GB2273792A

GB2273792A - Procedure for a three-level controller

Info

Publication number: GB2273792A
Application number: GB9325460A
Authority: GB
Inventors: Hans H Leszke
Original assignee: Anschuetz and Co GmbH
Current assignee: Raytheon Anschuetz GmbH
Priority date: 1992-12-22
Filing date: 1993-12-13
Publication date: 1994-06-29
Anticipated expiration: 2013-12-13
Also published as: DE4341979A1; NO306228B1; GB2273792B; DE4341979C2; NO934408D0; GB9325460D0; NO934408L

Abstract

The invention relates to a procedure for a three-level (ie left, right, stop) controller which controls the position of a lever, piston or a similar mechanical component and allows for adaptation to conditions changing with time. This is achieved by measuring and storing at specific intervals the position and overrun of the component, determining the overrun when the element has stopped and taking into account said overrun value in the generation of the next control signal. <IMAGE>

Description

Description: Procedure for a three-level controller The present invention relates to a three-level controller according to the first part of the first patent claim.

Machines and positioning devices are often operated today using three-level controllers. These controllers only feature three control levels, e.g. movement to the left or right, and stop. The movement to the left or right is started to bring a system or machine component into a new position. The current position (actual value) is often compared with the position required (nominal value) in an electronic circuit. If the values are different, a corresponding adjustment signal will be produced. If the values are identical, the signal will return to the stop status and the positioning device will stop.

Owing to the mass moved during this process, the above stop status cannot be achieved without any delay; instead, such a device will continue to move after the stop signal in the direction chosen by a certain amount. The extent of this overrun is dependent on various factors and can constantly change during operation. A well-known example of this phenomenon is the overrun of a rudder machine. Here, the extent of overrun is dependent on the number of pumps used, the water counterpressure and the design properties.

To ensure that this overrun which usually results in the nominal value being exceeded does not immediately lead to a signal in the opposite direction and thus to oscillation of the system, this type of controller is provided with a dead zone. Differences between the nominal and the actual values which remain within this tolerance range termed dead zone will therefore not result in a adjustment signal and inevitably cause incorrect positioning.

In addition, dead zones are empirically determined, taking into consideration the worst possible situation that may occur. This results in dead zones which are too large, further deteriorating the attainable setting accuracy.

It is the objective of the invention to establish a procedure for a three-level controller which adapts to changing ambient and operating conditions.

This objective is achieved by the means disclosed in the characterizing portion of the first patent claim.

Basically, the invention can be implemented in various ways.

The use of analog technology, switched logics, or a microcontroller (commercially available) including suitable programs is possible.

The microcontroller receives its input signals from suitable sources, i.e. sensors, which provide the necessary nominal and actual values. The output signals for adjustment are directed to circuit breakers suitable for the application which switch the components required for adjustment.

In the most simple version, the overrun of the mechanical system can be assumed to be identical in two directions. In this case, it is only the amount without sign which is important for the measurement and processing of signals, and the implementation of the invention is simplified accordingly.

The invention offers a number of advantages over state-of-theart technology.

Firstly, a manual, empirical setting of the overrun compensation with its inherent error possibilities is no longer necessary. Secondly, when operating conditions have altered or are altering (caused by general wear, ageing of elastic components) and are thus changing the overrun, optimum overrun compensation is always achieved. The exact setting of the overrun compensation which dynamically adapts to the actual situation makes it possible to obtain the required nominal value practically without any residual errors.

When an error integrator is used (block 9 in the diagram), minor changes in the nominal value or residual errors which are smaller than the overrun values are compensated by adjustment signals at appropriate intervals. During this process, an error summation is performed and a adjustment signal will only be produced if the adjustment signal to be executed exceeds a certain minimum value. Present-day technology does not take these deviations into account, resulting in permanent errors in an external control loop.

Further advantages and details of the procedure for an adaptive three-level control according to the invention follow from the subsequent description of several embodiments and the figures 1-3 enclosed. In the following text, the invention is implemented in the form of a digital unit including a microcontroller. But the invention also can be realized without a microcontroller only with mechanical components.

