HK1220751B - Method for determining hydraulic parameters in a displacement pump - Google Patents

Description

Method for determining hydraulic parameters in a positive displacement pump

The invention relates to a method for determining hydraulic parameters in a positive displacement pump. The displacement pump has a movable displacement element which delimits a metering chamber which is connected to a suction line and a pressure line via valves, whereby the pumped fluid can be alternately sucked into the metering chamber via the suction line and pressed out of the metering chamber via the pressure line by means of an oscillating movement of the displacement element. The displacement pump also has a drive for the oscillating movement of the displacement element.

There are, for example, electromagnetic driven diaphragm pumps in which the displacement element is a diaphragm that can be moved back and forth between two extreme positions, wherein in a first extreme position the volume of the metering chamber is at a minimum and in a second extreme position the volume of the metering chamber is at a maximum. Thus, when the diaphragm moves from its first position to its second position, the pressure in the metering chamber will drop, whereby the pumped fluid is sucked into the metering chamber through the suction line. In the case of a backward movement, i.e. a movement from the second position to the first position, the connection to the suction line is closed, the pressure of the pumped fluid will increase due to the volume reduction in the metering chamber, as a result of which the valve to the pressure line opens and the pumped fluid is conveyed into the pressure line. By the oscillating movement of the diaphragm, the pumped fluid is alternately sucked from the suction line into the metering chamber and the pumped fluid is sucked from the metering chamber, and the pumped fluid is conveyed from the metering chamber to the pressure line. The flow rate of the pumped fluid into the pressure line is also referred to as the metering curve. The metering curve is essentially determined by the movement curve of the displacement element.

In the case of an electromagnetic driven membrane pump, the membrane is connected to a thrust member, which is usually mounted at least partially pretensioned in a spring-loaded manner within the electromagnet. As long as the electromagnet has no current flowing through it and thus no magnetic flux is formed inside it, the spring-loaded pretension ensures that the thrust member and thus the diaphragm are held in a predetermined position, e.g. the second position, i.e. the position in which the metering chamber has the largest volume. If an electric current is now applied to the electromagnet, a magnetic flux is formed which urges a correspondingly formed thrust member within the electromagnet from its second position to its first position, which causes the pumped fluid located in the metering chamber to be conveyed from the metering chamber into the pressure line.

When the electromagnet is activated there is a substantially sudden stroke of the metering disc and thus the metering diaphragm from the second position to the first position.

Typically, such an electromagnetic driven diaphragm pump is used if the volume of fluid to be metered is significantly greater than the metering chamber volume, so the metering speed is essentially determined by the frequency or time of the current through the electromagnet. For example, if the metering speed is doubled, the current tends to flow through the electromagnet briefly twice at the same time, which in turn results in a shorter and often twice occurring movement period of the diaphragm pump.

Such a magnetometer pump is described, for example, in EP 1757809.

However, when only low metering speeds are required, the use of the magnetic metering pump reaches its limits, so that a sudden metering of the entire stroke is not advisable.

In the mentioned EP 1757809, it has therefore been proposed to provide a position sensor by means of which the position of the thrust member or the diaphragm connected thereto can be determined. By comparing the actual position of the thrust member with the predetermined target position of the thrust member, a control of the movement can follow, so that the magnetometer volume pump can also be used to deliver a significantly lower volume of fluid, since the stroke movement no longer occurs abruptly, but in a controlled manner.

In practice, it is difficult to find suitable control parameters. In practice, different control parameters are empirically determined for different thrust member position states in each case and stored in memory, so that the pump can retrieve and use the corresponding control parameters depending on the position of the thrust member.

However, determining the control parameters is very laborious. Furthermore, it is heavily dependent on the environment in the metering chamber, such as, for example, the density and viscosity of the pumped fluid. Thus, the control functions satisfactorily only when the system roughly corresponds to the desired state. In particular, when there are pressure fluctuations on the suction line and/or the pressure line, when cavitation occurs, when air accumulates in the metering chamber or when there is a change in the density of the pumped fluid, the control parameters stored in the memory are not suitable and the control accuracy decreases, so that the actual metering curve differs significantly from the desired metering curve. However, this is not desirable in particular in the case of continuous metering of very small quantities, such as for example in the case of chlorination of drinking water.

Control accuracy may be improved, for example, by measuring the density and/or viscosity of the pumped fluid and using the measurements to select control parameters.

However, for such a measurement at least one additional sensor is necessary, which would increase the sales price of the positive displacement pump and additionally require maintenance and repair. So that density and viscosity changes have hitherto not been taken into account in control.

