Abstract: An apparatus and method is disclosed for detecting, identifying, and classifying faults occurring in sensors measuring a process. A variety of process models can be used such as first principles models, dynamic multivariable predictive control models, from data using statistical methods such as partial least squares (PLS) or principal component analysis. If faults are identified in one or more sensors, the apparatus and method provide replacement values for the faulty sensors so that any process controllers and process monitoring systems that use these sensors can remain in operation during the fault period. The identification of faulty sensors is achieved through the use of a set of structured residual transforms that are uniquely designed to be insensitive to specific subsets of sensors, while being maximally sensitive to sensors not in the subset. Identified faults are classified into one of the types Complete Failure, Bias, Drift, Precision Loss, or Unknown.

Abstract: An apparatus and method is disclosed for detecting, identifying, and classifying faults occurring in sensors measuring a process. If faults are identified in one or more sensors, the apparatus and method provide replacement values for the faulty sensors so that any process controllers and process monitoring systems that use these sensors can remain in operation during the fault period. The identification of faulty sensors is achieved through the use of a set of structured residual transforms that are uniquely designed to be insensitive to specific subsets of sensors, while being maximally sensitive to sensors not in the subset. Identified faults are classified into one of the types Complete Failure, Bias, Drift, Precision Loss, or Unknown.

Abstract: A system and method for automatically tuning a fuzzy logic process controller is presented. A process under control is automatically assessed to determined process control characteristics including, for example, ultimate gain, ultimate period and time delay. Then, these quantities are used to calculate tunable control parameters which include scaling factors, which are applied to the fuzzy logic controller in order to optimize the performance of the fuzzy logic controller when connected to control the process.

Abstract: A recursive on-line wavelet data compression technique which may be used in, for example, a process control network, compresses a stream of data points on-line or in real-time (e.g., as the data arrives or is generated) without requiring the storage of a large amount of uncompressed data. The data compression system includes a data receiver that receives the data points in a sequential manner and a compression tree computation device that determines approximation coefficients and detail coefficients of a multi-layer wavelet compression tree from the received data points. The compression tree computation device determines all of the coefficients of the higher level layers of the compression tree that can be determined after the data receiver receives each of the data points to thereby perform on-line or real-time data compression.

Abstract: A system and method for automatically tuning a PID controller resident within a PID control loop. The PID control loop includes a PID controller and a process. The process supplies a process variable which is compared to the loop input. The result of the comparison is supplied to the PID controller, and the PID controller drives the process. A relay is applied to the loop input. The relay compares a set point value to the process variable. If the set point value is greater than the process variable, the relay drives the loop input with a first amplitude value. If the set point value is less than the process variable, the relay drives the loop input with a second amplitude value. In response to the set point relay, the process variable develops a sustained oscillation. The period and amplitude of the sustained oscillation are measured. A new set of PID controller parameters are calculated from the period and amplitude of sustained oscillation.

Abstract: A non-linear process controller drives a process variable to be substantially equivalent to a set point based on the expected load disturbances within the process and on a measurement of the actual magnitude of set point changes. The controller uses control parameters developed in an optimal manner to control the process in response to expected load disturbances when no set point change has occurred. Whenever a set point change is detected, the controller derives a set of set point change control parameters and uses these set point change control parameters in responding to the set point change. Each of the set point change control parameters may be developed as a function of the magnitude of the actual change in the set point and according to a set of process characteristics which may include the time delay and the dominant time constant of the process.

Abstract: A method and apparatus is provided for detecting a faulty sensor within a process control system having a set of sensors, each of which produces an associated sensor output signal. The method and apparatus produce a set of sensor estimate signals from the sensor output signals using principal component analysis and then determine a validity index for each of the sensors as a ratio of two residuals, wherein each of the residuals represents a different measure of the difference between the sensor output signals and the sensor estimate signals. The determined validity indexes are then used to detect a failure of one of the sensors and/or to identify which one of the sensors has failed.