WO2017051032A1

WO2017051032A1 - A method for estimating the need for maintenance of a component

Info

Publication number: WO2017051032A1
Application number: PCT/EP2016/072829
Authority: WO
Inventors: Søren THEODORSEN; Lars Dybdahl NIELSEN
Original assignee: Northern Vo Aps
Priority date: 2015-09-24
Filing date: 2016-09-26
Publication date: 2017-03-30

Abstract

The present disclosure relates to a method for determining one or more reference levels corresponding to the maintenance condition of a component of machinery, such as a vehicle, comprising the steps of: a) providing predefined physical limits for i) an intact level corresponding to a physical limit below which the component is assumed to be functional; and ii) a broken level corresponding to a physical limit above which the component is assumed to be broken; b) acquiring monitoring data parameters in a time period (indicative of movement, acceleration and/or angular orientation of the component and/or the vehicle) corresponding to physical impact and/or stress of the component and calculating the physical condition of the component based on the data parameters; c) evaluating the physical condition of the component after said time period, and updating the intact level to the calculated physical condition if the component is intact and the calculated physical condition exceeds the intact level; or updating the broken level to the calculated condition if the component is broken and the calculated physical condition is below the broken level. The disclosure further relates to a method for predicting the need for maintenance of machinery, such as a vehicle or a fleet of vehicles, comprising the steps of: providing monitoring data parameters acquired over a time period corresponding to physical impact and/or stress of one or more components of said machinery and calculating the physical condition of said component(s) based on the data parameters; providing at least two reference levels of said components of said machinery thereby obtaining at least an intact level and a broken level of each component, comparing the physical condition of each component with the corresponding intact level and broken level of said component, and predicting the need for maintenance of the machinery by determining the intactness of each component.

Description

A method for estimating the need for maintenance of a component

The present invention relates to estimating a maintenance condition of a component of machinery, such as a vehicle, by determining one or more reference levels based on monitoring data parameters and predefined physical limits of the component. The invention further relates to a method for predicting the need for maintenance of machinery based on monitoring data and reference levels.

Background of invention

The purpose of maintenance is to avoid or mitigate the consequences of failure of equipment. Run-to-failure maintenance is when components are replaced when they fail. This may be improved by applying planned and condition based maintenance wherein worn components are replaced before they actually fail in order to preserve and restore equipment reliability. This may further be improved by preventive maintenance that includes partial or complete overhauls at specified periods, oil changes, lubrication and so on. The ideal preventive maintenance program would prevent all equipment failure before it occurs to avoid corrective maintenance, i.e. repair.

A proper implementation of predictive maintenance requires the right information in the right time about the condition of the equipment, in reality predicting the future trend of the equipment's condition. However, by knowing which equipment needs maintenance, maintenance work can be better planned (spare parts, people, etc.) and what would have been "unplanned stops" are transformed to shorter and fewer or better planned stops. To evaluate equipment condition nondestructive testing technologies performing periodic or continuous (online) equipment condition monitoring can advantageously be applied, preferably while the equipment is in service, thereby minimizing disruption of normal system operations.

Typically, when estimating the need for maintenance, measured data is monitored and compared against a threshold. There are other possibilities of estimating the need for maintenance, such as measuring the time a component is used or recognizing a pattern that is known to cause failure. A limitation of such methods is that they do not indicate the probability for false alarms and the risk of not alerting of a real risk of failure. They typically also fail to take into account information from other components. Summary of invention

The present disclosure relates to a method and system for determining one or more reference levels corresponding to the maintenance condition of a component of machinery, such as a vehicle. In a first embodiment the method comprises the steps of: a) providing predefined physical limits for i) an intact level and ii) a broken level. The intact level may for instance correspond to a physical limit below which the component is assumed to be functional. The broken level may for instance correspond to a physical limit above which the component is assumed to be broken. The method may further comprise the step of acquiring monitoring data parameters in a time period (e.g. indicative of movement, acceleration and/or angular orientation of the component and/or the vehicle) corresponding to physical impact and/or stress of the component and preferably calculating the physical condition of the component based on the data parameters. The physical condition of the component may be evaluated after said time period. Consequently the intact level can be updated to the calculated physical condition if the component is intact and the calculated physical condition exceeds the intact level, or the broken level can be updated to the calculated condition if the component is broken and the calculated physical condition is below the broken level.

The predefined physical limits for an intact level and a broken level can be said to define at least three levels of estimations of the need for maintenance. If the calculated physical condition is above the broken level, it is estimated that the component is in need of maintenance. If the calculated physical condition is below the intact level, it is estimated that the component is not in need of maintenance. If the calculated physical condition is between the intact level and the broken level, there are typically no assumptions about the need for maintenance.

The presently disclosed method introduces a dynamic model that takes into account an evaluation of the physical condition of the component, which could be for example a manual inspection by a mechanic or other technical checks that serve as "true" data, which can be used to update the intact level and broken level, and increase the validity and the reference parameters. By combining data from measurements of the component and other data from evaluation of the physical condition, the present dynamic model can increase both the accuracy and the validity of the prediction of maintenance of a component. The presently disclosed method may in particular be useful in combination with orthogonal transformation data analysis, e.g. principal components analysis (PCA), as described in pending application WO 2015/139709. By for instance selecting one or more subsets of said monitoring data parameters; applying an orthogonal

transformation of at least one of said subsets thereby obtaining a set of eigenvectors for said subset; computing a multi-dimensional status model of the vehicle, such as by forming an ellipsoid of said set of eigenvectors; and evaluating the physical condition of the component by comparing the status model to a reference model of the vehicle, and applying the steps of the presently disclosed method, the potential of the PCA is exploited more efficiently.

Furthermore, in the presently disclosed system there is not necessarily a need to provide an accurate threshold for predicting the need for maintenance from the beginning as is the case in conventional systems for maintenance predictions. The broken level can initially be set to 0 and the intact level to 'infinite', which basically means that the system/method can be used to find the right levels and at the same time increase the validity accordingly. Alternatively, it is possible to set the initial levels to a level that is assumed to be a suitable starting point, for example based on other vehicles' levels or other vehicles' level with an offset added or subtracted. Possibly, before any intact levels are available, the component can be used or enabled for a certain amount of time that serves as a reference before starting to update the intact level. A reference period, wherein the component is considered to work properly, can be used to set an initial intact level. For a vehicle this can be achieved by driving for an amount of time and setting the intact level to the highest calculated level corresponding to a physical condition of a vehicle. This is also useful since the parameters that are taken into account in the step of acquiring monitoring data parameters in a time period can be different in terms of complexity and therefore not equally simple/difficult to provide a threshold for. In this sense the presently disclosed method can improve the validity of a prediction by enabling the method/system to learn from actual evaluation of the physical condition of the component. Similarly, data from more than one

component, for example components from a fleet of vehicles, can be used to increase the accuracy and validity.

The present disclosure further relates to a method for predicting the need for maintenance of machinery, such as a vehicle or a fleet of vehicles. In a first

embodiment the method comprises the steps of: providing monitoring data parameters acquired over a time period (e.g. indicative of movement, acceleration and/or angular orientation of the component and/or the vehicle), preferably corresponding to physical impact and/or stress of one or more components of said machinery and calculating the physical condition of said component(s) based on the data parameters. Accordingly at least two reference levels (e.g. intact level and broken level) of said components of said machinery may be provided, for instance according to the method described above, thereby obtaining at least an intact level and a broken level of each component, comparing the physical condition of each component with the corresponding intact level and broken level of said component, and predicting the need for maintenance of the machinery by determining the intactness of each component. Consequences of the abovementioned improvements may be lower costs for maintenance of components, vehicles or a fleet of vehicles. By applying the described method(s), the stops can be better planned - maintenance costs may be reduced with 10-20%, possibly

approaching 30% reduction for special purpose high-cost machinery, e.g. military vehicles.

Counters, e.g. mechanical damage counters, as described in further detail herein, may advantageously be applied to the presently disclosed method based, in particular in evaluating the physical condition and assessing the reference levels.

A further embodiment relates to the implementation into a system that can be installed in a vehicle, e.g. in the form of an on-board sensor system, e.g. in the form of system for prediction the need of maintenance for attachment to a vehicle and for monitoring the evaluated condition of said vehicle and/or predicting the need for maintenance, comprising

- at least one inertial measurement unit configured to measure the triple-axis proper acceleration, velocity and angular orientation of the chassis of the vehicle sampled over a time period,

- at least one GPS receiver for measuring the location of the vehicle,

- a computer comprising memory and a processing unit, configured for executing the method as herein described for assessing the condition of said vehicle.

