MAGNETORHEOLOGICAL VALVE TECHNICAL FIELD The present invention relates to the field of valves and in particular to so-called magnetorheological valves. More specifically, the present invention concerns a valve comprising a housing having a peripheral structure made of a magnetic field conductive material and a core disposed in said housing, the core and the peripheral structure delimiting a first chamber disposed on one side of the core, a second chamber disposed on another side of the core, and at least one channel formed between the peripheral structure and said core, this channel connecting said first chamber to said second chamber, the housing containing a magnetorheological fluid (MRF). The present invention further concerns a control system comprising a valve as defined above. The invention also concerns a method for controlling a valve as defined above. PRIOR ART Power transmission is omnipresent in engineering. It involves the conversion of energy into different forms and its distribution to different systems. This can be achieved using electrical, mechanical or fluid technologies. Among these, fluid power transmission systems offer the advantages of a high power-to-weight ratio, the ability to handle high magnitude forces and torques compared to mechanical transmissions, and increased heat transfer capability. There are two main types of fluid systems; hydraulic which use incompressible liquids such as oil, water or other types of fluids; and pneumatic which use neutral gases such as air. Because of these aforementioned advantages, power transmission based on fluid systems is used in a wide range of applications. In addition to the field of building and driving mobile machinery, hydraulic systems are key components in robotic manipulators and play an important role in the automation of industrial activities. In
automotive systems, clutches or vehicle suspensions, which are vital for comfort and safety, are frequently actualized by fluidic actuators. Primarily used to control large movements, pneumatic and hydraulic soft actuators are becoming increasingly miniaturised. A great showcase is minimally invasive surgery and rehabilitation systems, made possible by fluidic technologies using miniaturized valves. Valves are an integral component for controlling the direction and flow rate in fluidic systems. Over the last few decades, numerous attempts have been made to further improve the capabilities of valves by using different principles. Solenoid valves have been used as a cost effective and accurate solution, as well as active mechanical valves based on thermal or elastomeric principle. More recently, valves based on Dielectric Elastomer Actuators (DEAs) have been developed to provide lightweight and compact fluidic regulation systems. Furthermore, electrorheological fluid (ERF) and magnetorheological fluid (MRF), two types of smart materials, have been used as pressurized liquids in fluidic valves for applications such as automotive or medical. ERFs and MRFs are characterized by a controllable, reversible change in viscosity when an external electric or magnetic field respectively is applied. As a result, they allow the creation of valves with simple structure and fast response when continuously supplied with DC current and voltage. MRF is a class of smart materials, comprising micron-sized iron particles dispersed in oil and enhanced by surfactants. A conventional magnetorheological (MR) valve is cylindrical and consists of several parts. First, a coil is wound around a ferromagnetic rod. This part is covered at the top and bottom by two ferromagnetic discs . Additionally, a cylindrical ferromagnetic structure surrounds the first part. A gap is formed between the two structures, allowing the vertical flow of the MRF. The internal structure acts as an electromagnet. When a DC current is applied to the coil, it imposes a magnetic flux density B in the active regions of the gap (See Fig.1). The flux density is perpendicular to the fluid flow in the active regions. Accordingly, the particles form chain-like structures, aligned with the magnetic field lines. As a consequence, the MRF viscosity η increases in the active regions and the fluid flow rate Q decreases. This results in an increase in the pressure difference ∆P developed between the two sides of the valve.
The valve is highly efficient due to its ability to withstand high loads. However, in order to maintain the valve closed, the coil must be permanently supplied with a DC current. This results in high energy consumption, which limits the applications in which such valves can be used. For example, the use of these values may not be possible in some applications where the space available for an energy source is limited. A conventional MRF valve can be in open or closed state. When closed, the valve sustains a given pressure, which means that if a pressure below a threshold value is applied to the MRF on one side of the valve, the MRF will not flow to the other side of the valve. If the pressure applied to the MRF is above the threshold value however, the MRF will flow to the other side of the valve. Depending on the application, the threshold value should be as high as possible. In other applications however, it is important that the valve opens when a given pressure is reached. In this type of valves, the threshold pressure is fixed and given by the design of the valve. It would be interesting to create a valve where this threshold pressure can be adjusted according to the specific needs of the application. The development of MRF valves has influenced the mechatronic applications due to the ease of interface between electronic controls and mechanical components. One of the most challenging aspects of MRF valves is the difficulty in miniaturising the device while at the same time improving the achievable performance. Therefore, there is a need for a valve that can withstand high loads in a small volume and has a power consumption as low as possible. DESCRIPTION OF THE INVENTION The present invention proposes to overcome the drawbacks of the valves of the prior art by proposing a new valve capable of withstanding high loads in a small volume and in which the power consumption is very low. Moreover, the threshold pressure, i.e., the pressure above which the valve in closed state opens, can be set or selected by a user. These goals are achieved by a valve as described in the preamble and characterized in that the core comprises an electropermanent magnet (EPM).
