Determination of the position of a movable member relative to a fixed member

Abstract

In order to determine the position of a movable member (2) having a plurality of position magnets (Pk) relative to a stationary member (1) having a plurality of position sensors, it is stipulated that: for a set of position sensors in the region of the movable member (2), the sensor response is detected; from a sensor model (SM), the sensor model responses of the set of position sensors are determined for a plurality of assumed different relative positions of the movable member (2) relative to the stationary member (1), the sensor model response is compared with the sensor response and the relative position of the assumption having the smallest deviation between the sensor model response and the sensor response is used as the relative position of the movable member (2).

Description

Technical Field

[0001] The present invention relates to a method for determining the position of at least one movable component relative to a stationary component, wherein a device comprising a plurality of position magnets (positioning magnets) having a pole pitch is provided on the movable component, and a device comprising a plurality of position sensors having a sensor pitch is provided on the stationary component, wherein the magnetic field of the position magnets in the region of the position sensors is detected in the form of a sensor response by the position sensors. The present invention also relates to a motion system comprising a movable component and a stationary component, wherein the position of the movable component relative to the stationary component is determined. Background Art

[0002] A linear motor or planar motor is characterized by a movable part (mover) that moves relative to a fixed part (stator) due to interacting magnetic fields. To this end, a drive magnet (electromagnet or permanent magnet) is provided on one of the two parts. This drive magnet interacts with a magnetic field generated by a current-carrying drive coil on the other part to generate a propulsive force. If a voltage is applied to the drive coil, a magnetic field is generated that interacts with the magnetic field of the drive magnet, thereby generating a force on the movable part that moves it. To move the mover, a moving magnetic field is generated by correspondingly controlling the drive coil. This basic electromotive principle is well known and does not require further explanation here. In principle, it does not matter whether the drive coil is located on the movable part (mover) or the fixed part (stator).

[0003] To regulate the movement of a linear motor's mover, knowing its current position relative to the stator is crucial for correctly manipulating the drive coils to generate the moving magnetic field. Therefore, determining the stator's position is crucial for the operation of a linear motor. However, determining the mover's current position when the linear motor is switched on is particularly difficult because its location at the time of switching on is unknown. Various methods have been proposed for positioning or determining the position of a linear motor when it is switched on.

[0004] For example, US 7,932,684 B2 describes a linear motor which, for positioning purposes, additionally includes a position magnet arranged on the mover and a stationary position sensor (e.g., arranged on the stator). If the mover is moved, the position magnet moves relative to the position sensor, and the current position of the mover relative to the stator can be determined. The position magnet includes a first row of a plurality of permanent magnets arranged side by side and cooperating with an incremental sensor, and a second row of a plurality of permanent magnets arranged side by side and cooperating with an absolute sensor. The absolute sensor, such as a Hall sensor, is designed such that this absolute sensor provides only two states, and the state changes at a defined position of the mover. The incremental sensor, such as a magnetoresistive sensor, is designed such that this incremental sensor provides a number of repeated sensor cycles, and within one sensor cycle, the position can be determined very precisely. When powered on, a "homing" operation must first be performed, i.e., based on a pre-given known zero position. For this purpose, the mover is moved until the absolute sensor detects a state change, thereby determining the zero position. Then, starting from this zero position, by counting the number of sensor cycles and precisely determining the position within the sensor cycle, the current position of the mover can be determined step by step. Therefore, in US 7,932,684 B2, a reference run, i.e., moving the mover, is required to determine the position of the mover when powered on. However, this type of position detection can only be reasonably implemented in a relatively limited mover movement area. For many applications, especially for linear motors with a large movement space or for long-stator linear motors with a large number of movers, this type of positioning is of course completely unsuitable.

[0005] A method for determining the absolute position can be obtained from US 7,994,742 B2, even when the linear motor is powered on. Here, an elongated position magnet is arranged on the mover over the possible movement space, and this position magnet is arranged such that a position-dependent offset is generated transversely. On a stationary structure, such as on the stator, a position sensor for detecting the magnetic field of the position magnet is arranged. Due to the offset, a unique magnetic field is generated at each position of the mover, and this magnetic field is detected by the position sensor. Therefore, even when the linear motor is powered on, the current position of the mover can be immediately inferred without moving the mover. However, the movement space is of course limited to the length of the position magnet here, and is therefore very limited. For many applications, especially for linear motors with a large movement space or for long-stator linear motors with a large number of movers, this type of positioning is of course completely unsuitable.

[0006] US 6,876,107 B2 describes a known long-stator linear motor as a linear motor. Such a long-stator linear motor includes a plurality of drive coils which are fixedly arranged side by side and form the stator of the long-stator linear motor. A plurality of movers which can move along the stator can be arranged along the stator. Each mover carries a drive magnet. In order to move the movers, the drive coils which cooperate with just one mover are energized respectively. In this way, the individual movers can be moved along the stator independently of one another. Such long-stator linear motors are typically used in flexible conveying systems, for example in production processes or in conveying technology. In addition, US 6,876,107 B2 also describes the determination of a true absolute position, which makes it possible to immediately determine the exact position of a mover when the long-stator linear motor is switched on, without having to perform a reference positioning for this purpose (for example by a reference run of the mover). This is of course very advantageous, especially when considering that it is not uncommon for there to be several hundred movers in a long-stator linear motor at the same time. For this purpose, an additional position magnet is provided exactly on one mover, and a plurality of position sensors, such as magnetoresistive sensors, are arranged along the stator, and the magnetoresistive sensors detect the magnetic field of the position magnet. However, the position sensors must be arranged close enough at the same time to ensure that at least one position sensor can detect the magnetic field of the position magnet at each point in time. When the long-stator linear motor is switched on, at least one position sensor is started for each mover, so that positioning can be carried out even without a reference positioning of the mover. The disadvantage here is that an additional position magnet is required and the position sensors must be arranged very closely, which requires a large number of such position sensors.