EP 2557287 a2 describes a method for metering a reducing agent from a metering device into an exhaust gas treatment device. In the international conference of "IEEE network, sensing and control, 2009" in okangshan, kasai et al, the article "Modeling of Novel Type Diaphragm Pump" proposes a Novel Diaphragm Pump model.

Starting from the state of the art, it is therefore an object of the present invention to provide a method that allows determining hydraulic parameters such as, for example, the density and viscosity of the pumped fluid without the need for additional sensors.

This is achieved according to the invention by a method of determining a hydraulic parameter in a displacement pump, wherein the displacement pump is connected to suction and pressure lines, wherein the displacement pump has a movable displacement element which delimits a metering chamber, which metering chamber is connected to the suction and pressure lines by means of a valve, whereby pumped fluid can be alternately sucked into the metering chamber through the suction line and pressed out of the metering chamber through the pressure line by means of an oscillating movement of the displacement element, wherein a drive for the oscillating movement of the displacement element is provided, characterized in that a physical model with hydraulic parameters is built for a hydraulic system, the force exerted by the displacement element on the fluid located in the metering chamber or the pressure in the metering chamber and the position of the displacement element are determined, and at least one hydraulic parameter is calculated by means of an optimization calculation that best describes the position of the displacement element and the applied force or pressure in the metering chamber determined using the built physical model as a basis.

The hydraulic parameter refers to any parameter of the hydraulic system that affects the flow of pumped fluid through the metering chamber other than the position of the displacement element.

The hydraulic parameter is thus, for example, the density of the pumped fluid in the metering chamber and the viscosity of the pumped fluid in the metering chamber. Further hydraulic parameters are for example the hose or tube length and diameter of the hose and tube connected at least sometimes to the metering chamber.

The required determination of the position of the displacement element can be made by means of a position sensor which is normally present in any case. The speed and acceleration of the displacement element may be determined from the position of the displacement element.

If, in a preferred embodiment, the method according to the invention is used for electromagnetically driving a metering pump, preferably for electromagnetically driving a metering pump, the current through the electromagnetic drive can be measured and the force exerted by the displacement element on the fluid located in the metering chamber is determined by the measured current and the measured displacement element position. In this way, a separate pressure sensor is not required. However, the method can of course also be used with a separate pressure sensor.

An inherent characteristic of the displacement element is that the hydraulic system always varies significantly when one of the valves by which the metering chamber is connected to the suction line and the pressure line is opened or closed.

It is easiest to model the system for the case where the valve to the suction line is open and the valve to the pressure line is closed. That is, a flexible hose terminating in a tank at ambient pressure is typically mounted on a valve to the suction line.

This situation exists during the so-called suction stroke, i.e. during the displacement element moving from the second position to the first position. The hydraulic system can be described, for example, by means of a nonlinear Navier Stokes equation which takes into account laminar and turbulent flows. In addition to the density and viscosity of the pumped fluid, the diameter of the hose connecting the suction valve to the tank, the length of the hose and the height difference over which the fluid in the hose has to cross will also be considered as hydraulic parameters.

More meaningful assumptions can be made depending on the system used. By means of optimization calculations, which may occur, for example, by the known gradient method or the Levenberg-Marquardt algorithm, the hydraulic parameters contained in a physical model can be determined which best describes the pressure development in the metering head and the movement of the thrust member or the velocity and acceleration determined therefrom.

Optimization calculations refer to any calculation by which the best parameters of the system are found. The optimal parameter is the parameter that best describes the system, i.e. for which the difference between the model and the measured value is at a minimum.

The determination method according to the invention can be carried out essentially simply by repeated analysis of the suction stroke behavior.

Alternatively, however, the physical model of the hydraulic system may also be considered for the case in which the valve to the suction line is closed and the valve to the pressure line is open. However, since the pump manufacturer usually does not know initially what environment the metering pump is used in, and therefore does not know the pipe system attached to the pressure valve connecting the pressure line to the metering chamber, only general assumptions can be made here. Without information about the line system attached to the pressure valve, the constructed physical model cannot therefore be constructed as accurately as in the case of the described simplest form of the hydraulic system during the suction stroke.

In a particularly preferred embodiment, two physical models of the hydraulic system are used and then the valve opening time is measured or determined, and the respective appropriate physical model is selected as a function of the determination of the valve opening time. Basically, the method according to the invention is then carried out separately for the suction stroke and the pressure stroke. In both cases, values of hydraulic parameters such as, for example, the density and viscosity of the pumped fluid are obtained, which in practice are not completely identical. In principle, it will thus be possible to average the different values, wherein it may have to be taken into account that the values obtained during the suction stroke are weighted more heavily in averaging than the values obtained during the pressure stroke, since the actual situation is better described by the physical model during the suction stroke.

