Method for regulating the sequence of a continuous technical process or process step

(42) Date of commencement of the patent: 15. 7.1995 (45) Date of issue: 25. 3.1996 (51) Int.Cl.⁶ : G05D 27/02

<td>(56) Documents:</td><td>(73) Patent owner:</td>

<td>US 4469026A US 3933764A</td><td>ANDRITZ PATENT MANAGEMENT COMPANY MBH A-8045 GRAZ, STYRIA (AT). (72) Inventor: CADEK WALTER DIPL.ING. GRAZ, STYRIA (AT).</td>

(54) PROCEDURE FOR REGULATING THE PROCESSING OF A CONTINUOUS TECHNICAL PROCESS OR BZW. PROCEDURE (57) A method is proposed for controlling the operation of a continuous technical process or method step. eg a solid / liquid separation, that of internal, changeable but unchanging, and from external, depends on variable parameters, by varying at least one of its variable parameters during the process, with the license plate, that the actual value of the parameter is changed stepwise in order to approach a best value, the size of the change, the direction of change, the duration of the change and / or the time interval between the changes are each controlled as a function of the effect on the sequence achieved with the preceding change step.

<img file="AT400771B_D0001.tif" />

QQ

AT 400 771

Bffl 0078318

AT 400 771 Β

The invention relates to a method for controlling the flow of a continuous technical process or process step, eg a solid / liquid separation, which is dependent on internal, variable but unchangeable, and external, variable parameters, by varying at least one of its variable parameters during the process. As examples of solid / liquid separations mechanical dehydration and drying are mentioned.

Previous methods for controlling process steps lead to problems when approaching a parameter from an actual value to a best value if a function is present which has a more or less pronounced optimum. The characteristic of such a function is that deterioration occurs both when the optimum is exceeded and not reached. It can also be said that such a function reverses its effective direction at its optimum. With the known methods, only a minimal or maximum value or such a value of a continuously rising or falling function within a corresponding range. A reversal of the effective direction usually leads in such systems to oscillations or to instability.

It was therefore an object to provide a control method in which the best value is not exceeded as possible, so that there is no need to reverse in the reverse direction.

To achieve this object, the invention proposes to gradually approach the respective best value along the function curve.

The inventive method is therefore characterized in particular by the fact that the actual value of the parameter is changed stepwise to approximate a best value, wherein the size of the change, the direction of change, the duration of change and / or the time interval between the changes respectively depending on the with preceding change step have been impacted on the process.

An essential aspect is that the course of a continuous process is regulated, which depends on given internal parameters, which can change constantly uninfluenced, and on external, variable parameters, the latter being varied in the direction of optimizing the process sequence.

This is explained using the example of dewatering a solid suspension. Internal, unchangeable but changeable parameters of the solid suspension are eg water content, particle size and composition of the solids. External, changeable parameters are, for example, flocculant dose, residence time between flocculant addition and dehydration, and drainage rate. These external parameters are varied to control the drainage process.

In this case, the best value of each varied parameter is unknown, so it is not a question of controlling in a known manner to a desired value, for example a filling level: with the method according to the invention, the effect of changing internal parameters can be continuously compensated

The US 44 69 026 A relates to optimizing the drying function of a printing device and the separation behavior of the printed material from the transport roller. Basically, this is a one-time dry process in which the influencing print parameters are detected beforehand, but no detection of the result of the drying is provided. It is not the drying process seized and then controlled or regulated, but only temporally before the drying process parameters lying. Depending on these parameters, the energy supply of the drying device or set the residence time. These parameters are created in advance for each sheet to be dried. A procedure in which a process parameter is gradually changed to approximate an end result to be achieved thereby, the change being tuned to the result obtained with the preceding approximation step, is neither disclosed nor suggested.

US 39 33 764 A relates to a process for the recovery of polysulfone resins dissolved in organic solvents by dispersing these solutions in water and subsequent treatment with an aliphatic hydrocarbon. Details of how this procedure is to be regulated are not provided.

According to the invention, starting from an actual value of a parameter, the step direction which leads to the best value is determined as a function of whether a change in the actual value results in an improvement or deterioration of the result associated with the actual value. Then, the actual value is changed stepwise to the best value, wherein the step size is adjusted in dependence on the slope of the function curve. That is, for example, that at high slope of the function curve, so if a large change in the result is achieved per approach step, the step size can be reduced, while in flat curve, where to expect a small steady change in the result per approximation step is, can be increased. It is essential that step duration and step interval are chosen so that a stable result is achieved, ie a state of equilibrium between the parameter value and the result obtained before

AT 400 771 Β the next approximation step is carried out.

According to the invention, the respective optimum can be achieved quickly and easily in this way, whereby the optimum can be regarded as having been reached when approaching the associated turning point of the function curve, or the approximation steps are extremely shortened.