The systems and methods disclosed herein may in particular be applied to vehicles as described above, but they may also be applied to equipment, machinery and/or parts in general. Employing the herein described system and method for monitoring equipment and machinery may lead to a functional implementation of predictive maintenance. Once operational the next step may be reliability (or risk) centered maintenance (RCM), where the equipment is scheduled for maintenance based on monitoring the condition and what the users require in the present operating context. Compared with predictive maintenance, RCM may further reduce maintenance cost and increase fleet reliability and availability. The presently disclosed system and method may also help to prevent misuse of machinery by ensuring that the each unit in a fleet is operated within predefined operational limits, e.g. in terms of wear, speed, surface, etc. It may also help to improve safety for the drivers that can be warned of potential hazards before they occur, because the condition of the vehicles is monitored. This monitoring may be provided in each vehicle and also centrally monitoring the entire fleet.

These and other aspects of the invention are set forth in the following detailed description if the invention.

Description of drawings

The invention will in the following be described in greater detail with reference to the accompanying drawings. The drawings are exemplary and are intended to illustrate some of the features of the presently disclosed method and system, and are not to be construed as limiting to the presently disclosed invention. Fig. 1 shows a diagram of physical condition of a vehicle suspension based on the data parameters, intact and broken levels, a number of incidents and the validity of the prediction of maintenance of a component.

Fig. 2 shows a diagram of physical condition, in this example corresponding to the wear of a vehicle component, intact and broken levels, and the validity of the prediction of maintenance of a component.

Fig. 3 shows an overview of the scheduling of maintenance and status of a component of a number of vehicles.

Fig. 4 shows the status of one vehicle having a number of components.

Fig. 5 shows one example of an overview of a number of vehicles and their status with respect to wear, incidents and stand-still.

Fig. 6 shows another example of an overview of a number of vehicles and their status with respect to wear, incidents and stand-still.

Fig. 7 shows one embodiment overview of the herein disclosed system for predicting the need for maintenance of machinery.

Fig. 8 shows a theme configuration i.e. grouping of related functions. Fig. 9 shows shows an example of a system for prediction the need for maintenance in the context of difference reference models.

Fig. 10 shows one embodiment of a system for predicting the need for maintenance according to the present disclosure.

Fig. 11 A illustrates the concept of a counter by showing the case of constant continuation (top curve) and linear interpolation (bottom curve).

Fig. 11 B shows a number of recorded counters forming a time series that forms the basis for predicting the future need for maintenance.

Fig. 12 shows plots of the same counter from 13 different vehicles, the mean and the upper and lower quantiles.

Fig. 13A shows an example of predicting the future trend of a counter value, in this example the counter value indicates that the current failure probability is just below 50%.

Fig. 13B in continuation of fig. 13A, fig. 13B exemplifies what can be deducted from a combination of prediction and failure probabilities.

Fig. 14 illustrates the difference between fracture and fatigue, where fig. 14A illustrates that failure of a component can occur when the load exceeds the yield strength of that component. Fig. 14B illustrates that failures due to fatigue can occur well below yield strength load.

Fig. 15 illustrates the process flow followed when assessing mechanical damage counters.

Detailed description of the invention

The present disclosure relates to a method for determining one or more reference levels corresponding to the maintenance condition of a component of machinery, such as a vehicle, comprising the steps of: a) providing predefined physical limits for i) an intact level corresponding to a physical limit below which the component is assumed to be functional; and ii) a broken level corresponding to a physical limit above which the component is assumed to be broken; acquiring monitoring data parameters in a time period (indicative of movement, acceleration and/or angular orientation of the component and/or the vehicle) corresponding to physical impact and/or stress of the component and calculating the physical condition of the component based on the data parameters; evaluating the physical condition of the component after said time period, and updating the intact level to the calculated physical condition if the component is intact and the calculated physical condition exceeds the intact level; or updating the broken level to the calculated condition if the component is broken and the calculated physical condition is below the broken level. In this regard it can be noted that the intact and broken levels may be inversely defined in the sense that for some measurements a high value may represent an intact state and a low value may represent a broken level. The presently disclosed method shall be construed such that the physical limits for i) an intact level may, in a further embodiment, correspond to a physical limit below above which the component is assumed to be functional and ii) a broken level corresponding to a physical limit below which the component is assumed to be broken, wherein the intact level is updated to the calculated physical condition if the component is intact and the calculated physical condition is below the intact level, and the broken level is updated to the calculated condition if the component is broken and calculated physical condition exceeds the broken level. As stated, the broken level and intact level can, in combination with a calculated physical condition be used to estimate the need for maintenance. The presently disclosed method can be said to involve a dynamic model that takes into account an evaluation of the physical condition of the component, which could be for example a manual inspection by a mechanic or other technical checks. The method can also be regarded as a way of tuning the predictive

maintenance. Furthermore, the acquiring of monitoring data and calculation of physical condition can be made in substantially real-time.

Vehicle and component

The "vehicle" as referred to herein may be machinery in general, in particular vehicles used for transport or movement of personnel or goods, such as vehicles on tracks and wheel, such as trains, cars, trucks, off-road vehicles, military vehicles, motorcycles, helicopters, planes, harvesters, combat vehicles, and equipment like excavators, forwarders, loaders, tractors, harvesters, ships, vessels, ferries, tankers, i.e. any moving machinery etc. but also other types of machinery, e.g. stationary machinery that does not change location geographically but have moving parts that needs maintenance and which can be monitored, such as wind turbines, etc.

The component of the machinery may be one or more single elements such as tires, brake shoes or oil filters, and/or it may be a subsystem of the machinery, such as brake, steering, suspension, engine, gear, or any other machinery component of a vehicle. Principal component analysis

Eigen-decomposition, or sometimes spectral decomposition, may be seen as the factorization of a matrix into a canonical form, whereby the matrix is represented in terms of its eigenvalues and eigenvectors. Only diagonalizable matrices can be factorized in this way. The eigen-decomposition of a symmetric positive semidefinite (PSD) matrix yields an orthogonal basis of eigenvectors, each of which has a nonnegative eigenvalue. The orthogonal decomposition of a PSD matrix is used in multivariate analysis, where the sample covariance matrices are PSD. This orthogonal decomposition is often referred to as principal components analysis (PCA). PCA studies linear relations among variables. PCA is performed on the covariance matrix or the correlation matrix (in which each variable is scaled to have its sample variance equal to one). For the covariance or correlation matrix, the eigenvectors correspond to principal components and the eigenvalues to the variance explained by the principal components. Principal component analysis of the correlation matrix provides an orthonormal eigen-basis for the space of the observed data: In this basis, the largest eigenvalues correspond to the principal components that are associated with most of the covariability among a number of observed data.

Principal component analysis (PCA) is thus one of several eigenvector-based multivariate analyses, where a statistical procedure uses orthogonal transformation to convert a set of observations of possibly correlated variables into a set of values of linearly uncorrelated variables (the principal components in PCA). This transformation is defined in such a way that the first principal component has the largest possible variance (that is, accounts for as much of the variability in the data as possible), and each succeeding component in turn has the highest variance possible under the constraint that it is orthogonal to (i.e., uncorrelated with) the preceding components. For PCA the rank of the data matrix decides the maximal number of principal components. In practice, the variance explained by the first several principal components may take a fairly big percentage of the total variance. Then only the first several principal components may be kept as the extracted new features to be used in the model comparison. PCA can therefore be thought of as revealing the internal structure of the data in a way that best explains the global variance in the data. If a multivariate dataset is visualized as a set of coordinates in a high-dimensional data space (1 axis per variable), PCA can supply the user with a lower-dimensional picture, a projection or "shadow" of this object when viewed from its (in some sense; see below) most informative viewpoint. This is done by using only the first few principal components so that the dimensionality of the transformed data is reduced.