Of course, there are also more complex applications of the hydraulic system during the suction stroke.

After the hydraulic pressure parameter has been determined in the manner according to the invention, the physical model constructed can be used together with the hydraulic pressure parameter thus determined to thereby determine the pressure in the metering chamber.

This information may in turn be used to improve the motion control of the thrust member. In a preferred embodiment, a model-based control, in particular a non-linear model-based control, is provided for driving the displacement element.

In the case of model-based control, a model of the process dynamics that is as complete as possible is developed. In short, using this model, it is possible to predict where the system variables will move the next moment.

Then, suitable manipulated variables may also be calculated from the model. The characterizing feature of such model-based control is thus the necessary manipulated variable which is constantly calculated from the measured variables using the system variables given by the model.

Basically, the underlying physical system is approximately described mathematically by modeling. The mathematical description is then used to calculate the manipulated variables from the measured variables obtained. Unlike known metrology curve optimization methods, the drive is therefore no longer considered a "black box". Rather, known physical relationships are used to determine the manipulated variables.

In this way, a significantly better quality of control can be achieved.

In a preferred embodiment, the position of the displacement element and the current through the electromagnetic drive are measured, and a state space model is used for model-based control using the position of the displacement element and the current through the magnetic coil of the electromagnetic drive as measured variables.

In a particularly preferred embodiment the state space model is devoid of any further measured variables to be detected, i.e. the model is developed to make a prediction of the follow-up movement of the thrust member solely on the basis of the detected position of the thrust member and the detected current through the magnetic coil.

In a preferred embodiment, the determined hydraulic parameters are used.

State space models generally refer to a physical description of the current system state. For example, the state variable may describe the energy content of an energy storage element comprised in the system.

For example, a differential equation of the displacement element may be used as a model for model-based control. For example, the differential equation may be a motion equation. The equation of motion refers to a mathematical equation describing the spatial and temporal motion of the displacement element under the influence of an external influence. In a preferred embodiment, the displacement pump specific forces acting on the thrust member are modeled with equations of motion. Thus, for example, the force exerted by the spring on the thrust member or its spring constant k and/or the magnetic force exerted by the magnetic drive on the thrust member can be simulated. The force exerted by the pumped fluid on the thrust member may then be considered a disturbance variable. In a particularly preferred embodiment, the disturbance variable can then be simulated again using the determined hydraulic pressure parameter.

With the aid of such a state-space model, predictions can be made about the behavior of the following system when the measured variables are detected.

If the immediate behavior thus predicted deviates from the desired predetermined behavior, the system is influenced in a corrective manner.

In order to calculate what a suitable application of influence looks like, in the same model, the influence of the available manipulated variables on the controlled variables can be simulated. The optimal control strategy can then be adaptively selected using known optimization methods. Alternatively, it is also possible to determine the control strategy once from the model and then use it on the basis of the measured variables detected.

In a preferred embodiment, the nonlinear state-space model is therefore selected as the state-space model and the nonlinear control takes place by a control-Lyapunov function, by a smoothing-based control method with a smoothing-based feed-forward control, by an integral-back-push method, by a sliding-mode method or by a predictive control. Non-linear control by means of a lyapunov control function is preferred.

The five methods are all known in the mathematical art and are therefore not explained further here.

The Lyapunov control function is, for example, a generalized description of the Lyapunov function. An appropriately chosen lyapunov control function results in a stable behavior on the model framework.

In other words, a correction function is calculated which results in a stable solution of the model in the base model.

Generally, there are a large number of control possibilities that cause the difference between the actual curve and the target curve in the base model to become smaller.

In a preferred embodiment, the model forming the basis of the model-based control is used to formulate an optimization problem, wherein, as a quadratic condition for the optimization, the voltage in the electric motor and thus the energy supplied to the metering pump are as small as possible, but at the same time the actual curve approaches the target curve as quickly as possible and has as small an overshoot as possible. Furthermore, it may be advantageous if the measurement signal is filtered using a low-pass filter before being processed in the base model to reduce noise effects.

In a further particularly preferred embodiment, it is provided that during a suction-pressure cycle, a difference between the detected actual position curve of the displacement element and a desired target position curve of the displacement element is detected, and the target position curve corresponding to the desired target position curve minus the difference is used for the next suction-pressure cycle.