[0007] EP 3 376 166 A1 again describes the position detection of the mover of a long stator linear motor when the motor is switched on. In this case, the drive magnets on the mover are used for position detection, and no additional position magnets for position detection are required. The position sensors can also be arranged further apart from each other in this case, so that the number of position sensors required can be reduced. For position detection, first, the rough position of the mover is determined by determining the edge position of the drive magnet setting unit on the mover. A position sensor is determined for one of the drive magnets, and this position sensor detects the magnetic field of this drive magnet. Then, the relative position of the mover relative to this position sensor can be determined using this position sensor. Since the installation position of the position sensor is known, the absolute position of the mover can be determined using this known installation position and the determined relative position. Although this method is very reliable, it also has weaknesses. In particular, if two movers are very close to each other or even in contact with each other, this method may fail because the edge region cannot be determined reliably or with difficulty, or because multiple movers adjacent to each other closely cannot be identified.

[0008] Similar problems also occur in position measurement systems in which position magnets are arranged on a moving part and the magnetic fields of these position magnets are detected by position sensors on a fixed part. Here, there is also the basic problem of how the moving part, or possibly multiple moving parts, is positioned relative to the fixed part. Summary of the Invention

[0009] Therefore, the object of the present invention is to provide a method for determining the position of a movable member having a plurality of position magnets relative to a fixed member having a plurality of position sensors. In particular, this method can also identify multiple movable members that are close to each other or completely in contact with each other, and reliable position determination can also be achieved for such close or completely in-contact movable members.

[0010] This object is achieved by a method for determining the position of at least one movable member relative to a stationary member according to the present invention, wherein a setting unit having a plurality of position magnets with pole pitches is provided on the movable member, and a setting unit having a plurality of position sensors with sensor pitches is provided on the stationary member, and the position sensors detect the magnetic field of the position magnets in the area of the position sensors in the form of sensor responses. According to the present invention, for a group of a plurality of position sensors considered for position determination in the area of the movable member, sensor responses are detected; a sensor model is provided, which provides a sensor model response related to the relative position of the at least one movable member relative to a position sensor for the at least one movable member and a position sensor; the sensor model responses of all position sensors of the group of position sensors are determined for different relative positions of the movable member relative to the stationary member assumed; the sensor model responses are compared with the sensor responses detected by the group of position sensors; and the relative position having the smallest deviation between the sensor model response from the sensor model and the detected sensor response among the different relative positions assumed is used as the relative position of the movable member, or the relative position of the movable member is determined therefrom or the absolute position of the movable member relative to a stationary reference point is determined.

[0011] This object is also achieved by a motion system according to the invention, which comprises at least one movable member that is movably arranged relative to a stationary member, wherein a setting unit with a plurality of position magnets having a pole pitch is arranged on the movable member, and a setting unit with a plurality of position sensors is arranged on the stationary member, these position sensors having a sensor pitch, and the magnetic field of the position magnets in the region of the position sensors can be detected in the form of a sensor response by means of the position sensors, wherein a calculation unit is provided for determining the position of the movable member relative to the stationary member, and this calculation unit processes the sensor responses of the position sensors. According to the invention, the calculation unit detects the sensor responses for a group of position sensors considered for position determination in the region of the movable member; a storage unit is provided, in which a sensor model is stored, and this sensor model provides a sensor model response related to the relative position of the movable member relative to the position sensor for the at least one movable member and a position sensor; the calculation unit determines the sensor model responses of the group of position sensors for a plurality of different assumed relative positions of the movable member relative to the stationary member by means of the sensor model; the calculation unit compares the sensor model responses with the sensor responses detected by means of the group of position sensors; and the calculation unit determines the relative position of the movable member as the relative position having the smallest deviation between the sensor model response from the sensor model and the detected sensor response among the plurality of different assumed relative positions, or determines the relative position of the movable member or the absolute position of the movable member relative to a stationary reference point accordingly.

[0012] The unknown position of the movable member can be found by using a sensor model by comparing the sensor model response with the actually measured sensor response while changing the assumed position of the movable member. This method is very robust on the one hand, making reliable positioning possible even in the case of deviations due to inaccurate mounting positions of the position sensors or deviations in the position magnets. The sensor model describes the sensor response curve of the position sensor when the position magnet setting unit moves past a position sensor. Therefore, the sensor model only needs to be determined once and can then be used for each position sensor. On the other hand, by using the sensor model, even movable members that are close to or completely adjacent to each other can be identified and distinguished. This method can not only identify where the movable member is located, but also (if there are multiple different movable members) which movable member is located where. Description of the Drawings

[0013] The following will refer to the attached Figures 1 to 9The present invention will be described in detail. The accompanying drawings exemplarily, schematically and non - limitatively show advantageous structural designs of the present invention. In the drawings:

[0014] Figure 1 A motion system is shown, which has a stationary member and a member movable relative to the stationary member;

[0015] Figure 2 A typical magnetic field of a position magnet setting unit on the movable member is shown;

