Package Usage
The package is installed in the MATLAB-5 environment. The package calculates the highest quality model, using the time response data acquired in open or in closed loop identification experiment. Based on this model, optimal parameters of controller are automatically calculated. The above time response is caused by arbitrary deterministic input test signal in open or in closed loop.
Process Model Identification
The identification represents dynamics of loops or complicated systems for behavior analysis purposes and control loops tuning.
The highest quality process identification of five different model types are automatically generated.
The above process modeles are created in bandwidth of the real process by satisfying the direct Quality index . Transfer function of the best model and time responses for this model and a real process are reported as the identification results.
Example on the right: condenser level process of 75MW unit
Model Validation
Unique model validation process is fulfilled both in the time and in the frequency domains.
Validation in the time domain uses the data part not used in model identification (in green colored area) for forecasting process behavior by the identified optimal model.
Validation in the frequency domain defines the frequency interval where the phase shift between frequency responses of the real process and its model does not exceed 10 deg (the second sub-window). In addition, this model gain-frequency characteristic is shown in the above frequency interval (the third sub-window).
Example on the right: main air flow process of 75MW unit
Noise Filter Design
The CSTP intelligently solves problems of noise filtering based on an analysis of noise spectral density in dialog regime
The CSTP calculates the Filter Quality which is the generalized estimate of the "filter delay" . In addition, the noise standard deviation after the designed filter is calculated also.
Simulated filter time responses to real noise and to real process value are reported as filtering results.
The CSTP checks the effectiveness of any filter given by the User
Example on the right: condensator level process of 75MW unit
PI/PID Controller Synthesis
Cost-effective loop robust tuning is automatically fulfilled for multi-optimal performance.
Robust Tuning provides: a) optimal PI/PID parameters, increasing time period between revisions; b) object-oriented optimal responses.
The CSTP uses direct estimates in optimization criteria ( Quality Index - Qual, Overshoot -OV, Rise Time - RT and Settling Time -ST etc.) and restrictions for real actuator (permissible acceleration, limited work zone, acceptable amplitude fluctuation).
Set Point (SP) Filter is automaticaly designed in order to eliminate overshoot on SP chenges without reduction of ST.
Tunig results include PI and PID - controllers optimal parameters (individually and in combitation) satisfied four different optimization criteria. These parameters with time responses are reported as tuning results.
Additional option: interactive dialog selection of PI/PID controller parameters from the database, providing OV, RT and ST chosen by the User.
Example on the right: condensator level process of 75MW unit
Return
to Home page