DE2915424A1 - Lift car speed control for min. journey times - calculates required speed curve in dependence on path to be covered - Google Patents

Abstract

The lift car speed control regulates the speed of the car according to a given programme. This regulation is obtained via successive integration. Pref, a correction value is added to the initial value of the required speed of the car, obtained from the difference between the speed requirement value and a speed value dependent on the distance covered by the car. The car acceleration and the variation in the acceleration are subjected to max. limitation and the required speed is pref. calculated continuously during the movement of the car.

Description

The aim of regulating cars in elevator systems is to regulate the speed in such a way that the car is moved within a very short time from a specific stop to another specific stop with a given speed profile. The variables: set speed, acceleration and jerk must not exceed certain adjustable values. The task of a controller is to use the entered travel path, the specified travel curve, which results from the set limit values, to produce the target speed curve. The controlled system with setpoint generator and the functions involved are the subject of Fig. A.

With the time-dependent control with integrators, which is already known for this, the acceleration or deceleration process is purely time-dependent, i.e. the speed setpoint curve is determined solely by the integration time constants and determines the course of the control variable "jerk".

Disturbances from zip lines, rope stretching, etc. cannot be corrected. During the assembly, a comparison of flags is necessary for each floor, the distances from the stopping point to define the maximum possible deceleration distance.

The course of the control signal "delay" (delay), which effects the division of the delay process, and the setpoint speed (V [low] setpoint), plotted against time, results in the following diagrams.

By adding a measuring system, the time-dependent version could be improved to the extent that the necessary total deceleration path is calculated and the deceleration is initiated accordingly later. Since this would not eliminate the disadvantages of the time-dependent version, the effort cannot be justified. The following representations would result as corresponding diagrams.

The invention is now based on the consideration that if one calculates the distances covered during a certain "jerk" phase from the given jerk, the maximum acceleration and the target speed, a control method can be derived from this: By comparing the calculated distances with the distance covered, the control variable "jerk" can be determined by a simple threshold value decision.

The advantage of this version according to the invention is that no hardware changes are necessary during the transition from the known time-dependent to the path-dependent control according to the invention. However, there remains the disadvantage that disturbances become so noticeable that the range of the advancing speed is reached too early or not at all. This situation can be represented in diagrams as follows, where Dec is the actual driving speed, V is the target speed, S is the distance covered and t is the time. If the advancement speed is reached correctly, this is shown in the following diagrams.

If the advancement speed V [low] n is not reached, this is shown in the following diagrams:

Finally, reaching the advancement speed V [low] n too early is shown in the following diagrams.

The mentioned disadvantage of this version according to the invention can be remedied by additive superimposition of a correction signal on the output variable V [low] should. This correction signal is calculated from the difference between the target speed value and the speed value calculated back from the way. Each disturbance variable can be successively corrected by repeating the calculation process. Detrimental to the accuracy this only affects the length of the math expressions.

In the further embodiment of the invention, a large jump in the control quality can be achieved by direct non-linear feedback in the final phase of the deceleration process: the integrators are decoupled within a certain distance from the stop and the speed signal is calculated from the remaining distance. The requirement for defined and finite jerks and acceleration is dropped in favor of precise control in the area of the bus stop. This solution leads to the following block diagram.

The situation is shown in diagrams as follows: In point A of the diagram immediately below, the switchover from time-dependent control to direct path-dependent control takes place with a constant delay.

If one now has a control unit, in particular a microprocessor, calculate the desired speed variable not only in the stop area but during the entire driving process, then a path-dependent control can be implemented. The direct feedback, which allows a high control quality to be expected, is advantageous. The variables jerk, acceleration and distance determine the course of the speed setpoint curve. The following results in diagram form using the values that have already been defined:

The following situation arises in the form of a block diagram, where µP is the symbol for the microprocessor to be used.