PCA may be seen as equivalent to the following analysis techniques: discrete

Karhunen-Loeve transform (KLT), the Hotelling transform, proper orthogonal decomposition (POD), singular value decomposition (SVD), eigenvalue decomposition (EVD), factor analysis, canonical correlation analysis (CCA), Eckart-Young theorem, Schmidt-Mirsky theorem, empirical orthogonal functions (EOF), empirical eigenfunction decomposition, empirical component analysis, quasiharmonic modes, spectral decomposition, and empirical modal analysis. The methods and systems employing PCA as described herein may therefore in further embodiments apply the

abovementioned, at least partly analogous, analysis techniques to obtain the same results. E.g. PCA is closely related to factor analysis, wherein factor analysis typically incorporates more domain specific assumptions about the underlying structure and solves eigenvectors of a slightly different matrix. PCA is for example also related to canonical correlation analysis (CCA). CCA defines coordinate systems that optimally describe the cross-covariance between two datasets while PCA defines a new orthogonal coordinate system that optimally describes variance in a single dataset. A PCA transformation is thus a special orthogonal transformation that transforms the data to a new coordinate system such that the greatest variance by some projection of the data comes to lie on the first coordinate (called the first principal component), the second greatest variance on the second coordinate, and so on. Hence, PCA seeks the linear combinations of the original variables such that the derived variables capture maximal variance. The principal components are therefore uncorrelated. Furthermore, the derived principal components sequentially capture the maximum variability among the data vectors thereby providing minimal information loss;

The directions of the eigenvectors indicate the correlation between the measured variables. So they "form new variables" of the form ax + by + cz, where x,y,z, are the measured ones and a,b,c, are constants. The aim of the PCA is to find these new variables and thereby the dependencies between the measured ones. This

transformation is defined in such a way that the first principal component has the largest possible variance, and each succeeding component in turn has the highest variance possible under the constraint that it is orthogonal to and thereby uncorrelated with the preceding components. The lengths of the eigenvectors thus indicate the importance of the corresponding variable. Very short eigenvectors may therefore be ignored, which will lead to a reduction in the number of variables to be treated. The choice of the number of components is therefore adaptive, based on the result of the PCA.

Consider a data matrix, X, with column-wise zero empirical mean (the sample mean of each column has been shifted to zero), where each of the n rows represents a sample, and each of the p columns corresponds to e.g. sensor output. Mathematically, the transformation is defined by a set of p-dimensional loading vectors W(t) = {w ..., w_p )_(k) that map each row vector x_(i) of X to a new vector of principal component scores tit) = (^ , .,., ί^ )₍._} , given by t_k(i) = JC© · W(t) in such a way that the individual variables of t considered over the dataset successively inherit the maximum possible variance from x, with each loading vector w constrained to be a unit vector. Hence, loading vectors of the principal components are the eigenvectors of the variance-covariance matrix X^TX of the data matrix X, which is assumed to be centered by columns.

The principal components and the corresponding loading vectors are orthogonal in their vector spaces, and thus uncorrelated in statistics, which means the variance explained by each principal component is only from itself. The information contained in each principal component is not overlapped with the other principal components. So the cumulative variance explained by the first several principal components can be calculated directly by summing up the variance explained by each of them. The orthogonality of the principal components comes from the orthogonality of their loading vectors and eigenvectors with different eigenvalues are therefore orthogonal.

In one embodiment of the presently disclosed method, the acquired data is triple-axis proper acceleration, angular orientation, velocity and location of the vehicle sampled over a time period, and a further embodiment of the presently disclosed method further comprises the steps of: selecting one or more subsets of said monitoring data parameters; applying an orthogonal transformation of at least one of said subsets thereby obtaining a set of eigenvectors for said subset; computing a multi-dimensional status model of the component, such as by forming an ellipsoid of said set of eigenvectors; and evaluating the physical condition of the component by comparing the status model to a reference model of the component. By generating a reference such that incoming component monitoring data can be evaluated against a reference model representing the "normal" condition of the component, predictive maintenance can be improved. The presently disclosed method, involving this kind of collection of triple-axis and computing a multi-dimensional status model of the component, present an efficient means of improving predictive maintenance.

In one embodiment the orthonormal transformation is a principal component analysis (PCA) and the eigenvectors correspond to principal components of said PCA. A further possibility is that an abnormal condition in relation to the reference model is when the volume of the status model is greater than the volume of the reference model or when at least a part of the status model diverges from the reference model. A further possible condition for detection an abnormal condition may be the length of one or more of the eigenvectors or loading vectors of the status model exceeding the length of the corresponding eigenvector(s) of the reference model or the direction / orientation of one or more of the eigenvectors of the status model diverging from the direction / orientation of the corresponding eigenvector(s) of the reference model. A further possible condition for detection an abnormal condition may be the ratio of two of the eigenvectors of the status model diverging from the ratio of the two corresponding eigenvectors of the reference model. These conditions have been shown to be efficient means of improving the predictive maintenance in combination with the steps of the presently disclosed method. The time period (time window) for sampling data and subsequently performing a PCA may also be varied. The window size may be varied to analyse the data from different with different goals and approaches to find anomalies of a unit. This finding may be further linked to e.g. the monitoring data being segmented into segments with a length corresponding to a predefined duration as described below.

Counters

As used herein a counter is a variable which contains a numerical value. The value can be integer or continuous. The following can be assumed about counters:

- The initial value of the counter is (at least) 0.

- Counters never decrease. Hence, the value of a counter can be assumed to be non-negative.

- The value of a counter can only be changed at discrete, evenly spaced points in time.

Even though a counter cannot be decreased it can be reset to zero. In the following counters are explained with basis in time, but other measures could also be used correspondingly, e.g. mileage, such as mileage of a vehicle.

Let C_j(t) denote the value of counter y^' at time t. Let t₀, t t₂, . . . denote the distinct times at which a counter can be changed. If a counter C is changed at time /, and t_K where k > but not at the times /,·, j < i < k, then the values between / and /c are either the same as at j (C(i) = C(j)) or linearly interpolated ( C(i) = C(j) + (C(k) - C(j)) /(i-j)/(k-j) ). The choice depends on what makes most sense for the counter. Fig. 1 1 A shows an example for the case of constant continuation (top curve) and linear interpolation (bottom curve). Measurements are shown as circles, the interpolations as diamonds. It can therefore be assumed that for all counters Qand all time points /_/ the value C_j(ti) is defined. Examples of counters can be the number of incidents experienced by a component up to the given time or the accumulated (sum) measurement of the absolute acceleration or a more complicated combination of measurements.

Given counters C_0, C_h ... ,C^and a time point / the average (mean) Aft) of the values of Co_, C ... ,C_kat / is defined by A(t) = C₀(t)+ d(t)+ ... +C_k(t)

The variance of the values of C_0, C_h ... ,C_kdX ί is defined by

V(t) = [ (A(t)-Co(t)f+ (Α(1)-0 ²+ ... + (A(t)-C_k(t)f ]/k Given (a family of) counters C_0, C ... ,C_k which measure the same feature, i.e. it is assumed that the Qhave the same distribution at a given time i . This could for example be the wear of k+1 different vehicles. Consider time points t₀, t t₂, . . . , t_s. and let A(tj) and V(tj) be the average and variance at the time points. From the

measurements of the C, we would like to determine, for every time /_/ in which range [rrij MJ Vne values of the measured feature fall with a certain probability a.

In order to do this, we have to know - or at least assume we know - the distribution of the C_j. Then we can compute the (a/2) and (1 -a/2) quantiles. As an example we use a = 0.8 and assume that the Qare normally distributed with N(A(t,) , V(t,)). The 0.1 and 0.9 quantiles for the standard normal distribution N(0, 1 ) are -1 .281 and 1 .281 .

Transforming to N(A(tj) , V(tj)) this results in m, = A(ti) - 1.281 * (sqrt(V(ti)) and M, = A(t + 1.281 * (sqrt(V(t))

With a probability of 0.8 observations at time f,- will be in the interval [m, MJ. Fig. 12 shows the plot of the same counter from 13 vehicles (black lines), the mean (red) and the upper and lower quantiles /W and m (blue). Most of the counters stay in the middle 80% region all the time, but one line is almost constantly above and one below. The two represent abnormally hard respectively soft treatment of the corresponding vehicle.

From a statistical perspective, the recorded counter values may be seen as forming a time series - this can be exploited for predicting the future outcome. One on the key tools in time series analysis is model fitting, which means fitting a statistical model to the observed values. There is a vast number of potential models types: linear regression, polynomial regression, AR-, MA-, ARMA-, ARIMA-, GARCH-models, etc. The task is to find a model type which describes the observed time series well and to determine the parameter of a specific model of this type. Model fitting can be performed as follows: Assume that the time series cleaned form standstills and idle times. Then the dimension on the x-axis is usage time or mileage.