Such an electropermanent magnet can be defined as a magnet whose polarity can be selected by an electric pulse and which maintains its polarity as long as an electric pulse of opposite direction is not received by this magnet. The goals of the invention are also achieved by a control system as defined in the preamble and characterized in that said control system comprises a control module arranged to transmit to said coil, an electric pulse having a direction chosen so as to open or close said valve. The goals of the invention are further reached by a method for controlling a valve as described in the preamble and characterized in that said method comprises the steps consisting of: - choosing whether the valve should be open or closed; - apply an electric pulse to the coil surrounding the core of the EPM so that the core allows the passage of the MRF through the channel, from one of said first or second chamber to the other chamber in the event that the valve must be opened; and - applying an electrical pulse to the coil surrounding the core of the EPM such that the core prevents the passage of the MRF through the channel, from one of said first or second chambers to the other chamber if the valve needs to be closed. The present description proposes a miniaturized MRF valve with high efficiency in terms of sustained pressure and energy consumption. The proposed valve is designed to control hydraulic systems using MRF as filling fluid. In this valve, the Magnetorheological fluid (MRF) is used both as an actuation fluid and as control medium. The function of the valve is binary, operating either in the on-state, completely blocking the flow rate, or in the off-state, allowing the fluid to flow. The pressure sustaining capability of the valve is however tuneable. The goals of the invention are achieved by a valve using a magnetorheological fluid (MRF) and an electropermanent magnet (EPM), instead of an electromagnet to impose the magnetic flux density in the active regions of the valve. This is achieved by replacing the inner iron rod used in a conventional MRF valve, with a magnetic rod with tunable magnetization.
In the proposed EPM valve, the coil is used in a pulsed manner to magnetize or demagnetize the inner magnetic rod, closing or opening the valve respectively. A careful choice of magnetic material is necessary to construct an EPM. In this application, it is made of a ferromagnetic material with a high remanent magnetization and a low magnetizing field (i.e., soft ferromagnetic material). Whether a ferromagnetic material is a hard or a soft magnet depends on the strength of the magnetic field needed to align the magnetic domains. This property is characterized by the coercivity H
c. Hard magnets have a high coercivity (H
c), and thus retain their magnetization in the absence of an applied field, whereas soft magnets have low values. Thus, soft magnetic materials are easily magnetised and demagnetised. They typically have intrinsic coercivity less than 1000 Am
-1. The main parameter, often used as a figure of merit for soft magnetic materials, is the relative permeability (μr, where μr = B/μoH), which is a measure of how readily the material responds to the applied magnetic field. The other main parameters of interest in addition to the coercivity are the saturation magnetisation and the electrical conductivity. Technical properties of interest for soft magnets include: (i) high permeability (μ=B/H); (ii) low hysteresis loss, and (iii) low eddy current and anomalous losses. The valve comprises an EPM which is added in the traditional structure of a MRF valve with the goal of decreasing its power consumption. This enables achieving a low volume design, but with high sustained pressure capability. The above features make the proposed device an attractive solution for various applications such as control of soft bending actuators or medical devices with MR modules for plantar pressure offloading in diabetic patients. SHORT DESCRITPION OF THE DRAWINGS The present invention and its advantages will be better understood with reference to the enclosed drawings and to the detailed description of several embodiments, in which: ^ Figure 1 is a sectional view of an annular MR valve; this figure illustrates a valve of the present invention as well as a valve of the prior art;
^ Figure 2a is a cross section view of a valve of the prior art, in the embodiment of Figure 1; ^ Figure 2b is a cross section view of a valve of the invention, in the embodiment of Figure 1; ^ Figure 3a illustrates the operating principle of a MRF valve having an electromagnetic core, based on an ideal M-H characteristic and shows a first way of demagnetizing the core; ^ Figure 3b is similar to Figure 3a and shows a second way of demagnetizing the core; ^ Figure 4 represents a 2-D axisymmetric simulation of the magnetic flux density in the MRF valve with EPM rod; ^ Figure 5a illustrates hydraulic scheme of an experimental setup of the valve of the invention; ^ Figure 5b illustrates an experimental setup for the validation of the device and prototype of valve with an EPM in the inner rod of the structure; ^ Figure 5c illustrates a detail of the valve of the setup of Fig.5b; ^ Figure 6 is a graph showing the M-H and B-H characteristics for magnetization, B- H characteristic for demagnetization of an AlNiCo-5 rod, and the Brod and Hrod for each current value obtained analytically and with simulation; ^ Figure 7 is a graph illustrating the pulsed voltage over the coil inside the EPM valve; ^ Figure 8 represents the piston displacement with respect to the pressure difference between the chambers, for an open valve and a closed valve with different magnetization currents; and ^ Figure 9 shows the pressure difference and piston displacement over time in an EPM valve. WAY OF CARRYING OUT THE INVENTION The invention will be described below in more details, with reference to the drawings. Figure 1 illustrates a magnetorheological fluid (MRF) valve of the prior art as well as a MRF valve according to the invention. The components of both the valve of the prior art and the valve of the invention are shown in the same way. Figure 2a specifically
