Search

EP-4423581-B1 - SYSTEM AND METHOD FOR DETECTING ANOMALOUS SYSTEM BEHAVIOUR

EP4423581B1EP 4423581 B1EP4423581 B1EP 4423581B1EP-4423581-B1

Inventors

  • HORRY, Samuel
  • MAI, Hans-Heinrich
  • NICKLIN, Richard Edward John

Dates

Publication Date
20260506
Application Date
20221027

Claims (14)

  1. A computer-implemented method for predicting failure of an engineering asset based on real-time data, the method comprising: receiving a data record comprising data on the engineering asset collected from a plurality of sensors (50, 52, 54, 56, 58) at time t ; generating, using a trained machine learning algorithm (32), a probability P F that the received data record indicates that the engineering asset is in a faulty state; determining what number of data records received in a look-back time L t are indicative of the engineering asset being in a faulty state, wherein the look-back time L t is a time period occurring before the time t at which the data record was collected; predicting a probability of the engineering asset failing during a horizon time H t , wherein the horizon time H t is a time period after time t at which the data record was collected; wherein the predicting step comprises implementing a Bayes forecasting model to predict the probability of failure within the horizon time H t based on the generated probability P F and the number of data records which were determined to be faulty within the look-back time L t comparing the predicted probability to a failure threshold and when the predicted probability exceeds the failure threshold, outputting at least one of an alert to act as a decision aid to a user and a signal to trigger an automated self-protection protocol (34) within the asset, wherein the self-protection protocol adjusts one or more components within the engineering asset.
  2. The method of claim 1, further comprising generating a fault value F i,t for each data record i at time t , from F i , t = 0 , P F < P th 1 , P F ≥ P th where P F is the probability of the data record being faulty and P th is a probability threshold.
  3. The method of claim 1 or claim 2, further comprising grouping the received data record with multiple previously received data records to form an envelope of data records, wherein each envelope has an envelope length (e L ) and contains a sequence of all data records recorded during a time interval equal to the envelope length.
  4. The method of any one of the preceding claims wherein determining what number of data records received in the look-back time are indicative of the engineering asset being in a faulty state comprises determining what number of envelopes received in the look-back time are indicative of the engineering asset being in a faulty state.
  5. The method of claim 3 or claim 4, comprising determining that an envelope is indicative of the engineering asset being in a faulty state by calculating a sick rate of the data envelope from a normalised sum of each fault value for each data record within the envelope; comparing the sick rate to a sick rate threshold; and when a sick rate of the data envelope equals or exceeds the sick rate threshold determining that the data envelope is indicative of running in a faulty state.
  6. The method of any one of the preceding claims, wherein the Bayes forecasting model is defined as: P F H t n Lt = P n Lt F H t × P F H t P n Lt where F Ht is the failure of the system within the time period equal to H t following the moment of evaluation t, n Lt is the number of faulty records or envelopes within the look-back time L t prior to the moment of evaluation t , P(F Ht |n Lt ) is the probability of failure within a time period time H t after the moment of evaluation t given n records or envelopes classed as faulty in the time L t prior the moment of evaluation t , P(n Lt |F Ht ) is the probability of n records or envelopes classed as faulty in look back time L t prior to the moment of evaluation t given that a failure is known to have occurred within time H t after the moment of evaluation t , P(F Ht ) is the probability of failure in any time equivalent in length to the horizon time H t and P(n Lt ) is the probability of n records or envelopes classed as faulty in any time equivalent in length to the look-back time L t .
  7. The method of any one of the preceding claims, wherein the Bayes forecasting model comprises a plurality of sub-models, one for each of a plurality of different look-back times ( L 0 , L 1 , L 2 ,... L j ) and the method comprises, for each sub-model: determining what number of data records or envelopes ( n L0 , n L1 , n L2 ,... n Lj ) received in the corresponding look-back time ( L 0 , L 1 , L 2 , ... L j ) are indicative of the engineering asset being in a faulty state.
  8. The method of any one of the preceding claims, further comprising training the machine learning algorithm by receiving multiple data records for at least one engineering asset which corresponds to the engineering asset for which failure is to be predicted, wherein the multiple data records comprise data previously collected from the plurality of sensors at a sequence of times prior to failure of the least one engineering asset; and classifying each of the multiple data records as either indicative of the at least one engineering asset running in an acceptable state or running in a faulty state.
  9. The method of claim 8, wherein classifying each of the multiple data records as indicative of running in a faulty state comprises obtaining a minimum remaining running time, MRRT, which is a time period before failure of the at least one engineering asset; and classifying each of the multiple data records which is within the minimum remaining running time as indicative of running in a faulty state.
  10. The method of claim 9, wherein classifying each of the multiple data records as indicative of running in an acceptable state comprises obtaining a minimum running time, MRT, which is a time period after the at least one similar engineering asset has been started; and classifying each of the multiple data records which is after the minimum running time and before the minimum remaining running time as indicative of running in an acceptable state.
  11. The method of any one of the preceding claims, further comprising training the machine learning algorithm by receiving multiple data records for at least one engineering asset which corresponds to the engineering asset for which failure is to be predicted, wherein the multiple data records comprise data previously collected from the plurality of sensors at a sequence of times prior to failure of the least one engineering asset; grouping the multiple data records into a plurality of data envelopes each having an envelope length (e L ); and targeting correct classification of each data record within each data envelope as indicative of running in a faulty state or running in a not-faulty state.
  12. The method of claim 11, wherein targeting correct classification comprises determining that a data envelope is indicative of running in a faulty state when a sick rate of the data envelope equals or exceeds a sick rate threshold and that a data envelope is indicative of running in a non-faulty state when a sick rate of the data envelope is lower than the sick rate threshold; wherein the sick rate of each data envelope is equal to the sum of each fault value for the multiple data records within the data envelope, divided by the number of data records within the data envelope, and wherein . the fault value for each of the multiple data records is set to 0 or 1 and the fault value for each of the multiple data records is equal to 1 when a probability that the data record is faulty exceeds a probability threshold and 0 otherwise.
  13. A computer-readable medium comprising processor control code which when running on a system causes the system to carry out the method of any one of claims 1 to 12.
  14. A system for predicting failure of an engineering asset based on real-time data, the system comprising: a plurality of sensors for measuring data on the engineering asset and a processor which is configured to receive a data record comprising data on the engineering asset collected from the plurality of sensors at time t ; generate, using a trained machine learning algorithm, a probability P F that the received data record indicates that the engineering asset is in a faulty state; determine what number of data records received in a look-back time L t are indicative of the engineering asset being in a faulty state, wherein the look-back time L t is a time period occurring before the time t at which the data record was collected; and predict a probability of the engineering asset failing during a horizon time H t , wherein the horizon time H t is a time period after time t at which the data record was collected; wherein prediction implements a Bayes forecasting model to predict the probability of failure based on the generated probability P F and the number of data records which were determined to be faulty within the look-back time L t ; compare the predicted probability to a failure threshold and when the predicted probability exceeds the failure threshold, output at least one of an alert to act as a decision aid to a user and a signal to trigger an automated self-protection protocol within the asset, wherein the self-protection protocol adjusts one or more components within the engineering asset.