- Given is series, i.e. a sequence of observations (counter values) y₀, yi,y₂,■■ ·, y_k.

- Analyse the series in order to identify suitable model type (or types).

- For the type identified in the last step, compute a model, i.e. find the model parameters, which best match the observed data. Let z₀, z z₂, . . ., z_k be the values computed by the model.

- Compute the errors (y, - z and their distribution on the given data.

- Use the model to predict the future values z_k+1, z_k+1, ... and the confidence

bands for a given level confidence level a. That is, for each t > k compute £>_fand S_fSuch that b_t < z_t < B_t with probability a . For example the "the counter value at time t is between 120 and 150 with probability 95%".

Fig. 1 1 B shows an example thereof.

The concept of a counter is very useful for determining the probability for a component to fail or a certain incident to happen. When monitoring a number of identical vehicles, one records the value of all counters whenever a component fails (or an incident occurs). From this data a number of statistical parameters (mean, variance, skewness) of the distribution of the counter values at failure can be computed. This can also be applied for updating and/or determining the intact level and/or the broken level of the corresponding component or vehicle as used herein.

For a visualization the horizontal lines for the failure probabilities are shown and the recorded values of the counter of a vehicle. This leads to statements like "The probability or the component to break now is 5%." Often the future trend of the counter value can be predicted as explained above and this can be applied for predicting the maintenance as disclosed herein. Fig. 13A shows an example, where the counter value indicates that the current failure probability is just below 50%.

Fig. 13B exemplifies what can be deducted from a combination of prediction and failure probabilities. At time f the expected average failure probability is 90%. The graphic also shows that there is a chance of at least a (upper confidence line), a failure probability of 90% is already reached at time V.

The concept of counters can advantageously be applied to the presently disclosed systems and methods, for example for estimating the physical condition of a component, thereby helping to estimate the need for maintenance of a component, sub-system and/or machinery. Modelling of mechanical damage counters

Mechanically based "damage counters" are useful when assessing load-based maintenance and warranty, in particular for mechanical damage accumulation, e.g. for land vehicles. Since different components and sub-systems have different criticalities as well as failure modes and mechanisms, the damage accumulation models vary and the damage counters may therefore differ. The workflow described in this paragraph is illustrated in fig. 15.

A decomposition of the system into for example component or group(s) of components can be advantageous because it allows a differentiated analysis of the system's condition. A decomposition can follow existing module definitions, as often the case, or functional sub-systems. The purpose of decomposition is to be able to relate different causes of system failure to specific modules/sub-systems and their use or load history.

Specific knowledge about the failure modes and mechanisms related to the individual sub-systems (e.g. individual components or groups of components) enables targeted measurements and models to estimate the condition and risk for failure of the subsystems / components.

Failures due to a) fracture and b) fatigue are usually distinguished from each other. A fracture of a component typically occurs when the load exceeds the yield strength of that component. Load as well as strength may vary due to variation in for example use patterns, design (load paths), secondary failures (changing load paths), material properties and environmental conditions, as exemplified in fig. 14A. On the other hand, failures due to fatigue can occur well below yield strength load. The damage accumulates over the use with loads between the yield strength and the endurance strength as exemplified in fig. 14B. Failures due to fatigue as well as fracture are naturally subject to uncertainties. Continuous updating and feedback from failures occurring in the field can reduce these uncertainties. However, failures caused by for example misuse, design or production mistakes, aging, (wearout) or environmentally induced failures like for example corrosion, typically cannot be predicted by use and load monitoring.

Based on the knowledge of the failure modes related to the individual sub-systems the influencing use, load and environment factors are typically derived. As for finding these failure modes and mechanisms this part may be subject to continuous revision and updating to improve the understanding of what and how to measure to most accurately damage and rest-life for components and their sub-system.

Subsequent to identification of influencing factors, they can be ranked, for example ranked by importance, measurement difficulty and cost, to reduce the total number of parameters that need to be measured and monitored. This is typically followed by a final decision regarding which sensors are necessary, their specifications and where they should be located. One purpose is to take advantage of synergies to obtain the most accurate data from as few sensors as possible. The translation and post- processing of measurement data to serve the information gathering for multiple subsystems is typically provided to ensure data relevance and quality while striving for an economic solution.

Based on the available measurement data on the influencing parameters for individual sub-systems, different damage counters can be derived. Various damage accumulation models exist which are based on different paradigms. Among those are for example: - Energy-based models

- Models based on fatigues strength tests, for example Wohler / Miner rule

Empirical models for standard machine elements

- Crack initiation and propagation models

Damage accumulation models try to estimate the fatigue of a component and subsequently of the associated sub-system. These theoretical as well as empirical models have limitations and assumptions which include for example test environment, load direction, kind of load (tensile, compressive, shear, etc.). Together with the probabilistic nature of loads and component strengths in general as described under point 2, the verification and validation need large amounts of data and continuous feedback and improvement. Statistical failure distributions can be fitted to evaluate the risk for a failure. Damage models are typically based on load spectrums, and the different loads in a continuous load-time-function are typically classified using a specific counting method. Those methods are described in the standards DIN 45667 and ISO 121 10-2.

The concept of mechanical damage counters can advantageously be applied to the presently disclosed systems and methods for estimating the need for maintenance of a component, sub-system and/or machinery.

Application approaches

In practice there are a number of approaches of applying and using the presently disclosed method that have turned out to be useful in different ways. By repeating steps b) - c) (i.e. acquiring monitoring data, evaluating the physical condition and updating the broken/intact limits) the reliability and precision can be further improved continuously. In particular, if a level representing the validity of the estimation of maintenance condition and/or prediction of need for maintenance is introduced, the validity level may be controlled and changed according to e.g. the significance that is given to the evaluation. Therefore, in one embodiment of the method, the validity level is increased when the intact level and/or the broken level is updated and in a further embodiment the validity level is increased after the physical condition of the component has been evaluated. A further possibility is that the validity level is influenced/calculated based on data from several components in order to improve the statistical basis, possibly from a fleet of vehicles.

Fig. 1 shows a diagram of the physical condition (indicated curve 4) of a vehicle suspension based on data parameters, intact and broken levels, a number of incidents and the validity of the prediction of maintenance of a component. In this example it can be seen that the intact level is set to 0 and the broken level is set to 'infinite'. This can be said to represent an initial situation in which there is no information available of what the component can sustain without breaking. Before any intact levels are available it is possible to use or enable the component for a certain amount of time that serves as a reference before starting to update the intact level. A reference period, wherein the component is considered to work properly, can be used to set an initial intact level. For a vehicle this can be achieved by driving for an amount of time and setting the intact level to the highest calculated level corresponding to a physical condition of a vehicle. As incidents occur, there may be a need for evaluating the physical condition of the component. As an example, at some point in time an incident[1 ] occurs (the incident detection could be e.g. an incident detection to detect isolated incidents that normally clearly exceed normal use). This is represented in fig. 1 by a measured peak 1 (and in the same way as intact[2] at peak 2). At this point the physical condition can be evaluated, e.g. manually. If it is deemed that the component is still intact, the intact level can be increased to the level of the incident. This is represented by the reference level intact[1 ] in the diagram. Similarly, if, upon an incident, it is deemed that the component is broken, the broken level is decreased, which is shown in the diagram as broken[3] at peak 3. The measured physical condition may not always be consistent - for example it is possible that in one evaluation the component is deemed to be broken and later, for the same level, the component is deemed to be intact. Therefore, the algorithm may apply different strategies such as averaging or disregarding isolated incidents when updating the limits. Other alternatives are including data and/or a predefined physical limit for the intact level is set to an intact level of another component and/or the predefined physical limit for the broken level is set to the broken level of another component. A further possibility is to include a third level in addition to the intact and broken levels, having a softer meaning. Preferably, the additional third level represents a level between the intact and broken levels, which basically means in a region where there are normally no assumptions about the need for maintenance. The additional level can for example indicate that a component is likely to be broken but not with the same certitude as the broken level.