illustrates a valve 10 of the prior art while Fig.2b specifically illustrates a valve 10 of the invention. The valve 10 comprises a housing 11 having a peripheral structure 12 made of a magnetic field conductive material. The valve 10 further comprises a core 13 disposed in said housing 11. The core 13 and the peripheral structure 12 delimit a first chamber 14 disposed on one side of the core 13, a second chamber 15 disposed on another side of the core 13 and at least one channel 16 formed between the peripheral structure 12 and the core 13. This channel 16 connects said first chamber 14 to said second chamber 15. The housing 11 further contains a magnetorheological fluid (MRF). The core 13 comprises a coil 17 that is wound around a ferromagnetic rod 18. This part is covered by a first ferromagnetic disc 19 on the top and by a second ferromagnetic disc 20 on the bottom. The ferromagnetic peripheral structure 12 forms the sidewalls of the valve, surrounding the core 13. A gap formed between the core 13 and the peripheral structure 12 forms the channel 16 which allows the MRF to flow between the chambers 14, 15. In a prior art MRF valve, as illustrated by Fig. 2a, the inner structure acts as an electromagnet. When the coil 17 is supplied by a DC current, it imposes a magnetic flux density B in the active regions of the channel 16. The flux density is perpendicular to the fluid flow in the active regions and the valve is closed. The MRF in the active regions is shown in Fig.1 with and without a magnetic field. In the valve of the invention as illustrated by Figure 2b, the core 13 does not comprise a coil and a ferromagnetic rod acting as an electromagnet, but comprises a électropermanent magnet (EPM) 21. This EPM comprises an inner magnetic rod 18 surrounded by the coil 17. This coil is used in a pulsed manner to magnetize or demagnetize the inner magnetic rod 18, thus respectively closing or opening the valve. As mentioned above, a careful selection of the magnetic material is required to construct an EPM. In this application, it is made of a ferromagnetic material with a high remanent magnetization and a low magnetizing field (i.e., soft ferromagnetic material). The fulfilment of the above specifications led to the choice the material AlNiCo-5 for the device of the invention. The latter offers the possibility of storing magnetic energy, without a continuous power supply. The ability of the valve to sustain pressure up to 1010 kPa for a volume of 353 mm
3 is experimentally demonstrated. The fluid flow rate
when the valve is open is 459 mm
3/s for a pressure difference of 993 kPa. The corresponding power consumption is negligible in steady-state condition, while 15.3 mJ is consumed when the valve is activated and 6 mJ when it is deactivated. The experimental results also confirm the possibility of tuning the pressure sustaining capability of the valve. Figures 3a and 3b illustrate the operation principle of a valve 10 according to the invention, based on an ideal M-H characteristic of a device having a core 13 made of AlNiCo-5. The magnetic flux of the valve is shown in the magnetized and demagnetized states. Figure 3a illustrates a first way of demagnetizing the inner magnetic rod 18. According to this first way, a decaying oscillating external field is applied until the remanence coercivity of the rod is reached. Figure 3b illustrates a second way of demagnetizing the rod 18. According to this second way, a negative pulsed external field is applied until the remanence coercivity is reached. The white arrow represents the magnetization direction of the rod. The magnetizing field H, external to the material, is the one generated by the coil 17 and is proportional to the ampere-turn product: N · I, with N the number of turns of the coil, and I the electric current flowing through it. The resulting magnetization of the material M is defined as follows: M^= B^− µ0H^= µ0[µr(H) − 1] · H, (1) with µ0 the magnetic constant, and µr the relative magnetic permeability of the material. Figures 3a and 3b also show two possible modes of operation of the valve 10 which are illustrated by the magnetization curve (M-H characteristic) of the AlNiCo-5. The non-linear relationship between permeability and magnetizing field µ
r(H), results in different magnetization for the same magnetizing field. Thus, such material exhibits hysteresis (i.e., the magnetization depends on its history). This, usually undesirable behavior is exploited here to operate the electropermanent magnet. The inner magnetic rod 18 in AlNiCo-5 is initially demagnetized (H = 0, M = 0). There is no magnetic field blocking the MRF and thus, the valve 10 is open. When a positive
electric current is passed through the coil 17, the external magnetic field, in which the AlNiCo-5 is placed (i.e., the magnetizing field), increases until it reaches saturation (H = Hsat, M = Msat). When the current supply to the coil 17 is stopped, there is no more magnetizing field and the material reaches its remanence (H = 0, M = M
rem). The remanent magnetization in the material causes it to act as a permanent magnet. In this state, the valve 10 is closed. As mentioned above, there are two ways of demagnetizing a magnetic material. The first one is illustrated by Fig.3a and consists in applying a decaying oscillating external magnetic field, which enables reaching the origin of the M-H curve despite the hysteresis of the material and for any initial state of the material (its history). The second method is illustrated by Figure 3b and consists in applying a negative external magnetic field so as
to reach the remanence coercivity of the material, leading to a demagnetized state when the external field is removed. This should not be confused with the intrinsic coercivity which corresponds to zero magnetization of the material under a
specific external magnetic field ). Indeed, the hysteretic behavior of the material would result in a non-null residual magnetization when the external magnetic field is removed. The remanence coercivity is a theoretical state that is difficult to reach precisely in practice and as a result, a perfect demagnetization is not achieved. Nevertheless, the remaining magnetization can be considered small enough in many applications such as this one. The second method is preferred in this work because it requires a single demagnetization pulse implemented with a simple electronic design. If partial magnetization of the valve is desired (tunability of the sustained pressure as explained later), the first method should be considered as it is more robust and independent of the magnetization state. Many documents have presented electropermanent magnets comprising two different ferromagnetic materials (i.e., magnets) placed in parallel with each other and between two pieces of iron. In these cases, there is a hard material, such as NdFeB, and a soft material such as AlNiCo-5. The current pulse through the coil only affects the soft material. The latter is always fully magnetized, either in one direction (the same as the hard material), or in the other (opposed to the hard rod). The result is a device that respectively appears magnetized or demagnetized from the outside. The aforementioned combination of the two types of magnets is necessary to prevent self-
