EP3039289B1 - Method for determining hydraulic parameters in a positive displacment pump - Google Patents

Description

The present invention relates to a method for determining hydraulic parameters in a positive displacement pump. The positive displacement pump has a movable displacement element which delimits the metering chamber, which is connected via valves to a suction and pressure line, so that pumped by an oscillating movement of the displacement alternately conveying fluid via the suction line into the metering chamber and are pressed via the pressure line from the metering can. Displacement pumps additionally have a drive for the oscillating movement of the displacer element.

For example, there are electromagnetically driven diaphragm pumps in which the displacer element is a membrane which can be moved back and forth between two extreme positions, wherein in the first extreme position the volume of the metering chamber is minimal, while in the second extreme position the volume of the metering chamber is maximal , Therefore, if the membrane is moved from its first position into the second, the pressure in the dosing chamber will drop, so that the conveying fluid is sucked into the dosing space via the suction line. In the return movement, i. in the movement from the second to the first position, the connection to the suction line is closed, the pressure of the conveying fluid will increase due to the decreasing volume in the metering chamber, so that the valve is opened to the pressure line and the conveying fluid is conveyed into the pressure line. Due to the oscillating movement of the membrane, conveying fluid is alternately sucked from the suction line into the metering space, and conveying fluid is sucked out of the metering space, and conveying fluid is conveyed from the metering space into the pressure line. The conveying fluid flow into the pressure line is also referred to as dosing profile. This dosing profile is essentially determined by the movement profile of the displacement element.

In electromagnetically driven diaphragm pumps, the membrane is connected to a pressure piece, which is usually spring biased at least partially mounted within an electromagnet. As long as the electromagnet is not traversed by a current, so that no magnetic flux is built up in its interior, the resilient bias ensures that the pressure piece and thus the membrane in a predetermined position, for example, the second position, ie the position in which the Dosing chamber has the largest volume remains. If a current is impressed on the electromagnet, then a magnetic flux is formed, which forms the correspondingly formed pressure element within the electromagnet from its second position brings in the first position, whereby the conveying fluid located in the dosing chamber is conveyed from the dosing into the pressure line.

With the activation of the electromagnet, therefore, there is substantially abruptly a stroke of the metering piece and thus of the metering membrane from the second to the first position.

Typically, such electromagnetically driven diaphragm pumps are used when the volume of fluid to be metered is significantly greater than the Dosierraumvolumen, so that the metering rate is determined essentially by the frequency or the timing of the current flow through the electromagnet. If, for example, the dosing speed is to be doubled, then the electromagnet is twice as frequently flowed through with a current in the same time, which in turn means that the movement cycle of the diaphragm is shortened and takes place twice as often.

Such a magnetic metering pump is for example in the EP 1 757 809 described.

However, the use of these Magnetdosierpumpen then reaches its limits when only low dosing speeds are required, so that the abrupt dosing of a whole stroke is not desired.

In the mentioned EP 1 757 809 Therefore, it has already been proposed to provide a position sensor with which the position of the pressure piece or the associated membrane can be determined. By comparing the actual position of the pressure piece with a predetermined desired position of the pressure piece, a control of the movement can then take place, so that magnetic metering pumps can also be used to deliver significantly smaller amounts of fluid, since the lifting movement no longer takes place abruptly, but regulated.

In practice, it is difficult to find suitable control parameters. In fact, for different pressure piece position states, different control parameters are respectively empirically determined and stored in a memory, so that the pump can retrieve and use the corresponding control parameters as a function of the position of the pressure piece.

However, the determination of the control parameters is very expensive. In addition, it depends strongly on the conditions in the metering chamber, such as the density and the viscosity of the conveying fluid from. Therefore, the control works satisfactorily only when the system is approximately in the desired state. In particular, in the case of pressure fluctuations on the suction and / or pressure line, the occurrence of cavitation, the accumulation of air in the dosing chamber or changes in viscosity in the conveying fluid are the control parameters stored in the memory unsuitable and the control accuracy decreases, so that the actual metering profile differs significantly from the desired metering profile. However, this is particularly undesirable in the continuous metering of very small quantities, such as in drinking water chlorination.

The control accuracy can be improved, for example, by measuring the density and / or viscosity of the delivery fluid and using the measurement result for the selection of the control parameters.

For such a measurement, however, at least one additional sensor is necessary, which would increase the sales price of the positive displacement pump and is also prone to maintenance and repair. Therefore, so far density and viscosity changes are not taken into account in the scheme.