In can be further noted in the diagram of fig. 1 that the validity (indicated 5) is increased for each incident that is evaluated. By combining data from measurements of the component and other data from evaluation of the physical condition, the present dynamic model can increase both the accuracy and the validity of the prediction of maintenance of a component. When an incident occurs, the intact and/or broken levels may or may not be updated based on the results. Independent of the result of such an evaluation it can be assumed that some data has been taken into account and that the validity increases. Incident

The incidents that are referred to above could be any type of incident affecting the physical condition of the component. Generally, the concept is developed under the assumption that an incident is detected based on the data parameters exceeding a predefined incident threshold and that the evaluation of the physical condition is triggered by detected isolated incidents. However, the incident may also be referred to as an event triggered by for example a visual observation or another part of the system communicating data.

In one embodiment the incident corresponds to exceeding a predefined threshold in an x, y, or z coordinate of the component. The orthonormal transformation like the PCA is typically applied to monitoring data to determine abnormal patterns in the driving or behaviour. Incident detection on the other hand is applied to detect isolated incidents that normally clearly exceed normal use. Incidents can be related to e.g. accelerations and angular rates of the vehicle.

For accelerations and angular rates incidents can be triggered (detected) in both negative and positive directions. The magnitude of an incident is defined as the value normalized with the incident threshold value to provide a dimensionless relative parameter where the sign indicated the direction of the incident. Threshold values are provided for both positive and negative directions, e.g. for both positive and negative thresholds for x,y,z acceleration and for pitch, roll, yaw. The positive and negative threshold values may be symmetrical, e.g. x_min = -x_max-

The monitoring data may e.g. be segmented into segments with a length corresponding to a predefined duration. In one embodiment the predefined duration is between 1 second and 1 minute, or between 1 minute and 60 minutes, or between 15 minutes and 60 minutes, or between 30 minutes and 2 hours, or between 1 hour and 4 hours, or between 1 hours and 8 hours, or between 1 hour and 12 hours, or between 1 day and 2 days, or longer than 2 days, such as 1 , 2, 3, 4, 5, 6, 7, 8, 9 or 10 minutes. Incidents can then be detected for each segment by analyzing each segment and identifying values exceeding the threshold values specified in the reference model. If a value is found that exceeds one of the threshold values, an incident is detected. The detected incidents may then be categorized according to their magnitude. The signal with the largest value relative to the threshold value is identified as the signal of the incident. A new incident can preferably not be triggered at least 1 second after a threshold value is exceeded.

The sample frequency for acquiring data is preferably at least 50 Hz, more preferably at least 100 Hz, or most preferably at least 200 Hz, or 1 -50 Hz, or at least 10 Hz, or at least 20 Hz, or at least 30 Hz, or at least 40 Hz,or at least 300 Hz, or at least 400 Hz, or at least 500 Hz.

The monitoring data, e.g. three acceleration and rate signals, may be high-pass filtered initially, e.g. using a filter with a cut-off frequency of 0.1 Hz, in order to remove any dc content. A reference model with threshold values can e.g. be generated based on the statistical distribution of acceleration and angular rate peaks in reference data, which is representative for normal use of the vehicle. Peaks of each of the acceleration and angular rate signals (both local max and min) in the reference data can be extracted. The absolute peak values may then be collected in a vector for each of the three acceleration and angular rate axes. The assembled peak vectors may be fitted with a cumulative exponential probability distribution P(acc). The threshold values can then be defined as the peak acceleration giving P(acc)=0.9999. I.e. in that case 1 out of 10.000 acceleration cycles is expected to trigger an incident during normal operation. Wear and accumulated wear

In one embodiment of the presently disclosed method the calculation of the physical condition further comprises the step of accumulating an estimated wear of the component possible based on principal component analysis including a reference model as described above. The idea is similar to that of the incident based evaluation described above but opens for additional parameters and combinations hereof that can be used to further improve and increase both the accuracy and the validity of the prediction of maintenance of a component. The presently disclosed method can be said to be developed around the fact that the wear of a component or vehicle is nonlinear. The calculated accumulated wear takes into account this non-linearity.

Furthermore, in one embodiment the method further comprises the step of introducing a predefined physical limit for a worn level above which the component is assumed to be worn, which is explained in the following example. Counters, in particular mechanical damage counters, as described herein can advantageously be applied for estimation of wear and accumulated wear.

Fig. 2 shows a diagram of physical condition, in this example corresponding to the wear of a vehicle component, intact and broken levels, and the validity of the prediction of maintenance of a component. The broken level is supplemented by a "worn" level in this case. The acquired data is used to calculate the physical condition of the component, which is then accumulated (shown as a curve in the diagram), i.e.

illustrated by means of counters. At some point (t₀) the physical condition of the component is evaluated. The component is deemed to be intact and the intact level intact[t₀] is therefore increased. At another point in time (ti) the physical condition of the component is evaluated again and considered to be worn but not broken. For this reason the component is changed and therefore the accumulated wear is reset to 0. At another point in time t₂ the component is changed, possibly for another reason such as an incident, age etc. Also in this case the accumulated wear is reset to 0. The intact level does not have to be updated; however the validity 5 can be updated since it has been confirmed that a certain level of accumulated wear does not correspond to a worn or broken component. At t₃ the calculated physical condition of the component exceeds the worn level but no inspection is made. Possible reasons for not inspecting when a component is calculated to be worn could be that other vehicles are prioritized or that the consequences of failure are not severe. At t₄ the component breaks and is therefore replaced. The broken level is updated to the calculated physical condition.

The accumulated wear as illustrated in fig. 2 can be seen as forming time series - this can be exploited for predicting the future outcome by modeling, e.g. fitting a statistical model to the observed values as disclosed previously and exemplified in figs. 1 1 B, 13A and 13B, i.e. be used for estimating the future failure or incident probability by calculating the probability of the counter value being within an certain counter value interval at time t. Thereby a quantitative measure of the need for maintenance of the corresponding machinery is provided.

Examples of estimation of load and wear are furthermore disclosed in pending application WO 2015/139709.

Further parameters

In relation to the accumulated wear it can be noted that other parameters than wear could be handled in the same way, or combined with the wear. One such factor that can be taken into account, either as a separate parameter or in combination with the wear, is the time the component or vehicle is not used, referred to as the stand-still parameter or accumulated stand-still parameter. This can also be estimated by means of counters. In some cases it is more damaging when a component or vehicle is not used, especially if the component/vehicle is not used at all. Vehicle components that are not used may be more susceptible for corrosion and may suffer from slow leakage of e.g. lubricant. For this reason the presently disclosed method may be applied on parameters corresponding to measured stand-still time. In a further embodiment, the stand-still parameter is combined with other parameters such as abovementioned wear. In some cases the actual condition of a component may be best reflected by adding the accumulated wear and longest consecutive stand-still period. Detection of stand-still, on-road and off-road driving are exemplified in pending application WO 2015/139709.

Similarly, the acquired monitoring data parameters can be used to detect particular activities or states that can be grouped into different categories that add to the wear of the component. As an example, the monitoring data parameters could be used to calculate and accordingly determine that the vehicles is used for on-road or off-road driving. Off-road driving can be calculated based on the principal component analysis. Instead of accumulating the exact wear, the collected data divides the use into these categories which are given individual score/significance/weight to be taken into account when accumulating wear. In one embodiment the calculation of the physical condition of the component takes into account a part of use of the vehicle corresponding to off- road use, one part corresponding to on-road use and the remaining part corresponding to stand-still. Another parameter that has an influence of some parts of a vehicle is the number of cold starts. In another embodiment the calculation of the physical condition of the component takes into account the periods in which the component is not used as a parameter.

All of the above parameters may be associated with individual weights. This enables a possibility to give certain parameters higher significance than others. Typically this approach involves statistical feedback in the sense that if some of the parameters are shown to have a more significant impact to the failure of a component.

Prediction of the need for maintenance

In line with the abovementioned concept, the present disclosure further relates to a method for predicting the need for maintenance of machinery, such as a vehicle or a fleet of vehicles, comprising the steps of: providing monitoring data parameters acquired over a time period (possibly indicative of movement, acceleration and/or angular orientation of the component and/or the vehicle) corresponding to physical impact and/or stress of one or more components of said machinery and calculating the physical condition of said component(s) based on the data parameters; providing at least two reference levels of said components of said machinery according to the method above, thereby obtaining at least an intact level and a broken level of each component, comparing the physical condition of each component with the

corresponding intact level and broken level of said component, and predicting the need for maintenance of the machinery by determining the intactness of each component. The method can be seen as a way of improving conventional preventive maintenance, which is primarily based on data collection, fixed limits, prioritizing and common sense. The presently disclosed method involves multidimensional analysis, dynamic modeling and a progressive way of approaching the challenges of preventive maintenance.