The invention will now be described in detail below. The abbreviations used have the following meanings:

The object of the invention is to create a method for improving the controller quality of a path-dependent setpoint generator which calculates the output variable of the setpoint generator directly as a function of the deceleration path within a driving cycle. The target course of the output variable should be specified.

Test drives with conventional elevators show that their setpoint generators do not fully meet the requirements with regard to control quality. The range of the possible braking distance, the rounding errors due to the scaling and format changes, the quantization errors due to the analog-digital and digital-analog conversion, as well as errors caused by external disturbances such as zip lines, cause the so-called advancement speed was reached too early or not at all.

To improve this driving behavior, a path-dependent control is therefore proposed, which calculates the target speed directly as a function of the braking distance still to be covered. The implementation of this invention in practice requires a far-reaching change in the hardware and software previously used. However, the practical demands on the control quality of elevators are increasing to such an extent that they also justify a considerable change in previous solutions using the latest technical findings.

I. Regulatory procedure

In previously implemented methods, the process of the output variable of the controller (V [low] setpoint) is controlled via the manipulated variable (jerk). The set values jerk, max.acceleration and max.set speed become before the start of the deceleration process calculates the waypoints at which the jerk switch should take place in the event of an undisturbed driving cycle.

During the braking process, the distance covered is checked and the jerk is controlled accordingly when a calculated position marker is reached. This procedure cannot compensate for errors caused by zip lines, etc. An improvement of this method through a continuous correction of the target speed leads to very computationally time-intensive control algorithms. Since the derivatives of the actual braking distance are not available according to the time, an observer would be necessary to record these variables. A realization of this concept would only be possible with hardly justifiable effort due to the necessary computing time.

The already mentioned path-dependent control is proposed to remedy these disadvantages. The time-dependent courses of the variables: jerk, acceleration, target speed and braking distance are specified using the integral equations. If the time is eliminated from this system of equations, a system is obtained in which the braking distance is only a function of the boundary conditions and the current target speed, or the target speed is a function of the boundary conditions and the braking distance.

This mathematical relationship represents a control that is very easy to implement on a control unit, in particular a microprocessor (µP) system. The control unit or the µP either calculates the output variable V [low] set on-line, ie during the driving process, or reads it the values from a data memory. The actual path of the elevator car recorded by the copier is used as the variable.

Two variants can be implemented with the table method: The route read in is used as an address pointer for the data memory. The content of a data cell represents the target speed associated with a certain path. This path-constant table method is shown in Fig. 1. With this method, the change in the target speed - only this is noticed by a passenger - is path-dependent. The cost of data storage cells is relatively high for this "path constant" solution, since the necessary resolution for the path must be very fine. For the application, it would therefore be expedient in a further embodiment of the invention to keep the speed increase constant to keep. To implement this method, the set speed is used as an address pointer. This address is the waypoint at which a switch to the next speed level is to take place. If the entered path is identical to the saved waypoint, the target speed is reduced by a fixed amount. At the same time, the address pointer is incremented and points to a cell in which the marker for the next switchover is stored. This rate-constant table method is shown in Fig. 2. This process is repeated until the calculated braking distance has been covered. The advantage of this method is that only a minimal number of storage locations is required and that the changes in the setpoint speed are constant. The "voltage jumps" at the digital-to-analog converter can be rounded off by a downstream low-pass filter, so that the permissible speed increase can be further increased. The latter procedure was implemented for the travel-dependent setpoint generator.

With the on-line calculation method, this path is used in the formulas for the setpoint speed. The variables: jerk, acceleration, initial speed and final speed are implicitly contained in the formula as factors. Both the path constant as well as the rate constant solution possible.

I.1 Calculation of the state variables as a function of time and

I.2 Driving mode with limited acceleration.