The EP 2 557 287 A2 describes a method for metering a reducing agent from a metering device into an exhaust gas treatment device. In the article " Modeling Novel Type Diaphragm Pump ", Kasa et al., IEEE, International Conference on Networking, Sensing and Control, 2009, pp. 647-652, Okayama, Japan a model of a new type of diaphragm pump is proposed.

Starting from the described prior art, it is therefore an object of the present invention to provide a method which allows the determination of hydraulic parameters, such as e.g. the density or viscosity of the delivery fluid is allowed without the need for additional sensors.

This is achieved by a method according to claim 1 according to the invention.

Hydraulic parameters are understood to mean any parameter of the hydraulic system, apart from the position of the displacer, which influences the flow of the fluid through the metering chamber.

Hydraulic parameters are therefore, for example, the density of the delivery fluid in the metering chamber and the viscosity of the fluid in the metering chamber. Further hydraulic parameters are, for example, hose or pipe lengths and diameters of hoses and pipes which are at least temporarily connected to the metering space.

The necessary position determination of the displacer element can take place via the usually existing position sensor. From the position of the displacement element, the speed and the acceleration of the displacement element can be determined.

If the method according to the invention is used in an electromagnetically driven metering pump, and best in the case of an electromagnetically driven diaphragm pump, then so In a preferred embodiment, the current through the electromagnetic drive can be measured and the force exerted by the displacement element on the fluid in the dosing space can be determined from the measured current and the measured position of the displacement element. In this case, no separate pressure sensor is necessary. Of course, however, the present method can also be used with a separate pressure sensor.

It is an inherent characteristic of the positive displacement pump that the hydraulic system always changes significantly when one of the valves, via which the metering chamber is connected to the suction and pressure line, is opened or closed.

The easiest way to model the system is in case the valve to the suction line is open and the valve to the pressure line is closed. Usually, namely, a flexible hose is mounted on the valve to the suction line, which ends in a reservoir under ambient pressure.

This condition is during the so-called suction stroke, i. while the displacer moves from the second position to the first position. This hydraulic system could, for example, be described by means of the nonlinear Navier-Stokes equation taking into account laminar and turbulent flows. In addition to the density and viscosity of the conveying fluid then the diameter of the hose connecting the suction valve to the reservoir, the length of the hose and the difference in height, which must overcome the fluid in the hose, are to be considered as hydraulic parameters.

Depending on the system used, further meaningful assumptions can be made. By means of an optimization calculation, which can take place, for example, via the known gradient method or Levenberg-Marquardt algorithms, it is possible to determine the hydraulic parameters contained in the physical model, the pressure curve in the dosing head and the movement or the speed determined therefrom Best describe acceleration of the pressure piece.

An optimization calculation is any calculation that finds the optimum parameters of the system. Optimal parameters are the parameters that best describe the system, i. where the difference between model and measured situation becomes minimal.

In principle, the determination method according to the invention could be carried out solely by a repeated analysis of the suction stroke behavior.

Alternatively, however, the physical model of the hydraulic system can be considered in the case that the valve is closed to the suction line and the valve is open to the pressure line. However, since it is initially not known to the pump manufacturer in which environment the metering pump is used and therefore also the pipe system connected to the pressure valve which connects the pressure line to the metering chamber is not known, only a generalized assumption can be made here , The erected physical model can therefore be set up without knowledge of the piping system connected to the pressure valve, not in the accuracy as it is in the described simplest form for the hydraulic system during the suction stroke.

In a particularly preferred embodiment, physical models are used for both described hydraulic systems, and then the valve opening times are measured or determined and the respective applicable physical model is selected depending on the result of the determination of the valve opening time. In essence, the process according to the invention is then carried out separately for the suction stroke and the pressure stroke. In both cases one obtains for the hydraulic parameters, such as e.g. the density and the viscosity of the pumped fluid values that do not exactly match in practice. In principle, it would therefore be possible to average the different values, whereby it may be necessary to take into account that due to the better description of the actual situation by the physical model during the suction stroke, the value obtained during the suction stroke is weighted more heavily than the averaging value determined during the print stroke.

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

After the hydraulic parameters have been determined in the manner according to the invention, the set-up physical model with the hydraulic parameters determined in this way can be used, in turn, to determine the pressure in the dosing space.

This knowledge can in turn be used to improve the movement control of the pressure piece. In a preferred embodiment, it is provided that a model-based control, in particular a non-linear model-based control, is used to drive the displacement element.

In a model-based control, the most complete model of process dynamics is developed. Simplified speaking, with the help of this model it is possible to make a prediction of where the system sizes will move in the next moment.

From this model, a suitable manipulated variable can then be calculated. Characteristic of such a model-based control is therefore the constant calculation of the necessary control variable on the basis of measured variables using the system variables given by the model.