Description

TECHNICAL FIELD The present invention relates generally to a system and method for predicting failure of an engineering asset, for example in complex engineering assets including those found in maritime vessels. BACKGROUND Prognostics and Health Management (PHM) is an active and well subscribed to field of study. Companies managing complex distributed or safety-critical engineering assets want to understand the status of their portfolio in order to maximise efficiency and minimise downtime. Approaches such as health scores, anomaly detection, failure prediction and remaining useful life (RUL) calculation are applied across various domains in order to advise maintenance and repair schedules. The published literature in this field will typically use synthetic data or carefully collected lab-based data to demonstrate the proficiency of a technique, however the use of "real-world" applications is less common. The challenges of developing PHM algorithms to run in a real-world environment are discussed and postulated, but tangible examples are not widespread. For example, see as a reference "Prognostics and Health Management for Maintenance Practitioners-Review, Implementation and Tools Evaluation" by Atamuradov et al. published in International Journal of Prognostics and Health Management in 2017. Most known PHM applications apply to engineered systems operating in and on a well-defined single-state and without consideration of prior usage and age. Especially, but not exclusively, in the maritime domain, engineering systems' operating envelopes are manifold by design and operating environment, and can include rapid, anomalous transient events (the term may be used interchangeably with "transients"). These transients can pose problems for classic machine learning (ML) algorithms, in that during these events a system may exhibit behaviour consistent with impending failure before recovering to "healthy" behaviour, leading to a false-positive which results in wasted repair efforts and a loss of operator confidence in the prognostic algorithm. This problem is further compounded by the fact that such transient events can cause low-level damage to a device, resulting in an increased risk of failure at a later date. ML algorithms performing anomaly-based fault detection are rarely capable of estimating the remaining useful life (RUL) of a system. However, it is of enormous benefit to a maintainer to have such an estimate and a probability value attached to the estimate in order to permit maximum efficiency in task planning and supply chain optimisation. Furthermore, existing non-deep learning ML anomaly-based detectors, especially in the maritime domain, tend to trigger far in advance of system failure, provoking excessively early intervention by a maintainer, denying the operator the economic benefits of extracting maximum use from an asset. EP-3255588A1 discloses methods and systems for condition-based maintenance of a structural component exhibiting a physical defect. US-20160097698A1 discloses a system, a computer-readable storage medium storing at least one program, and a computer-implemented method of remaining life estimation. KORDESTANI MOJTABA ET AL: "A New Hybrid Fault Prognosis Method for MFS Systems Based on Distributed Neural Networks and Recursive Bayesian Algorithm", IEEE SYSEMS JOURNAL, IEEE, US, vol. 14, no. 4, 12 May 2020 (2020-05-12), pages 5407-5416, discloses a hybrid prognosis method to predict a remaining useful lifetime of multi-functional spoiler systems. It is an example aim of the present invention to at least partially overcome or avoid one or more problems of the prior art, whether identified herein or elsewhere, or to at least provide an alternative to existing systems and related methodologies. SUMMARY According to first and second aspect of the present inventions, there is provided a system and method for predicting failure of an engineering asset as set out in in the independent claims. Further and optional features are described in the dependent claims. We describe a method for a computer-implemented method for predicting failure of an engineering asset based on real-time data, the method comprising: receiving a data record comprising data on the engineering asset collected from a plurality of sensors at time t; generating, using a trained machine learning algorithm, a probability PF that the received data record indicates that the engineering asset is in a faulty state; determining what number of data records received in a look-back time Lt are indicative of the engineering asset being in a faulty state, wherein the look-back time Lt is a time period occurring before the time t at which the data record was collected; and predicting a probability of the engineering asset failing during a horizon time Ht, wherein the horizon time Ht is a time period after time t at which the data record was collected; wherein the predicting step comprises implementing a Bayesian model to predict the prob