WO2006033165A1

WO2006033165A1 - Armature movement detection apparatus and armature position estimation apparatus for an elevator brake

Info

Publication number: WO2006033165A1
Application number: PCT/JP2004/014422
Authority: WO
Inventors: Forrai Alexandru; Takaharu Ueda; Takashi Yumura; Masanori Yasue
Original assignee: Mitsubishi Denki Kabushiki Kaisha
Priority date: 2004-09-24
Filing date: 2004-09-24
Publication date: 2006-03-30
Also published as: DE112004002963T5; CN1997578A; CN100546895C; DE112004002963B4

Abstract

In an armature movement detection apparatus for an elevator brake, an electric current caused to flow through a brake coil is detected by a current detector. A voltage applied to the brake coil is detected by a voltage detector. An abnormal voltage drop occurring in a constant voltage source for energizing an electromagnet is detected by a voltage variation detector. A movement detector detects movement of an armature relative to the electromagnet by comparing an estimated induced electromotive force and an estimated instantaneous electromagnetic power with set threshold levels and by judging whether or not the abnormal voltage drop has been detected by the voltage variation detector.

Description

DESCRIPTION

ARMATURE MOVEMENT DETECTION APPARATUS AND ARMATURE POSITION ESTIMATION APPARATUS FOR AN ELEVATOR BRAKE

Technical Field

The present invention relates to an electromagnetic brake device forcontrollingtherotationofadrivingsheaveofanelevator.

Background art

In a known electromagnetic brake for an elevator, the moving armature's position is detected using position, speed, or acceleration sensors (from now called mechanical sensors) or using a mechanical switch.

Another approach disclosed in U.S. Patent by Michael Page (Patent No. 5,241,218, Patent date: Aug. 31, 1993) relates to a circuit fordetectingthemovement ofa solenoidarmature (forexample a solenoid valve) whereby correct or incorrect operation of the armature, thereforethevalvemaybemonitoredfroma remote location.

In accordance with the disclosed patent above the momentary reduction (or otherwise) in the current flowing through the solenoid coil is utilized to monitor correct (or incorrect) operation of the solenoid.

Moreover, the U.S. Patents by Masami Nomura (Patent No. 4. ,974,703, Patentdate: Dec.4, 1990 aswellas PatentNo.4., 984, 659, Patent date: Jan. 15, 1991) relate to an elevator control apparatus, which enhances the elevator riding quality during starting and stopping operation such that:

- the elevator cage is started only after the brake has beenpulled-up avoiding the undesirable case when the elevator motor generates torque while the braking force of the brake is still acting;

- after the brake has been released the elevator motor torque is set to zero avoiding the undesirable case when the elevator motor generates torque while the braking force of the brake is applied.

The operation of the electromagnetic brake is detected by monitoring the brake coil current.

The methods proposed for armature movement detection based on momentary variation of current might be applied for elevator brake systems but has the following limitations: the current variation (current reduction during armature pull-up and current increase during armature release) might be due to voltage oscillations of the power source and not only due to armature movement, which can lead to erroneous operation; the current increasemightbeduetothebrake control apparatus applied upon brake application for impact noise reduction; the current variation (reduction or increase) depends on the constructive variant of the solenoid and is influenced by armature stroke and inductance of the solenoid, therefore precision of the method limits its application field.

In case of currently used elevator brake systems the armature stroke is usually less than 1 mm and due to the limited space (compact design) the electromagnetic actuator is highly saturated (magnetic saturation is present) .

Disclosure of the Invention

The present invention has been made to solve the above-mentioned problems, and it is an object of the invention to provide an armature movement detection apparatus for an elevator brake and an armature position estimation apparatus for an elevator brake, which can improve the precision with which the armature position is detected.

An armaturemovement detection apparatus for an elevatorbrake according to the present invention is an apparatus for detecting movement of an armature of an elevator brake, the elevator brake including: a brake rotor; a brake shoe for frictionally braking rotation of the brake rotor; a spring for urging the brake shoe to be pressed against the brake rotor; and a brake release section for releasing the brake shoe away from the brake rotor, the brake release section being provided with an electromagnet including a brake coil, and an armature that is attracted to the electromagnet against a resilient force of the spring upon energizing the electromagnet, the armaturemovement detectionapparatus including: a current detector for detecting an electric current that flows through the brake coil; a voltage detector for detecting a voltage appliedto the brake coil; a voltage variation detector for detecting an abnormal voltage drop occurring in a constant voltage source for energizing the electromagnet; and a movement detector for detecting movement of the armature relative to the electromagnet by comparing information obtained from the current detector and the voltage detectorwith set threshold levels andby judgingwhether or not the abnormal voltage drop has been detected by the voltage variation detector.

Anarmaturepositionestimationapparatus foranelevatorbrake according to the present invention is an apparatus for estimating a position of an armature of an elevator brake, the elevator brake including: a brake rotor; a brake shoe for frictionally braking rotation of the brake rotor; a spring for urging the brake shoe to be pressed against the brake rotor; and a brake release section for releasing the brake shoe away from the brake rotor, the brake release section being provided with an electromagnet including a brake coil, and an armature that is attracted to the electromagnet against a resilient force of the spring upon energizing the electromagnet, the armature position estimation apparatus including: a current detector for detecting an electric current that flows through the brake coil; a voltage detector for detecting a voltage applied to the brake coil; an armature position estimation section for estimating at least one of the position of the armature and a parameter dependent on the position of the armature based on information obtained from the current detector and the voltage detector; and a position indicator section for judging whether or not the position of the armature is normal based on an output from the armature position estimation section, at least one of a preset range of the position of the armature and a preset parameter, and information obtained from the current detector.

