Data processing: generic control systems or specific application – Generic control system – apparatus or process – Optimization or adaptive control
Reexamination Certificate
2000-04-03
2002-12-31
Black, Thomas (Department: 2121)
Data processing: generic control systems or specific application
Generic control system, apparatus or process
Optimization or adaptive control
C700S031000, C700S041000, C700S042000
Reexamination Certificate
active
06501998
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to a method for the closed-loop control of a time-delayed process having compensation, in particular for temperature control, as well as a control device for carrying out the method.
BACKGROUND INFORMATION
An adaptive controller for time-delayed processes having compensation is described in German Patent No. 39 29 615. The controller has the capacity to adjust its parameters autonomously to the characteristics, changing over time, of a controlled system. For this purpose, the response of the process to a step-like shift of the setpoint value is recorded and, in an iterative method, a PTn model is sought, using the Ptn model, this step-response can be simulated with the greatest possible accuracy. The PTn model is viewed as optimal if the error, calculated according to the method of the least error squares, between the step-responses of the PTn model and the real process is minimal. After the process identification is concluded, a controller is designed in accordance with the absolute-value optimum on the basis of the process model established. In German Patent No. 39 29 615, consideration is given to single-loop closed-loop control circuits having controllers of the PID type. The processes to be controlled demonstrate a method having compensation, i.e., they represent a controlled system in which the step-response flows into a new steady state.
A disadvantage of this method is that when the conventional adaptive controller is placed into operation, an often time-consuming step-change test must be carried out before the controller can be optimized. Also disadvantageous are the creeping time constants, i.e., poles that are located very close to the origin of the Laplace plane and which lead to an integrator-like behavior of the process, so that after a manipulated-variable step-change, no steady state can be achieved for a long time.
In the German Patent Application No. 195 48 909.8, a method for closed-loop controlling a time-delayed process having compensation is described along with a control device for carrying out the method, in which, when a controller is placed into operation, the time-consuming step-change test can be dispensed with. In this context, however, it is disadvantageous that storage of the measuring data and a time-discrete parameter estimator, which uses a cumbersome calculation method, are necessary.
SUMMARY
An object of the present invention is to provide a method for the closed-loop control of a time-delayed process having compensation, in particular for temperature control, using which the above-mentioned disadvantages are avoided. Another object of the present invention is to provide a control device for carrying out the method.
To achieve this objective, the new method of the type mentioned above has the features indicated in the characterizing part of Claim
1
. Refinements of the method are described in Claims
2
through
7
. A control device having the features mentioned in Claim
8
is suited for carrying out the method.
An advantage of the present invention is that when placing a controller into operation, a time-consuming step-change test can be completely dispensed with. During the first phase of an adjustment procedure in accordance with a setpoint value step-change, a simple IT1 model approximates the process, and the first setpoint value can be started in a manner that is virtually time-optimal and without overshooting. This has a particular advantageous effect in the closed-loop control of temperatures at a specified setpoint value using an adjustable heating system, in which a controller is to maintain the temperature at a constant value against external influences and, in the event of changes in the setpoint value entry, e.g., in a chemical process, is to set the desired new temperature as rapidly as possible. The adaptive controller adapts itself automatically to any time-delayed processes having compensation without manual parametrization and without prior knowledge of the process. The controller is optimized during the first start-up after the control device has been installed in the control circuit, without it being necessary that the operator provide inputs for this purpose. As criterion for reaching a steady state after the setpoint value step-change, the speed of the controlled variable changes and the system deviation are evaluated and monitored when the latter are below a specifiable limiting value using a PI controller that is parametrized for an IT1 model. Subsequently, the process is identified on the basis of a more exact model, in particular a PT2 model. This model is therefore available after the first adjustment procedure and can, at this point, be retrieved for setting the controller.
The systems that come into question in temperature control processes can generally be controlled using simple PI or PID controllers, which demonstrate very good performance with respect to interference. In response to setpoint value changes, they can be supported, for example, by the following supplemental open-loop control measures: in the event of positive setpoint value step-changes beginning from a specifiable minimum value, a switchover in the structure takes place in which the integral-action component of the PI or the PID controller is switched off, i.e., a P or a PD controller is used; in the vicinity of the setpoint value, the integral-action component is again connected to the controller in a surge-free manner. For systems having very large delay times, in the event of setpoint value step-changes the following control measure can advantageously be applied: after the setpoint value step-change in an open-loop controlled operation, the manipulated variable that is required to be steady for the new setpoint value is output (read out) constantly until the setpoint value is nearly reached; subsequently, there is again a switchover to the PI or PID controller. Although this strategy is less than optimal with respect to time, it is not dependent on the precision of estimation in the system time constants, unlike the time-optimal open-loop control.
By using an IT1 model for identifying the process in an early phase of a first step-change response, the disadvantage of more exact models having compensation is avoided in that at this time point no dependable estimations regarding the steady process gain are yet possible. Rather, the maximum rate of rise is dependent on the ratio of the gain to the dominating time constant, for example, of a PT2 model. However, the IT1 model contains a reference to the delay time, which corresponds to the small time constant of a PT2 model, and a reference to the rate of rise. Thus, sufficient information is made available to parametrize a PI controller using an initially cautious setting, in which overshoots are substantially avoided and yet a sufficiently rapid control performance is obtained.
A control device for carrying out the method can be realized advantageously using an internal sequence control system having a plurality of different phases, which are differentiated in structures of the controller as well as in the models upon which the process identification is based. The control device can be implemented equally as a hardware circuit or as a processor having a program memory into which a suitable operating program has been loaded.
Physical models for temperature control systems can be designed on the basis of elementary PT1 models for the individual partial systems having their heat capacity and their corresponding heat transfers. On the whole, in this way PTn models are obtained of a usually low order. With real processes, it is seldom that more than three relevant time constants can be demonstrated. The typical gains of the processes, depending on the design of the heating system, are between 0.5 and 30, the time constants being between one and several thousand seconds. In general, temperature control systems are themselves not capable of overshoots, i.e., in the open closed-loop control circuit, they only have real pole positions and they usual
Black Thomas
Pham Thomas
Staas & Halsey , LLP
LandOfFree
Method for controlling a time-lagged process with... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for controlling a time-lagged process with..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for controlling a time-lagged process with... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2976690