Fig. 1 illustrates the schematic design of a control system using a direction-independent adaptive three-level controller; Fig. 2 illustrates the schematic design of a control system using a direction-dependent adaptive three-level controller; Fig. 3 illustrates the schematic design of a control system using a three-level controller including an error integrator; and Fig. 4 illustrates the schematic design of a control system according to Figs 1, 2, or 3 which has been implemented using a microcontroller.

The individual functions of the control system for a direction- independent adaptive three-level controller are schematically shown in the block diagram of Fig. 1. The individual blocks represent individual sections of an electronic circuit built using state-of-the-art technology and are formed by hardware components (logic modules) for signal processing. Input signals, such as the nominal value and the actual value, are supplied by suitable sources (e.g. memories or sensors) which are not shown here and which, depending on the concrete use of the invention, are different for different applications. The same applies to the processing of the adjustment signals actually moving the mechanism concerned.

The blocks shown have the following functions: Block 1 performs digitization. During this process, an arriving or stored analog signal for a nominal value is converted into digital data that can be read by the microcontroller. If the nominal value signal is already provided in digital form, this block will not be needed.

Block 2 is also used for digitization. During this process, an arriving analog signal (e.g. from a sensor not shown here) for an actual value is converted into digital data that can be read by the microcontroller. If the actual value signal is already provided in digital form, this block will not be needed.

Block 4 performs a difference computation. The signals or the data sets corresponding to the signals output by blocks 1 and 2 are supplied to block 4. This block computes the difference D between the nominal and the actual value at fixed time inter-vals, with the sign of the values being taken into account. The fixed time intervals are defined by a clock/timer circuit (which also contains an oscillating quartz for the generation of a time base) which produces a signal at fixed time intervals and to which block 4 is connected via a data line.

If the difference between the nominal and actual value is below the limit value defined by the particular application, block 5/6 will produce a logic signal (stop) for "actual value standstill". Here, the change of the actual value with time is monitored using a clock/timer signal.

Block 7 performs the overrun computation. To enable this, the block 7 is connected to block 4 and block 5/6. This block 7 is activated by the transition of a adjustment signal (left/ right) at the output of block 8 to a stop status or by a adjustment signal in the opposite direction. Using the signals provided by block 5/6, block 7 checks when the actual value for the direction of motion last applicable has come to a standstill. At this moment (and at this moment only) when this status of inactivity has been reached, the difference present between the nominal value and the actual value (signal D) will be added to the sum of all differences obtained before in the same way. The result (signal N) will be stored in a memory for the overrun of the machine.

The addition of the (residual) differences to the overrun values already determined ensures that these values adapt dynamically to changing situations. For example, a standstill before the nominal value is reached will lower the previous overrun value and an exceeding of the nominal value will increase the previous overrun value.

To achieve a certain filtering effect during the adaptation process, it is possible to use only a specific portion of the difference for addition. The adaptation process will then be performed with a time delay.

Block 8 now computes the adjustment command. At the data input of block 8, both the output signal from block 4 and the output signal from block 7 are present. The necessary adjustment direction is thereby derived from the polarity (sign) of the difference between the nominal value and the actual value (signal D).

The overrun value (signal N) is then added to the difference, taking into account the polarities. If the sum thus obtained is > 0 in one direction or < 0 in the other direction, a corresponding signal S1 or Sr will be the output. Otherwise, the adjustment signals will remain in the stop status.

This method of computation ensures that a adjustment signal will only be generated if the adjustment required is equal to or greater than the overrun.

Hence, the function of this block 8 corresponds to that of a classic window comparator, the difference being that the width of the window is not constant.

The two data outputs of block 8 are connected to drivers (not shown in the diagram) which control the mechanical system.