Basically, modeling the underlying physical system is described approximately mathematically. This mathematical description is then used to calculate the manipulated variable on the basis of the measured quantities obtained. In contrast to the known Dosierprofiloptimierungsverfahren thus the drive is no longer considered a "black box". Instead, the known physical relationships are used to determine the manipulated variable.

As a result, a significantly better control quality can be achieved.

In a preferred embodiment, the position of the displacer and the current through the electromagnetic drive are measured, and for model-based control, a state space model is used which uses the position of the displacer and the current through the solenoid of the electromagnetic drive as measures.

In a particularly preferred embodiment, the state space model has no further measured variables to be detected, i. H. the model is designed to predict the immediate movement of the pad based only on the detected pad position and the sensed current through the solenoid coil.

In this case, in a preferred embodiment, the specific hydraulic parameters are used.

A state space model is usually understood to be the physical description of a current system state. For example, the state variables can describe the energy content of the energy storage elements contained in the system.

For example, a differential equation of the displacer may be used as a model for the model-based control. For example, the differential equation may be an equation of motion. An equation of motion is understood to mean a mathematical equation which describes the spatial and temporal movement of the displacer element under the influence of external influences. In this case, in a preferred embodiment, displacement-pump-specific forces acting on the pressure piece are modeled in the equation of motion.

For example, the force exerted by a spring on the pressure element, or its spring constant k, and / or the magnetic force exerted on the pressure element by the magnetic drive can be modeled. The force exerted by the conveying fluid on the pressure piece force can then be treated as a disturbance. This disturbance can then also be modeled in a particularly preferred embodiment using the particular hydraulic parameters.

By such a state space model, when the measured quantities are detected, a prediction for the immediately following system behavior can be made.

If the immediately following behavior predicted in this way deviates from the desired given behavior, then the system is influenced in a corrective way.

In order to calculate what a suitable influence looks like, the influence of the available manipulated variables on the controlled variable can be simulated in the same model. With the aid of known optimization methods, the currently best control strategy can then be selected adaptively. Alternatively, it is also possible to determine a control strategy once on the basis of the model and then to apply this as a function of the acquired measured variables.

In a preferred embodiment, therefore, a nonlinear state space model is selected as the state space model and the nonlinear control takes place either via control Lyapunov functions, via flatness-based control methods with flatness-based feedforward control, via integrator backstepping methods, via sliding-mode methods or via predictive control. Nonlinear control over control Lyapunov functions is preferred.

All five methods are known from mathematics and will therefore not be explained in more detail here.

For example, Control Lyapunov functions are a generalized description of Lyapunov functions. Appropriately chosen control Lyapunov functions lead to a stable behavior within the model.

In other words, a correction function is calculated which leads to a stable solution of the model in the underlying model.

In general, there are a variety of control options that cause the difference between the actual profile and the target profile to be smaller in the underlying model.

In a preferred embodiment, the model underlying the model-based control is used to formulate an optimization problem in which, as a constraint of the optimization, the electrical voltage at the electric motor and thus the energy supplied to the metering pump is as small as possible, but at the same time as fast and as little as possible overshooting approach of the actual profile to the target profile is achieved. In addition, it may be advantageous if the measured signals are low-pass filtered prior to processing in the underlying model to reduce the influence of noise.

In a further particularly preferred embodiment it is provided that during a suction-pressure cycle, the difference between the detected actual position profile of the displacer and a desired desired position profile of the displacer is detected and for the next suction-pressure cycle, a desired position profile is used, which corresponds to the reduced by the difference desired desired position profile.

Basically, a self-learning system is realized here. Although the model-based control according to the invention has already led to a significant improvement in the control behavior, it can nevertheless lead to deviations between the desired profile and the actual profile. This is unavoidable especially in the case of energy-minimizing selection of the control intervention. In order to further reduce this deviation, at least for subsequent cycles, the deviation is detected during one cycle and the detected deviation at the next cycle is at least partially subtracted from the desired setpoint position profile.

In other words, a following pressure-suction cycle is intentionally given a "wrong" setpoint profile, and the "wrong" setpoint profile is calculated from the experience gained in the previous cycle. If, in the following suction-pressure cycle, exactly the same deviation between the actual and desired profile occurs as in the previous cycle, then using the "wrong" setpoint profile results in the actual desired setpoint profile is reached.

Although it is basically possible and due to the periodic behavior of the system in some applications also sufficient to perform the described self-learning steps only once, that is to measure the difference in the first cycle and to correct the setpoint profile accordingly from the second and in all other cycles , However, it is particularly preferred if the difference between the actual and desired profile is determined at regular intervals, preferably at each cycle, and taken into account accordingly in the subsequent cycle.