Predictive maintenance can be said to include direct measurement of the equipment and the equipment is scheduled for maintenance based on monitoring of the equipment condition. Predictive maintenance is thus designed to help determine the condition of in-service equipment in order to predict when maintenance should be performed and when implemented properly it can provide substantial cost savings and higher equipment reliability. In the presently disclosed method for predicting the need for maintenance of machinery, the need for maintenance is predicted by introducing at least two reference levels that are continuously improved by evaluating them against actual data from inspections or other checks. The intactness of a component may therefore be expressed as the probability of the component being intact in relation to the corresponding intact level and broken level. A validity level may be used to provide an indication of the reliability of the prediction of maintenance; the validity level can be increased when the intact level and/or the broken level is updated, or after the physical condition of the component has been evaluated. In one embodiment the method comprises the step of evaluating the validity of the intact level(s) and the broken level(s) before predicting the need for maintenance. Also in the present method the different components may have different impact of the prediction of maintenance, and therefore, in one embodiment, the method further comprises the step of evaluating the significance of each component before predicting the need for maintenance of the machinery.

As can be seen in table 1 and 2 the method is useful when data for need for maintenance is collected and presented for a number of vehicles or a fleet of vehicles. In table 1 it can be noted that there is a physical condition calculated according to the presently disclosed method. The calculated physical condition may be for example principal component analysis as described above (and/or accumulated wear as in table 2). As can be seen the user of the system will in this embodiment be able to see not only the physical condition of a component/vehicle and an estimation whether it is intact or broken, but also the probability that the evaluation is correct and a validity. The probability can in this context be linked to e.g. a specific event associated with the physical condition. For example if the physical data corresponds to a triple-axis analysis of the suspension, one level or pattern may correspond to driving into a large hole or bump or a shock of some kind. If this event is known for causing failure to the component/vehicle the probability of the component/vehicle being broken is higher. One embodiment of the presently disclosed method further comprises the step of evaluating the severity of each component before predicting the need for maintenance of the machinery. In the example in table 1 the two reference levels "broken" and "intact" are supplemented by a third level "possibly broken" according to the description above. It can also be noted that there is a validity of the assumed condition, which is based on the history of evaluation of the physical condition and updates of intact, broken and "possibly broken" levels according to the presently disclosed method. As can be seen there is also a proposed action to take based on the displayed information - therefore, in one embodiment the method further comprises the step of providing a recommendation regarding a recommended use of the machinery and/or a

recommended action item of the machinery.

Table 1 : Overview of the suspensions for a number of vehicles and the probability and validity that the suspension is broken

Raw

Assumed

calculated Probability Validity Action condition

wear

Vehicle 1 200 97% Broken 60% Change

Vehicle 2 192 83% Broken 60% Change Possibly

Vehicle 3 160 50% 60% Inspect

broken

Vehicle 4 120 25% Intact 60% None

Vehicle 5 50 20% Intact 60% None

Vehicle 6 34 15% Intact 60% None

Vehicle 7 15 7% Intact 60% None

Table 2: Overview of the components for a number of vehicles and the probability and validity that the component is broken.

System

The methods disclosed herein may in particular be applied to vehicles as described above, but they may also be applied to equipment, machinery and/or parts in general. Therefore, a further embodiment relates to a system for incorporation into equipment, such as a vehicle, e.g. in the form of a system for predicting the need for maintenance, for attachment to a craft / vehicle and for monitoring the condition of said vehicle, comprising at least one inertial measurement unit configured to measure the triple-axis proper acceleration, velocity and angular orientation of the chassis of the vehicle sampled over a time period, at least one GPS receiver for measuring the location of the vehicle, a computer comprising memory and a processing unit configured for executing any of the methods described herein for assessing the condition of said vehicle.

An inertial measurement unit is a sensor unit and/or an electronic device that is configured to measure velocity, orientation and gravitational forces or acceleration and/or a change in orientation of a moving object whereto the unit is attached, typically using a combination of movement detectors in the form of accelerometers and gyroscopes, sometimes also magnetometers.

In general data processing and data analysis is not limited to the use of basic data as input to the model generation. It is also possible to use the results and/or

interpretations produced by a first layer evaluation as the input for a second-layer evaluation, which may provide interpretation on a higher level of abstraction. One example is the use of labelled data, where the data first have been classified (and thereby) labelled and secondly reference models and/or vehicle condition models can be generated based on labelled data. A third step may then be to identify outlier clusters and use these as input for further evaluation, e.g. in combination with data from additional sensors on the vehicle or CAN bus data from electronic control units in the vehicle.

The presently disclosed methods for determining one or more reference levels corresponding to the maintenance condition of a component of machinery and for predicting the need for maintenance of machinery may be combined with a vehicle monitoring system configured to provide the condition of the vehicle in real-time, e.g. during driving of the vehicle. This may be employed to provide the driver with real-time information about the vehicle, e.g. by displaying the condition of the vehicle in a display in the vehicle. This may be a strong tool to prevent misuse of vehicles and machinery because the online feedback may ensure that the vehicle is operated within predefined operational limits, limits that may be determined by feeding the appropriate reference model(s) to the monitoring system. In one embodiment the system comprises three main components; the real time system, the offline analysis tool and the parameter calculation module. The real time system is the system which is installed in the vehicle. The system consists of a sensor module for measuring the herein described monitoring data, a processing module, a local storage module and a driver interface module. The measured data is stored in the local storage module, and processed in the processing module. The processing module may calculate indicators and alarms related to wear, fuel consumption etc. These indicators and alarms can be displayed in real time to the driver by the driver interface module, and indicate to the driver how to change his/her behavior. The calculation of the indicators and alarms are based on vehicle specific threshold parameters. Thus, the parameters are determining at which levels a given alarm or indication is triggered. Hence, the driver behavior can be adjusted by adjusting the parameters, such that a required behavior is achieved. The parameters are fed to the real time system from the parameter calculation module. The offline analysis system is a tool configured for analyzing single vehicle and vehicle fleet performance. The offline analysis system displays the data that is stored in the database in structured manner. The system enables different stakeholders to assess the performance of the fleet and make operational decisions. A wide range of stakeholders may be using the offline analysis tool, i.e. executive staff, workshop managers, logistic managers, driving instructors etc. Some of the stakeholders will have privileges to request a behavior change, i.e. request a behavior that increases the service intervals. This is done using the parameter calculation module.

The parameter calculation module is configured to translate qualitative behavioral changes requested by the stakeholders to quantitative changes in the parameters that can be fed back to the real time system. The parameter module allows the user to adjust certain performances around a baseline level, i.e. the stakeholder can by increase or decrease the service interval around the baseline level. The system can be configured to immediately show the consequences of one performance change on the remaining performances, i.e. if the frequency of service intervals is increased the expected mean velocity is increased. The required qualitative changes are translated to variations of the parameters that can be fed back to the real time system.

Fig. 9 shows an illustration of the driver and fleet manager advice that can be provided by the presently disclosed system and method. The monitoring system provided vehicle monitoring data that can be analyzed and incorporated into a context of reference models representing different uses of the vehicle, e.g. careful use, normal use, outside normal use and excessive use and extreme events. Different tours or sequences of driving events (indicated by the worm-like lines) can be analyzed in the context of the different reference models. If most of the tour is within the normal use of the vehicle, it can be deemed to be acceptable. However, if most of the tour is outside normal use it may pose a problem.

