Data processing: structural design – modeling – simulation – and em – Simulating nonelectrical device or system – Mechanical
Reexamination Certificate
1999-11-05
2002-09-17
Hilten, John S. (Department: 2863)
Data processing: structural design, modeling, simulation, and em
Simulating nonelectrical device or system
Mechanical
C702S189000, C702S116000, C702S104000, C702S183000, C714S736000
Reexamination Certificate
active
06453279
ABSTRACT:
TECHNICAL FIELD OF THE INVENTION
This invention relates to the field of scheduled maintenance in a process control system and, in particular, to a method and system for determining, on the basis of its performance, whether a system component requires maintenance.
BACKGROUND
In a process control system, a controller typically sets the value of one or more manipulated variables with the objective of maintaining the value of a controlled variable at a specified setpoint. For example, a controller can maintain a fluid level (the controlled variable) in a tank, by altering the flow rates (the manipulated variables) of fluid entering and/or leaving the tank. Since the appropriate value of the manipulated variable typically depends on the value of the controlled variable, a process control system almost invariably includes a measurement sensor for monitoring the value of the controlled variable.
The correct operation of the measurement sensor is of paramount importance in the operation of a process control system. It is easy to see, in the foregoing example, that a measurement sensor that under-represents the fluid level can cause the controller to increase the flow rate into the tank, thereby causing the tank to flood. It is thus desirable, in a process control system, for measurement sensors to function with utmost reliability.
Because of the nature of its function, a measurement sensor is typically in intimate contact with the process under control. In many cases, this results in exposure of the measurement sensor to extremes of pressure and temperature in the presence of corrosive fluids. In some cases, the mechanical agitation used to mix process constituents results in prolonged vibration of the measurement sensor. Indeed, the environment faced by a measurement sensor in its day-to-day operation can be more hostile than that faced by interplanetary probes.
Given the cumulative effect of the conditions under which they operate, it is not surprising that even the most rugged measurement sensors require periodic maintenance to assure their optimal functioning. This periodic maintenance generally entails shutting down the process and testing the operation of the measurement sensor. The loss of production time resulting from shutting down the process makes it desirable to avoid unnecessary maintenance of the measurement sensor.
What is difficult to determine, however, is when the periodic maintenance should occur. Periodic maintenance that occurs more frequently than necessary is economically wasteful. However, periodic maintenance that is too infrequent can lead to catastrophe. Because the maintenance requirements of a particular measurement sensor depend on its cumulative exposure to wear, it is difficult to predict with certainty when failure is likely to occur.
A particular difficulty with accounting for a measurement sensor's cumulative exposure to wear is that, in some cases, the variable that the measurement sensor measures, is itself one of the factors contributing to wear on the measurement sensor. If, as is often the case, exposure of the measurement sensor to a hostile environment gradually reduces the accuracy of the sensor measurement, and if one is also relying on that sensor's measurements to predict the need for maintenance of that sensor, then one's predictions can be flawed.
For example, in order to determine when a pressure sensing device might need replacement, one can monitor the performance of the pressure sensing device. However, it is not generally possible to determine whether the measured pressure is the actual pressure, in which case the device is working correctly, or whether the measured pressure and the actual pressure differ, in which case the device may need maintenance. This poses a bootstrapping quandary, since the only way to know the actual pressure is to read the measured pressure using the very pressure sensing device whose accuracy is in question.
It is not a satisfactory solution to provide two measurement sensors since any discrepancy would simply indicate that at least one of the two sensors is unreliable without specifying which it is. Although additional measurement sensors could be provided, this would dramatically increase equipment costs.
SUMMARY
A system incorporating the principles of the invention determines the probability that a measurement sensor, which generates a measurement signal, requires maintenance. The system includes a reference sensor adapted to generate a reference signal that is at least partially correlated with the measurement signal generated by the measurement sensor. This reference signal, together with the measurement signal generated by the measurement sensor, are provided to an error signal generator that generates, from the measurement signal and the reference signal, an error signal. A trend spotter, in communication with the error signal generator, monitors the error signal to detect the occurrence of a triggering event. This triggering event is pre-selected to be indicative of the probability that the measurement sensor requires maintenance.
The method of the invention infers the probability that a measurement sensor requires maintenance on the basis of an error signal indicative of a difference between a measurement signal generated by that sensor and an independently obtained reference signal that is correlated with, but distinct from, the measurement signal. Although the reference signal is not identical to the measurement signal, the correlation between these two signals assures that when the sensor is functioning correctly, the two signals track each other over time. The method of the invention monitors this error signal to detect a triggering event indicative of a mismatch between the reference signal and the measurement signal. The occurrence of the triggering event is indicative of a probability of a malfunction in the sensor and hence indicates a need for maintenance.
A “triggering event” is intended to include composite events formed by unions and intersections of events. For example, a triggering event can be that the reference signal and the measurement signal differ by more than a selected threshold. A triggering event can also be the event that the reference signal and the measurement signal differ by more than a selected threshold a pre-selected number of times during a pre-selected interval. In both cases, the triggering event is indicative of the probability that the measurement sensor is malfunctioning. By suitably defining the triggering event, the method of the invention reduces the likelihood that spurious differences between the measurement signal and the reference signal incorrectly indicate a need for maintenance.
The step of generating the error signal includes the step of generating a difference signal representative of the difference between the reference signal and the measurement signal. This is followed, in one embodiment, by the step of rectifying, or evaluating the absolute value of, the difference signal. The absolute value of the difference signal, also referred to as the “rectified difference signal,” is then monitored to determine the extent to which it deviates from a constant value.
In another embodiment, the error signal is generated by passing the difference signal in parallel through two different first-order filters having different time constants. The outputs of these two first-order filters are then subtracted from each other and the resulting difference is used to generate the error signal. To the extent that the difference between the outputs of these two filters is close to zero, the method of the invention reports that the measurement sensor is functioning correctly. In this embodiment, a relatively constant non-zero or monotonically changing value indicates that one of the two inputs generating the difference signal is steadily moving away from the second input as a function of time.
In yet another embodiment, the error signal is the time rate-of-change of the rectified difference signal. This error signal is monitored to determine whether its
DeMoura David N.
Prasad Mohan
Foley & Hoag LLP
Hilten John S.
Oliver Kevin A.
The Foxboro Company
Vo Hien
LandOfFree
Statistical trend generator for predictive instrument... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Statistical trend generator for predictive instrument..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Statistical trend generator for predictive instrument... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2894338