Of course, it is also possible to use only a fraction of the detected difference as a profile correction for the following or following cycles. This may be particularly advantageous in cases where the detected difference is very large in order not to create instability of the system due to the sudden change in the desired value.

Furthermore, it is possible to determine the size of the fraction of the detected difference, which is used as a profile correction, based on the actual difference between the desired and actual profile.

It is also possible that the difference between actual and nominal profile over several cycles, eg. B. 2, and from this a mean difference is calculated, which is then subtracted at least in part from the target profile of the following cycles.

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

The modeling of the invention may be used in another preferred embodiment to determine a physical quantity in the positive displacement pump. For example, the fluid pressure in the dosing chamber can be determined.

The equation of motion of the displacer takes into account all forces acting on the displacer. In addition to the force applied by the drive to the displacement element, this is also the counterforce applied by the fluid pressure in the dosing space to the membrane and thus to the displacement element.

Therefore, when the force applied by the drive to the displacer force is known, conclusions about the fluid pressure in the dosing head can be drawn from the position of the displacer element or from the speed or acceleration of the displacer element derived therefrom.

For example, when the actual fluid pressure reaches or exceeds a predetermined maximum value, a warning signal may be issued and the warning signal may be sent to an automatic cut-off device which shuts down the metering pump in response to the receipt of the warning signal. Should therefore for some reason a valve does not open or the pressure on the pressure line rise sharply, this can be determined by the inventive method without using a pressure sensor and the pump can be shut off for safety's sake. Basically, the displacement element with the associated drive additionally assumes the function of the pressure sensor.

In a further preferred embodiment of the method, a target fluid pressure curve, a desired position curve of the displacer element and / or the desired current profile are stored by the electromagnetic drive for a movement cycle of the displacement element. In this case, the actual fluid pressure with the desired fluid pressure, the actual position of the displacer with the desired position of the displacer and / or the actual current can be compared by the electromagnetic drive with a desired current through the electromagnetic drive and, if the differences between the actual value and the target value fulfill a predetermined criterion, a warning signal is output.

This method step is based on the idea that certain events, such as gas bubbles in the hydraulic system or cavitation in the pump head cause a recognizable change in the expected fluid pressure and therefore conclusions about the above events can be drawn from the determination of the fluid pressure.

The warning signal can activate, for example, a visual display or an acoustic display. Alternatively or in combination, however, the warning signal can also be provided directly to a control unit, which takes appropriate action in response to the receipt of the warning signal.

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

However, to avoid the possible error events, such as Not only to detect gas bubbles in the dosing chamber or occurrence of cavitation, but also to distinguish between them, it is possible to define a separate criterion for each fault event.

In a preferred embodiment, a weighted sum of the relative deviations from the target value may be determined and the criterion selected such that a warning signal is output when the weighted sum exceeds a predetermined value.

The different error events can be assigned different weighting coefficients. Ideally, when a miss event occurs, exactly one criterion is met so that the fault event can be diagnosed.

The described method therefore makes it possible to determine the pressure in the dosing head without resorting to a pressure sensor, and conclusions can be drawn from the pressure determined in this way be pulled to certain states in the dosing, which in turn can trigger the initiation of certain measures.

With the method according to the invention, pressure changes can be determined very precisely.

In a further embodiment, therefore, the temporal gradient of a measured or determined variable is determined and, if this exceeds a predetermined limit value, the valve opening or the valve closure is diagnosed.

In an alternative embodiment, the mass m of the displacement element, the spring constant k of the spring biasing the displacement element, the damping d and / or the electrical resistance R _{Cu of} the electromagnetic drive are determined as the physical variable.

In a particularly preferred embodiment, even all of the variables mentioned are determined. This can be done for example by a minimization calculation. All of the above-mentioned variables, with the exception of the pressure in the metering chamber, represent constants which can be determined experimentally and generally do not change during pumping operation. Nevertheless, it can lead to fatigue of the different elements that change the value of the constant. For example, the measured pressure-path curve can be compared with an expected pressure-path curve. The integrated over a cycle difference from both gradients can be minimized by varying the constant sizes. If one puts thereby for example If the spring constant has changed, a faulty spring can be diagnosed.

Such minimization could also be achieved in the unpressurised state, i. if there is no fluid in the metering chamber, be carried out.