[0016] Figure 3 A typical sensor signal of a position sensor is shown;

[0017] Figure 4 A typical sensor response of the position sensor obtained from the sensor signal is shown;

[0018] Figure 5 A typical curve showing the relationship between the sensor response and the relative position of the movable member with respect to the position sensor is shown, as an example of a sensor model;

[0019] Figure 6 An alternative storage method of a sensor model is shown;

[0020] Figure 7 Possible situations during positioning are shown;

[0021] Figure 8 Possible curves showing the relationship between the cost function value and the assumed relative position are shown; and

[0022] Figure 9 Positioning in the area of a splitter in a long - stator linear motor (as an embodiment of the motion system) is shown. Detailed Description of the Invention

[0023] As shown in Figure 1 The present invention starts from a motion system in which a plurality of position sensors Sn (n > 1) are arranged side - by - side on a stationary member 1 and a number of position magnets Pk (k > 1), i.e., permanent magnets or electromagnets, are arranged side - by - side on a movable member 2. The movable member 2 moves relative to the stationary member 1. The movable member 2 is arranged such that the position sensors Sn can detect the magnetic field of the position magnets Pk. Generally, this means that the position magnets Pk are arranged on the movable member 2 facing the stationary member 1, while the position sensors Sn - with their respective active surfaces - are arranged facing the movable member 2. The position magnets Pk are not necessarily as in Figure 1arranged to be directly adjacent to each other as shown. The distance between two adjacent position magnets Pk is referred to as the pole pitch Tp. This pole pitch Tp can, but does not have to, be equal to the position magnet width p. The pole pitch Tp can, but does not have to, be smaller than the sensor pitch s between two adjacent position sensors Sn. The position magnets Pk can be arranged on the movable member 2 with alternating polarities (see also Figure 2 ). In the case of a planar motor, such a setting unit is produced in one plane along two directions.

[0024] Such a motion system occurs in many applications. An example is a linear motor, a long-stator linear motor, or a planar motor, in which a plurality of drive coils As (s > 1) are additionally arranged on the stationary member 1, as shown in dashed lines in Figure 1 . In the case of a planar motor, the drive coils As are arranged in one plane, while in the case of a linear motor or a long-stator linear motor, the drive coils are arranged side by side along the direction of motion. In this case, the stationary member 1 forms the stator of the motor, and the movable member 2 can move along this stator. A plurality of drive magnets are arranged on the movable member 2, which forms the mover of the motor in this application. Here, the drive magnets can simultaneously be the position magnets Pk (as described, for example, in EP 3 376 166 A1), or the drive magnets are arranged separately from the position magnets Pk. If the drive coils As in the region of the movable member 2 are energized (by applying a voltage), for example, under the control of a drive controller (not shown), a drive magnetic field is generated, which interacts with the magnetic field of the drive magnets on the movable member 2 to generate a propulsive force on the movable member 2. The drive magnets can be permanent magnets or electromagnets here. In the case of an electromagnet, permanent magnets can also be arranged on the stationary member 1 instead of the drive coils As. This motor principle is well known and will not be elaborated on here.

[0025] Another application is the pure position detection in a motion system that includes a member 2 movable relative to the stationary member 1, such as a carriage on a machine tool. In this application, the movable part 2 is moved by additional mechanical means, and the position of the movable part 2 relative to the stationary member 1 is to be determined.

[0026] The position sensor Sn used measures a magnetic field M, for example, the magnitude and / or direction of the magnetic field of a position magnet Pk in the area of the corresponding position sensor Sn. Possible position sensors Sn include magnetostrictive sensors, magnetoresistive sensors (such as anisotropic magnetoresistive sensors, tunnel magnetoresistive sensors, or giant magnetoresistive sensors), or Hall sensors. Of course, other sensor types capable of detecting a magnetic field M are also possible. Different sensor types can also indicate the magnitude of the magnetic field M (if such a magnitude is detected) in different ways. For example, a Hall sensor provides a value representing the magnetic field strength (e.g., Tesla), although Hall sensors generally do not operate in a saturated state. In contrast, a magnetoresistive sensor provides a value within a specific value range, for example, between 0 and 1, although such sensors generally operate in a saturated state. This means that such a sensor provides a value of, for example, 1 for a wide range of magnetic field strengths. Therefore, such a sensor generally cannot determine the absolute value of the magnetic field strength of the magnetic field M; instead, it can only be used to determine whether the corresponding sensor type is in the saturation range.

[0027] exist Figure 2 The figure shows, by way of example, the magnetic field M of the arrangement of the position magnet Pk on the movable component 2, in the form of magnetic lines of force. It can be seen that the magnetic field M is strongest at the center of the magnetic pole and rapidly decreases toward the edges. It can also be seen that the magnetic field M decreases significantly and relatively quickly in the surrounding area, that is, at a greater distance from the position magnet Pk. Therefore, it is beneficial to improve measurement reliability if the air gap between the position sensor Sn and the position magnet Pk does not exceed half the pole pitch Tp.