Another characteristic of the invention is that the one process parameter is selected for optimization whose incremental change provides the highest response of the process. This results in a particularly efficient regulation.

In systems which depend on several parameters, according to the invention, after the optimum result of the change of the at least one parameter has been reached, at least one further parameter is changed.

In this case, after a single change of the further process parameter, the at least one process parameter can be changed again to optimize the result, or first the further process parameter can be changed in an analogous manner as the at least one process parameter and then the at least one process parameter can be changed again.

In this case, a step interim can be run after each change.

It is possible to change several process parameters, in particular cyclically, step by step.

The invention is described below by way of example with reference to the drawing, wherein FIG. 1 shows a possible dependency of the result on a parameter, FIG. 2 shows a flow chart for a possible control and / or regulation of a process stage, and FIG. 3 shows a further flow chart for a possible one Control and or or regulation of a process step represents.

In Fig. 1 is shown a possible function of a result as a function of a process parameter, with an optimum exists. In the case of a deviation from the optimum, the result is worse both with a lower and also with a higher value of the parameter. Thus, depending on the value of the parameter, a change can result in both an improvement and a deterioration of the result. But since it is not clear from the outset on which side of the optimum one is currently located, the step direction must be determined. It is thus the actual value of the examined parameter eg first reduced. If the result improves, the next change of the parameter will be in the same direction. If this procedure is successful once or several times, the step size is increased in order to get to the new optimum faster. However, the adjustment has a negative effect, ie the result becomes worse, so the direction of the change step is reversed. If there is no improvement even if the step direction is reversed, the step size is reduced since the optimum is very likely to be between the points tested. Subsequently, the search continues with a smaller step size until the change of the result is within the desired limits and a further change no longer yields a corresponding improvement. This method is particularly advantageous for regulating or Controlling a process stage applied where the result function changes more or less strongly depending on a process parameter by uninfluenced changes at the beginning of the process stage, the influence of the change can be determined purely empirically.

FIG. 2 shows the flowchart of a possible embodiment of the invention control and or regulatory procedure of a process step. When the process step is started, a first result of the process step is determined on the basis of a start value for the individual parameters, which may be respectively the last stored value of the corresponding parameter or a value which is as optimal as possible. Then the parameter is incremented, which in turn can be the last stored increment, eg reduced. This is especially done when, for example Savings result. In other cases, an increase of the parameter eg Bring savings, so here is the first step will be an increase. The same applies to any other parameters.

After the change, depending on the change of the parameter (increase, decrease) and the result (better / worse) the parameter is increased or reduced. This will be clarified by means of an example. If the parameter was increased before the test (+ delta), follow the right arrow of fig. Query. Was the result successful, ie if a better result was achieved, one follows at the 2. Query also the right arrow and the parameter is increased again by one step. Was the result unsuccessful, ie the result became worse, you follow at the 2. Query the left arrow and the parameter is reduced (- delta). In one embodiment of the invention, the step size can then also be reduced here since it is assumed that the optimum is between the last two values ​​and that a reduction by the same delta value would only lead back to the previous result. After changing the parameter, a certain waiting time passes to take into account the dwell time of the process step. The result will be repeated

AT 400 771 Β and the change of the result is checked for improvement, followed by a further change of the parameter. If a new value results within the given limits, a parameter 2 is optimized according to the present flowchart. This takes place until parameter 2 is within predetermined limits. Subsequently, parameter 1 is again optimized. This is particularly necessary if the input conditions of the process stage and thus the dependence of the result of the individual parameters vary greatly.

Alternatively to the flowchart of FIG. 2, it is also possible to proceed according to FIG. 3. in the

In contrast to FIG. 2, once the parameter 2 has been changed once again, the optimum of the result is sought by changing the parameter 1. The starting point for the queries can be either the last jump in general or the last jump of the corresponding parameter.

The invention is not limited to the examples. Rather, within the scope of the claims, for example, a parameter can only be reversed in its direction after several unsuccessful steps or changed in its increment.

As a practical example of the application of the method according to the invention, the control of the dewatering of a solid-liquid mixture on a screen with the addition of a flocculant is given.

The result to be optimized is the dewatering rate or the degree of dewatering. The process parameters which can be changed to obtain an optimum result and thus serve as manipulated variables are, above all, the flocculant metering and the screen speed, ie the speed with which the loaded screen belt passes through the system.

In this case, the Fiockungsmitteldosierung can be regarded as the paramount parameter, since it gives a rapid response of the system and is otherwise costly; it is thus selected as an adjustment parameter and changed step by step. The resulting due to the first change in the amount of flocculant resulting change in the drainage rate or the degree of dewatering is used to determine the direction of change, in which the flocculant dosage is then changed step by step for optimization. Subsequently, the screen speed is changed step by step for optimization with constant flocculant metering. Of course, as an optimal result, a compromise can be considered taking into account the flocculation costs.