Brief Description of the Drawings

Fig. 1 is a schematic view showing the entire construction of a brake system of an elevator including an armature movement detection apparatus according to the present invention.

Fig.2 is agraph showingatypicalrelationshipbetweenapplied voltage and time, armature displacement and time as well as coil current and time, when an electromagnet is energized and de-energized.

Fig.3 is a graphshowingatypical relationshipbetweenapplied voltage and time, armature displacement and time as well as induced electromotive force (E.M.F.) and time, when the electromagnet is energized and de-energized.

Fig.4 is a constructional view showing one example of armature movement detection apparatus based on electromotive force (E.M.F. ) estimation and monitoring according to the present invention. Fig. 5 is an explanatory view of the operation of the armature movement detection apparatus based on electromotive force (E.M.F.) estimation and monitoring according to the present invention.

Fig.6 is an explanatory graph of the operation of the armature movement detection apparatus based on electromotive force (E.M. F. ) estimationandmonitoringuponbrake release accordingtothepresent invention.

Fig.7 is an explanatory graph of the operation of the armature movement detection apparatus based on electromotive force (E.M.F. ) estimation and monitoring upon brake application with or without armature control according to the present invention.

Fig.8 is agraph showingatypical relationshipbetweenapplied voltage and time, armature displacement and time as well as instantaneous electromagnetic power (P) and time, when the electromagnet is energized and de-energized.

Fig. 9 is a structural diagram showing one example of armature movementdetectionapparatusbasedoninstantaneous electromagnetic power (P) estimation and monitoring according to the present invention.

Fig. 10 is an explanatory, diagram of the operation of the armature movement detection apparatus based on instantaneous electromagnetic power (P) estimation and monitoring according to the present invention.

Fig.11 is an explanatorygraph of the operation ofthe armature movement detectionapparatus basedoninstantaneous electromagnetic power (P) estimation and monitoring upon brake release according to the present invention.

Fig.12 is anexplanatorygraph ofthe operation ofthe armature movement detectionapparatusbasedoninstantaneous electromagnetic power (P) estimation and monitoring upon brake application without armature control according to the present invention.

Fig. 13 is an explanatory graph (extension of the explanatory graph shown in Fig. 12) of the operation of the armature movement detection apparatus based on instantaneous electromagnetic power (P) estimation and monitoring upon brake application with armature control according to the present invention.

Fig. 14 is a explanatory view of armature current control applied during armature pull-up and armature hold.

Fig. 15 is a graph showing a typical relationship between armature (coil) current and time, armature displacement and time as well as applied voltage and time, when the electromagnet is energized (under current control) and de-energized, according to the present invention.

Fig. 16 is a graph showing a typical relationship between applied voltage and time, armature displacement and time, applied voltage derivate and time as well as induced electromotive force andtime, whentheelectromagnet is energized (undercurrent control) and de-energized. Fig. 17 is an explanatoryview of the operation of the armature movement detection apparatus, when the electromagnet is energized (under current control) according to the present invention.

Fig.18 is anexplanatorygraphofthe operationofthe armature movement detection apparatus based on applied voltage or control signal monitoring upon brake release application with armature current control according to the present invention.

Fig. 19 is a schematic view showing the entire construction of a brake system of an elevator including an armature position estimation apparatus according to the present invention, composed of an armature position estimation section and a normal and abnormal position indicator section.

Fig. 20 is a graph showing a typical relationship between applied voltage and time, armature displacement and time as well as coil current andtime, when the electromagnet is energized (during armature pull-up and armature hold) and de-energized (during armature release) .

Fig.21 is a graph showing a typical variation ofthe inductance with air-gap.

Fig. 22 is an explanatory diagram of the parameter estimation principle based on signal injection.

Fig. 23 is an explanatory diagram of current control under hysteresis control loop.

Fig. 24 is an explanatory diagram of the principle of the switching frequency estimation.

Fig. 25 is a graph showing a typical relationship between appliedvoltage andtimeaswellas coilcurrentandtime. The current is not controlled during the armature pull-up and provides a resistance estimation section. The current is under hysteresis control during armature hold and after armature release andprovides an inductivity estimation section.

Fig. 26 is an explanatory diagram of the armature position estimation section of the armature position estimation apparatus and shows armature position estimation according to the gradient method.

Fig. 27 is a block graph showing the armature position estimation according to the reference model-based switching frequency estimation method according to the present invention.

Fig. 28 is a graph showing the operation principle of a trend estimator according to the present invention.

Fig. 29 is an explanatory diagram showing a recursive implementation trend estimating section.

Fig. 30 is an explanatory algorithm in a pseudo-programming language of the normal and abnormal position indicator section according to the present invention.

Fig. 31 is an explanatory graph showing estimated inductivity for different armature positions. Best Mode for carrying out the Invention First Embodiment

Fig. 1 shows the construction of an entire brake system of an elevator. A car 1 of the elevator is hung, togetherwith a counter weight 4, by a main rope 3 wrapped around a driving sheave 2 in a well bucket: fashion.

A brake rotor (such as a brake drum or brake disc) 6 driven by a hoist motor 5 is generally installed on an axle that couples the hoist motor 5 and the driving sheave 2 with each other. The brake shoe 8 is urged into engagement with the brake rotor 6 under the action off the resilient force of a spring 7 thereby to provide a braking force due to the friction. When abrake coil 10 consisting of an electromagnet is energized using a drive circuit 9 supplied by a constant voltage source 11, an armature 12 attached to the brake shoe 8 is attracted to the brake coil 10 while overcoming theresilient forceofthe spring7. Abrakerelease sectioncontains the electromagnet including the brake coil 10 and the armature 12.

