Optimizing a prognostic-surveillance system to achieve a user-selectable functional objective
The disclosed embodiments relate to a system that optimizes a prognostic-surveillance system to achieve a user-selectable functional objective. During operation, the system allows a user to select a functional objective to be optimized from a set of functional objectives for the prognostic-surveillance system. Next, the system optimizes the selected functional objective by performing Monte Carlo simulations, which vary operational parameters for the prognostic-surveillance system while the prognostic-surveillance system operates on synthesized signals, to determine optimal values for the operational parameters that optimize the selected functional objective.
1 . A method for optimizing a prognostic-surveillance system to achieve a user-selectable functional objective, comprising:
receiving a selection of a functional objective to be optimized, wherein the functional objective to be optimized is selected from a set of functional objectives configured for the prognostic-surveillance system, wherein the functional objective comprises a quality of information objective; and
optimizing the selected functional objective, wherein the optimizing the selected functional objective comprises:
performing Monte Carlo simulations using a synthetic data pump,
wherein the synthetic data pump generates time-series signals that vary operational parameters for the prognostic-surveillance system in response to defects in a monitored asset,
wherein the Monte Carlo simulations vary the operational parameters for the prognostic-surveillance system while the prognostic-surveillance system operates on synthesized signals;
determining optimal values for the operational parameters that optimize the selected functional objective; and
determining, based on the Monte Carlo simulations, a quality of sensors, and based on the quality of the sensors, adjusting a sampling rate.
2 . The method of claim 1 , wherein the set of functional objectives for the prognostic-surveillance system comprises:
fast anomaly detection, which can be evaluated in terms of an average sample number (ASN) for a detection decision;
high prognostic accuracy, which can be evaluated in terms of false alarm probabilities (FAPs) and missed alarm probabilities (MAPs); and
low compute cost, which can be evaluated in terms of required computational operations and memory usage.
3 . The method of claim 1 , wherein the prognostic-surveillance system uses an inferential model while detecting incipient anomalies, and wherein the operational parameters include one or more of the following:
a number of signals in the inferential model;
a number of samples for each signal;
signal-to-noise ratios for the signals; and
a number of training vectors for the inferential model.
4 . The method of claim 3 , wherein the prognostic-surveillance system uses a sequential probability ratio test (SPRT) while detecting incipient anomalies, and wherein the operational parameters include one or more of the following SPRT parameters:
a desired false alarm probability parameter α;
a desired missed alarm probability parameter β;
a variance parameter V; and
a sensitivity parameter M.
5 . The method of claim 1 , wherein unselected functional objectives in the set of functional objectives become subordinate objectives and/or constraints while performing the Monte Carlo simulations.
6 . The method of claim 1 , wherein the method uses a stochastic gradient-descent technique to optimize the selected functional objective while performing the Monte Carlo simulations.
7 . The method of claim 1 , wherein the method further comprises performing digital filtering operations on signals received by the prognostic-surveillance system to reduce signal-to-noise ratios of the signals.
8 . The method of claim 1 , wherein the method further comprises performing intelligent data preprocessing operations on signals received by the prognostic-surveillance system to improve signal quality.
9 . The method of claim 1 , wherein the method further comprises configuring the prognostic-surveillance system to use the optimal values for the operational parameters during subsequent prognostic-surveillance operations.
10 . The method according to claim 1 , wherein the prognostic-surveillance system determines subordinate objectives based on the received functional objective to be optimized.
11 . The method according to claim 1 , wherein the Monte Carlo simulations are performed for a user in a single instance.
12 . A non-transitory computer-readable storage medium storing instructions that when executed by a computer cause the computer to perform a method for optimizing a prognostic-surveillance system to achieve a user-selectable functional objective, the method comprising:
receiving a selection of a functional objective to be optimized, wherein the functional objective to be optimized is selected from a set of functional objectives configured for the prognostic-surveillance system, wherein the functional objective comprises a quality of information objective; and
optimizing the selected functional objective, wherein the optimizing the selected functional objective comprises:
performing Monte Carlo simulations using a synthetic data pump,
wherein the synthetic data pump generates time-series signals that vary operational parameters for the prognostic-surveillance system in response to defects in a monitored asset,
wherein the Monte Carlo simulations vary the operational parameters for the prognostic-surveillance system while the prognostic-surveillance system operates on synthesized signals;
determining optimal values for the operational parameters that optimize the selected functional objective; and
determining, based on the Monte Carlo simulations, a quality of sensors, and based on the quality of the sensors, adjusting a sampling rate.
13 . The non-transitory computer-readable storage medium of claim 12 , wherein the set of functional objectives for the prognostic-surveillance system comprises:
fast anomaly detection, which can be evaluated in terms of an average sample number (ASN) for a detection decision;
high prognostic accuracy, which can be evaluated in terms of false alarm probabilities (FAPs) and missed alarm probabilities (MAPs); and
low compute cost, which can be evaluated in terms of required computational operations and memory usage.
14 . The non-transitory computer-readable storage medium of claim 12 , wherein the prognostic-surveillance system uses an inferential model while detecting incipient anomalies, and wherein the operational parameters include one or more of the following:
a number of signals in the inferential model;
a number of samples for each signal;
signal-to-noise ratios for the signals; and
a number of training vectors for the inferential model.
15 . The non-transitory computer-readable storage medium of claim 14 , wherein the prognostic-surveillance system uses a sequential probability ratio test (SPRT) while detecting incipient anomalies, and wherein the operational parameters include one or more of the following SPRT parameters:
a desired false alarm probability parameter α;
a desired missed alarm probability parameter β;
a variance parameter V; and
a sensitivity parameter M.
16 . The non-transitory computer-readable storage medium of claim 12 , wherein unselected functional objectives in the set of functional objectives become subordinate objectives and/or constraints while performing the Monte Carlo simulations.
17 . The non-transitory computer-readable storage medium of claim 12 , wherein the method further comprises performing digital filtering operations on signals received by the prognostic-surveillance system to reduce signal-to-noise ratios of the signals.
18 . The non-transitory computer-readable storage medium of claim 12 , wherein the method further comprises performing intelligent data preprocessing operations on signals received by the prognostic-surveillance system to improve signal quality.
19 . A system that optimizes a prognostic-surveillance system to achieve a user-selectable functional objective, comprising:
a computing system with one or more processors and one or more associated memories; and
an optimization processor that executes on the computing system, wherein during operation, the optimization processor:
receives a selection of a functional objective to be optimized, wherein the functional objective to be optimized is selected from a set of functional objectives configured for the prognostic-surveillance system, wherein the functional objective comprises a quality of information objective, and
optimizes the selected functional objective, wherein the optimizing the selected functional objective comprises:
performing Monte Carlo simulations using a synthetic data pump,
wherein the synthetic data pump generates time-series signals that vary operational parameters for the prognostic-surveillance system in response to defects in a monitored asset,
wherein the Monte Carlo simulations vary the operational parameters for the prognostic-surveillance system while the prognostic-surveillance system operates on synthesized signals;
determining optimal values for the operational parameters that optimize the selected functional objective; and
determining, based on the Monte Carlo simulations, a quality of sensors, and based on the quality of the sensors, adjusting a sampling rate.
20 . The system of claim 19 , wherein the set of functional objectives for the prognostic-surveillance system comprises:
fast anomaly detection, which can be evaluated in terms of an average sample number (ASN) for a detection decision;
high prognostic accuracy, which can be evaluated in terms of false alarm probabilities (FAPs) and missed alarm probabilities (MAPs); and
low compute cost, which can be evaluated in terms of required computational operations and memory usage.