Essentially, a self-learning system is implemented here. Although the model-based control according to the invention has achieved a significant improvement in the control behavior, there may still be a deviation between the target curve and the actual curve. In particular, this cannot be avoided in the case of an energy minimization option for control interventions. To further reduce such deviations for at least subsequent cycles, deviations are detected during one cycle and the detected deviations are at least partially subtracted from the target position profile in the next cycle.

In other words, the subsequent pressure-suction cycle is intentionally provided with a "wrong" target value curve, wherein the "wrong" target value curve is calculated from the information obtained in the previous cycle. That is, if in a subsequent suction-pressure cycle there is exactly the same deviation between the actual curve and the target curve as in the previous cycle, the use of a "wrong" target value curve results in the actual said target value curve being achieved as a result.

Although it is basically possible and in some applications also sufficient due to the periodic behavior of the system to carry out the self-learning step only once, i.e. to measure the difference in the first period and to correct the target value curve correspondingly from the second and in all further periods, it is particularly preferred if the difference between the actual curve and the target curve is determined periodically, preferably in each period, and is taken into account correspondingly in the subsequent periods.

Of course, it is also possible to use only a small part of the detected differences as a curve correction for the subsequent period or periods. This may be advantageous in order not to generate system instability caused by sudden changes of the target values, especially in cases where the detected differences are very large.

Further, the current difference between the target curve and the actual curve can be used to determine the magnitude of the detected difference that is used as part of the curve correction.

It is also possible that the difference between the actual curve and the target curve is measured over several cycles, for example 2 cycles, and that an average difference is calculated from this difference, which average difference is then at least partially subtracted from the target curve of the subsequent cycle.

In another alternative embodiment, any function dependent on the detected difference may be used to correct the next target position curve.

In another preferred embodiment, the modeling according to the invention can be used to determine physical variables in a positive displacement pump. Thus, for example, the fluid pressure in the metering chamber may be determined.

The equation of motion of the displacement element takes into account all forces acting on the displacement element. In addition to the force applied to the displacement element by the driver, this is also a reaction force applied to the diaphragm and hence the displacement element by the fluid pressure in the metering chamber.

Thus, if the force applied by the drive to the displacement element is known, conclusions regarding the fluid pressure in the metering head may be made from the position of the displacement element or from the velocity or acceleration of the displacement element that may be derived therefrom.

For example, if the actual fluid pressure reaches or exceeds a predetermined maximum value, a warning signal may be issued and sent to an automatic shut-off switch that shuts off the metering pump in response to receiving the warning signal. Thus, if for any reason the valve does not open or the pressure on the pressure line suddenly increases, this can be determined by the method according to the invention without the use of a pressure sensor, and as a precaution the pump can be shut down. Basically, the displacement element with the associated drive also functions as a pressure sensor.

In a further preferred embodiment of the method, a target fluid pressure curve, a target position curve of the displacement element and/or a target current progression through the electromagnetic drive is stored for a movement cycle of the displacement element. The actual fluid pressure may be compared to a target fluid pressure, the actual position of the displacement element compared to a target position of the displacement element and/or the actual current through the electromagnetic drive compared to a target current through the electromagnetic drive, and a warning signal may be issued if the difference between the actual value and the target value meets a predetermined criterion.

The basic idea forming the method step is that certain events, such as, for example, air bubbles in a hydraulic system or cavitation in a pump head, cause identifiable changes in the desired fluid pressure, and conclusions about the mentioned events can thus be made from the determination of the fluid pressure.

The warning signal may for example trigger an optical indicator or an audible alarm. Alternatively or in combination therewith, however, the warning signal may also be made directly available to the control unit, which takes appropriate measures in response to the received warning signal.

In the simplest case, a difference between the actual value and the target value is determined for one or more measured variables or determined variables, and a warning signal is emitted if one of the differences exceeds a predetermined value.

However, in order not only to detect the occurrence of possible error events, such as e.g. bubbles or cavitation in the metering chamber, but also to discern the differences between them, separate criteria can be defined for each error event.

In a preferred embodiment, a weighted sum of the relative deviations from the target values may be determined and the criterion may be chosen such that a warning signal is emitted if the weighted sum exceeds a predetermined value.

Different weighting coefficients may be assigned to different error events. Ideally, when an error event occurs, a criterion is precisely met, and the error event can be diagnosed.

Using the method, it is thus possible to determine the pressure in the metering head without resorting to a pressure sensor, and conclusions about certain conditions in the metering head can be drawn from the pressure thus determined, which can in turn trigger the taking of certain measures.

With the method according to the invention, the pressure change can be determined very accurately.

In another embodiment, a time gradient of the measured variable or of the determined variable is thus determined, and if the time gradient exceeds a predetermined threshold value, the valve opening or the valve closing is diagnosed.