A current detector 13 and a voltage detector 14 detect the electric curxent as well as the applied voltage on the brake coil 10 (electromagnet) . A voltage variation detector 15 detects an abnormal voltage drop of the constant voltage source 11. When the voltage leveX is less or over a well-defined threshold, amonitoring signal (logical signal) denoted here by VD is set to zero (VD = 0) . In case of normal operation its value is set to 1 (VD = 1) . The armature movement detection is performed in a movement detectorandmovement indicatorunit 16 accordingtothresholdlevels specified in a threshold level setting section 17. The threshold levels settings specified in the threshold level setting section 17 are represented with THl and TH2 for the brake release period and TH3 and TH4 for the brake application period.

Fig. 2 shows a typical relationship of applied voltage (u) and time (t) (Fig. 2a), armature displacement (x) and time (t) (Fig. 2b) as well as coiIL current (i) and time (t) (Fig. 2c), when the electromagnet is energized and de-energized.

When the current is first switched on (time point Tl on a graph of Fig. 2a, and time point A on a graph of Fig. 2c) it gradually rises until the strength of the magnetic field generated by the coil becomes sufficient to pull-up the armature. At this point of time, due to the armature movement, a current (i) flowing through thecoilmomentarilydrops (pointB onthe graphof Fig.2c) . Finally, the current reaches its steady-state value during armature holding (time point T2 on the graph of Fig. 2a, point C on the graph of Fig. 2c) .

When the current is first switched off (time point T3 on the graph of Fig. 2a, point D on the graph of Fig. 2c) it gradually decreases until the force generated by the magnetic field of the coil becomes less ^"than the force of the spring and the armature is released. Atthd_s point, due to the armaturemovement, the current (i) flowing through the coil momentarily increases (point E on the graph of Fig. 2c) and finally reaches its steady-state value during armature release (time point T4 on the graph of Fig. 2a, point F on the graph of Fig. 2c) .

Armature movement detection based on induced electromotive force (E.M.F.) estimation and monitoring

Hereinafter, reference will be made to one example of armature movement detection method based on electromotive force estimation and monitoring according to the first embodiment of the present invention.

Fig. 3 is an explanatory/ view of the basic operation of the armature movement detection apparatus according to the first embodiment of present invention. Fig. 3 (a) shows a voltage given to the brake coil 10, Fig.3 (b) shows the displacement of the armature 12, and Fig. 3 (c) shows the induced electromotive force. In Fig. 3, when the brake is released, an attraction voltage is applied to the brake coil 10 at time point Tl, so that the electromagnet provided with the brake coil 10 attracts the armature 12. In the firstphase the induced electroinotive force (Fig.3 (c) ) is a constant value (theoretically zero) due to the sensor offsets and when the electromagnetic attraction force is overcoming the force generated by the spring 7, the armature 12 starts to move and the induced electromotive force is increasing. After the moving armature 12 hits the fixed armature the induced electromagnetic force starts to decrease. The armature movement ends at time point T2.

When the brake is applied the app]_ied voltage on the brake coil 10 is made zero from the attraction "voltage at time point T3, and as a consequence the brake current st arts to decrease and when the electromagnetic attraction forcebecomes smaller thanthe spring force, the armature 12 starts falling orr moving toward the brake rotor 6, and the induced electromotive force is decreasing as shown in (c) of Fig. 3. At time point T4 the armature 12 ends its falling operation, as shown in (b) of Fig. 3.

Fig.4 is a constructional view show÷Lng one example of armature movement detection apparatus based on electromotive force (E.M.F.) estimation and monitoring according to the present invention.

The induced electromotive force is estimated in an EMF estimating section 18 by measuring an applied voltage (u) and the current (i) using the voltage detector 14 and the current detector 13. The armature movement is detected by a movement detection algorithm A section 19 according to the threshold level setting section 17 and considering the signal VD provided by the voltage variation detector 15.

A movement indicator 20 signals visually (for example, a LED is turned on or off if the armature 12 moved or did not move) and/or electronically (a digital signal is sent to a supervisory unit) the armature movement. Now, reference will be made to electromotive force (E.M.F.) estimation (shown in Fig. 5) according to the first embodiment of the present invention.

The voltage equation of an electromagnetic actuator can be written as:

u = R i + dΨ/dt (1) where (u) is the applied voltage, (i) is the current, R is the coil

resistance, and Ψ is the total magnetic flux. The total magnetic flux Ψ =Ψ(i,x) is dependent on the current (i) and the armature displacement (x) .

Therefore, from the above equation we obtain:

u = R i + dΨ/dt = R i + dΨ/d± di/dt + dΨ/dx dx/dt (2) The above equation can be approximated as:

u M R i +L(i) di/dt + e (3) where e is the induced electromotive force

e = dΨ/dx dx/dt (4) and

L (i) di/dt « dΨ/d± di/dt (5)

If there is no magnetic saturation then L(i) = X = const.

According to Equation (3) we compute the induced electromotive force as:

e « u - R i - L(i) di/dt (6)

Assuming that the Laplace transform of the current signal (i) detected by the current detector 13 is represented by I(s), the section 21 performs filtering with a time constant ti. The section 21 calculates the filtered current signal represented by i_f ( and itsLaplacetransformrepresentedbyI_f(s) ) accordingtothe following equation;

I_f(s) ={l/(τ_lS+l) }I(s) (7)

The filtered and amplified electromotive force sicjnal represented by β_f (and its Laplace transform by E_f(s) ) is obtained by the following equation:

E_f(s) = K₁ (U(S) - R I_f(s) - L{s/(τ₂s+l) }I_f(s) } (8) where U (s) is the Laplace transform of the applied voltage (u) sensed by the voltage detector 14.