Figur 1

eine schematische Darstellung der an die Verdrängerpumpe angeschlossenen Saugleitung,

Figuren 2a-2e

Beispiel hydraulischer Parameter und deren zeitabhängige Entwicklung,

Figur 3

eine schematische Darstellung eines idealen Bewegungsprofils,

Figur 4

eine schematische Darstellung der Selbstlernfunktion,

Figur 5

eine schematische Darstellung eines Druck-Weg-Diagramms und eines Weg-Zeit-Diagramms für den Normalzustand und

Figur 6

eine schematische Darstellung eines Druck-Weg-Diagramms und eines Weg-Zeit-Diagramms für einen Zustand mit Gasblasen in der Dosierkammer.

Further advantages, features and applications of the present invention will become apparent from the following description of a preferred embodiment and the accompanying drawings. Show it:

FIG. 1

a schematic representation of the suction line connected to the positive displacement pump,

FIGS. 2a-2e

Example of hydraulic parameters and their time-dependent development,

FIG. 3

a schematic representation of an ideal movement profile,

FIG. 4

a schematic representation of the self-learning function,

FIG. 5

a schematic representation of a pressure-path diagram and a path-time diagram for the normal state and

FIG. 6

a schematic representation of a pressure-path diagram and a path-time diagram for a state with gas bubbles in the metering chamber.

By designing a physical model, in particular a non-linear system description of the hydraulic process in the dosing chamber or in the line of an electromagnetic dosing pump system connected to the dosing space, it is possible to use model-based identification methods in real time. For this, the hydraulic parameters, i. the state variables of the hydraulic models are evaluated and the system dynamics as well as the parameters of the hydraulic process are determined.

The measured variables or external variables to be determined are the position of the displacer element or the speed and acceleration of the displacer element determined therefrom, and the pressure in the metering space, which can be determined via the force exerted by the membrane on the conveying fluid.

Since usually in the said displacement pumps, the suction line consists of a hose which connects the suction valve with a reservoir can, for the suction stroke, ie while the pressure valve is closed and the suction valve is opened, the hydraulic system will be described in simplified FIG. 1 is shown. The suction line consists of a hose with the diameter Ds and the hose length L. The hose bridges a height difference Z.

The non-linear Navier-Stokes equations can be simplified if it is assumed that the suction line has a constant diameter and is not stretchable and that an incompressible fluid is used.

With the aid of known optimization methods, such as By using the gradient method or the Levenberg-Marquardt algorithms, the hydraulic parameters are now determined which, based on the established model, can best describe the measured position of the pressure piece as well as the measured or specific pressure in the dosing space.

In the FIGS. 2a to 2e Here, using the example of glycerol as conveying fluid, a hydraulic parameter (dotted line) as well as the values resulting from the method according to the invention (solid line) over time are shown.

So shows, for example FIG. 2a the density of the delivery fluid. This is about 1260 kg / m ³ (dotted line). It can be seen that the method according to the invention is able to determine the density within about 100 seconds. Although at time t = 0 seconds is the certain value still well below the actual value. Due to the continuous optimization, however, the density value determined by the method according to the invention approaches the true value very rapidly (solid line).

The same applies to the hose length L (see FIG. 2b ), the height difference Z (see Figure 2c ), the hose diameter (see Figure 2d ) and the viscosity (see FIG. 2e ).

The parameters determined by the method according to the invention can in turn be used, together with the established physical model, to determine the force exerted by the hydraulic system on the pressure piece.

This information can be used for the regulation. In particular, if model-based non-linear control strategies are used to control the movement of the pressure piece, the model developed here can physically map the influence of the hydraulic system and take this into account in the form of a feedforward control.

The inventive method has been developed in connection with a Magnetdosierpumpe. In a preferred embodiment, such a magnetic metering pump has a movable pressure piece with a push rod fixedly connected thereto. The pressure piece is axially movably mounted in a longitudinally fixed in the pump housing in the magnetic jacket, so that the pressure piece is drawn with push rod in the electrical control of the magnetic coil in the magnetic jacket against the action of a compression spring in a bore of the magnetic shell and the pressure piece after deactivation of the magnet returns to the starting position by the compression spring. This has the consequence that the pressure piece and actuated by this membrane with continued activation and deactivation of the magnetic coil performs an oscillating movement in the arranged in the longitudinal axis of the dosing in cooperation with an outlet and inlet valve to a pumping stroke (pressure stroke) and a Intake stroke (suction stroke) leads. Activation of the solenoid is done by applying a voltage to the solenoid. The movement of the pressure piece can thus be determined by the time course of the voltage at the magnetic coil.

It goes without saying that the pressure stroke and suction stroke do not necessarily have to last the same length. Since no dosing takes place during the suction stroke, but only the dosing is filled again with conveying fluid, it is on the contrary advantageous to perform the suction stroke in each case as quickly as possible, while still making sure that there is no cavitation in the Compression chamber is coming.