In alternative embodiments, the mass m of the displacement element, the spring constant k of the spring pretensioning the displacement element, the damping d and/or the resistance R of the electromagnetic drive_CuIs determined as a physical variable.

In a particularly preferred embodiment, virtually all of the variables mentioned are determined. This is done, for example, by calculation of a minimum value. All the variables mentioned, with the exception of the pressure in the metering chamber, represent constants which can be determined experimentally and are generally not changed during the operation of the pump. Nevertheless, fatigue symptoms of different elements may occur, which change the value of the constant. For example, the measured pressure-path progression may be compared to an expected pressure-path progression. The difference of the two gradients integrated over one period can be minimized by changing the constant variable. If, for example, it is determined that the spring constant has changed, the spring can be diagnosed as defective.

This minimization can also be carried out in the unpressurized state, i.e. when there is no fluid in the metering chamber.

Further advantages, features and application possibilities of the invention will become apparent from the following description of preferred embodiments and the accompanying drawings. There are shown:

figure 1 is a schematic view of a suction line attached to a positive displacement pump,

figures 2a-2e are examples of hydraulic parameters and their development in relation to time,

figure 3 is a schematic view of an ideal motion curve,

figure 4 is a schematic diagram of the self-learning function,

FIG. 5 is a schematic diagram of a pressure-path diagram and a path-time diagram in a normal state, an

Fig. 6 is a schematic of a pressure-path diagram and a path-time diagram with a state of a bubble in the metering chamber.

By designing a physical model of the electromagnetic metering pump system, in particular a non-linear system description of the hydraulic process in the metering chamber or in a line connected to the metering chamber, a model-based identification method can be used in real time. For this purpose, hydraulic parameters, i.e. state variables of the hydraulic model, are evaluated and system dynamics and parameters of the hydraulic process are determined.

The position of the displacement element or the velocity and acceleration of the displacement element determined therefrom, and the pressure in the metering chamber, which may be determined by the force exerted by the diaphragm on the pumped fluid, are used as measurement variables or external variables to be determined.

Generally, in the mentioned positive displacement pumps, the hydraulic system can be described in a simplified manner for the suction stroke, i.e. when the pressure valve is closed and the suction valve is open, as shown in fig. 1, since the suction line consists of a hose connecting the suction valve to the tank. The suction line is composed ofHaving a diameter D_SAnd a hose length L. The hose bridges the height difference Z.

The non-linear navier-stokes equation can be simplified if it is assumed that the aspiration line has a constant diameter and is not expandable and an incompressible fluid is used.

Using known optimization methods, such as, for example, the gradient method or the levenberg-marquardt algorithm, the hydraulic parameters which can best describe the measured position of the thrust member are now determined, or the pressure in the metering chamber is determined on the basis of the constructed model.

In fig. 2a to 2e herein, glycerol is used as an example of the fluid pumped, in each case representing the hydraulic parameters (dashed line) and the values resulting from the method according to the invention (solid line) over time.

Thus, for example, fig. 2a shows the density of the fluid being pumped. This is approximately 1260kg/m3 (dashed line). It can be seen that the method according to the invention enables the density to be determined in about 100 seconds. Although at the time point when t is 0 the determined value is still significantly smaller than the actual value, continuous optimization leads to the density value determined with the method according to the invention approaching the actual value (solid line) very quickly.

As are the hose length L (see fig. 2b), the height difference Z (see fig. 2c), the hose diameter (see fig. 2d) and the viscosity (see fig. 2 e).

The parameters determined by the method according to the invention can then in turn be used together with the constructed physical model to determine the force exerted by the hydraulic system on the thrust member.

This information can be used for control. In particular, when a model-based nonlinear control strategy is used to control the motion of the thrust member, the model developed herein may physically simulate the effect of the hydraulic system and take it into account as a form of feedforward disturbance variable.

The method according to the invention has been developed in relation to a magnetometer pump. In a preferred embodiment, such a magnetometer pump has a movable thrust member having a linkage firmly connected thereto. The thrust member is mounted axially displaceable on a longitudinal axis in a magnetic cage which is firmly anchored in the pump housing, so that when the magnetic coil in the magnetic cage is electrically activated, the thrust member with the connecting rod is sucked into the bore of the magnetic cage against the action of the pressure spring and, after deactivation of the magnet, the thrust member is returned to the starting position by means of the pressure spring. As a result of this, when the magnetic coil is continuously activated and deactivated, the thrust member and the diaphragm thus actuated perform an oscillating movement which, in the metering head arranged on the longitudinal axis, cooperates with the outlet valve and the inlet valve, resulting in a pumping stroke (pressure stroke) and an intake stroke (suction stroke). Activation of the magnetic coil occurs by a voltage applied to the magnetic coil. The movement of the thrust member may thus be formed by the development of the voltage over time on the magnetic coil.