Needless to say, blocks 1 and 2 can also be located in front of the input of a microcontroller described at the beginning which performs all computing and storing processes and which provides the two left/right signals at its data output which after a conversion into analog signals, if necessary - can be fed to the appropriate drivers.

The block diagram shown in Fig. 2 represents a schematic of a direction-dependent adaptive three-level control.

Block 4 performs a difference computation. The signals or the data sets corresponding to the signals output by blocks 1 and 2 are supplied to the inputs of block 4. This block computes the difference D between the nominal and the actual value at fixed time intervals, with the sign of the values being taken into account. The fixed time intervals are defined by a clock/timer circuit (which also contains an oscillating quartz for the generation of a time base) which produces a signal at fixed time intervals and to which block 4 is connected via a data line.

Block 5 controls the function "Change left < limit value.

Here, the change of the actual value with time in one direction (towards the left in this case) is monitored using a clock/timer signal (produced by a clock/timer circuit). If a value is below the limit value defined by the particular application, the block will produce a logic signal (stop L) for "actual value standstill".

The same process happens in block 6, the difference being that the function "Change right < limit value" in the opposite (right-hand) direction is monitored.

Block 7 performs the overrun computation. To enable this, the block 7 receives its input signals from blocks 4, 5 and 6.

This block 7 is activated by the transition of a adjustment signal (left/right) at the output of block 8 to a stop status or by a adjustment signal in the opposite direction. Using the signals provided by block 5 and 6, block 7 checks when the actual value for the direction of motion last applicable has come to a standstill. At this moment (and at this moment only) when this status of inactivity is reached, the difference present between the nominal value and the actual value (signal D) will be added to the sum of all differences obtained before in the same way. The result (signal N) will be stored in a memory for compensation of the overrun of the machine in the corresponding direction.

This method produces two independent overrun values for one and the other direction (N1 and Nr).

Block 8 now computes and generates the adjustment command. The input of block 8 is connected to the two signal outputs of block 7 and the output of block 4. The output of block 8 is connected to the two drivers (not shown) for operation of the mechanical system. The necessary adjustment direction is first derived from the polarity (sign) of the difference between the nominal value and the actual value (signal D).

The overrun value valid for this direction is then added to the difference, taking into account the polarities. If the sum thus obtained is > O in one direction or < 0 in the other direction, a corresponding signal S1 or Sr will be output.

Otherwise, the adjustment signals will remain in the stop status.

This method of computation ensures that a adjustment signal will only be generated if the adjustment required is equal to or greater than the overrun in the respective direction.

Hence, the function of this block 8 corresponds to that of a classic window comparator, the difference being that, firstly, the window can be asymmetrically positioned relative to zero due to unequal overrun values and that, secondly, the width of the window is not constant.

This three-level controller can also be implemented using a software program with a microcontroller (description related to Fig. 1).

An embodiment of the invention including an error integrator is shown in Fig. 3 in the form of a block diagram. Blocks 3 and 9 are new; in addition, block 8 takes the integrated error signal into account.

The blocks have the following functions: Block 1 performs digitization. During this process, an arriving or stored analog signal for a nominal value is converted into digital data that can be read by the microcontroller. If the nominal value signal is already provided in digital form, this block will not be needed.

Block 3 checks the change of the nominal value using a clock/timer signal and performs the following functions: If the change of the nominal value remains slight during a certain period of time, the current nominal value will be stored as a basis (basis value). This storage process prevents further storage processes until the next RESET signal is given.

If the nominal value changes independently of time and/or sign by more than a fixed limit value from the above basis value, a logic RESET signal for block 9 will be generated. The magnitude of the limit value is dependent on the respective application.

If the nominal value changes as a function of time (differential quotient) by more than a specific second limit value, a RESET-signal will also be produced for block 9. The magnitude of this second limit value is also dependent on the respective appli- cation.

Block 5 controls the function '1change left < limit value".