The pressure stroke, however, can last very long, especially in applications where only very small amounts of fluid are to be dispensed. This has the consequence that the pressure piece moves only gradually in the direction of the dosing. To achieve a movement of the pressure piece, as it is idealized in FIG. 3 is shown, the movement of the pressure piece must be regulated. Usually, only the position of the pressure piece and the size of the current through the magnetic coil are available as a measured variable.

According to the invention, therefore, a (non-linear) model is developed, which describes the state of the magnetic system.

wobei

m:

Masse des Druckstücks

Φ:

magnetischer Fluss

K_L (δ)Φ²:

magnetische Kraft

N ₁:

Windungszahl

u:

Spannung

d:

Dämpfung

k:

Federkonstante

F_vor :

Kraft auf Druckstück durch Federvorspannung

F_p :

Kraft auf Druckstück durch Fluiddruck im Förderraum

R_mges (δ,Φ) :

magnetischer Widerstand

R_Cu :

ohmscher Widerstand der Spule

x:

Position des Druckstückes

δ :

Spaltgröße zwischen Anker und Magnet

For a preferred embodiment, the following model results: $$\dot{x}=\left[\begin{array}{c}\dot{x}\\ \ddot{x}\\ \mathrm{\Phi }\end{array}\right]=\left[\begin{array}{c}\dot{x}\\ \frac{1}{m}\left(-d\dot{x}-\mathit{kx}-F_{\mathit{in\; front}}+F_p+K_{\mathrm{<figure-callout\; id="2"\; label="L\; \delta \; \Phi "\; filenames="imgf0002.png,imgf0004.png"\; state="\{\{state\}\}">L\; </figure-callout>}}\left(\delta \right){\mathrm{\Phi }}^{}{\mathrm{}}^2\right)\\ \frac{1}{N_1}\left(-R_{\mathit{cu}}\frac{R_{m_{\mathit{ges}}}\left(\delta ,\mathrm{<figure-callout\; id="1"\; label="\Phi \; N"\; filenames="imgf0004.png"\; state="\{\{state\}\}">\Phi \; </figure-callout>},\right)}{N_{}}\mathrm{\Phi }+u\right)\end{array}\right]$$

in which

m:

Mass of the pressure piece

Φ:

magnetic river

K _L ( δ ) Φ ² :

magnetic force

N ₁ :

number of turns

u:

tension

d:

damping

k:

spring constant

F _before :

Force on thrust piece by spring preload

F _p :

Force on thrust piece by fluid pressure in the delivery chamber

R _{m ges} ( δ , Φ):

magnetic resistance

R _Cu :

Ohmic resistance of the coil

x:

Position of the pressure piece

δ :

Gap size between armature and magnet

This is a non-linear system of differential equations, which, starting from a starting point, allows us to predict the behavior of the system immediately following it.

With the aid of this model, it is therefore possible to detect future or actually existing deviations between the desired curve and the actual curve. In addition, the model can be used to calculate the anticipated influence of a control intervention.

Therefore, it is determined in real time how the system is likely to evolve based on the measurement of current and position of the pressure pad. In addition, it can be calculated by which control intervention, that is to say by which voltage change on the magnet coil, the system can be moved again in the desired direction.

Of course, there are a variety of ways to intervene in the system. Thus, stable solutions of the dynamic system can be searched for at any time. This computation step is constantly repeated, that is, as often as the available computing power permits, in order to obtain an optimal control.

In the model proposed here, it is generally not necessary to determine new stable solutions of the dynamic system at any time. As a rule, it is sufficient to use the appropriate correction function as a function of the measured quantities, i. depending on the position of the pressure piece and the voltage at the magnetic drive, to determine once and use this the correction function from now on for the control.

Despite this rule, there will inevitably be deviations between the setpoint and the actual value, since the chosen model always represents idealization. In addition, the recorded measured variables are always faulty (noise).

In order to further reduce the difference between actual and nominal profile, this difference is measured during a pressure-suction cycle and used as the desired profile for the following cycle, the sum of the measured difference and the desired nominal profile. In other words, it is exploited that the pressure-stroke cycle repeats itself. Thus, in the case of the following cycle, a setpoint profile is deviated which deviates from the actual desired setpoint profile.

For clarification, this self-regulating principle is in FIG. 4 shown schematically. Shown is the position of the pressure piece on the Y axis and the time on the X axis.