Equation (8) above is calculated by a differentiating section

22, a filtering section 23 (with time constant 1₂) , a brake coil resistance value 24, a coil inductance value 25, specified by an inductance adjusting section 26, and an amplifying section 27 (with gain Ki) .

The operation of the inductance adjusting section 26 will be described below. The inductance L = L(i) is obtained beforehand, and the relation between the brake coil current (i) and inductance L is tabulated. The movement detector and movement indicator umit 16 calls orpicks up the inductance L fromthis table based on filtered signal of the current detector 13, and changes the inductance L in the electromotive force estimating section 18.

Then, a filtered electromotive force signal e_f(s) 28 is used inthemovement detectionalgorithitiAsection 19 forarmaturemovement detection according to the threshold levels specified in the threshold level setting section 17, when the abnormal voltage variation is detected by the voltage variation detector 15.

As a result of variation of electromotive force, the armature movement detection algorithm (represented by Algorithm A.1) is as indicated in Fig. 6 in case of armature pull-up and in Fig. 7 in case of armature release with or without armature control

(represented by Algorithm A.2) .

Now, the operation of this embodiment will be describedbelow.

In Fig. 6 (represented by Algorithm A.I) the filtered electromotive force signal 28 represented by β_f is compared with the threshold level represented by THl specified in the threshold level setting section 17. If the signal 28 β_f is always less than the threshold level THl, that means the electromotive force did not increase and implicitly the armature did not move. Therefore, a logical signal represented by SETl, which detects the armature movement during pull up, is set o zero. SETl = 0 (9)

If the signal 28 represented by β_f becomes bigger than thethreshold level THl, and after a while becomes smaller than the threshold level TH2 specified in the threshold level setting section 17, that means the estimated electromotive force increased. The next step is to test if this is due or not to the abnormal voltage variation of the constant voltage source 11. According to the operation of the voltage variation detector 15, if VD = 0 that means abnormal voltage variation occurred and the signal SETl is set to

0.

SETl = 0 (10)

If VD = 1 that means the electromotive force variation is due tothearmaturemovementandnotduetotheabnormalvoltagevariation. Therefore, the logical signal SETl is set to 1. SETl = 1 ^' (11)

Furthermore, if the signal 28 represented by β_f becomes bigger thanthe threshold level THl anddoes not decrease underthe threshold level TH2, that means the estimated electromotive force increased due to voltage increase and not due to the armature movement. Therefore, the logical signal SETl is set to 0. SETl = 0 (12)

Therefore, armature movement upon brake release is detected by the logical signal SETl. The armature has been moved if SETl = 1 and has not moved if SETl = 0.

The armature movement detection during armature release with or without control is indicated in Fig. 7. Now, the operation of this embodiment will be described below. In Fig. 7 (represented by Algorithm A.2) the estimated electromotive force β_f is compared with a threshold level representedby TH3 specified in the threshold level setting section 17. If the signal 28 represented by e_f is always bigger than the threshold level TH3 that means the electromotive force has not been induced and implicitly the armature did not move. Therefore, a logical signal SET2, which represents the armature movement during release, is set o zero. SET2 = 0 (13)

If the signal 28 represented by e_f becomes smaller than the thresholdlevel TH3, andafterawhilebecomes biggerthanathreshold level represented by TH4 specified in the threshold level setting section 17, that means the estimated electromotive force has been decreased. The next step is to test if this is due or not to the abnormal voltage variation of the constant voltage source 11. According to the operation of the voltage variation detector 15, if VD = 0 that means abnormal voltage variation occurred and the signal SET2 is set to 0. SET2 = 0 (14)

If VD = 1 that means the electromotive force variation is due tothearmaturemovementandnotduetotheabnormalvoltagevariation. Therefore, the logical signal SET2 is set to 1. SET2 = 1 (15)

If the signal 28 represented by e_f becomes smaller than the threshold level TH3 and does not increase over the threshold level TH4, that means the estimated electromotive force changed due to voltage drop and not due to the armature movement. Therefore, the logical signal SET2 is set to 0. SET2 = 0 ( 16 )

Therefore, armature movement upon brake application is detected by the logical signal SET2. The armature has been moved if SET2 = 1 and has not moved if SET2 = 0.

As been mentioned earlier, during brake application the brake shoe hits the drum and generates undesirable noise, which can be reduced using a brake control apparatus.

In this case the presented algorithm (shown in Fig. 7) also can be applied taking into account that the applied voltage on the brake coil 10 is measured using the voltage detection section 14.

Second Embodiment

Armature movement detection based on instantaneous electromagnetic power estimation and monitoring

Hereinafter, reference will bemade to one example of armature movement detectionapparatus basedoninstantaneous electromagnetic power estimation and monitoring according to a second embodiment of the present invention.

Fig. 8 is an explanatory diagram of the basic operation of the armature movement detection apparatus according to the second embodiment of present invention. Fig. 8 (a) shows a voltage given to the brake coil 10, Fig.8 (b) shows the displacement of the armature 12, and Fig. 8(c) shows the instantaneous electromagnetic power change of the electromagnet. In Fig. 8, when the brake is released, an attraction voltage is applied.to the brake coil 10 at time point Tl, so that the electromagnetprovidedwiththebrake coil 10 attracts the armature 12. In the first phase the instantaneous power (Fig. 8 (c) ) stored into the electromagnetic field increases and when the electromagnetic attraction force is overcoming the spring 7 the armature starts to_.move and the instantaneous power drops and after a while increases again. The armature movement ends at time point T2.