The system may furthermore comprise a wireless transmitter. This may be provided to transmit the acquired data and/or processed data and/or the condition of the vehicle to a central server and/or database and/or data analysis center. Data may be transmitted continuously or whenever the vehicle is within range of a plurality of hotspots forming a wireless data collecting system. Whenever a vehicle comes close to a hotspot, the data from the vehicles memory is transferred to the hotspot. The data collected by the wireless data collection system may be transferred a database where it may be long- term stored for further handling. The system may further comprise one or more data modelling systems configured to analyze the stored data in the database with respect to different criteria, e.g. vehicle wear, driver performance, fleet availability etc. Each application may require one or more reference models, possibly labelled reference models. Thus, the relevant data may be retrieved from the database for processing. An exemplary vehicle monitoring system is illustrated in fig. 10. The NVO core system comprises the a vehicle support unit (VSU) comprising a vehicle monitoring system with sensors and an onboard computer including a wireless transmitter for transmitting monitoring data to a wireless data collection system / hotspot storage wherefrom the data is distributed via a cloud service. From the cloud data validation, processing and storage and a number of reference models and vehicle condition models can be computed. Stored data can also be retrieved by a data acquisition and transport system (DATS) for generation of models and assessing the condition of the vehicle to provide a result that can be shown to a user. Another option is that the vehicle comprises a vehicle support unit (VSU) and a DATS for real-time generation and interpretation of models such that results can be shown to a user in the vehicle, e.g. the driver.

Processing may include analysis, evaluation and interpretation. E.g. a reference model is computed and evaluated against incoming data for online / real-time analysis and/or evaluated against data stored in the database. Evaluations may include statistical analysis of a fleet, a single vehicle or driver or a group of drivers, a specific description of the condition of vehicle, the registration of abnormal events and their severity, and/or a real-time advice to the driver as a reaction to an incident.

In a further embodiment the monitoring system that is used in combination with the method for determining one or more reference levels corresponding to the

maintenance condition of a component of machinery and for predicting the need for maintenance of machinery further comprises one or more additional movement detectors, such as accelerometer, gyroscope, or initial measurement unit, mounted on the chassis of the vehicle or on one or more internal moving parts of the vehicle for measuring the movement, acceleration and/or angular orientation of said part(s), e.g. the engine, bearings, suspension, etc. The monitoring system may further be adapted to acquire data from the vehicle's internal electronic control units such as the engine control unit, the powertrain control module, the transmission control unit, antilock braking control unit, cruise control unit, or power steering unit. This type of data acquisition are typically standardised via the CAN bus standard or normal Ethernet based communication. Other types of input could be video imaging the road or manual input provided by the driver. The herein described detailed modelling may be more precise if the acquired data can be defined according to the same reference coordinate system. That typically requires a very low drift in the data outputted from the sensors in the car. Drift in orientation typically arises from temperature variations around the sensor. Many consumer electronic devices comprise both accelerometers and gyroscopes, but they also typically account for an unacceptable drift if used for the presently disclosed purpose of vehicle monitoring. The present inertial measurement unit(s) may therefore

advantageously be temperature controlled. Further, it may be provided with a static accuracy of < ±1 °, preferably < ±0.5°, with regard to pitch and/or roll. Furthermore, the inertial measurement unit(s) preferably has a dynamic accuracy of < ±3 °, preferable < ±2.0° with regard to pitch and roll. Furthermore, the inertial measurement unit(s) furthermore has a repeatability of < 0.4° or < 0.3 ° or < 0.2 °, and/or a resolution less than < 0.3° or < 0.2° or < 0.1 °. The long term drift of the present inertial measurement unit is therefore preferably neglectable.

The presently disclosed methods and systems may for part of a new type of monitoring, surveillance and maintenance of machinery which can be termed iHUMS for intelligent health and usage monitoring system. iHUMS can be applied to everything from the single automobile to a large fleet of vehicles or the wind turbines of a wind turbine farm. iHUMS provides three levels of analysis. The first level relates to the single unit which can be monitored by monitoring 1 ) movement from triple-axis acceleration, angular orientation and optionally velocity and location, e.g. from an external sensor mounted on the unit, 2) internal data provided directly from the unit, i.e. data that is generated by internal sensors of the unit, e.g. CANBUS data, etc., and 3) estimation of load and wear as described previously. Data from 1 ), 2) and 3) can be assembled and analyzed, e.g. in real time, and features can be extracted to generate status models within 1 ), 2) and 3) and compared to corresponding reference models within 1 ), 2) and 3), respectively, showing e.g. normal behavior and anomalies of the single unit.

The second level relates to comparison of data from the single unit across 1 ), 2) and 3) thereby possibly revealing additional features and patterns across the collected datasets. The third level relates to comparison of a plurality of units providing surveillance and monitoring of an entire fleet of units. Logistics and maintenance can then be optimized to obtain large cost reductions in the fleet management.

Detailed description of drawings

Fig. 1 shows a diagram of the physical condition of a vehicle suspension based on the data parameters, intact and broken levels, a number of incidents and the validity of the prediction of maintenance of a component. Fig. 2 shows a diagram of physical condition, in this example corresponding to the wear of a vehicle component, intact and broken levels, and the validity of the prediction of maintenance of a component. The details of fig. 1 and fig. 2 are explained in the detailed description of the invention above.

Fig. 3 shows an overview of the scheduling of maintenance and status of a component of a number of vehicles. The bars that are present in some of the fields may represent various parameters and relation between different types of use e.g. off-road vs. on-road driving. The inspections (which may serve as evaluation of physical condition according to the presently disclosed method) are also shown in some of the fields of the overview, as well as indications of when components have been changed. In fig. 4 the details of one specific vehicle is shown. The figure discloses details about incidents and other driving details such as on-road and off-road driving. There are also details about specific components of the vehicle. Furthermore there are information about the maximum stand-still (longest consecutive stand-still period), driving distance, accumulated wear, number of incidents, number of hard (severe) incidents and specific information about the most severe incident.

Fig. 5 and fig. 6 show examples of an overview of a number of vehicles and their status with respect to wear, incidents and stand-still. It can be seen that there are information regarding the maximum stand-still (longest consecutive stand-still period), driving distance, accumulated wear, number of incidents, number of hard (severe) incidents and specific information about the most severe incident that may be taken into account in the predicting of need for maintenance. The diagram also comprises information about when the status was last updated and when the last inspection took place. Fig. 7 shows one embodiment overview of the herein disclosed system for predicting the need for maintenance of machinery. In this example, data is collected from one or several of NVO sensors, GPS, MMI, vehicle HUMS or other signals. The NVO core system comprises a vehicle support unit (VSU) and a data acquisition and transport system (DATS) as well as a (long term) storage unit. In this embodiment the system is linked to a support system (l-TLS) that is responsible of grouping functionalities that together provides the user with a specific capability within one of the NVO themes. The module can involve a number of users. Fig. 8 shows a theme configuration i.e. grouping of related functions. The illustration can be said to explain inter alia the hierarchical relationship between themes, modules, functionality and functions. A theme in the context of fig. 8 is a grouping of modules that are naturally related. A module in the context of fig. 8 is a grouping of

functionalities that together provides a user with a specific capability within one of the NVO themes. The module can involve users or decision makers from different parts of the costumer organisation. A functionality within the context of fig. 8 provides the ability for a user to perform a specific task. A functionality ensures that all relevant and necessary functions are integrated in the functionality. A functionality can be used individually, but is typically of most relevance in combination with other functionalities within the module. A within the context of fig. 8 provides the user with the ability to perform a single action that is relevant to the user, e.g. by providing a specific piece of decision support information, by sorting information in a specific order, by enabling input of a specific piece of information or by enabling drill-down to more detailed information.

Claims

A method for predicting the need for maintenance of machinery, such as a vehicle or a fleet of vehicles, comprising the steps of:

- providing monitoring data parameters acquired over a time period

corresponding to physical impact and/or stress of one or more components of said machinery, wherein the monitoring data is indicative of movement, acceleration and angular orientation of the component,

- calculating the physical condition of said component(s) based on the data parameters;

- providing at least two reference levels of said components of said

machinery thereby obtaining at least an intact level and a broken level of each component, wherein the at least two reference levels are obtained by calculating the physical condition of the component based on the data parameters evaluating the physical condition of the component after said time period, and

• updating the intact level to the calculated physical condition if the component is intact and the calculated physical condition exceeds the intact level;

or

• updating the broken level to the calculated condition if the

component is broken and the calculated physical condition is below the broken level,

- comparing the physical condition of each component with the corresponding intact level and broken level of said component, and

predicting the need for maintenance of the machinery by determining the intactness of each component.

The method of claim 1 , wherein the intactness of a component is the probability of said component being intact in relation to the corresponding intact level and broken level.

The method of any of preceding claims 1 -2, further comprising the step of evaluating the validity of the intact level(s) and the broken level(s) before predicting the need for maintenance.