In the first cycle, a reference profile used for the control is shown with a dashed line. This desired profile corresponds to the desired nominal profile, which is shown for comparison in the third cycle as a reference profile. Despite the model-based control according to the invention the actual profile deviates from the target profile. In the first cycle of FIG. 4 is therefore an example of an actual profile shown by a solid line. The deviations between actual and nominal profile for clarity are more pronounced than they occur in practice.

In the second cycle, the difference between the actual profile of the first cycle and the reference profile is subtracted from the desired profile used for the first cycle, and the difference is used as a target profile for the control during the second cycle. The so-obtained target profile is shown in dashed lines in the second cycle.

Ideally, in the second cycle, the actual profile deviates to the same extent from the desired profile used, as was observed in the first cycle. This results in an actual profile (drawn with a solid line in the second cycle), which corresponds to the reference profile.

By measuring the position of the pressure piece and the current through the magnetic drive F _p , ie the force on the pressure piece by the fluid pressure in the pumping chamber, the only unknown size. Therefore, using this model, the force on the pressure piece can be determined by the fluid pressure in the delivery space. Since the surface of the pressure piece, which is acted upon by the fluid pressure is known, can be calculated from the force of the fluid pressure.

The described design of a non-linear system description of the electromagnetic metering pump system makes it possible to use model-based diagnostic methods. For this, the state variables of the system models are evaluated and the pressure in the pump head of the electromagnetic dosing pump is determined. The necessary current and position sensors are already installed in the pump system for control purposes, so that the information is already available without the structure of the metering pump must be supplemented. Based on the temporal change of the state variables and the pressure in the dosing head of the pump, the diagnostic algorithms can then be executed.

Thus, for example, the model-based diagnosis of process-side overpressure and the automated pump shutdown can be realized.

The detection of the valve opening and valve closing times can be done, for example, by determining and evaluating temporal gradients of coupled state variables of the system model. An overshoot or undershoot of the state gradients can be detected by means of predetermined barriers, which leads to the detection of the valve opening and valve closing times.

Alternatively, the pressure in dependence on the position of the pressure piece can be determined and the valve opening and valve closing times are derived from an evaluation. A corresponding pressure-path diagram is in FIG. 5 shown on the left. In FIG. 5 on the right is the associated path-time diagram. The path-time diagram shows the time-dependent movement of the pressure piece. It can be seen that the pressure piece first moves forward from an initial position 1 (x = 0 mm) and reduces the volume of the dosing chamber (pressure phase). The pressure piece goes through a maximum at time 3 and then moves back to the starting position (suction phase).

In the FIG. 5 on the left the associated pressure-distance diagram is shown. It is traversed in a clockwise direction, starting at the point of origin, where the pressure piece is in position 1. During the pressure phase, the pressure in the dosing chamber will initially rise sharply until the pressure is able to 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 marked with the number 2. From this point in time also in FIG. 5 is entered on the right, it comes to a dosage. With each further movement of the pressure piece dosing fluid is pumped into the pressure line. As soon as the pressure piece has reached the maximum position (time 3), the movement of the pressure piece turns over, the pressure valve closes immediately and the pressure in the dosing chamber drops again. Once a minimum pressure is reached (time 4) opens the suction valve, which connects the dosing with the suction line, and dosing fluid is sucked into the dosing until the starting position is reached.

The valve closing times can be determined from the path-time diagram, since they are on the maximum travel of the pressure piece. The times 2 and 4, ie the valve opening times, are not so easy to determine, especially since in practice the pressure-path diagram has rounded "corners". Therefore, for example, starting from position 1 in the pressure-path diagram when reaching 90% of the maximum pressure (known from position 3), the path can be read and the slope of the pressure-displacement diagram between points 1 and 2 are determined. The 90% curve is dotted. The resulting straight line intersects the curve p = p _max at the valve opening time. In the same way, the time 4 can be determined. This determination can be made in each cycle and the result used for a later cycle. As a result, changes in the opening times are detected.

By comparing the desired and actual trajectories of the individual state variables of the system models, gas 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 units can be diagnosed. Especially then, if between the target and actual trajectories a predetermined error limit is exceeded, this can trigger a warning signal and appropriate action.

An example is in FIG. 6 shown. Again, the pressure-distance diagram is shown on the left and the path-time diagram on the right. The right figure is identical to the corresponding diagram of FIG. 5 , If there are gas bubbles in the hydraulic system which are compressible, this will cause the pressure valve to open only at time 2 'and the suction valve to open only at time 4'. A clear shift of the valve opening times can thus be used to diagnose the condition "air in the metering chamber". In the case of cavitation, only the valve opening timing 4 'but not the valve opening timing 2 shifts so that such a behavior for diagnosing the "cavitation" state can be used.

