a road severity index is not available for that location, the durability control module 302 uses a default value. The default value may be based on the overall accumulated damage and the total vehicle mileage.

Based on the accumulated damage value and the applicable road severity index, the durability control module 302 derives an adaptation signal for an active control system of the vehicle. The adaptation signal is intended to control the active control system in question to preserve the life of the component in question. Where more than one active control system is present in the vehicle, the durability control module 302 may be adapted to generate an adaptation signal for some or all of active control systems.

Where the components being monitored have been estimated to have accumulated expected levels of fatigue damage then no adaptation signal maybe generated by the durability control module 302, as the component and active system appear to be acting as designed. Where a single component or a subset of components have accumulated fatigue damage to a level higher than expected, then the durability control module may apply adaptations only to the relevant active control systems for those components.

The durability control signal may comprise an adaption enable component to control whether an active control system should apply an adaptation, and a control level component to control the level of the adaptation. Depending upon the active system under control and the desired effect, the relationship between the road severity index and the control level may include no adaptation of the active control system, scaling applied to a subset of calibration power parameters of the active control system, or discrete switching of a subset of calibration part of the calibration parameters.

The adaptation of the active control systems may be controlled to achieve a target level of damage accumulation. In this way if a component is predicted to not achieve its intended life then the appropriate active control system or systems can be adapted to attempt to prolong the component life closer to the target lifetime. Alternatively, if the damage accumulation is expected to achieve a longer than expected component life, either no adaption could be applied, or an adaption could be applied to improve an alternative attribute performance at the expense of durability.

The durability control module 302 may additionally generate an alert signal to the user of the vehicle if the monitoring system predicts that a component is nearing the end of its intended life. The monitoring system 300 may determine that a component is nearing the end of its intended life if its combined fatigue value is above a certain threshold.

In an example where the data repository is remote from vehicle, the durability control module 306 may comprise a local copy of the data repository containing the fatigue damage component and the road severity indexes for that vehicle.

The monitoring systems as described herein may be implemented either as an on-board system on a vehicle in use; implemented remotely for example using a cloud-based infrastructure; or it may be implemented in a split manner such that it may be partially implemented on the vehicle and partially remotely. In some examples, the monitoring system 100, 200, 300 comprises remote data storage, which may be used for storing data related to the monitoring system. In an example. the data storage used by the monitoring system may be referred to as a data repository. The use of a cloud-based infrastructure provides scalability and flexibility over time.

The data generated by the vehicle sensors is not required to be analysed in real time. This allows the data to be buffered and/or stored temporarily so as to maximise computational resource utilisation. For example, carrying out heavy load computational work at a time when computational demand from other sources may be low due to the time of day.

In an example, the vehicle sensor data 112 is received at the input means 102 on a vehicle. The vehicle sensor data is consolidated at the vehicle and then sent to a remote processor, which may be a cloud-based processor, for further processing to obtain the fatigue damage value. The remote processor may also produce other data such as the road severity index. The monitoring system may comprise consolidation means to carry out the consolidation. The consolidation may comprise a triage analysis process to identify data indicating actions that would likely lead to fatigue damage to the components.

Referring now to Figure 8, there is shown a flow diagram of a fatigue monitoring method 400 according to an embodiment of the invention. At 402, the method comprises receiving vehicle sensor data. The vehicle sensor data may then be consolidated. At 404, the method comprises determining the wheel centre actions from the vehicle sensor data. At 406, the method comprises defining a plurality of digital twins representing the manufacturing variability of the component or components to be monitored. The digital twins are defined based on the nominal value and tolerance of the rating for the components in question. At 408, the method comprises estimating a plurality of component level forces, using data from the plurality of digital twins, such that each digital twin results in a value for the component level forces. At 410, the method comprises deriving a fatigue damage value from the component level forces.

Each of the set of component level forces may be used to derive a provisional fatigue damage value.

The final fatigue damage value may then be obtained from the provisional fatigue damage values. At 412, the method comprises outputting the final fatigue damage value. Outputting the fatigue damage value may comprise storing the data in a data repository.

Referring to Figure 9, there is shown a vehicle in accordance with an embodiment of the present invention, the vehicle being indicated generally by the reference numeral 900. The vehicle 900 may comprise the automatic speed control system described herein, and/or may be adapted to implement one or more of the methods for automatically controlling the speed of the vehicle.

It will be appreciated that various changes and modifications can be made to the present invention without departing from the scope of the present application.