This block 7 is activated by the transition of a adjustment signal (left/right) at the output of block 8 to a stop status or by a adjustment signal in the opposite direction. Using the signals provided by block 5 and 6, block 7 checks when the actual value for the direction of motion last applicable has come to a standstill. At this moment (and at this moment only) when this status of inactivity has been reached, the difference present between the nominal value and the actual value (signal D) will be added to the sum of all differences obtained before in the same way. The result will be stored in a memory for compensation of the overrun of the machine in the corresponding direction.

The addition of the (residual) differences from the overrun values already determined ensures that these values adapt dynamically to changing situations. For example, a standstill before the nominal value is reached will decrease the previous overrun value and an exceeding of the nominal value will increase the previous overrun value.

Block 9 is used to improve the behaviour of the system. To achieve this, this block contains an error integrator. The input of the block is connected to the outputs of the blocks 3, 4, 7 and 8; the output is connected to block 8. The function of block 9 consists in adding up any residual errors remaining after adjustment or any differences caused by minor subsequent changes of the nominal value and supplying them to block 8 (signal Di).

Block 8 takes this added-up error comming from block 9 into account and triggers a adjustment signal at suitable intervals when the given limit values are exceeded, even if the current difference between the actual and the nominal value would not generate a readjust- ment signal. The time average obtained from the positions of the mechanical system thus set then corresponds to the nominal value required.

To adapt the integrator to the operating conditions and the machine to be controlled, additional measures are envisaged.

These can be implemented singly or in combination, depending on the requirements specified.

1. The added-up error will be set to zero if the signal from block 3 displays a major change (i.e. a limit value has been exceeded) of the nominal value (RESET signal).

2. The integrator can be switched on or off or otherwise changed by an external status signal (status signal on signal line 10), depending on general operating states (e.g. manual/ automatic).

3. The summation process is suspended for as long as the adjustment signal is active.

4. The time constant of the integrator is coupled to the overrun values determined. This means that summation will be faster for high overrun values and slower for low overrun values. Thus, the correcting adjustment signals are generated (via block 8) at about constant intervals.

5. The time constant of the integrator is also changed as a function of the frequency of the RESET signals. Frequent RESET-signals are to decrease the time constant.

As already described, the adjustment signals are computed in block 8. The use of the error integrator makes it necessary to extend this function as follows. Block 8 receives its input signals from blocks 4, 7 and 9. The output of this block is connected to drivers operating the mechanical system via signal lines 11 and 12 for the S1 and Sr signals which also supply an input signal to block 9.

In block 8, the added-up error (signal Di) is also added to the current difference between the nominal value and the actual value (signal D), and the resulting sum is then used further.

Block 8 takes this added-up error into account and triggers a adjustment signal at suitable intervals when the given limit values are exceeded, even if the current difference between the actual and the nominal value would not generate a adjustment signal. The time average obtained from the positions of the machine thus set then corresponds to the nominal value required.

As in the previous block diagrams, the block functions described here can also be implemented using a microcontroller.

Fig. 4 illustrates an embodiment of the invention using a microcontroller (20).

The actual value of the piston (22a) of a mechanical system (positioning motor 22) is recorded by a sensor (21) and fed to an analog-digital converter (ADC I) via a signal line (23).

This converter converts the analog signals to digital signals.

Together with the analog nominal values digitized by the analog-digital converter (ADC II), the actual values reach the CPU (27). Status signals (15) are read in via the input port (25).

In the CPU (27), a program stored in the ROM (29) performs the block functions described in Figs 1, 2 or 3. Intermediate values of these computations are stored in the RAM (28) of the microcontroller (20). The required time base is provided by a clock/timer cirucit (30) which contains an oscillating quartz.

In addition, the output ports (26) are connected to two drivers (31a, b) operating the mechanical system (22).

Claims

Patent claims:

1. Procedure for a three-level controller for the adjustment of at least one mechanical system, characterized by the fact that the positions and the overrun of the mechanical system are measured and stored at specific intervals by a suitable sensor (21) and taken into account in the generation of the next adjustment signal so that the controller has an adaptive behaviour.