demagnetization of the soft material (AlNiCo-5). In this invention, the small air gap and the negligible change of magnetic permeability coming from the MRF flow prevent the self-demagnetization of the soft material. In this sense, there is no need for a hard material, but the use of a single soft material (AlNiCo-5 in our case) is sufficient. To the authors’ knowledge, there are only a few works using a single AlNiCo-5 magnet inside an electropermanent magnet structure. There, despite taking advantage of the full magnetization curve, the large air-gap and varying magnetic environment will lead to its self-demagnetization, jeopardizing a reliable operation. The strategy proposed in this paper is robust and highlights a great application of single-rod electropermanent magnets. An analytical model of the device is described below. The goal is to determine the magnitude of the current pulse supplying the coil 17 and to estimate the maximum pressure the valve 10 can sustain. The chosen strategy to demagnetize the rod 18 is to use a single negative current pulse reaching the remanence coercivity of the material (as shown in Fig. 3b). However, magnet manufacturers do not provide the characteristics of this state because it is based on highly non-linear phenomena and depends on the magnetization state as well as on the rod environment. Thus, the negative current I
coeri required to achieve the intrinsic coercivity is therefore derived. It will be used as an upper limit to experimentally determine the correct demagnetizing current I to approach the remanence coercivity (I i demag demag < Icoer ). Due to the numerous non-linearity phenomena occurring in the valve, some assumptions are made. First, the core 13 made of iron is assumed to be a perfect magnetic conductor. Then, the MRF flow is assumed to be laminar with a homogeneous particle distribution in the carrier medium. These assumptions lead to the simplified version of Ampere’s law: Hrod · lrod + 2Hgap · g^= N^· I, (2) with Hrod, the magnetic field in the inner rod 18 and Hgap the magnetic field in the gap (i.e., MRF channel 16) respectively, l
rod the height of the inner rod, and g the width of the gap or channel 16 respectively. The dimensions are shown in Fig. 1 and their values in Table I. The conservation of magnetic flux links the two magnetic flux densities Brod and Bgap in the inner rod and the channel respectively, taking into account the flux fringes:
(3)
with r
rod, the radius of the inner rod 18 and r
gap the mean radius of the MRF channel 16. tgap + 2 · tfr is the length of the active region. It includes the length of the gap or channel 16 tgap and the flux fringes tfr. The relationship between the flux density B and the magnetic field H in the materials (AlNiCo-5 and MRF) are non-linear due to saturation effects. A polynomial interpolation of the positive B-H characteristic of the MRF, presented in the MRF132DG datasheet (L. Corporation, “MRF-132DG magneto- rheological fluid,” https://lordfulfillment.com/pdf/44/DS7015_MRF- 132DGMRFluid.pdf), is performed to extract the following relationship: ^ ^
The relationships between the magnetic field and magnetic flux density in the channel 16 can be deduced:
. An exponential curve fitting is performed separately for the magnetization and demagnetization parts of the B-H characteristic of the AlNiCo-5 rod 18 obtained experimentally by the German company MagnetPhysik (shown in Fig.3a). They are represented with the following equations:
TABLE I: Characteristics of MRF valve and AlNiCo-5 material properties. Parameter Value Parameter Value Parameter Value g (mm) 0.18 lrod (mm) 1.9 Mrem (T) 1.15 rgap (mm) 4.09 hvalve (mm) 4.55 Brem (T) 1.15 rvalve (mm) 5 V (cm
3) 3.53 (kA/m) 52.2 rrod (mm) 2 Lc (µH) 61 Hsat (kA/m) 200 t
gap (mm) 1.325 R
c (Ω) 0.46 M
sat (T) 1.25 tfr (mm) 0.32 N 55 Bsat (T) 1.50 TABLE II: Curve fitting coefficients. C1 7.40 · 10
-6 α1 1.498 α2 1.30 C
2 -8.51 · 10
-10 β1 1.475 · 10
-7 β2 7.71 · 10
-7 C3 -2.35 · 10
-10 γ1 -44.56 γ2 -5.71 · 10
-2 C4 1.86 · 10
-13 δ1 -9.01 · 10
-4 δ2 -6.20 · 10
-5
The coefficients used to interpolate the material properties of the MRF and AlNiCo are given in Table II. By combining Eq. (2), (3), and (4), the current through the coil 17 can be expressed as a function of the magnetic field and magnetic flux density in the magnetic rod I = I :
The current required for a complete magnetization Imag should reach the saturation point of the M-H curve resulting in I
mag = I(H
sat,B
sat) with B
sat = M
sat + µ
0H
sat. As mentioned above, the current that reaches the remanence coercivity of the material cannot be derived analytically. Thus, the current reaching the intrinsic coercivity
is used to provide an upper bound to the demagnetization current that will be determined experimentally. Thus, Idemag can be derived from (7) as follows:
An analytical estimation of the pressure sustained by the valve 10 when closed (i.e., in the no-flow condition) is proposed here. It corresponds to the sum of the field- dependent ∆P
τ and viscosity-dependent ∆P
η pressure drop. It is expressed as follows, if flux fringes are not taken into account:
with Q the flow-rate, η the fluid viscosity, and c a fluid parameter dependent on the ratio D^= lies between 2 (when D ≈ 1) and 3 (D > 100). In the case of a closed valve, ∆P
τ>>^∆P
η. Thus, the maximum pressure difference sustained by the valve 10 is approximated by ∆P
max ≈ ∆P
τ, and c lies close to 3. The yield stress function τ(H
MRF) is given in the datasheet of the MRF. It shows linear characteristics: BMRF(HMRF) = aHMRF, and τ(HMRF) = bH
MRF with a^= 0.0047 and b^= 0.2588, for B
MRF <^0.5T. However, in our case, the quantity of MRF surrounding the valve is large, leading to flux fringes that need to be accounted for in our model. The distribution of the flux density in the channel 16 is necessary. For
this purpose, a 2-D axisymmetric magneto-static simulation of the valve is performed using the FEMM 4.2 software. This is shown in Fig.4, which corresponds to the fully magnetized state of the rod in AlNiCo-5 (using the corresponding B-H characteristic) without current in the coil 17. The distribution of the magnetic flux density along the dotted line (see Fig.4) in the middle of the channel in the active regions can thus be evaluated. This allows Eq. (8) to reformulate as follows, with l, the position in the channel along the aforementioned line: ^ ⋅ ^ ^^^