For time-optimal behavior, the course of the target speed is calculated according to the Pontryagin minimum principle as the result of a double integration of the jerk. The amplitude of the jerk, the maximum acceleration and the start and end speed are specified as boundary conditions. This results in a course of the setpoint speed with limited or without limited acceleration. According to Fig. 3, these two driving modes differ in the amplitude of the acceleration. In Fig. 3, first the time curves of the state variables are shown for a driving mode with limited acceleration, then a time diagram for a deceleration cycle without limited acceleration. If the maximum permitted acceleration is reached, a certain distance is traveled with this acceleration. (Area II). In the other case, area II is skipped, ie the final values from area I are used as the starting values of area III. The decision as to whether the driving mode is limited or unlimited can be derived from the following formula:

If the condition is met, the driving mode with limited acceleration and the like applies.

The following boundary conditions and equations apply to area I:

From this it follows for the values at time T2

The following relationships apply to Area II:

The point in time T3 results from the boundary conditions in area III to:

The following relationships apply to area III:

I.1.3. Driving mode with unlimited acceleration.

The timing diagrams shown in Fig. 4 serve to explain the driving mode with unlimited acceleration.

The following boundary conditions and equations apply to area I:

Time T2 results from the target speed:

The boundary conditions apply to area III:

1.2 Mathematical derivation of the control algorithm

Within a range, the progressions of the state variables are a function of time and the initial conditions, which correspond to the end conditions of the previous range. The relationship between the individual equations is established by the integral. If you solve one of these equations for time and insert this expression into another equation, the direct dependence on time disappears and the resulting new expression describes the connection with the variable that was solved after the variable time.

In the following two control methods are to be calculated for a travel cycle and a constant speed driving cycle.

The constant path mode describes the target speed profile as a function of the distance covered: As soon as a certain distance is reached, a new output variable is specified. The interval between these waypoints is constant. In constant-speed mode, the setpoint speed is a function of the distance covered calculated. This enables a constant speed increase: As soon as the path belonging to the next speed level has been reached, this is output at the output. The advantage of this method is the minimal number of storage spaces required.

I.2.1 With constant path change

I.2.1.a Driving mode with limited acceleration

Area I.

The solution to this cubic equation is calculated as follows:

Inserted into the output variable results in:

This expression describes the relationship between setpoint speed and distance. However, the expressions cos [high] 2 (X) and arc cos (Z) must be split up into simpler terms by means of a mathematical transformation, e.g. the power series expansion.

The trigonometric expressions can be expanded into the following series:

If the series expansion of arc cos z is inserted into the series expansion of cos [high] 2 x, the expression can be further summarized. The required accuracy determines the degree of final expression from the shape

The computing time for the series can be further shortened if the constants C [deep] V are calculated in advance as far as possible and the equation is solved according to the Horner scheme:

Area II:

Written as a polynomial and resolved results in:

The calculated relationship represents the relationship between target speed and path valid in area II. No series expansion is necessary for the square root expression, as it is one of the standard routines of every µP arithmetic program.

Area III:

This expression, used in (1), describes the dependence of the target speed on the path in area III. Similar to area I, the hyperbolic functions can be developed in series:

Inserting this expression into equation 1 results in an expression that is favorable for µP programming:

The dependence of the target speed and the time on the path with limited acceleration is shown in Fig. 5.

Method with constant path change.

I.2.1.b Driving mode without limited acceleration

Area I.

Equation (2) solved for t and inserted into (1) gives:

This printout can be resolved as described above.

Area III

Equation (2) solved for t and inserted into (1) gives:

If you insert this expression in 1, you get the relationship between setpoint speed and distance. The expansion of the expression into a polynomial of the nth degree must be carried out in the manner described above.

The dependence of the target speed and the time on the path with unlimited acceleration is shown in Fig. 6.

I.2.2 Method with constant speed change

I.2.2.a Driving mode with limited acceleration

Area I.

If you solve (1) for t and insert it into (2), you get the path as a function of the target speed.

This expression can be realized without difficulty with a µP program.