It should be understood that the pressure stroke and the suction stroke do not necessarily have to last for the same amount of time. Conversely, since no metering takes place during the suction stroke, but the metering chamber is only refilled with pumped fluid, it is advantageous to carry out the suction stroke as quickly as possible in each case, taking care that there is still no cavitation in the pressure chamber.

On the other hand, the pressure stroke may last a long time, especially in applications where only a small amount of fluid is to be metered. This results in the thrust member only gradually moving in the direction of the metering chamber. In order to achieve the movement of the thrust member as shown in an idealized manner in figure 3, the movement of the thrust member must be controlled. Only the position of the thrust member and the magnitude of the current through the magnetic coil can generally be used as measurement variables.

According to the invention, a (non-linear) model describing the state of the magnetic force system is thus developed.

The following model was generated in the preferred embodiment:

wherein the content of the first and second substances,

m mass of thrust member

Phi magnetic flux

K_L(δ)Φ²Magnetic force

N₁Number of turns

u is voltage

d damping

k is spring constant

F_vorThe force acting on the thrust member being pre-tensioned by a spring

F_pForce acting on the thrust member by fluid pressure in the delivery chamber

Magnetic resistance

R_CuOhmic resistance of coil

x position of thrust member

Delta size of gap between anchor and magnet

This is a system of nonlinear differential equations that allows prediction of the immediate behavior of the system from a starting point.

Using this model, deviations between the target curve and the actual curve that are in the future or actually already exist can thus be identified. Furthermore, the modules may be used to calculate the likely effect of a control intervention.

Based on the current strength and the measurements of the position of the thrust member, it is determined in real time how the system may develop. Furthermore, the system can be moved back to the desired direction by the control intervention calculation, i.e. by the voltage change calculation on the magnetic coil.

Of course, there are a number of possibilities for intervention on the part of the system in terms of control. At each point in time, a stable solution can thus be sought for the dynamic system. This calculation step is repeated constantly, i.e. as frequently as the available computing power allows, to obtain an optimal control.

In the case of the model proposed here, it is generally not necessary to determine a new stable solution for the dynamic system at each point in time. In general, it is sufficient to determine a suitable correction function once from the measured variables, i.e. from the position of the thrust member and the voltage on the magnetic drive, and to use this correction function for control purposes.

Since the selected model always represents an ideal state, there will inevitably be a deviation between the target value and the actual value regardless of the control. Furthermore, the measured variable detected always contains errors (noise).

To further reduce the difference between the actual curve and the target curve, the difference is measured during the pressure-suction cycle, and the measured difference and the desired target curve are used as the target curve for the subsequent cycle. In other words, the fact that the pressure stroke cycle repeats is utilized. In the subsequent cycle, a target value curve different from the actual one is thus specified.

For clarity of explanation, this self-control principle is schematically illustrated in fig. 4. The position of the thrust member is represented on the y-axis and time is represented on the x-axis.

In the first period, the target curve for control is indicated by a broken line. The target curve corresponds to a desired target curve that is modeled as a reference curve for comparison in the third cycle. Regardless of the model-based control according to the invention, the actual curve deviates from the target curve. In the first period of fig. 4, the actual curve is thus represented, for example, by a solid line. For clarity of explanation, the deviations between the actual curves and the target curves are expressed in a more pronounced manner than they occur in practice.

In the second period, the difference between the actual curve and the reference curve of the first period is then subtracted from the target curve for the first period and the difference is used as the target curve for control during the second period. The target curve thus obtained is indicated by a dashed line in the second cycle.

Ideally, in the second cycle, the actual curve deviates from the target curve used to the same extent as observed in the first cycle. This results in an actual curve (drawn with a solid line in the second period) corresponding to the reference curve.

By measuring the position of the thrust member and the current through the magnetic drive, Fp, i.e. the force acting on the thrust member by the fluid pressure in the delivery chamber, is the only unknown variable. Using this model, the force acting on the thrust member by the fluid pressure in the delivery chamber can thus be determined. Since the surface area of the thrust member to which the fluid pressure is applied is known, the fluid pressure can be calculated from the force.

The design described by the non-linear system of the electromagnetic metering pump system enables the use of model-based diagnostic methods. For this purpose, state variables of the system model are evaluated and the pressure in the pump head of the electromagnetic metering pump is determined. For control purposes, the current sensors and position sensors necessary here are already built into the pump system, so that information is already available without the need to construct metering pumps which have to be supplemented. Using the state variables and the time variation of the pressure in the metering head of the pump, a diagnostic algorithm may then be executed.