^^^^^^
^^
Although flux fringes are undesirable in magnetic devices, in our case, they lead to an increase in the sustained pressure, improving the performance of our device. The characteristics of the proposed valve and the key properties of the M-H and B-H characteristics of the AlNiCo-5 are presented in Table I. The total volume V of the valve is 353.25 mm
3, which classifies it among the miniaturized versions of valves. The setup presented in Figures 5a, 5b and 5c is used to perform the experimental validation of the device of the invention. This setup comprises a cylindrical structure 21 with a first chamber 14’ and a second chamber 15’ that are separated by the MR valve 10. The first chamber 14 of the valve 10 is in fluidic communication with the first chamber 14’ of the cylindrical structure 21 while the second chamber 15 of the valve 10 is in fluidic communication with the second chamber 15’ of the cylindrical structure 21. Both first and second chambers 14, 15 of the valve 10 are filled with the MRF (MRF132DG by Lord Corp.). The second chamber 15’ is open and, thus, in atmospheric pressure. A pneumatic piston 22 is inserted into the first chamber 14’ of the cylindrical structure. The movement of the piston 22 can reduce the volume of the first chamber 14’ of the cylindrical structure. This increases the pressure in the latter when the valve is closed or creates fluid flow when the valve is open. A sectional view of the setup shows the internal part of the chambers, as well as the arrangement placement of the valve. Two pressure sensors 23 (PFT510 Miniature Pressure Sensor by Futek) measure the pressure in each chamber 14’, 15’ of the cylindrical structure 21. A laser displacement
sensor 21 (LK-G3000 by Keyence) measures the displacement of the pneumatic piston 22 over time. Wires of the coil 17 exit the second chamber 15 and are connected to the electronics that perform valve switching. A current probe (AP015 by LeCroy) measures the current flowing in the coil 17. Activation and deactivation of the valve 10 are achieved by applying current pulses to the coil 17. A positive pulse magnetizes the AlNiCo-5 rod 18 and closes the valve 10, while a smaller negative pulse demagnetizes the rod 18 and opens the valve 10. A full- bridge arrangement of MOSFETs allows the coil to be supplied with current in both directions. The IPB054N08N3-G MOSFETs from Infineon Technologies are used for faster switching (Nchannel type), with maximum voltage and current ratings of 80V and 80A respectively. Two gate drivers ICs FAN7842 from Fairchild Semiconductor Corp. (one per pair of MOSFETs) are used to drive the gates of the transistors, increasing the switching speed. The gate drivers are controlled with a F401-RE Nucleo board from ST. The M-H and B-H characteristics acquired experimentally by the German company Magnet-Physik are plotted together in Fig.6. The light gray dotted curve 30 represents the magnetization M over H and the dark dotted curve 31 represents the magnetic flux density B over H. These curves represent the AlNiCo-5 rod from a fully demagnetized condition to the saturation point, after imposing an external field. The third curve 31 with continuous line represents the demagnetization of the AlNiCo-5 rod, from the fully magnetized condition to B = 0. From Eq. (7), the current estimated to reach magnetization of the EPM is I
mag = 7.4 A. Fig. 9 shows that by decreasing the magnetization current at 5.1 A and, thus, decreasing also the energy consumed, it is possible to reach magnetization close to the saturation M
sat. The marked points depicted in Fig. 6 have been obtained analytically and numerically according to the simulation of Fig. 4 for the different fixed current values (0 A, 5.1 A, 7.4 A). The analytical and simulation results match with an estimation error of less than 6%. Fig. 7 shows the voltage across the coil 17 and the current measured during magnetization and demagnetization of the rod. Positive values result in the magnetization of the inner magnetic rod 18, while negative values result in its demagnetization. Both magnetization and demagnetization are achieved with a single
current pulse for each one. More specifically, the current for magnetizing the EPM and closing the valve 10 is 5.1 A and is reached within t
mag = 0.5 ms. This current value was used experimentally instead of 7.4 A in order to reduce the power consumption as explained above. The negative demagnetizing current achieving the remanence coercivity (from a fully-magnetized EPM) is determined experimentally: I
demag = -2.5 A. It can be confirmed that it is lower than the upper limit determined analytically (Idemag < Icoeri = -1.8 A). It is important to point out that the demagnetization of the rod 18 is determined experimentally when the opening of the valve 10 is detected (i.e., no pressure difference can be maintained). However, perfect demagnetization cannot be assured with this method. A small magnetization of the rod 18 could still remain while not being strong enough to block the flow in the valve. Nevertheless, in this application, perfect demagnetization of the rod is not critical. The first demagnetization method proposed in Fig. 3a could be used to overcome this limitation. The demagnetizing current is reached within t
demag = 0.72 ms, after the current starts to flow in the coil. Thus, the switching frequency is 0.85 kHz for 50% duty cycle and 1.25 kHz for 90% duty cycle. These values are high considering different types of fast switching valves, where the switching frequencies are lower than 1 kHz. According to these values, the energy consumption of the valve is 15.3 mJ for activation and 6 mJ for deactivation. Two experimental scenarios are presented here. The first one examines the comparison between the open and closed state of the valve 10. In the closed state, the inner magnetic rod 18 is magnetized with the piston pressure, and consequently, the pressure in the first chamber 14, is set and controlled manually with a manometer. The displacement of the piston and the pressure difference between the two sides of the valve (i.e., between the first and second chambers 14 and 15) are measured while the piston is moving towards the valve. When the EPM is fully magnetized at 5.1 A, we manually increase the piston pressure using the manometer up to the limit of our system, which is 1000 kPa. As shown in Fig.8 (curve with circles), there exists an initial abrupt increase of displacement of maximum ∆x = 10 mm, which remains constant as the pressure difference increases. The abrupt increase of displacement at the beginning of the process is explained by the air trapped in the chamber, which is initially