2. Procedure for a three-level controller acccording to claim 1, characterized by the fact that the overrun of the mechanical system is measured separately for two directions, that the values determined are stored in a data memory and separately taken into account in the generation of the next adjustment signal.

3. Procedure for a three-level controller according to claim 1 or 2, characterized by the fact that an error integrator additionally adds up the remaining residual errors and the added-up error is taken into account in the generation of the next adjustment signal(s).

4. Procedure for a three-level controller according to claim 3, characterized by the fact that the error integrator is reset by signals (RESET signal), that the summation process is stopped as long as a adjustment signal is active, and that the time constant of the summation process is changed as a function of other signals and operational sequences, with the influences specified either acting singly or in combination on the integrator.

5 Procedure for a three-level controller according to claim 1, characterized by the fact that the actual value is digitized (block 1) and compared with a digitized nominal value (block 2), that the difference between the actual value and the nominal value is determined (block 4), that the change of the actual value is compared with a limit value (block 5/6) and that a logic stop signal will be generated for the overrun computation (block 7) if the change of the actual value is below this limit value, that, during the overrun computation (block 7), the existing difference between the nominal value and the actual value is added to the sum of all differences obtained in the same way at the moment at which the actual value in the direction of motion last applicable has come to a standstill, that the sum thus obtained is stored (RAM) and used together with the result of the difference computation to compute a adjustment signal (block 8), with the necessary adjustment direction first being derived from the polarity of the difference between the nominal value and the actual value, that the overrun value applicable for the respective direction is added to the difference during the computation of the adjustment signal (block 8), taking into account the polarity of the value, and that a corresponding signal (S1, Sr) will be the output if the result of the addition differs from zero, otherwise the signals will remain in the stop status.

6. Procedure for a three-level controller according to claim 5, characterized by the fact that the differences between the actual value and the nominal value for the signal to the left and the signal to the right are each compared with a separate limit value (block 5 and block 6) and that a stop signal (left) and/or a stop signal (right) will be generated and incorporated in the overrun computation if the difference is smaller than a limit value (stop 1, stop r).

7. Procedure for a three-level controller according to claim 6, characterized by the fact that the amount of the change of the nominal value is compared with a limit value (block 3), that the current nominal value will be used as a basis if the change of the nominal value remains slight for a specific period of time.

8. Procedure for a three-level controller according to claim 7, characterized by the fact that an integrator (block 9) is used which adds up the residual errors remaining after adjustment or the differences caused by small subsequent changes of the nominal value (added-up errors) and that this added-up error is added to the current difference between the nominal value and actual value and that this sum results in an output signal (block 8a).

9. Procedure for a three-level controller according to claim 8, characterized by the fact that a RESET signal will be generated (block 3) if the nominal value changes markedly (exceeding the limit value) from the value used as a basis and that the value of the added error is set to zero (block 9).

10. Procedure for a three-level controller according to one of the claims 8 or 9, characterized by the fact that the integrator (block 9) is switched on or off or changed in any other way by external status signals (from signal line 10) depending on the general operating condition (e.g. manual/ automatic operation).

11. Procedure for a three-level controller according to one of the claims 8, 9 or 10, characterized by the fact that the integrator (block 9) is stopped as long as a adjustment signal is active.

12. Procedure for a three-level controller according to one of the claims 8, 9, 10 or 11, characterized by the fact that the summation process in the integrator (block 9) runs faster for high overrun values and slower for low overrun values.

13. Procedure for a three-level controller according to one of the claims 8-12, characterized by the fact that the time constant of the integrator (block 9) is changed in dependence of the frequency of the RESET signals (from block 3), with frequent RESET signals decreasing the time constant and vice versa.

14. Procedure for a three-level controller according to one of the claims 1-13, characterized by the fact that the three-level controller includes a sensor (21) determining the position of a mechanical system (22, 22a) and a microcontroller (20) which features at least one A/D converter (ADC I, ADC II), one CPU (27), one output port (26) and one clock/timer circuit (30).