Thus, for example, model-based diagnostics of excess pressure in the process and automated pump shut-down may be achieved.

The valve opening time and the valve closing time can be identified, for example, by determining and evaluating the time gradient of the associated state variable of the system model. It is possible to detect when the state gradient overshoots or undershoots by means of predetermined limits, which leads to the identification of the valve at the opening time and the valve closing time.

As an alternative embodiment, the pressure can also be determined from the position of the thrust member, and the valve opening time point and the valve closing implementation can be derived from the evaluation. The corresponding pressure-path diagram is shown on the left side of fig. 5. The associated path-time diagram is shown on the right of fig. 5. The path-time diagram shows the motion of the thrust member in relation to time. It can be seen that the thrust member is first moved forward from the starting position 1(x ═ 0mm) and the volume of the metering chamber is reduced (pressure phase). At time point 3, the thrust member passes through a maximum value and then moves back to the starting position (suction phase).

The corresponding pressure-path diagram is shown on the left side of fig. 5. It will start at the origin of the coordinates where the thrust member is located at position 1, proceeding in a clockwise direction. During the pressure phase, the pressure in the metering chamber first increases sharply until the pressure can open the valve to the pressure line. Once the pressure valve is opened, the pressure in the metering chamber remains substantially constant. The opening point is indicated with reference numeral 2. Metering is started from this point in time, which is also recorded on the right side of fig. 5. With each further movement of the thrust member, the metering fluid is pumped into the pressure line. Once the thrust member has reached the maximum position (time point 3), the movement of the thrust member reverses, the pressure valve immediately closes and the pressure in the metering chamber drops again. As soon as the minimum pressure is reached (point in time 4), the suction valve connecting the metering chamber to the suction line opens and the metering fluid is sucked into the metering chamber until the starting position is reached again.

The valve closing time point may be determined from the path-time map when it is at the displacement maximum of the thrust member. The time points 2 and 4, i.e. the valve opening time points, are not easy to determine, especially since in practice the pressure-path diagram has rounded "corners". Starting from position 1 in the pressure-path diagram, for example when 90% of the maximum pressure value is reached (known from position 3), the path can therefore be read out and the increase of the pressure-path diagram between points 1 and 2 can be determined. The 90% curve is drawn with a dashed line. The straight line thus formed and the curve p ═ p_maxAt the valve opening time point. Time point 4 may also be in the same directionAnd (4) determining the formula. This determination may occur in each cycle and the results used for the next cycle. A change in the point in time can thus also be detected.

By comparing the target and actual trajectories of the individual state variables of the system model, it is possible to diagnose air bubbles in the hydraulic system, cavitation in the pump head of the metering unit and/or valve opening and valve closing times of the metering unit. In particular, when a predetermined error limit is exceeded between the target trajectory and the actual trajectory, this may trigger a warning signal and corresponding measures.

One is shown in fig. 6. Here, too, the pressure-path diagram is shown on the left, while the path-time diagram is shown on the right. The right drawing is the same as the corresponding drawing of fig. 5. If there are compressible bubbles in the hydraulic system, this will result in the pressure valve opening only at point 2 'and the suction valve opening only at point 4'. A significant shift in the valve opening time point can therefore be used to diagnose the state "air in the metering chamber". In the case of cavitation, only the valve opening time point 4' is shifted and the valve opening time point 2 is not shifted, so that this behavior can be used to diagnose the state "cavitation".

Through analysis of individual associated system state variables, the proposed model-based methodology enables a substantially more comprehensive and useful diagnosis than has heretofore been achieved.

Furthermore, this can be achieved at low cost in terms of sensors and high reliability and reliability. With a higher diagnostic quality, the field of use of electromagnetic metering pumps can be expanded, since the metering accuracy can now be greatly increased.

Claims (24)

1. Method for determining hydraulic parameters in a displacement pump which is connected to suction and pressure lines, wherein the displacement pump has a movable displacement element which delimits a metering chamber which is connected to the suction and pressure lines via valves, whereby pumped fluid can be alternately sucked into the metering chamber via the suction line and pressed out of the metering chamber via the pressure line by means of an oscillating movement of the displacement element, wherein a drive for the oscillating movement of the displacement element is provided, characterized in that a physical model with hydraulic parameters is constructed for a hydraulic system, the force exerted by the displacement element on the fluid located in the metering chamber or the pressure in the metering chamber and the position of the displacement element are determined, and at least one hydraulic parameter is calculated by means of an optimization calculation, the optimization calculation best describes the position of the displacement element and the applied force or pressure in the metering chamber determined using the constructed physical model as a basis.