compressed. After that, the absence of displacement proves that the valve 10 remains closed. The maximum pressure measured is 1010 kPa. Higher pressures could not be tested because the compressed air supply used for these experiments is limited to about 1000 kPa. The value derived from the analytical and simulation results is 949 kPa. The experimental and numerical results are in agreement with an estimation error of less than 6.4%. The proposed system allows the possibility of tuning the sustained pressure. In fact, the AlNiCo rod could be partially magnetized and reach lower values than M
rem when the current is stopped. Thus, the magnetic field in the MRF would also be lower, reducing the pressure that can be sustained. Two further examples are shown in Fig.8, with magnetizing currents of 1.8 A (curve with triangles) and 3.5 A (curve with squares). In the case of I = 1.8 A, the manometer value is set to 100 kPa. At a pressure of 75 kPa, the displacement starts to increase until the stroke limit of 50 mm is reached. In the case of I = 3.5 A, once the air trapped in the chamber is compressed, the displacement also increases to the limit of the system at a pressure of 180 kPa. These results demonstrate the possibility of tuning the sustained pressure. In the open condition (dashed curve), the pressure on the pneumatic piston is again manually set at 1000 kPa. An increase in piston displacement of ∆x = 15 mm occurs until the pressure measured in chamber 15 reaches 993 kPa. After that, the MRF begins to flow through the valve at a constant flow rate. As a result, the piston displacement continues to increase while the pressure difference remains constant. The piston has a stroke limit of 50 mm. This explains the range of Piston Displacement axis in Fig.8. The flow rate in this case is calculated as Q = v ·Arod = 459 mm
3/s, where v is the piston’s velocity and Arod is the piston’s surface. v is calculated from the experimental measurement of the piston’s displacement over time. Eq.8 is validated using the above experimental values of Q and ∆P, as well as with using η = 0.112Pa · s and ∆Pτ = 0 because the valve is open and there is no magnetic field. Fig.9 shows the second experimental scenario. It shows the displacement ∆x of the piston and the pressure difference ∆P over time of a pressure of 1000 kPa. The transition of the valve from the closed to the open state is shown. In an initial state, the valve is fully-magnetized after applying a magnetization pulse. At t = 2 s the piston starts to push and continues throughout the whole experiment. The valve remains
closed until t = 3 s when the demagnetization pulse is applied and the valve is open. During this time, there is no piston displacement, except for the initial abrupt displacement associated with the compression of the air. This means that the valve remains closed. When the valve is open, the displacement ∆x begins to increase linearly, confirming that the valve is open. For a pressure difference of 1000 kPa the stroke limit of 50 mm is reached within one second. The valve is initially closed. At t = 3s the demagnetization takes place and the valve opens, allowing flow and piston displacement. In conclusion, valves constitute an essential component of the control of fluidic power transmission systems. In this paper, we have presented the design, implementation and experimental validation of a miniaturized MRF valve that uses a single rod EPM. The EPM is magnetized and demagnetized with a single current pulse. The valve can sustain a maximum of 1010 kPa, while being activated and deactivated within 0.5 ms and 0.72 ms respectively. At the same time, it consumes negligible power in the steady state condition, with 15.3 mJ for activation and 6 mJ for deactivation. The current work is an improvement compared to the state-of-the-art in terms of power consumption and maximum sustained pressure. However, smaller valves exist, with much lower capabilities with regard to pressure. In addition, our valve has the property that the sustained pressure can be tuned, which can provide flexibility to the applications that use it. This valve is of particular interest in several fields such as for example robotics applications or medical application such as medical applications for plantar pressure relief in diabetic patients. The MRF valve can be used to control the stiffness of medical footwear.
to reach the remanence coercivity of the material, leading to a demagnetized state when the external field is removed. This should not be confused with the intrinsic coercivity which corresponds to zero magnetization of the material under a
with rrod, the radius of the inner rod 18 and rgap the mean radius of the MRF channel 16. tgap + 2 · tfr is the length of the active region. It includes the length of the gap or channel 16 tgap and the flux fringes tfr. The relationship between the flux density B and the magnetic field H in the materials (AlNiCo-5 and MRF) are non-linear due to saturation effects. A polynomial interpolation of the positive B-H characteristic of the MRF, presented in the MRF132DG datasheet (L. Corporation, “MRF-132DG magneto- rheological fluid,” https://lordfulfillment.com/pdf/44/DS7015_MRF- 132DGMRFluid.pdf), is performed to extract the following relationship: ^ ^
The relationships between the magnetic field and magnetic flux density in the channel 16 can be deduced:
. An exponential curve fitting is performed separately for the magnetization and demagnetization parts of the B-H characteristic of the AlNiCo-5 rod 18 obtained experimentally by the German company MagnetPhysik (shown in Fig.3a). They are represented with the following equations:
TABLE I: Characteristics of MRF valve and AlNiCo-5 material properties. Parameter Value Parameter Value Parameter Value g (mm) 0.18 lrod (mm) 1.9 Mrem (T) 1.15 rgap (mm) 4.09 hvalve (mm) 4.55 Brem (T) 1.15 rvalve (mm) 5 V (cm3) 3.53 (kA/m) 52.2 rrod (mm) 2 Lc (µH) 61 Hsat (kA/m) 200 tgap (mm) 1.325 Rc (Ω) 0.46 Msat (T) 1.25 tfr (mm) 0.32 N 55 Bsat (T) 1.50 TABLE II: Curve fitting coefficients. C1 7.40 · 10-6 α1 1.498 α2 1.30 C2 -8.51 · 10-10 β1 1.475 · 10-7 β2 7.71 · 10-7 C3 -2.35 · 10-10 γ1 -44.56 γ2 -5.71 · 10-2 C4 1.86 · 10-13 δ1 -9.01 · 10-4 δ2 -6.20 · 10-5
The coefficients used to interpolate the material properties of the MRF and AlNiCo are given in Table II. By combining Eq. (2), (3), and (4), the current through the coil 17 can be expressed as a function of the magnetic field and magnetic flux density in the magnetic rod I = I :