When the brake is applied the applied voltage on the brake coil 10 is made zero from the attraction voltage at time point T3, and as a consequence the brake current starts to decrease and implicitly the instantaneous power (Fig. 8 (c) ) drops. When the electromagnetic attraction force becomes smaller than the spring force, the armature 12 starts falling or moving toward the brake drum, and the instantaneous power is increasing as shown in Fig. 8 (c) . At time point T4 the armature 12 ends its falling operation, as shown in Fig. 8 (b) .

Fig.9 is a structural diagramthat shows the armature movement detection apparatus. The inventor of the present invention has focused attention on the fact that the instantaneous power stored into the electromagnetic field changes when the armature starts to move.

When the armature is pulled-up, part of the energy stored into the magnetic field is transformed into kinetic energy, and thus the instantaneous power stored into the electromagnetic field will decrease.

When the armature is released, part of the moving armature's kinetic energy is transformed into magnetic energy so the instantaneous power stored into the electromagnetic field will increase.

The instantaneous power stored into the electromagnetic field of the brake coil 10 (electromagnet) supplied by the drive circuit 9 is detected by an instantaneous electromagnetic power estimation section 29, when the electric current is detected by the current detector 13. Comparing the output signal of the instantaneous electromagnetic power estimation section 29 with threshold levels (specified into the threshold level setting section 17) by the movement detection algorithms (specified into a movement detection algorithm B section 30), the armature movement is detected.

Themovement indicatorsection20 signalsvisually (forexample an LED is turned on or off if the armature moved or did not move) and/or electronically (the digital signal is sent to the supervisory unit) the armature movement.

Now, reference will be made to the instantaneous electromagnetic power estimation section 29 (shown in Fig. 10) according to the second embodiment of the present invention.

Assuming that the current which flows through the coil 10 is represented by (i) and the instantaneous power stored into the electromagnetic field of brake coil 10 is represented by P, the relationbetweeninstantaneouspower storedintotheelectromagnetic field P and current (i) is represented by the following equation; P = L(i) *i* (di/dt) (17)

The instantaneous power stored into the electromagnetic field is proportional to the product between current and first-order derivate of the current.

Now, referencewillbemade to an instantaneous power detection apparatus accordingto the secondembodiment ofthepresent invention. AssumingthattheLaplacetransformofthe current signal (i) detected by current detector 13 is represented by I (s) , a section 31 performs

filtering with a time constant τi. The section 31 calculates the filtered current signal represented by i_f (and its Laplace transform represented by I_f(s)) according to the following equation;

I_f(s) = {l/(τls+l) }I(s) (18)

The filtered and amplified instantaneous power signal represented by P_f (and its Laplace transform by P_f (s) ) is obtained by the following equation:

Pf (s)=K₂*L*{I_f (s) }{s/ (τ2s+l) }I_f(s) ={L/ (τls+1) ^Λ2} {s/ (τ2s+l) }I^Λ2(s)

(19)

Equation (19) above is calculatedby a differentiating section

32, a filtering section 33 (with time constant τ2) , a coil inductance value 34 specified by an inductance adjusting section 35, and an amplifying section. 36 (with gain K₂) .

The operation, of the inductance adjusting section 35 is similar to the inductance adjusting section 26. The inductance L = L(i) is obtained beforehand, and the relation between the brake coil current (i) and inductance L is tabulated. The movement detector and movement indicator unit 16 calls or picks up the inductance L from this table based on filtered signal of the current detector 13, andchanges the ±nductance L inthe instantaneous electromagnetic power estimating section 29. The filtered and amplified instantaneous power signal is noted by 37.

As a result of variation of instantaneous power stored into the electromagnet!c field, the armaturemovementdetectionalgorithm (represented by ALgorithm B.I) is indicated in Fig. 11 in case of armature pull-up and in Fig. 12 in case of armature release (represented by Algorithm B.2) . Moreover, when brake application (armature release) is performed under control the armature movement detection algorithm indicated in Fig. 12 is extended with the algorithm (represented by Algorithm B.3) indicated in Fig. 13.

Now, the operation of this embodiment will be describedbelow. In Fig. 11 (represented by Algorithm B.1) , a filtered instantaneous power signal 37 represented by P_f is compared with the threshold level representedbyTHl specifiedinthresholdlevel setting section 17. If the signal 37 P_f is always bigger than the threshold level THl, that means the instantaneous power did not decrease and implicitlythe armature did. not move. Therefore, the logical signal represented by SETl, which detects the armature movement during pull up, is set o zero. SETl = 0 (20)

If the signal 37 represented by P_f becomes smaller than the threshold level THl, and after a while becomes bigger than the threshold level TH2 specified in threshold level setting section 17, that means the instantaneous power decreased due to the armature movement and after the aαrmature stops starts to increase again. Obviously, instantaneous power variation can be caused by abnormal voltage variation of the constant voltage source 11. Therefore, the next step is to test the signal VD, which detects abnormal voltage variation. According to the operation of the voltage variation detector 15, if VD = 0 that means abnormal voltage variation occurred and the signal SETl is set to 0.