The presented model-based methodology allows a much more comprehensive and higher-quality diagnosis by analyzing the individual coupled system state variables than has hitherto been realized.

In addition, this can be realized with low sensor costs and high reliability and security. Due to the higher diagnostic quality, the range of application of electromagnetic dosing pump systems can possibly be extended, since now the dosing accuracy can be extremely improved.

Claims (18)

1. Method for determining hydraulic parameters in a hydraulic system with a displacement pump, which is connected to a suction and pressure line, wherein the displacement pump has a movable displacement element, which bounds the metering chamber, which is connected to the suction and pressure line via valves, with the result that 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 motion of the displacement element, wherein a drive for the oscillating motion of the displacement element is provided, characterized in that a physical model having hydraulic parameters is constructed for the 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 is determined, and at least one hydraulic parameter is calculated by means of an optimization calculation, which describes the determined position of the displacement element as well as the exerted force or the pressure in the metering chamber on basis of the set physical model.
2. Method according to claim 1, characterized in that the density of the fluid in the metering chamber and/or the viscosity of the fluid in the metering chamber is determined as hydraulic parameter.
3. Method according to claim 1 or 2, characterized in that the displacement pump is an electromagnetically driven metering pump, preferably an electromagnetically driven diaphragm pump.
4. Method according to claim 3, characterized in that 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 from the measured current and the measured position of the displacement element.
5. Method according to claim 3 or 4, characterized in that the physical model is constructed for the case that the valve to the suction line is open and the valve to the pressure line is closed, and/or for the case that the valve to the suction line is closed and the valve to the pressure line is open, wherein if the physical model is constructed both for the case that the valve to the suction line is open and the valve to the pressure line is closed, and for the case that the valve to the suction line is closed and the valve to the pressure line is open, the valve opening time points are determined, and the physical model is selected in dependence on the result of the determination of the valve opening time points.
6. Method according to one of claims 1 to 5, characterized in that after the hydraulic parameter is determined, this and the physical model are used to determine the force exerted on the displacement element by the pumped fluid, and the force determined in this way is used in a control of the motion of the displacement element.
7. Method according to one of claims 1 to 6, characterized in that a model-based control is used for the drive to optimize the metering profile of the displacement pump.
8. Method according to claim 7, characterized in that a differential equation, and preferably an equation of motion, of the displacement element is used as model for the model-based control.
9. Method according to claim 7 or 8, characterized in that displacement pump-specific forces which act on the thrust member are modeled in the differential equation.
10. Method according to one of claims 7 to 9, characterized in that a nonlinear state-space model is chosen as state-space model, wherein the nonlinear control takes place either via control-Lyapunov functions, via flatness-based control methods with flatness-based feedforward control, via integrator backstepping methods, via sliding mode methods or via predictive control, wherein the nonlinear control is preferably via control-Lyapunov functions.
11. Method according to one of claims 7 to 10, characterized in that the difference between the detected actual position profile of the displacement element and a predetermined target position profile 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 profile is used as target value profile for the next suction-pressure cycle.
12. Method according to one of claims 7 to 11, characterized in that a physical variable in the displacement pump is determined using the differential equation or equation of motion.
13. Method according to claim 12, characterized in that the fluid pressure p of a pumped fluid located in a metering chamber of a displacement pump is determined as physical variable.
14. Method according to one of claims 12 to 13, characterized in that if the actual fluid pressure reaches or exceeds a predetermined maximum value, a warning signal is emitted and the warning signal is preferably sent to an automatic shutoff, which shuts off the metering pump in response to the warning signal being received.
15. Method according to one of claims 12 to 14, characterized in that a target fluid pressure curve, a target position curve of the displacement element and/or the target current progression through the electromagnetic drive is stored for a motion cycle of the displacement element, and the actual fluid pressure is compared with the target fluid pressure, the actual position of the displacement element with the target position of the displacement element and/or the actual current through the electromagnetic drive with a target current through the electromagnetic drive and, if the differences between actual and target value satisfy a predetermined criterion, a warning signal is emitted.
16. Method according to claim 15, characterized in that a weighted sum of the relative deviations from the target value is determined and the criterion is chosen such that a warning signal is emitted if the weighted sum exceeds a predetermined value.
17. Method according to claim 15 or 16, characterized in that several criteria are predetermined, an error event is assigned to each criterion and, if a criterion is satisfied, the assigned error event is diagnosed.
18. Method according to one of claims 7 to 17, characterized in that the mass m of the displacement element, the spring constant k of the spring which pretensions the displacement element, the damping d and/or the electrical resistance R_Cu of the electromagnetic drive is determined as physical variable.

Images