METHOD FOR DETERMINING A SWITCHING STATE OF A VALVE, AND SOLENOID VALVE ASSEMBLY

Abstract

A method for determining a switching state of a valve, an inductance variable being determined on the basis of current and voltage measurements and the switching state being determined on the basis of the inductance variable. Also disclosed is a solenoid valve assembly for carrying out such a method.

Claims

1. A method for determining a switching state of a valve which is actuated by a coil, the method comprising: respectively ascertaining a current flowing through the coil and a voltage applied to the coil at several times following one another with a specified time interval, calculating an inductance variable of the coil on the basis of the currents, the voltage and the time interval and, determining the switching state on the basis of the inductance variable.

2. The method as claimed in claim 1, wherein the inductance variable is calculated using the method of least squares.

3. The method as claimed in claim 2, wherein the method of least squares is used recursively with a forgetting factor.

4. The method as claimed in claim 1, wherein the inductance variable is ascertained by a linear equation, in which a first column vector is set equal to a matrix multiplied by a second column vector, wherein the times are numbered with index k, wherein the first column vector contains in row k a difference between the current at time k+1 minus the current at time k, wherein the matrix has two columns, wherein the first column of the matrix contains in row k a sum of the voltage at time k+1 and the voltage at time k, wherein the second column of the matrix contains in row k a sum of the current at time k+1 and the current at time k, and wherein the second column vector contains a first parameter in its first row and a second parameter in its second row.

5. The method as claimed in claim 4, wherein the inductance variable is calculated as the quotient of the time interval divided by the number two and divided by the first parameter.

6. The method as claimed in claim 1, wherein a resistance of the coil is also calculated on the basis of the currents, the voltages and the time interval.

7. The method as claimed in claim 4, wherein a resistance of the coil is calculated as the quotient of the second parameter divided by the first parameter.

8. The method as claimed in claim 1, which is executed continuously or continually repeatedly.

9. The method as claimed in claim 1, wherein the inductance variable is compared with a first end value and a second end value, wherein, when the inductance variable is at most a predetermined distance from the first end value, a first switching state is determined, and wherein, when the inductance variable is at most a predetermined distance from the second end value, a second switching state is determined.

10. The method as claimed in claim 9, wherein the switching states are end states of the valve.

11. The method as claimed in claim 1, wherein the switching state of the valve is detected from a test signal.

12. The method as claimed in claim 1, wherein calculations are carried out entirely or partly by fixed point arithmetic.

13. The method as claimed in claim 1, wherein the coil is controlled by pulse width modulation, wherein current and voltage are each averaged over a pulse width modulation period.

14. A solenoid valve assembly, comprising: a valved, a coil for actuating the valve, a control device for applying a current and/or a voltage to the coil, and a state determination device which is configured to carry out a method according to claim 1.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0037] A person skilled in the art will take further features and advantages from the exemplary embodiment described below with reference to the appended drawing, in which:

[0038] FIG. 1 shows a solenoid valve assembly,

[0039] FIG. 2 shows an equivalent circuit diagram of a coil,

[0040] FIG. 3 shows an inductance in dependence on a valve position,

[0041] FIG. 4 shows an inductance in dependence on a current,

[0042] FIG. 5 shows a profile of a resistance and

[0043] FIG. 6 shows a profile of an inductance.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0044] FIG. 1 shows a solenoid valve assembly 5 according to an exemplary embodiment of the invention. This is configured to carry out a method according to an exemplary embodiment of the invention.

[0045] It should be understood that the solenoid valve assembly 5 is only shown schematically here.

[0046] The solenoid valve assembly 5 has a valve 10. It also has an armature 20, which is connected to the valve 10 via an armature rod 25. It also has a coil 30, which surrounds the armature 20. Electrical current can be applied to the coil 30, whereby the armature 20 can be moved. This allows movement or actuation of the valve 10. In particular, the valve 10 can be switched between two end positions, specifically an open end position and a closed end position.

[0047] The solenoid valve assembly 5 also has a control device 40, which is designed to apply a current and/or a voltage to the coil 30. This is for driving as just mentioned. In addition, the solenoid valve assembly 5 has a state determination device 50, which is configured to carry out a method according to an aspect of the invention. The functionality will be discussed in more detail below.