If VD = 1 the armatiαre moved and the logical signal SETl is set to 1. SETl = 1 (21)

If the signal 37 represented by P_f becomes smaller than the threshold level THl and does not increase over the threshold level TH2, that means the instantaneous power decreased due to voltage drop and not due to the armature movement. Therefore, the logical signal SETl is set to 0. SETl = 0 (22) Therefore, armature movement upon brake release is detected by the logical signal SETl. The armature has been moved if SETl = 1 and has not moved if SETl = 0.

The armature movement detection during armature release is indicated in Fig. 12.

Now, the operation of this embodiment will be describedbelow. In Fig.12 (representedbyAlgorithm B.2) the estimatedinstantaneous power signal P_f is compared with the threshold level represented by TH3 specified in the threshold level setting section 17. If the signal 37 represented by P_f is always smaller than threshold level TH3 that means the instantaneous power stored into the electromagnetic field is decreasiLng (transformed into heat) and implicitly the armature didnotmove . Therefore, the logical signal SET2, which represents the armatiαre movement during release, is set to zero. SET2 = 0 (23)

If the signal 37 represented by P_f becomes bigger than the threshold level TH3, and after a while becomes smaller than the

( threshold level represented by TH4 specified in the threshold level setting section 17, that means the instantaneous power stored into the electromagnetic field increased due to the armature movement, and after the armature stopped started to decrease. This scenario is true if there is no abnormal voltage variation of the constant voltage source 11. Therefore, in the next phase the signal VD which detects abnormal voltage variation is tested. According to the operation of the voltage variation detector 15, ±f VD = 0 that means abnormal voltage variation occurred and the signal SET2 is set to 0.

If VD =1, that means the instantaneous power variation is due to the armature movement, and therefore, the logical signal SET2 is set to 1. SET2 = 1 (24)

If the signal 37 represented by P_f becomes bigger than the threshold level TH3 and does not decrease under the threshold level TH4, that means the instantaneous power increased due to voltage increaseandnotduetothearmaturemovement. Therefore, thelogical signal SET2 is set to 0. SET2 = 0 (25)

However, during brake application the brake shoe hits the drum and generates undesirable noise, which can be reduced using a brake control apparatus.

Therefore, the armature movement detection algorithm upon brake application according to the present inveation (shown in Fig. 12) is extended with the algorithm shown in Fig. 13 if a brake control apparatus is used for noise reduction. This is required in order to assure correct armature movement detection even under -un-proper armature control (control system fails or does not work properly) .

In Fig. 13 (represented by Algorithm B.3) , after the armature control period ends, the signal 37 represented by Pf is detected and the value of logical signal SET2 returned by the algorithm shown in Fig. 12 is stored.

If the signal 37 represented by P_f is negative, t riat means the instantaneous power stored into the electromagnetic field is decreasing. If the logical signal SET2 detected by the algorithm shown in Fig. 12 equals 1 that means the armature has been moved. SET2 = 1 (26)

Inadditionifthe logical signal SET2 detectedbythe algorithm shown in Fig. 12 equals 0 that means the armature has not moved and the instantaneous power decreased only due to the voltage drop. Therefore, SET2 = 0 <27)

If the signal 37 represented by P_f is positive that means the instantaneous power stored into the electromagnetic field is increasing, and if the signal 37 is zero that means the armature has not been released. Therefore, SET2 = 0 <28)

Therefore, armature movement upon brake application is detected by the logical signal SET2. The armature has been moved if SET2 = 1 and has not moved if SET2 = 0. Third Embodiment

Armature movement detection based on applied voltage or control signal monitoring

There are situationswhenit is desiredto control the armature current during the armature pull-up and hold. Armature current control is usually performed according to the control scheme presentedinFig.14, whereacontrollerK(s) usuallyhasthefollowing transfer function:

K(s) = Kp+Ki/s (29) where Kp is the proportional gain and Ki is the integral gain.

A control signal denoted by U_c(s) is given by: U_c(s) = {Kp+Ki/s}Err(s) (30)

Where an error signal is the difference between the current reference i* and the measured current i: Err(s) = I* (S) - I(s) (31)

Thepowerconvertercanbeconsideredanidealoneintheworking frequency range, therefore the applied voltage u is proportional tothecontrol signalu_c. Therefore, forarmaturemovementdetection both signals can be used.

Hereinafter, reference will bemade to one example of armature movement detectionapparatus duringarmaturepull-up (with armature current under control) based on applied voltage or control signal monitoring according to the third embodiment of the present invention .

During armature release, the previously described methods (based on electromotive force or instantaneous power estimation and monitoring) are applied without change.

Fig. 15 is an explanatory view of the basic operation of the armature movement detection apparatus according to the third embodiment of present invention. Fig. 15 (a) shows the brake coil 10 with current under control, Fig. 15 (b) shows the displacement of the armature 12, and Fig. 15 (c) shows voltage given to the brake coil 10 under control - solid line, curve 1, when the armature moves and dashed line, curve 2, when the armature does not move.

The current drop due to the induced electromotive force generated by the armature movement is sensed and compensated by the control system. Fig. 16 (a) shows the voltage given to the brake coil 10, which has a crest due to the control action, generated by the armature movement.

The simplest way to detect the armature movement is to monitor the derivate of the applied voltage u or the derivate of the control signal u_c shown in Fig. 16 (c) .

Now, reference will be made to the armature movement detection apparatus based on applied voltage or control signal monitoring according to the third embodiment of the present invention - shown in Fig. 17.

Assuming that the applied voltage is detected by a voltage detector 38 or the control signal u_c is directlyused that is computed according to Equation (30) , a section 39 performs filtering with

a time constant Ti. A section 40 derivates the filtered signal, which is amplified by a section 41 (with gain Ki) .