4. The method of any of preceding claims 1 -3, further comprising the step of evaluating the significance of each component before predicting the need for maintenance of the machinery.

5. The method of any of preceding claims 1 -4, further comprising the step of

evaluating the severity of each component before predicting the need for maintenance of the machinery.

6. The method of any of preceding claims 1 -5, further comprising the step of

providing a recommendation regarding a recommended use of the machinery and/or a recommended action item of the machinery.

7. The method according to any of preceding claims 1 -6, wherein the reference levels are obtained according to the method of any of claims 18-48.

8. A system for predicting the need for maintenance of machinery, such as a

vehicle, comprising

- at least one inertial measurement unit configured to measure the triple-axis proper acceleration, and angular orientation of the machinery or chassis of the vehicle sampled over a time period,

- at least one GPS receiver for continuously measuring the location of the machinery or vehicle,

- a computer comprising memory and a processing unit, configured for

executing the method according to any of preceding claims 1 -7 for predicting the need for maintenance of the machinery or vehicle.

9. The system according to claim 8, further comprising one or more additional movement detectors, such as accelerometer, gyroscope, or initial measurement unit, mounted on one or more internal moving parts of the vehicle for measuring the movement, acceleration and/or angular orientation of said part(s).

10. The system according to any of preceding claims 8-9, wherein the system is configured to continuously measure the velocity of the vehicle based on triple axis proper acceleration data and/or location data.

1 1 . The system according to any of preceding claims 8-10, wherein the inertial measurement unit(s) has a static accuracy of less than ±1 °, preferably less than ±0.5°, pitch and roll. 12. The system according to any of preceding claims 8-1 1 , wherein the inertial measurement unit(s) has a dynamic accuracy of less than ±3 °, preferable less than ±2.0° pitch and roll.

13. The system according to any of preceding claims 8-12, wherein the inertial measurement unit(s) has a neglectable long term drift.

14. The system according to any of preceding claims 8-13, wherein the inertial measurement unit(s) has a repeatability of less than 0.2 °. 15. The system according to any of preceding claims 8-14, wherein the inertial measurement unit(s) has a resolution less than 0.1 °.

16. The system according to any of preceding claims 8-15, further comprising a wireless transmitter.

17. The system according to any of preceding items 8-16, wherein the evaluated condition of the vehicle is transmitted to a server system by means of a wireless transmitter. 18. A method for determining one or more reference levels corresponding to the maintenance condition of a component of machinery, such as a vehicle, comprising the steps of:

a) providing predefined physical limits for

i. an intact level corresponding to a physical limit below which the component is assumed to be functional; and

ii. a broken level corresponding to a physical limit above which the component is assumed to be broken;

b) acquiring monitoring data parameters in a time period (indicative of

movement, acceleration and/or angular orientation of the component and/or the vehicle) corresponding to physical impact and/or stress of the component and calculating the physical condition of the component based on the data parameters;

c) evaluating the physical condition of the component after said time period, and

- updating the intact level to the calculated physical condition if the

component is intact and the calculated physical condition exceeds the intact level;

or

- updating the broken level to the calculated condition if the component is broken and the calculated physical condition is below the broken level.

19. The method according to claim 18, further comprising the step of iteratively performing steps b) - c)

20. The method according to any of the preceding claims 18-19, wherein a validity level is increased when the intact level and/or the broken level is updated.

21 . The method according to any of the preceding claims 18-20, wherein the

validity level is increased after the physical condition of the component has been evaluated.

22. The method according to any of the preceding claims 18-21 , wherein the

validity level is calculated based on statistical data from determining the reference levels of a plurality of components and/or a fleet of vehicles.

23. The method according to any of the preceding claims 18-22, wherein the

predefined physical limit for the intact level is 0 and/or the predefined physical limit for the broken level is set to 'infinite'.

24. The method according to any of the preceding claims 18-23, wherein the

predefined physical limit for the intact level is set to an intact level of another component and/or the predefined physical limit for the broken level is set to the broken level of another component.

25. The method according to any of the preceding claims 18-24, further comprising the steps of: - selecting one or more subsets of said monitoring data parameters;

- applying an orthogonal transformation of at least one of said subsets

thereby obtaining a set of eigenvectors for said subset;

- computing a multi-dimensional status model of the vehicle, such as by

forming an ellipsoid of said set of eigenvectors; and

- evaluating the physical condition of the component by comparing the status model to a reference model of the vehicle.

26. The method according to claim 25, wherein the orthonormal transformation is a principal component analysis (PCA) and the eigenvectors correspond to principal components of said PCA.

27. The method according to any of claims 25-26, wherein an abnormal condition of the vehicle is defined as the volume of the status model being greater than the volume of the reference model.

28. The method according to any of claims 25-27, wherein an abnormal condition of the vehicle is defined as at least a part of the status model diverging from the reference model.

29. The method according to any of claims 25-28, wherein an abnormal condition of the vehicle is defined as the length of one or more of the eigenvectors or loading vectors of the status model exceeding the length of the corresponding eigenvector(s) of the reference model.

30. The method according to any of claims 25-29, wherein an abnormal condition of the vehicle is defined as the direction / orientation of one or more of the eigenvectors of the status model diverging from the direction / orientation of the corresponding eigenvector(s) of the reference model.

31 . The method according to any of claims 25-30, wherein an abnormal condition of the vehicle is defined as the ratio of two of the eigenvectors of the status model diverging from the ratio of the two corresponding eigenvectors of the reference model.

32. The method according to any of the preceding claims 18-31 , wherein the time period is between 1 second and 1 minute, or between 1 minute and 60 minutes, or between 15 minutes and 60 minutes, or between 30 minutes and 2 hours, or between 1 hour and 4 hours, or between 1 hours and 8 hours, or between 1 hour and 12 hours, or between 1 day and 2 days, or longer than 2 days.

33. The method according to any of the preceding claims 18-32, wherein data are acquired with a predetermined sample frequency of preferably at least 50 Hz, more preferably at least 100 Hz, most preferably at least 200 Hz, or 1 -50 Hz, or at least 10 Hz, or at least 20 Hz, or at least 30 Hz, or at least 40 Hz,or at least 300 Hz, or at least 400 Hz, or at least 500 Hz.

34. The method according to any of the preceding claims 18-33, wherein the

evaluation of the physical condition is triggered by detected isolated incidents.

35. The method according to claim 34, wherein the incident is detected based on the data parameters exceeding a predefined incident threshold.

36. The method according to claim 34, wherein the incident is detected based on visual observation.

37. The method according to any of claims 34-36, wherein the incident corresponds to exceeding a predefined threshold in an x, y, or z coordinate of the

component.

38. The method according to any of the preceding claims 18-37, wherein the

physical condition is estimated by means of a counter or a combination of counters, a counter such as a mechanical damage counter.

39. The method according to any of the preceding claims 18-38, wherein

calculating the physical condition further comprises the step of accumulating an estimated wear of the component.

40. The method according to claim 39, wherein the wear is non-linear.

41 . The method according to any of claims 38-40, further comprising the step of introducing a predefined physical limit for a worn level above which the component is assumed to be worn.

42. The method according to any of the preceding claims 18-41 , wherein the

calculation of the physical condition of the component takes into account a part of use of the vehicle corresponding to off-road use (the remaining part corresponding to on-road use and/or stand-still).

43. The method according to any of the preceding claims 18-42, wherein the

calculation of the physical condition of the component takes into account the periods in which the component is not used as a parameter.

44. The method according to any of the preceding claims 18-43, wherein the

monitoring data parameters are assigned specific significances/weight.

45. The method according to any of the preceding claims 18-44, wherein the

detected incidents, the estimated wear of the component and the periods in which the component is not used are assigned specific significance/weight.

46. The method according to any of the preceding claims 18-45, wherein driving hours, idling hours and number of cold starts are taken into account.

47. The method according to any of the preceding claims 18-46, wherein the

component of machinery is a brake, steering, suspension, engine, gear, tires or any other machinery component of a vehicle.

48. The method according to any of the preceding claims 18-47, wherein the

vehicle is a vehicle for movement of personnel or goods, such as a vehicle on tracks or wheels selected from the group of: a train, a car, a truck, an off-road vehicle, a military vehicle, a motorcycle, a helicopter, an airplane, a harvester, a combat vehicle, or equipment such as an excavator, a forwarder, a loader, or a tractor.