[0028] A position sensor Sn provides, for example, a sine and cosine curve (in the form of a voltage measurement value u) as sensor signals, as in Figure 3 As shown in FIG. As is known, the position x of a single movable position magnet Pk proportional to the magnetic field angle γ of the magnetic field M can be determined linearly approximately from the sensor signal (e.g. in the form of a voltage u) (sine and cosine curves) based on the angle γ of the magnetic field M, i.e. the angle at which the magnetic field M impinges on the position sensor Sn, e.g. x=k * arctan(sin(2γ) / cos(2γ)), where k is a constant factor, for example k=p / (2π). However, the position sensor Sn can of course also directly provide the position x as a sensor response SA, or a sine and cosine curve that is then evaluated, or an angle information γ. The position sensor Sn can also (depending on the type of sensor) provide the absolute value |A| (amplitude of the sine and cosine curves) of the variable for the magnetic field M as a sensor response SA, for example as Therefore, the sensor response SA can also contain multiple variables, such as angle information γ and absolute value |A|.

[0029] If a position magnet Pk moves past a position sensor Sn, a typical sawtooth curve of the angular information γ between 0 and 2π (depending on the sensor value) and a curve of the absolute value |A| are generated according to the position x of the position magnet Pk relative to the position sensor Sn, as shown in Figure 4 . If the position sensor Sn operates in a saturated state, such as in the case of a magnetoresistive sensor, the absolute value |A| can mostly be constant here. In the case of other sensor types, such as Hall sensors, different curves of the absolute value |A| can also be generated. Different sensor types can also provide different sensor signals and / or sensor responses SA representing the measured magnetic field.

[0030] If the setting unit of the position magnet Pk (k > 1) on the movable member 2 (as shown, for example, in Figure 1 ) moves past a position sensor Sn, a sawtooth curve of the angular information γ and / or a curve of the absolute value |A| are thus generated regarding the relative position x of the movable member 2 with respect to the stationary member 1, as shown in R , for k = 5 position magnets Pk as shown in Figure 5 .

[0031] The value of the angular information γ can also be scaled by a constant factor. For example, it can be stipulated that the position sensor Sn provides the value π / 2 for the angular information γ at the center of the position magnet Pk. According to Figure 4 , the value π is expected. Thus, it is scaled by the factor -π / 2. When the sensor measures the edge of the position magnet Pk, 3π / 2 is thus generated, and the value 0 is provided at 1 / 4 of the position magnet Pk. Such a scaling is also the reason why the curve of the angular information γ contains six sawtooth vertices in the case of five position magnets Pk in Figure 5 . However, such a scaling is not important for the present invention, and it is only necessary to understand it and apply it uniformly.

[0032] In Figure 5 , a non-linear curve of the angular information γ can also be seen, especially in the edge region of the movable member 2. An expansion of the magnetic field at the edge of the setting unit of the position magnet Pk (wider sawtooth vertices are generated at the edge) can also be seen as a result of the magnetic field M as shown in Figure 2 .

[0033] Here, the relative position x between the movable member 2 and the position sensor Sn RIt is for an arbitrary point of the movable member 2, for example, for the center (observed along the movement direction) or an end of the setting unit of the movable member 2 or the position magnet Pk. In Figure 2 it, for example, the center of the setting unit of the position magnet Pk is selected as the zero point of the relative position x R as shown by the dotted line drawn by the movable member 2 for the movement position. Therefore, the relative position x R for the positive and negative values with respect to the selected zero point will be obtained according to the position of the movable member relative to the stationary position sensor Sn R The relative position x

[0034] If the position x(i) of the movable member 2 (whether as an absolute position or a relative position) is only known at the current detection time point (i) (in which time period the position is determined), then the current position x(i) can be easily obtained from the position x(i - 1) at the previous detection time point (i - 1) obtained, for example, from the known movement of the movable member 2. However, the main problem is to determine the position x(i) without relying on the information of the position x(i - 1) at the previous detection time point (i - 1), that is, for example, especially at the initial stage when the system is turned on or accelerating.

[0035] When the system is turned on or accelerating, although all the position sensors Sn on the stationary member 1 can be read, it is impossible to determine which position magnet Pk on the movable member 2 is detected by a certain position sensor Sn. Each position magnet Pk on the movable member 2 can cause the same sensor response SA, for example, the same angular information γ, in a certain position sensor Sn. The sensor response SA measured by a position sensor Sn is therefore non - one - to - one (non - unique or ambiguous). To solve this ambiguity and make reliable positioning possible even in such cases, the following operations are carried out according to the present invention.

[0036] For a predefined combination of a known setting unit of a position magnet Pk on a movable member 2 (the pole length p, pole pitch Tp, orientation, polarity, etc. of the position magnet Pk) and a certain known position sensor Sn on a stationary member 1 (sensor type, distance between the sensor and the position magnet, etc.), a sensor model SM can be determined, which determines the sensor model response SA R related to the relative position x * of the movable member 2 relative to the stationary member 1 or the position sensor Sn * such as the angular information γ * and / or the absolute value |A| * For the angular information γ* and the absolute value |A| * For example, the sensor model SM corresponds to the curves of these variables, as shown in Figure 5 The sensor model SM thus provides that, using a specific position sensor Sn, different relative positions x of a specific movable component 2 are detected. R Which values of the sensor response SA can be expected in the equation? Such a sensor model SM can be determined for a specific combination of a positioning unit of a position magnet Pk and a position sensor Sn, for example by measurement technology, by moving the positioning unit of the position magnet Pk past the position sensor Sn and simultaneously detecting and storing the sensor model response SA. * However, the sensor model SM can also be determined by simulation or calculation. The sensor model response SA can also be detected. * The sensor model SM is stored.