[0048] As already mentioned, the valve 10 is to be subjected to a switching state determination. For this purpose, a mathematical model of the electrical subsystem of the coil 30 will now be discussed before further details are described.

[0049] It should be understood that the mathematical details and formulas given below may also represent essential aspects that can also be used to restrict the claims.

[0050] A mathematical model of the coil 30 can be reproduced for the coil voltage u and the coil current i as well as the inductance L and the resistance R of the coil 30 in good approximation using the following relationship:

[00001]$\begin{array}{cc}\frac{d}{dt}\left(L\left(i,x\right)i\left(t\right)\right)=u\left(t\right)-Ri\left(t\right).& \left(1\right)\end{array}$

Here, the time t is given as a parameter.

[0051] It should be noted that this equation also applies in the case of a pulse width modulation-operated current control. In this case, however, u means the average voltage of a pulse width modulation period and i means the average current of a pulse width modulation period.

[0052] FIG. 2 shows an electrical equivalent circuit diagram, with inductance L, resistance R, voltage u and current i being shown.

[0053] Dynamic saturation effects play a subordinate role in typical systems and voltage ranges. Therefore, the above equation (1) can be simplified in good approximation to the following equation:

[00002]$\begin{array}{cc}L\left(x,i\right)\frac{d}{dt}i=u-Ri-vi\frac{\partial }{\partial x}L\left(x,i\right).& \left(2\right)\end{array}$

[0054] Here, v denotes the tappet speed of a tappet, not shown separately, of the valve 10. For example, the tappet can be the already mentioned armature 20.

[0055] This mathematical model can be used to determine the valve state. This will be explained below.

[0056] FIG. 3 shows a profile of the inductance with constant current against the tappet position of a typical solenoid valve, for example the valve 10. Furthermore, FIG. 4 shows a typical profile of the inductance against the current of an unswitched (x=0; reference symbol S1) and a switched (x=1; reference symbol S2) solenoid valve, for example of the valve 10. The switching state is shown in percent on the horizontal axis of FIG. 3, while the inductance is shown in arbitrary units on the vertical axis. The current is plotted relative to the maximum current on the horizontal axis in FIG. 4, while the inductance is plotted relative to the maximum inductance on the vertical axis in FIG. 4. It is easy to see that knowledge of the inductance and the current can be used to infer the switching state.

[0057] Since neither the tappet speed v nor the exact location dependency of the inductance L(x,i) is known, the estimation is however only easily possible when the valve tappet is not moving, i.e. when v=0 and L(x,i)=L(i) applies. This is at least always satisfied when the valve tappet is at one of the end stops.

[0058] If L.sub.0=L(0,i0) denotes the inductance in the unswitched state and L.sub.1=L(100,i100) the inductance in the switched state, in the event that the tappet is at one of the end stops then either the differential equation

[00003]$\begin{array}{cc}\frac{d}{dt}i=\frac{1}{L_0}u-\frac{R}{L_0}i& \left(3\right)\end{array}$

or the differential equation

[00004]$\begin{array}{cc}\frac{d}{dt}i=\frac{1}{L_1}u-\frac{R}{L_1}i& \left(4\right)\end{array}$

[0059] applies.

[0060] If there is a change in the current over time, the inductance L can be estimated, and thus the switching state can also be inferred directly by means of purely making comparisons of the size of the estimated inductance L with L.sub.0 and L.sub.1 and preferably also the knowledge of the current.

[0061] A method for inductance estimation is shown below.

[0062] If the voltage profile and current profile between two measuring points with the sampling time T.sub.s are approximated using a straight line, then the following relationship is obtained for the current profile between the sampling points t.sub.k=kT.sub.s and T.sub.k+1=(k+1)T.sub.s:

[00005]$\begin{array}{cc}i\left(t\right)\approx \frac{i_{k+1}-i_k}{T_s}\left(t­k{T}_s\right)+i_k,t\in \left[t_k,t_{k+1}\right].& \left(5\right)\end{array}$

[0063] The voltage profile is obtained analogously. By then integrating the differential equation (3) for L.sub.0=L using [t.sub.k,t.sub.k+1], together with equation (5), this then gives:

[00006]$\begin{array}{cc}\left(i_{k+1}-i_k\right)=\frac{1}{L}{T}_s\frac{u_{k+1}+u_k}{2}-T_s\frac{R}{L}\frac{i_{k+1}+i_k}{2}& \left(6\right)\end{array}$

or after a brief rearrangement

[00007]$\begin{array}{cc}\underset{\underset{=y}{︸}}{\left(i_{k+1}-i_k\right)}=\underset{\underset{=s^T}{︸}}{\left[u_{k+1}+u_k,i_{k+1}+i_k\right]}\underset{\underset{=p}{︸}}{\left[\begin{array}{c}\frac{T_s}{2L}\\ \frac{T_{s}R}{2L}\end{array}\right]}.& \left(7\right)\end{array}$

[0064] Thus, the determination of the parameters R, L is reduced to solving the linear least squares problem (7), and R, L can be determined from the substitute parameters p=[p.sub.1,p.sub.2].sup.T in the form

[00008]$\begin{array}{cc}L=\frac{T_s}{2p_1}R=\frac{p_2}{p1}.& \left(8\right)\end{array}$

[0065] The idea of the valve state estimation is to solve the least squares problem recursively with a forgetting factor. This allows the inductance to be estimated in the event of changes in the current, from which the switching state can be determined by comparing the size. It should be noted that, because of the dependence of the inductance on the position, a change in current occurs not only as a result of a change in the coil voltage but also when there is a movement of the tappet, as is shown for example by equation (2) above. This also allows unwanted valve switching, for example due to externally acting forces or flow forces, to be detected.

[0066] It should be noted that a recursive method of least squares can also be implemented in fixed point arithmetic by suitable implementation. Furthermore, such a method has a very low computational effort for two estimation parameters, which also allows direct implementation in hardware by simple logic gates.

[0067] The calculation just described is carried out by the state determination device 50. For this purpose, the corresponding voltages and currents are measured at respective times.

[0068] To validate the method, the duty cycles for a PWM-operated solenoid valve were set from 20% to 80% in 10% steps and at the same time the coil current averaged over the pulse width modulation period and the averaged coil voltage were measured. The algorithm described above was then applied to these voltage and current profiles. The resulting resistance and inductance profiles are shown in FIG. 5 (horizontal axis: time in arbitrary units; vertical axis: resistance in arbitrary units) and FIG. 6 (horizontal axis: time in arbitrary units; vertical axis: inductance in arbitrary units). It is easy to see that the algorithm for R and L converges after a short settling phase. It can also be seen from the profile of the inductance whether the valve closes or not. With a duty cycle of 20% and 30%, the current is insufficient to close the valve and the inductance values converge toward L0. With higher duty cycles, the current is sufficient to close the valve, which increases the inductance, which is detected very well by the estimate.

[0069] Up to now, the valve switching state dependent on the control was considered, i.e. it was assumed that a valve is switched when the controller requests it. It could not be ascertained whether a valve changes its switching state due to flow forces or other externally acting forces or whether for example a normally closed valve opens at all. There are already approaches to determining the switching state, but the resulting measurement signals are very small and prone to disturbance.

[0070] The estimation described above now makes it possible for the first time to make a precise and robust statement about the switching state of a digital valve. This means that functional safety requirements can be met better or for the first time. In addition, the fact that the state of the valve is known means that hydraulic safety measures such as check valves are not required. In contrast to already established methods, the resulting signals are rather large and, thanks to the filtering of the recursive least squares algorithm, not very sensitive to noise. In addition to the inductance, the algorithm also determines the current electrical resistance, almost as it were as a waste product. This can be used directly for availability issues.

[0071] The valve state estimation on the basis of the change in inductance is particularly advantageous. The use of the coil current and the coil voltage in connection with the mathematical model (3) or (4) as well as a suitable estimation method is decisive. For example, a recursive method of least squares can be used here, but other methods can also be used. Furthermore, the use of the coil current averaged over a pulse width modulation period and the averaged coil voltage is advantageous for systems operated by pulse width modulation. If the supply voltage is known, the duty cycle of a pulse width modulation period can also be used as an alternative to the averaged coil voltage.

[0072] The mentioned steps of the method according to an aspect of the invention may be carried out in the order given. However, they may also be carried out in a different order. In one of its embodiments, for example with a specific combination of steps, the method according to an aspect of the invention may be carried out in such a way that no further steps are carried out. However, in principle, further steps can also be carried out, even steps that have not been mentioned.

[0073] The claims which form part of the application do not represent a relinquishment of the attainment of further protection.