2. The method of claim 1, wherein the density of the fluid in the metering chamber and/or the viscosity of the fluid in the metering chamber is determined as a hydraulic parameter.

3. The method of claim 1, wherein the positive displacement pump is an electromagnetically driven metering pump.

4. The method of claim 3, wherein the positive displacement pump is an electromagnetic driven diaphragm pump.

5. The method of claim 4, wherein the current through the electromagnetic drive is measured and the force exerted by the displacement element on the fluid located in the metering chamber is determined by the measured current and the measured position of the displacement element.

6. A method as claimed in claim 3, characterized in that the physical model is configured for the case in which the valve to the suction line is open and the valve to the pressure line is closed and/or for the case in which the valve to the suction line is closed and the valve to the pressure line is open, wherein a valve-opening time point is determined if the physical model is configured not only for the case in which the valve to the suction line is open and the valve to the pressure line is closed but also for the case in which the valve to the suction line is closed and the valve to the pressure line is open, and the physical model is selected on the basis of the determination of the valve-opening time point.

7. Method according to claim 5, characterized in that the physical model is configured for the case in which the valve to the suction line is open and the valve to the pressure line is closed and/or for the case in which the valve to the suction line is closed and the valve to the pressure line is open, wherein a valve opening time point is determined if the physical model is configured not only for the case in which the valve to the suction line is open and the valve to the pressure line is closed but also for the case in which the valve to the suction line is closed and the valve to the pressure line is open, and the physical model is selected on the basis of the determination of the valve opening time point.

8. The method of one of claims 1 to 7, wherein after the hydraulic parameter is determined, the hydraulic parameter and the physical model are used to determine a force exerted by the pumped fluid on the displacement element, and the force so determined is used to control the movement of the displacement element.

9. Method according to one of claims 1 to 7, characterized in that a model-based control is used for the drive to optimize the metering curve of the positive displacement pump.

10. The method of claim 9, wherein a differential equation of the displacement element is used as a model for the model-based control.

11. The method of claim 10, wherein the differential equation is a motion equation.

12. The method of claim 10, wherein the differential equation is used to model a specific force of the positive displacement pump acting on a thrust member of the positive displacement pump.

13. The method of claim 9, wherein a non-linear state space model is selected as the state space model, wherein the non-linear control is performed by a lyapunov control function, by a smoothing-based control method with a smoothing-based feed forward control, by an integral back-push method, by a sliding mode method, or by a predictive control.

14. The method according to claim 9, characterized in that a difference between the detected actual position curve of the displacement element and a predetermined target position curve of the displacement element is detected during a suction-pressure cycle, and the difference of at least a part of the detected difference and the predetermined target position curve is used as the target value curve for the next suction-pressure cycle.

15. The method of claim 10, wherein a physical variable in the positive displacement pump is determined using the differential equation.

16. The method of claim 11, wherein physical variables in the positive displacement pump are determined using the equation of motion.

17. Method according to claim 15, characterized in that the fluid pressure p of the pumped fluid in the metering chamber of the positive displacement pump is determined as a physical variable.

18. Method according to one of the claims 15 to 17, characterized in that a warning signal is issued if the actual fluid pressure reaches or exceeds a predetermined maximum value.

19. A method according to claim 5, characterised in that a target fluid pressure curve, a target position curve of the displacement element and/or a target current progression through the electromagnetic drive is stored for a movement cycle of the displacement element, and the actual fluid pressure is compared with the target fluid pressure, the actual position of the displacement element is compared with the target position of the displacement element and/or the actual current through the electromagnetic drive is compared with the target current through the electromagnetic drive, and a warning signal is issued if the difference between the actual value and the target value meets a predetermined criterion.

20. A method according to claim 19, wherein a weighted sum of the relative deviations from the target values is determined and the criterion is chosen such that a warning signal is issued if the weighted sum exceeds a predetermined value.

21. A method according to claim 19 or 20, wherein a number of criteria are predetermined, an error event is assigned to each criterion, and if the criteria are met, the assigned error event is diagnosed.

22. A method according to claim 9, characterised in that the mass m of the displacement element, the spring constant k of the spring pretensioning the displacement element and/or the damping d are determined as physical variables.

23. The method of claim 5, wherein the resistance R of the electromagnetic driver_CuIs determined as a physical variable.

24. The method of claim 18, wherein the warning signal is sent to an automatic shut-off switch that shuts down the positive displacement pump in response to receiving the warning signal.