[0037] The sensor model SM is thus determined by detecting the sensor model response SA when the entire arrangement of the plurality of position magnets Pk moves past the position sensor Sn of the stationary member 1. * During the movement, the sensor response SA of the position sensor Sn is detected as the sensor model response SA * Therefore, the sensor model SM is disconnected from the arrangement of the position sensors Sn on the stationary component, since only the sensor model response SA for one position sensor Sn is detected. * .

[0038] If there are a plurality of different movable components 2, which, for example, have a different number of position magnets Pk, as is common, for example, in a long-stator linear motor or a planar motor, or when different position sensors Sn are arranged on the stationary component 1, then of course there is an associated sensor model SM for each possible combination of an arrangement unit of position magnets Pk and a position sensor Sn.

[0039] Of course, for a specific motion system, a sensor model SM or a plurality of sensor models SM only have to be determined once. Therefore, the sensor model SM, being known and existing, can be regarded as a prerequisite for carrying out the positioning method.

[0040] You can Figure 5 As shown in FIG, the sensor model SM is stored as a relative position x R Related sensor model response SA * The value of γ, i.e., the angle information γ * and / or the absolute value |A| *The form of the value. However, this would require a lot of storage space, especially if the sensor model SM should be stored at a high resolution with respect to the relative position x R and / or the sensor model response SA * is stored at a high resolution.

[0041] For this reason, it is also possible to store the sensor model response SA * of the sensor model SM in other forms, as described by means of Figure 6 As described. According to Figure 5 the sensor model SM is discretized with a position increment Δx, and the overflow (a quantity related to the number of position magnets Pk) of a sawtooth curve with an angular information γ of height z (e.g., 2π) is added at the relative position x * This results in R a discrete curve of the angular information γ1 in Figure 6 An as smooth as possible curve is approximated in this discrete curve, for example a sequence of polynomials (spline curve) or another suitable smooth curve. As is well known, a spline curve is a function that is composed of polynomials of a certain degree piecewise, where the discrete points can be the interpolation points of the spline curve. Then the angular information γ2 approximated by the curve can be stored as the mathematical description of the curve selected for the sensor model SM, which requires much less storage space. By using the modulo function modstu, the angular information γ * of the sawtooth curve can be easily calculated from the approximated angular information γ2, i.e., γ * (x R ) = γ2(x R ) mod z (e.g., z = 2π).

[0042] Similarly, the curve of the absolute value |A| * in the sensor model SM can be discretized with the position increment Δx to obtain interpolation points, in which then a smooth curve can be approximated, for example again a spline curve. Therefore, the curve of the absolute value |A| * can also be stored in the form of the mathematical description of the curve with storage space saved. In this case, the value of the absolute value |A| R for a certain relative position x * can be directly obtained from the stored curve.

[0043] In the case of other or additional sensor responses SA or the sensor model response SA * of the sensor model SM, the same operation can be carried out in the same way.

[0044] In Figure 7Shown is a movable part 2 having a number of position magnets Pk, the movable part being in a certain relative position x relative to a stationary member 1 having a position sensor Sn R In this relative position x of the movable member 2 R the magnetic field M of the position magnets Pk on the movable member 2 causes a sensor response SA of a certain number of the position sensors Sn in the area of the movable member 2. How many position sensors Sn respond naturally depends on the sensor spacing s and also on the magnetic field strength. For position determination, it is preferred to consider only the responding position sensors Sn. In this case, a limit value of a sensor response SA, for example the absolute value |A|, can be specified from which a sensor response SA of a position sensor Sn is considered as a measured value. However, it is also possible to always consider a predefined number of the position sensors Sn in the area of the movable member 2. The j position sensors Sn considered for position determination or positioning j are usually a partial number of the existing position sensors Sn.

[0045] If one observes a position sensor Sn, then this position sensor provides a corresponding sensor response SAn, for example angular information γn and absolute value |A|n, according to the position of the position magnet Pk measured by a magnetic field M relative to the position sensor Sn (as shown in Figure 4 ). The other position sensors Sn - 1, Sn + 1 considered likewise provide sensor responses, for example angular information γn - 1, γn + 1 and absolute values |A|n - 1, |A|n + 1. All j≥1 position sensors Sn considered j each provide a sensor response SA j .

[0046] Furthermore, a sensor model SM for this movable member 2 and the existing position sensors Sn is known. The task is currently to determine the relative position x of the movable member 2 R such that the sensor model responses SA j from the sensor model SM of the considered position sensors Sn * j match as well as possible the actually measured sensor responses SA j of the considered position sensors Sn j .

[0047] A determined relative position x of the movable member 2 R thus provides from the sensor model SM the sensor model responses SA j of the considered position sensors Sn * j .

[0048] These sensor models respond to SA * j and the sensor response SA actually measured using the considered position sensor Snj j For this purpose, the relative position x of the movable component 2 is changed R , and for each changed relative position x from the sensor model SM Rv Determine the sensor model response SA * j (x Rv ). Of course, this change is only hypothetical and the movable member 2 does not physically move on the stationary member 1 for this purpose. * j (x Rv ) and the actual measured sensor response SA j The relative position x of the minimum deviation between Rv It is then used to determine the relative position x of the movable component 2 relative to the stationary component 1 R .

[0049] Then, based on the determined relative position x R By means of the known geometrical shapes of the movable component 2 and the fixed component 1 and by means of the known installation position of the position sensor Sn, the absolute position x can be easily determined. A . Absolute position x A It refers to a predetermined, fixed reference point RP ( Figure 7 ), the reference point can be set at an arbitrary position, usually a point on the fixed component 1. The installation position x of the position sensor Sn Sn is known. Therefore, for example, it can be easily determined by x A =x Sn +x R Calculate the absolute position x A However, the absolute position x A Of course, the same may also apply to any other arbitrary point of the movable component 2 .