The current required for a complete magnetization Imag should reach the saturation point of the M-H curve resulting in Imag = I(Hsat,Bsat) with Bsat = Msat + µ0Hsat. As mentioned above, the current that reaches the remanence coercivity of the material cannot be derived analytically. Thus, the current reaching the intrinsic coercivity
is used to provide an upper bound to the demagnetization current that will be determined experimentally. Thus, Idemag can be derived from (7) as follows:
An analytical estimation of the pressure sustained by the valve 10 when closed (i.e., in the no-flow condition) is proposed here. It corresponds to the sum of the field- dependent ∆Pτ and viscosity-dependent ∆Pη pressure drop. It is expressed as follows, if flux fringes are not taken into account:
with Q the flow-rate, η the fluid viscosity, and c a fluid parameter dependent on the ratio D^= lies between 2 (when D ≈ 1) and 3 (D > 100). In the case of a closed valve, ∆Pτ>>^∆Pη. Thus, the maximum pressure difference sustained by the valve 10 is approximated by ∆Pmax ≈ ∆Pτ, and c lies close to 3. The yield stress function τ(HMRF) is given in the datasheet of the MRF. It shows linear characteristics: BMRF(HMRF) = aHMRF, and τ(HMRF) = bHMRF with a^= 0.0047 and b^= 0.2588, for BMRF <^0.5T. However, in our case, the quantity of MRF surrounding the valve is large, leading to flux fringes that need to be accounted for in our model. The distribution of the flux density in the channel 16 is necessary. For
this purpose, a 2-D axisymmetric magneto-static simulation of the valve is performed using the FEMM 4.2 software. This is shown in Fig.4, which corresponds to the fully magnetized state of the rod in AlNiCo-5 (using the corresponding B-H characteristic) without current in the coil 17. The distribution of the magnetic flux density along the dotted line (see Fig.4) in the middle of the channel in the active regions can thus be evaluated. This allows Eq. (8) to reformulate as follows, with l, the position in the channel along the aforementioned line: ^ ⋅ ^ ^^^^^^^^^^^
Although flux fringes are undesirable in magnetic devices, in our case, they lead to an increase in the sustained pressure, improving the performance of our device. The characteristics of the proposed valve and the key properties of the M-H and B-H characteristics of the AlNiCo-5 are presented in Table I. The total volume V of the valve is 353.25 mm3, which classifies it among the miniaturized versions of valves. The setup presented in Figures 5a, 5b and 5c is used to perform the experimental validation of the device of the invention. This setup comprises a cylindrical structure 21 with a first chamber 14’ and a second chamber 15’ that are separated by the MR valve 10. The first chamber 14 of the valve 10 is in fluidic communication with the first chamber 14’ of the cylindrical structure 21 while the second chamber 15 of the valve 10 is in fluidic communication with the second chamber 15’ of the cylindrical structure 21. Both first and second chambers 14, 15 of the valve 10 are filled with the MRF (MRF132DG by Lord Corp.). The second chamber 15’ is open and, thus, in atmospheric pressure. A pneumatic piston 22 is inserted into the first chamber 14’ of the cylindrical structure. The movement of the piston 22 can reduce the volume of the first chamber 14’ of the cylindrical structure. This increases the pressure in the latter when the valve is closed or creates fluid flow when the valve is open. A sectional view of the setup shows the internal part of the chambers, as well as the arrangement placement of the valve. Two pressure sensors 23 (PFT510 Miniature Pressure Sensor by Futek) measure the pressure in each chamber 14’, 15’ of the cylindrical structure 21. A laser displacement
sensor 21 (LK-G3000 by Keyence) measures the displacement of the pneumatic piston 22 over time. Wires of the coil 17 exit the second chamber 15 and are connected to the electronics that perform valve switching. A current probe (AP015 by LeCroy) measures the current flowing in the coil 17. Activation and deactivation of the valve 10 are achieved by applying current pulses to the coil 17. A positive pulse magnetizes the AlNiCo-5 rod 18 and closes the valve 10, while a smaller negative pulse demagnetizes the rod 18 and opens the valve 10. A full- bridge arrangement of MOSFETs allows the coil to be supplied with current in both directions. The IPB054N08N3-G MOSFETs from Infineon Technologies are used for faster switching (Nchannel type), with maximum voltage and current ratings of 80V and 80A respectively. Two gate drivers ICs FAN7842 from Fairchild Semiconductor Corp. (one per pair of MOSFETs) are used to drive the gates of the transistors, increasing the switching speed. The gate drivers are controlled with a F401-RE Nucleo board from ST. The M-H and B-H characteristics acquired experimentally by the German company Magnet-Physik are plotted together in Fig.6. The light gray dotted curve 30 represents the magnetization M over H and the dark dotted curve 31 represents the magnetic flux density B over H. These curves represent the AlNiCo-5 rod from a fully demagnetized condition to the saturation point, after imposing an external field. The third curve 31 with continuous line represents the demagnetization of the AlNiCo-5 rod, from the fully magnetized condition to B = 0. From Eq. (7), the current estimated to reach magnetization of the EPM is Imag = 7.4 A. Fig. 9 shows that by decreasing the magnetization current at 5.1 A and, thus, decreasing also the energy consumed, it is possible to reach magnetization close to the saturation Msat. The marked points depicted in Fig. 6 have been obtained analytically and numerically according to the simulation of Fig. 4 for the different fixed current values (0 A, 5.1 A, 7.4 A). The analytical and simulation results match with an estimation error of less than 6%. Fig. 7 shows the voltage across the coil 17 and the current measured during magnetization and demagnetization of the rod. Positive values result in the magnetization of the inner magnetic rod 18, while negative values result in its demagnetization. Both magnetization and demagnetization are achieved with a single