A filtered and amplified signal noted by 42 is compared with threshold levels (specifiedinto the threshold level setting section 17) by the movement detection algorithms (specified into a movement detection algorithm C section 43) and the armature movement is detected.

The movement indicator section 20 signals visually and/or electronically the armature movement. Now, the operation of this embodiment will be described below. In Fig. 18 (represented by Algorithm C) the filtered and derivated applied voltage or control signal is compared with the threshold level represented by THl specified in the threshold level setting section 17. If the signal 42 is always less than the threshold level THl, thatmeans the applied voltage or control signal has not been increased by the current controller so no current drop has been detected and implicitly the armature did not move. Therefore, the logical signal represented by SETl, which detects the armature movement during pull up is set to zero. SETl = 0 (32)

If the signal 42 becomes bigger than the threshold level THl, and after a while becomes smaller than the threshold level TH2 specified in the threshold level setting section 17, that means the applied voltage or control signal has been increased by the current controller due to the detected current drop.

Obviously, current drop can be caused by abnormal voltage variation of the constant voltage source 11. Therefore, the next stepistotestthe signalVD, whichdetectsabnormalvoltagevariation. According to the operation of the voltage variation detector 15, if VD = 0 that means abnormal voltage variation occurred and the signal SETl is set to 0.

If VD = 1 the armature moved and the logical signal SETl is set to 1. SETl = 1 (33)

If the signal 42 becomes bigger than the threshold level THl and does not decrease under the threshold level TH2, that means thevoltageincreasedbutnotduetothearmaturemovement. Therefore, the logical signal-SETl is set to 0. SETl = 0 (34)

It is easy to remark that though the previous algorithm is simple, it has certain disadvantages. The derivate of the applied voltage or the control signal might be affected by noise, which can lead to limited operating range or in the worst case erroneous operation .

Therefore, another approach is also proposed, which is based on electromotive force estimation and monitoring.

During armature pull-up, the current drop due to the induced electromotive force generated by the armature movement is compensated by the controller.

Therefore, the control signal is increasedproportionallywith thecurrentdrop (theintegraltermofthecontrollercanbeneglected, the armature movement is much faster than the integral term time constant) , which can be considered proportional to the induced electromotive force (e.m.f.) .

The induced electromotive force is approximately as follows (see also Equation (6)) :

e « u - R i - L(i) di/dt (35)

The signal used for armature movement detection is the electromotive force or any magnitude proportional to it.

If the armature does not move, the induced electromotive force is approximately zero. If the armature moves, the current drop is sensed by the current controller and is compensated by increasing the control signal and implicitly the applied voltage. In this case the induced electromotive force has different values from zero (in this case positive values) as shown in Fig. 16 (d) .

The armature movement detection algorithm is performed in an identical manner as described in the first embodiment of the present invention .

Fourth Embodiment

Fig. 19 shows the construction of an entire brake system of an elevator according to the fourth embodiment of the present invention.

The armature position estimation is performed in an armature position estimation section 51 and the normal and abnormal armature position is indicated by a normal and abnormal position indicator section 52. Other constructions are equivalent to the first embodiment.

Fig. 20 shows a typical relationship of: applied voltage (u) and time (t) (Fig. 20a), armature displacement (x) and time (t) (Fig. 20b) as well as coil current (i) and time (t) (Fig. 20c), when the electromagnet is energized and de-energized.

When the current is first switched on (time point Tl on the graph of Fig. 20a, point A on the graph of Fig. 20c), it gradually rises until the strength of the magnetic field generated by the coil becomes sufficient to pull-up the armature. At this point due to the armature movement, the current (i) flowing through the coil momentarily decreases (point B on the graph of Fig. 20c) . Finally, the current reaches its steady-state value during armature pull-up (time point T2 on the graph of Fig. 20a, point C on the graph of Fig.20c) . Afterthearmaturehasbeenpulled-up, theappliedvoltage is reduced to a lower level during armature hold (between time point T2 and time point T3) in order to reduce the Ohmic losses.

When the current is first switched off (time point T3 on the graph of Fig. 20a, point D on the graph of Fig. 20c) it gradually decreases until the force generated by the magnetic field of the coil becomes less than the force of the spring and the armature is released. At this point due to the armaturemovement, the current (±) flowing through the coil momentarily increases (point E on the graph of Fig.20c) and finally reaches its steady-state value during armature release (time point T4 on the graph of Fig. 20a, point F on the graph of Fig. 20c) .

Fig. 21 shows the inductance variation with the air-gap in case of a non-saturated electromagnetic actuator. This means that if the coil' s inductance or any parameter proportional to it is estimated then armature position estimation is possible.

Fig. 22 shows the basic idea of parameter estimation, where (u) is the applied input signal also called ^λinjected signal' and (±) the measured output signal.

In order to estimate the system's parameter the input signal must be ^persistently exciting' , condition described in cited document {Ljung, Astrom} . Incaseofelectromagneticactuatorsused inelevatorbrakestheinputsignalcanbegeneratedusingahysteresis control loop shown in Fig. 23.

There are different recursive (on-line) parameter estimation techniques, which can be applied for inductance estimation and implicitly for armature position estimation.

One of the well-known recursive parameter estimation method is the recursive least squares method (RLS) , which is described in [Ljung, A.strom] .

The idea is to minimize the quadratic loss function denoted by (V(θ)) - see Equation (36) - using the least squares method.

where (θ) istheparametervector, (e) istheerrorbetweenthemeasured output (y) and estimated output (y^A) .