Area II

Equation (1) solved for t and inserted into (2) gives:

For area II there is a quadratic relationship between distance and setpoint speed. Programming this expression is not a problem.

Area III

Equation (1) solved for t and inserted into (2) gives:

The expression can be further simplified by resolving according to the Horn scheme and calculating the constants in advance.

The dependence of the distance and the time on the target speed with limited acceleration is shown in Fig. 7.

Method with constant speed change

I.2.2.b Driving mode without limited acceleration

Area I.

Equation (1) solved for t and inserted into (2) gives:

The expression for the way can be realized without difficulties on the µP.

Area III

Equation (1) solved for t and inserted into (2) gives:

If one summarizes the constants in this expression and calculates them in advance, then the computing time can be drastically reduced.

The dependence of the distance and the time on the target speed with unlimited acceleration is shown in Fig. 8.

I.3. Hardware concept

To implement this setpoint generator concept, it is split into a control unit and a programming unit. The structure of the control unit can be seen in Fig. 9. For cost reasons, a minimal solution is sought, i.e. the function sequence (target speed path) is saved in a table.

An on-line calculation would presumably - if the control process requires it for reasons of resolution - require an arithmetic module as a slave processor, as well as a series of switches in each setpoint generator. Due to the division, these modules are only required once in the programming device. The structure of the programming device is shown in Fig. 10.

The control panel could also be designed as a keyboard.

When it comes to hardware implementation, value is placed on a cost-effective, but nevertheless universally applicable plug-in technology. The structure of the slide-in system and the use of certain µP types are not part of the invention.

The hardware concept for the setpoint generator is shown in Fig. 9.

The hardware concept for a setpoint generator programming device is shown in Fig. 10.

Claims (16)

1. Control of the car of elevators, characterized in that the control of the target travel speed of the car takes place according to a predetermined program as a function of the distance covered by the car. 2. Control according to claim 1, characterized in that the regulation takes place as a function of the path with integrators. 3. Control according to claim 2, characterized in that a correction signal is additively superimposed on the output value of the setpoint speed. 4. Control according to claim 3, characterized in that the correction signal results from the difference between the desired speed variable and the speed variable resulting from the path covered by the car. 5. Control according to claim 4, characterized in that each disturbance variable is successively corrected by repeating the arithmetic process. 6. Control according to claim 2, characterized in that the integrators are decoupled within a certain distance from the stop and the speed signal is calculated from the remaining distance. 7. Control according to claim 6, characterized in that the speed setpoint value is calculated not only in the bus stop area, but throughout the entire driving area. 8. Control according to claim 7, characterized in that the calculation of the speed setpoint value is carried out with a microprocessor. 9. Control according to one of claims 1 to 8, characterized in that the controlled system is represented by an integrator. 10. Control according to one of claims 1 to 9, characterized by a compulsory limitation of the acceleration and the change in acceleration of the car. 11. Control according to one of claims 1 to 10, characterized in that pulses coming from a copier are fed to a travel counter so that the read-in path of the car serves as an address pointer for the data memory, in which the data cell content is the target speed associated with a certain path represents (= constant path table method). 12. Control according to claim 11, characterized in that the content of the travel counter serves as a variable of an internal computer program, so that the target speed can be calculated therefrom during the driving process. 13. Control according to one of claims 1 to 10, characterized in that pulses coming from a copier are fed to a path counter, the target speed of the car being used as the address pointer and the address of the waypoint at which the switch to the next speed level is located should take place ("constant speed table procedure"). 14. Control according to claim 13, characterized in that the content of the path counter is compared with the content of a register in which the waypoint for the switchover is stored to the next speed level, the content of the register being calculated during the driving process according to an internal computer program. 15. Control according to claim 11 or claim 13, characterized in that the content of the memory cells is part of a fixed program or that the content of the memory cells is calculated by a program from the read path. 16. Control according to claim 11 or claim 13, characterized in that the data memory is followed by a low-pass filter.