[0050] In order to implement the method of positioning by different means, the relative position x of the movable component 2 can be changed as explained below. Rv. For the following description, the angular information γ and the absolute value |A| are assumed to be the sensor response SA. However, for the present invention, only one of these variables can also be used as the sensor response SA, or an additional variable provided by the position sensor Sn or other variables provided by the position sensor Sn can also be used. The sensor response SA used can be directly provided by the position sensor Sn, or it can also be obtained from the sensor signals provided by the position sensor Sn (as in, for example, Figure 3 ), for example, in a computing unit 5.

[0051] To determine the relative position x R (or the absolute position x A ) of a movable member 2, a group having several j position sensors Sn j is used. Of course, the position sensors Sn that provide the sensor response SAj are used j , and these position sensors are thus located in the region of a movable member 2. In the example according to Figure 7 , for example, the position sensors Sn-2, Sn-1, Sn, Sn+1, Sn+2 can be used. However, as an alternative, only the position sensors Sn-2, Sn-1, Sn or the position sensors Sn, Sn+1, or other position sensors can also be used. The position sensors Sn used j do not necessarily have to be directly adjacent position sensors Sn.

[0052] To determine the relative position x R of the movable member 2, the position sensors Sn used j thus provide the sensor response SA j , for example . For an arbitrary relative position x Rv of the movable member 2, a sensor model response SA j (x * j ) can be obtained from the sensor model SM for each position sensor Sn Rv , for example .

[0053] To determine the relative position x R of the movable member 2 relative to the stationary member 1, the sensor response SA Rv measured by the considered position sensor Sn j will be compared with the sensor model response SA j for * j "(x Rv ).

[0054] At the position sensor SN under consideration j the sensor model response SA * j (x Rv ) and the sensor response SA actually measured using the position sensor Sn under consideration j The relative position x of the change when they are closest j is used as the relative position x of the movable member 2 Rv , or for obtaining the relative position x R (for example when this relative position is for another point of the movable member 2) or the absolute position x of the movable member 9 R . A .

[0055] For the comparison between the sensor response SA j and the sensor model response SA * j (x Rv ), a cost function J can be used as a function of the sensor response SA j and the sensor model response SA * j (x Rv ), that is, J = f(SA j , SA * j (x Rv )) ∀ j. This cost function evaluates the deviation between the measured sensor response SA j and the sensor model response SA j of the position sensor Sn under consideration * j (x Rv ). The cost function J is preferably evaluated for all j position sensors Sn under consideration j . The cost function J can be easily formulated as, for example, a quantity of the difference, such as , or as the squared error, such as in the form of , or formulated in any other suitable form. If the sensor response SA is a vector composed of multiple variables, such as composed of the angle information γ and the absolute value |A| as above, then an arbitrary norm can be used, such as the Euclidean norm ‖ ‖2, that is, for example .

[0056] The relative position in the changed relative position that optimizes (i.e., either minimizes or maximizes) the value of the cost function is then the relative position x being sought R or for obtaining / determining the relative position x of the movable member 2R or the absolute position x A .

[0057] In the first design configuration, the relative position x can be changed in small position increments Δx R The smaller the position increment Δx Rv , the higher the resolution. In which region the relative position x is changed R can be predefined - for example, based on the known geometric dimensions of the movable member 2 and / or the known sensor spacing s. In Rv for example, the cost function J for such a change is shown. Then, the position at the minimum of the cost function J in the predefined region is the relative position x sought Figure 8 or for determining the relative position x of the movable member 2 R or the absolute position x R . A .

[0058] However, a possible second design configuration for positioning is more advantageous. If k represents the number of position magnets Pk on the movable member 2, then one of the position sensors Sn used j can detect the magnetic field of one of these k position magnets Pk. Therefore, in order to cover all possible positions of the movable member 2, it is only necessary to change the movable member 2 by an amount equal to k pole pitches Tp. Thus, the number of relative positions x that need to be changed can be significantly reduced Rv without compromising the positioning accuracy. Even in this case, a value for the cost function J is obtained for each of the changed relative positions x Rv . Then, the relative position x for which the value of the cost function J is optimal (either minimum or maximum) Rv is the relative position x sought R , or for determining the relative position x of the movable member 2 R or the absolute position x A .

[0059] The number of relative positions x that need to be changed can be further reduced Rv . For example, those relative positions x for which the attached absolute value |A| is below a certain limit can be ignored Rv .

[0060] In another possible design configuration, the cost function J can be optimized (minimized or maximized) in a mathematical optimization based on the relative position x Rv . Mathematically, J = f(SA j , SA * j (xRv )) This is represented in the form of. Here, the relative position x that changes is obtained by the optimization algorithm Rv , such as Newton's method, gradient method, evolutionary algorithm, etc. Through the optimization algorithm, the relative position x is iteratively changed in this way Rv such that the cost function J approaches the minimum or maximum value and is thus optimized. The first relative position x of the iteration Rv can be arbitrarily given or predetermined by the applied optimization method. For this purpose, a termination criterion is also predetermined in advance, that is, the optimization is iteratively repeated until the termination criterion is reached. A possible termination criterion is below (or above) a predetermined cost function threshold or below a predetermined deviation of the values of the cost function J of two successive iterations.