current pulse for each one. More specifically, the current for magnetizing the EPM and closing the valve 10 is 5.1 A and is reached within tmag = 0.5 ms. This current value was used experimentally instead of 7.4 A in order to reduce the power consumption as explained above. The negative demagnetizing current achieving the remanence coercivity (from a fully-magnetized EPM) is determined experimentally: Idemag = -2.5 A. It can be confirmed that it is lower than the upper limit determined analytically (Idemag < Icoeri = -1.8 A). It is important to point out that the demagnetization of the rod 18 is determined experimentally when the opening of the valve 10 is detected (i.e., no pressure difference can be maintained). However, perfect demagnetization cannot be assured with this method. A small magnetization of the rod 18 could still remain while not being strong enough to block the flow in the valve. Nevertheless, in this application, perfect demagnetization of the rod is not critical. The first demagnetization method proposed in Fig. 3a could be used to overcome this limitation. The demagnetizing current is reached within tdemag = 0.72 ms, after the current starts to flow in the coil. Thus, the switching frequency is 0.85 kHz for 50% duty cycle and 1.25 kHz for 90% duty cycle. These values are high considering different types of fast switching valves, where the switching frequencies are lower than 1 kHz. According to these values, the energy consumption of the valve is 15.3 mJ for activation and 6 mJ for deactivation. Two experimental scenarios are presented here. The first one examines the comparison between the open and closed state of the valve 10. In the closed state, the inner magnetic rod 18 is magnetized with the piston pressure, and consequently, the pressure in the first chamber 14, is set and controlled manually with a manometer. The displacement of the piston and the pressure difference between the two sides of the valve (i.e., between the first and second chambers 14 and 15) are measured while the piston is moving towards the valve. When the EPM is fully magnetized at 5.1 A, we manually increase the piston pressure using the manometer up to the limit of our system, which is 1000 kPa. As shown in Fig.8 (curve with circles), there exists an initial abrupt increase of displacement of maximum ∆x = 10 mm, which remains constant as the pressure difference increases. The abrupt increase of displacement at the beginning of the process is explained by the air trapped in the chamber, which is initially
compressed. After that, the absence of displacement proves that the valve 10 remains closed. The maximum pressure measured is 1010 kPa. Higher pressures could not be tested because the compressed air supply used for these experiments is limited to about 1000 kPa. The value derived from the analytical and simulation results is 949 kPa. The experimental and numerical results are in agreement with an estimation error of less than 6.4%. The proposed system allows the possibility of tuning the sustained pressure. In fact, the AlNiCo rod could be partially magnetized and reach lower values than Mrem when the current is stopped. Thus, the magnetic field in the MRF would also be lower, reducing the pressure that can be sustained. Two further examples are shown in Fig.8, with magnetizing currents of 1.8 A (curve with triangles) and 3.5 A (curve with squares). In the case of I = 1.8 A, the manometer value is set to 100 kPa. At a pressure of 75 kPa, the displacement starts to increase until the stroke limit of 50 mm is reached. In the case of I = 3.5 A, once the air trapped in the chamber is compressed, the displacement also increases to the limit of the system at a pressure of 180 kPa. These results demonstrate the possibility of tuning the sustained pressure. In the open condition (dashed curve), the pressure on the pneumatic piston is again manually set at 1000 kPa. An increase in piston displacement of ∆x = 15 mm occurs until the pressure measured in chamber 15 reaches 993 kPa. After that, the MRF begins to flow through the valve at a constant flow rate. As a result, the piston displacement continues to increase while the pressure difference remains constant. The piston has a stroke limit of 50 mm. This explains the range of Piston Displacement axis in Fig.8. The flow rate in this case is calculated as Q = v ·Arod = 459 mm3/s, where v is the piston’s velocity and Arod is the piston’s surface. v is calculated from the experimental measurement of the piston’s displacement over time. Eq.8 is validated using the above experimental values of Q and ∆P, as well as with using η = 0.112Pa · s and ∆Pτ = 0 because the valve is open and there is no magnetic field. Fig.9 shows the second experimental scenario. It shows the displacement ∆x of the piston and the pressure difference ∆P over time of a pressure of 1000 kPa. The transition of the valve from the closed to the open state is shown. In an initial state, the valve is fully-magnetized after applying a magnetization pulse. At t = 2 s the piston starts to push and continues throughout the whole experiment. The valve remains
closed until t = 3 s when the demagnetization pulse is applied and the valve is open. During this time, there is no piston displacement, except for the initial abrupt displacement associated with the compression of the air. This means that the valve remains closed. When the valve is open, the displacement ∆x begins to increase linearly, confirming that the valve is open. For a pressure difference of 1000 kPa the stroke limit of 50 mm is reached within one second. The valve is initially closed. At t = 3s the demagnetization takes place and the valve opens, allowing flow and piston displacement. In conclusion, valves constitute an essential component of the control of fluidic power transmission systems. In this paper, we have presented the design, implementation and experimental validation of a miniaturized MRF valve that uses a single rod EPM. The EPM is magnetized and demagnetized with a single current pulse. The valve can sustain a maximum of 1010 kPa, while being activated and deactivated within 0.5 ms and 0.72 ms respectively. At the same time, it consumes negligible power in the steady state condition, with 15.3 mJ for activation and 6 mJ for deactivation. The current work is an improvement compared to the state-of-the-art in terms of power consumption and maximum sustained pressure. However, smaller valves exist, with much lower capabilities with regard to pressure. In addition, our valve has the property that the sustained pressure can be tuned, which can provide flexibility to the applications that use it. This valve is of particular interest in several fields such as for example robotics applications or medical application such as medical applications for plantar pressure relief in diabetic patients. The MRF valve can be used to control the stiffness of medical footwear.