Claims

CLAIMS1. A monitoring system for a vehicle, the monitoring system comprising: input means configured to receive vehicle sensor data indicative of a state of at least one aspect of the vehicle; means configured to: determine, in dependence on the vehicle sensor data, at least one wheel centre action for at least one wheel of the vehicle; estimate, from the at least one wheel centre action, a set of component-level forces for at least one mechanical component of the vehicle, wherein the component-level forces are estimated using data from a plurality of digital twins associated with the vehicle; and derive, from the component-level forces, a fatigue damage value for the at least one mechanical component; and output means configured to output a signal indicative of the fatigue damage value.
2. A monitoring system according to claim 1 wherein the monitoring system comprises: determining means configured to determine the wheel centre action for the at least one wheel of the vehicle; estimating means configured to estimate the component-level forces for the at least one mechanical component; and deriving means configured to derive the at least one fatigue damage value.
3. A monitoring system according to claim 1 or claim 2 wherein the set of component level-forces comprises a component-level force for each of the plurality of digital twins.
4. A monitoring system according to any preceding claim wherein each of the set of component-level forces is used to derive a provisional fatigue damage value, and a final fatigue damage value is obtained from the plurality of provisional fatigue damage values.
5. A monitoring system according to any preceding claim wherein the sensor data relates to a specific journey on a specific road portion such that the fatigue damage value is associated with the specific road portion.
6. A monitoring system according to claim 5 wherein the system is configured to calculate at least one combined fatigue damage value for a combination of road portions.
7. A monitoring system according to claim 5 or 6 wherein the system is configured to define a road severity index for each road portion for which sensor data is received, wherein the road severity index is based on the fatigue damage value for the at least one component.
8. A monitoring system according to claim 7 wherein the road severity index is vehicle specific.
9. A monitoring system according to any of claims 6 to 8 wherein the system is configured to generate a warning if the combined fatigue value is above a predefined threshold.
10. A monitoring system according to any preceding claim wherein the system is configured to maintain a component database comprising fatigue damage values for a plurality of mechanical components of the vehicle.
11. A monitoring system according to any of claims 7 or 8, wherein the system is configured to maintain a road portion database comprising road severity index information for each road portion.
12. A monitoring system according to any preceding claim wherein the system is configured to derive a durability control signal from the fatigue damage value, and the output is configured to output the durability control signal to at least one control system of the vehicle.
13. A monitoring system according to claim 12 wherein the durability control signal is based on the road severity index.
14. A monitoring system according to any claim 12 or 13 wherein the durability control signal comprises an adaption enable component to control whether the at least one control system should apply an adaption, and a control level component to control the level of the adaptation.
15. A monitoring system according to any of claims 12 to 14, wherein the control systems include active control systems.
16. A monitoring system according to any preceding claim wherein at least a part of the monitoring system is located on the vehicle.
17. A monitoring system according to any preceding claim wherein at least a part of the monitoring system is located remote from the vehicle.
18. A monitoring system according to any preceding claim wherein the system is configured to consolidate the sensor data.
19. A monitoring system according to any preceding claim wherein the wheel centre actions comprise at least one of: wheel centre forces, wheel moments, and kinematic position of the suspension.
20. A monitoring system according to any preceding claim wherein the aspects of the vehicle whose states may be indicated by the sensor data include engine torque; wheel hub vertical acceleration; intelligent tyre; brake pressure; steering angle, speed and torque; body Inertial Measurement Unit (IMU); wheel speed and odometer; ride height; Global Navigation Satellite Systems (GNSS); Ambient temperature; and odometer.
21. A method for monitoring mechanical components within a vehicle, the method comprising: determining, in dependence on vehicle sensor data indicative of a state of at least one aspect of the vehicle, at least one wheel centre action for at least one wheel of the vehicle; estimating from the at least one wheel centre action, a set of component-level forces for at least one mechanical component of the vehicle, wherein the component-level forces are estimated based on data from a plurality of digital twins associated with the at least one vehicle; deriving, from the set of component-level forces, a fatigue damage value for the at least one mechanical component; and outputting a signal indicative of the fatigue damage value.
22. The method of claim 21 comprising defining a road severity index for a road portion for which sensor data is received, wherein the road severity index is based on the fatigue damage value for the at least one component.
23. The method of claim 21 or claim 22 comprising deriving a durability control signal from the fatigue damage value, and the output is configured to output the durability control signal to at least one control system of the vehicle.
24. A non-transitory computer readable medium comprising computer readable instructions that, when executed by a processor, cause performance of the method of any of claims 21 to 23.
25. A vehicle comprising a monitoring system according to any one of the preceding claims.