[0061] Multiple different groups with position sensors Sn j can also be used for positioning. The determined position sensor Sn j can also be included in multiple groups, however, these groups are distinguished by at least one position sensor Sn j respectively. In Figure 7 the example of, for example, the first group can include position sensors Sn - 1, Sn, Sn + 1, while the second group can include position sensors Sn, Sn + 1, Sn + 2. The number of position sensors Sn in the group j does not have to be the same, that is, the first group can include, for example, two position sensors Sn j , while another group can include three position sensors Sn j .

[0062] Therefore, a method can be used to optimize a cost function J for each group of position sensors Sn by changing the relative position x Rv as described above. Now, a group can be selected, and the relative position x for which the value of the cost function J is optimal (either minimum or maximum) for this group j (or generally, the relative position x at which the deviation between the sensor model response and the measured sensor response is the smallest Rv ) is used as the sought relative position x Rv or for obtaining or determining the relative position x of the movable member 2 R or the absolute position x R . As a result of optimizing the cost function using other groups, it can then be used to check the relative position x of the movable member 2 A (or the absolute position x R A ​The calculated confidence level (e.g., the deviation between two calculated positions shall not exceed a certain value. For example, the set with the minimum value of the cost function J can be selected, and other sets are not adopted or used for confidence check).

[0063] If there are multiple different sensor models SM, for example, due to different movable members 2, the above method can be implemented for each sensor model SM, i.e., optimization can be performed based on each sensor model SM. Therefore, this method not only provides the relative position x of the movable member 2 R , but also provides information about which movable member 2 it is. For example, in the case of optimizing a cost function J, the cost function J becomes minimum by using one of the sensor models SM, from which it can be concluded that the movable member 2 is the one on which this sensor model SM is based.

[0064] If there is a single movable member 2 or a movable member 2 far enough from other movable members, then the minimum value of the cost function J will be close to zero (depending on the accuracy of the sensor model SM and the accuracy of the installation position of the position sensor Sn j ).

[0065] However, even when two movable members 2 are close to each other or even in direct contact, the positioning method of the present invention still works. "Close" here refers to an arrangement in which the magnetic fields M of the two movable members 2 affect each other in terms of the measurement accuracy of the position sensor Sn. Since the magnetic field M decreases significantly outside the position magnet Pk, it can be assumed that the influence is only generated in the edge region of the movable member 2. Within the range of this influence, the sensor model response SA j from the sensor model SM (which is determined for an isolated position sensor Sn) for a position sensor Sn * will deviate more from the sensor response SA measured by using this position sensor. However, because multiple position magnets Pk are provided on the movable member 2 and multiple position sensors Sn j are always considered for position determination, there are such position sensors Sn j for which a sensor model response SA from the sensor model SM * can be found to be in good agreement with the measured sensor response SA. This agreement is sufficient to calculate / determine the relative position x R . Even when more than two movable members 2 are close to each other or in direct adjacency, this still works. Therefore, the relative position x of multiple movable members 2 close to each other or in direct adjacency can also be easily determined R(or the absolute position x A ). Since the characteristics of the movable member 2 are also represented by the sensor model SM, information on which movable members 2 are close to each other or directly adjacent can be additionally obtained.

[0066] As a supplementary solution, a plausibility check can also be implemented, especially in cases where the movable members 2 are close to each other or directly adjacent. If, from the relative position x R (or the absolute position x A ) determined for two adjacent movable members 2, it is found that these movable members 2 will overlap, then the error in the position determination is obvious. Since the geometry of the movable member 2 is known, such a plausibility check can be easily implemented. If multiple sets of position sensors Sn j are used for position determination, then, for example, the known dimensions of the movable member 2 can be used to select a set for position determination. For example, the known dimensions of the movable member 2 can be used to select a cost function J (which does not necessarily have a minimum value), and this cost function makes the result of the position determination credible.

[0067] In the case of the long stator linear motor 10, there are usually stator structures in which two stationary members 1 of the long stator linear motor 10 meet on a splitter W, or one stationary member 1 is divided into two stationary members 1 on a splitter W (depending on the movement direction of the movable member 2), as shown in Figure 9 . In the region of the splitter W, the movable member 2 can thus be guided on one of the two participating stationary members 1. With the method described, positioning can even be implemented in the region of the splitter W.

[0068] It can be expected that a sensor model SM of the movable member 2 can be more consistent with the measured sensor response SA on one stationary member 1 of the splitter W than on the other stationary member 1 of the splitter W, because the air gap between the position sensor Sn and the position magnet Pk is different in the region of the splitter W (which in turn results in different magnetic fields on the position sensor Sn). In the case of applying the cost function J, the cost function for one stationary member 1 participating in the splitter W will be smaller (or larger) than the cost function for the other stationary member 1 participating in the splitter W, because the measured sensor response SA will deviate more from the sensor model response SA due to the larger air gap * . In this way, with the present position determination method, the relative position x R (or the absolute position x A). As a supplementary solution, it is also possible to determine at the same time on which of the stationary components 1 of the branching unit W the movable component 2 is currently located.

[0069] The position determination method can be implemented in a computing unit 5. For this purpose, the computing unit 5 obtains the required sensor response SA from the position sensor Sn (as shown, for example, in Figure 1 ). The computing unit 5 can be part of the control system of the motion system or can also be integrated in such a control system. The computing unit 5 is preferably microprocessor-based, such as a computer, a DSP (Digital Signal Processor), a PLC (Programmable Logic Controller), etc., and the method for position determination can be implemented as software and stored on the computing unit 5 for execution. However, the computing unit 5 can also be an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array), in which a microprocessor can also be integrated. Any other suitable hardware can also be used as the computing unit 5. The required sensor model SM for implementation can be stored in a storage unit 6 of the computing unit 5. Other data that may be required, such as the geometry of the movable component 2, the installation position of the position sensor Sn, etc., can also be stored in the storage unit 6. The computing unit 5 can also include known input / output devices and / or a data interface for data communication with other components (such as the control system of the motion system), for example, via a fieldbus, such as Ethernet POWERLINK, DeviceNet, Profibus, CAN, etc., or other bus systems, such as Ethernet, etc.

Claims

1. A method for determining the position of at least one movable member (2) relative to a stationary member (1), wherein a setting unit having a plurality of position magnets (Pk) with a pole pitch (Tp) is provided on the movable member (2), and a setting unit having a plurality of position sensors with a sensor pitch (s) is provided on the stationary member (1), and the magnetic field (M) of the position magnets (Pk) in the region of the position sensors is detected in the form of a sensor response by means of the position sensors, characterized in that: For a group of a plurality of position sensors considered for position determination in the region of a movable member (2), sensor responses are detected; a sensor model (SM) is provided, which provides a sensor model response related to the relative position (x R ) of the movable member (2) relative to the position sensor for the at least one movable member (2) and one position sensor; the sensor model (SM) is used to determine the sensor model responses of all the position sensors of the group of position sensors for different relative positions of the movable member (2) relative to a stationary member (1); the sensor model responses are compared with the sensor responses detected using the group of position sensors; and the relative position having the smallest deviation between the sensor model response from the sensor model (SM) and the detected sensor response among the different relative positions of the plurality of assumptions is used as the relative position (x R ) of the movable member (2), or the relative position (x R ) of the movable member (2) is determined therefrom, or the absolute position (x A ) of the movable member (2) relative to a stationary reference point (RP).

2. The method according to claim 1, characterized in that: The sensor model (SM) is determined as follows: The curve of the sensor model response is measured when the setting unit of the plurality of position magnets (Pk) moves along the one position sensor past the stationary member (1).

3. The method according to claim 1, characterized in that: To compare the sensor model response with the sensor response detected using the set of position sensors, a cost function (J) is formulated that evaluates the deviation between the detected sensor response and the sensor model responses of all the position sensors in the set of position sensors, and the relative position among the different relative positions of the multiple hypotheses that optimizes the value of the cost function (J) is used as the relative position (x R ) of the movable member (2), or the relative position (x R ) of the movable member (2) or the absolute position (x A ) of the movable member (2) relative to a fixed reference point (RP) is determined accordingly.

4. The method according to claim 1 or 3, characterized in that: The different relative positions of the plurality of assumptions are determined as follows: The relative position of the movable member (2) is changed by a pre-given position increment.

5. The method according to claim 3, wherein: The different relative positions of the plurality of assumptions are determined as follows: The mathematical optimization of the cost function (J) is performed according to the relative position, wherein the relative position is changed by this optimization.

6. The method according to claim 1 or 3, characterized in that: The different relative positions of the plurality of assumptions are determined as follows: The relative position of the movable member (2) is changed by at least one pole pitch (Tp).

7. The method according to claim 6, wherein: The relative position of the movable member (2) is changed by a certain number of pole pitches (Tp), and this number is at most equal to the number of position magnets (Pk) on the movable member (2).

8. The method according to any one of claims 1 to 3, characterized in that: Using at least another set of position sensors and comparing the sensor model response with the sensor response detected using the at least another set of position sensors, and determining the relative position (x R of the movable member (2) or the absolute position (x A ) of the movable member (2) relative to a fixed reference point (RP) based on the set of position sensors that produces the minimum deviation between the sensor model response from the sensor model (SM) and the detected sensor response.

9. A motion system, which comprises at least one movable member (2) movably arranged relative to a stationary member (1), wherein a setting unit having a plurality of position magnets (Pk) with a pole pitch (Tp) is arranged on the movable member (2), and a setting unit having a plurality of position sensors is arranged on the stationary member (1), these position sensors having a sensor pitch (s), and the magnetic field (M) of the position magnets (Pk) in the region of the position sensors can be detected in the form of a sensor response by means of the position sensors, wherein a calculation unit (5) is provided for determining the position of the movable member (2) relative to the stationary member (1), and the calculation unit processes the sensor responses of the position sensors, characterized in that: The calculation unit (5) detects sensor responses for a set of position sensors considered for position determination in the area of the movable member (2); a storage unit (6) is provided, in which a sensor model (SM) is stored, which provides a sensor model response related to the relative position (x R ) of the movable member (2) relative to the position sensor for the at least one movable member (2) and one position sensor; The calculation unit (5) determines the sensor model responses of the group of position sensors for different relative positions of the movable member (2) relative to the stationary member (1) by using the sensor model (SM); The calculation unit (5) compares the sensor model response with the sensor response detected by the group of position sensors; and the calculation unit (5) determines the relative position with the smallest deviation between the sensor model response from the sensor model (SM) and the detected sensor response among the different relative positions of the plurality of hypotheses as the relative position (x R ) of the movable member (2), or determines the relative position (x R ) of the movable member (2) or the absolute position (x A ) of the movable member (2) relative to the fixed reference point (RP).

Images