The parameter estimation algorithm is written in recursive form using the matrix inversion lemma cited document {Kailath, Astrom} . Although this approach can offer good precision and fast convergence, due to its numerical complexity it is not suitable in many real_-time industrial applications.

Another approach known as the gradient method cited document {Astrom} is widely applied in adaptive control and is more suitable for real-time implementation, although its precision is less than the precision of the RLS.

The basic idea is to adjust the parameters in such a way that

the loss function (V(G)) is minimized.

Tomake (V(θ) ) small, it is reasonable to change theparameters in the direction of the negative gradient of (V(θ)), that is:

dθ dV de ,,_Q.

—= -γ—=-γe— (38 dt ^r dθ ^r dθ .

where (γ) is a posit±ve constant.

The presented algorithm can be written in different forms and are known as gradient or projection algorithms cited document {Astrom} . Moreover there are other alternatives, too, such as:

dθ de . ..

-*⁼ -rγ_θ ^s^ ⁽³⁹⁾

as well as

which are called sign—sign algorithms (sign is thewell-known signum function) .

Another approach to estimate the armature position is to estimate the switching frequency of the current, which is under hysteresis control. This approach is described in cited document {Noh,Mizuno} and it is shown that the inductance is inverse proportional to the switching frequency.

The inductance estimate is achieved with a high-pass filter

(to remove the low-frequency components of the current) followed by a rectifier and a low-pass filter (to demodulate the amplitude of the signal), illustrated in Fig. 24. as a sequence of signal operations.

This approach has the following main disadvantages:

- limited precision;

- limited application when the magnetic core is saturated.

Armature position estimation based on coil' s parameters estimation - gradient method

In the fourth embodiment off the present invention, reference will bemade to one example of armature position estimation apparatus applied for elevator brakes based on coil's parameter estimation using the gradient method.

Due to the particularities of electromagnetic brakes applied in elevators:

- the electromagnetic actuator during the pull-up time is highly saturated;

- after a while the pull-up current is reduced to the holding current level.

Therefore, the following parameter estimation method is considered. First, during armature pull-up the coil's resistance is estimated, thenthisvalueisusedtoestimatethecoil' s inductance or its inverse during armature hold and after armature release, when the current is under hysteresis control.

This parameter estimation approach - in two steps: first, resistanceis estimated, theninductance is estimated- canbeapplied for any electromagnetic actuator. The only disadvantage is that a small delay is introduced in the parameter estimation.

Inthefollowings, referencewillbema.detothemodelstructure considered during parameter estimation us±ng the gradient method.

Under static conditions: armaturedoes notmove and for a given current value (in this way it is possible to consider the magnetic saturation) the following model structure is considered: u=R+Ldildt (41) where (u) is the applied voltage, (i) is the current, (R) is the coil resistance and (L) is the coil inductance.

During the armature pull-up after "the current reaches the steady state, the coil's resistance is estimated recursively according to the following equation.

where the index (k) refers to the values at the moment (t_k) and (γ_R) is a positive constant.

After the current has been reduced to the holding level, the armature current enters under hysteresis control presented in Fig. 25b, and this hysteresis control loop provides the so-called ^Λinjected signal', which is used to estimate the inductance (L) or the inverse of the inductance (G = 1/IL) .

The inverse of the inductance is estimated according to the followingformula, whichisdeducedaccordingtothegradientmethod. r

G^G_k._x+γ_a{μ_k-BiJ^~G_k__ι{jt_k-Ri^\ (43) Moreover, it is possible to prove that the above equation after approximation can be written in the following forrm:

where (R) is the estimated resistance during the pull-up period,

the index (k) refers to the values at the moment (^~tk) and (γ_G) and (Y_L) are positive constants.

Depending on the constructive variant of the electromagnetic actuator, magnetic saturation levels and signal to noise ratio, one or other estimation method can be used.

Equations (8) and (9) canprovide accurate parameter estimate, which is armature position dependent, therefore constitute the core of armature position estimation method.

The relationship x = f(L), x = f(G) or in a general case x = f (p) between the estimated parameter (p) and armature position (x) , mightbeapproximatedbyalinearfunctionorforb_igherprecision the non-linear function (f) can be stored in the memory as a look-up table.

Fig. 26 shows the armature position estimation based on the gradient method. A recursive parameter estimation section 55 provides the coil' s parameters recursive estimation (in two steps) , where the input values are:

- (ui_c) is the applied voltage measured using the voltage detector section 14 or is the reference voltage given by the drive circuit section 9 ;

- (i]_c) is the filtered current provided by a low pass filter (LPF) section 53, whose input is provided by the current detector section 13;

- (dik/dt) is the current derivate provided by a derivator (DER) section 54, whose input is provided by the current detector section 13.

Using Laplace transform representations, the low-pass filter can be implemented as:

where ki and τi are positive constants. The derivator can be implemented as:

H_DER(s)=k₂—— (46)

^DER l+τ₂s

where k.2 and τ₂ are positive constants.

The output of recursive parameter estimation section 55 is low pass filtered by a low pass filtering section 56, which provides the input for a trend estimator section 57.

For relatively large air-gaps and non-saturated or lightly saturated magnetic core the algorithm can be applied for armature position estimation for any electromagnetic actuator.

Furthermore, the precision of the estimated parameter? is enhanced by a so-called ^Λtrend estimator' section 57. For a given armature position, the estimated inductance or its inverse incase ofthe electromagneticbrakes is atime-invariant parameter, but in practice small fluctuations around the mean value can be observed, which are due to sensor noise, estimation error, etc. Inorderto increasethe estimationprecisionoftime invariant parameter, a so-called ^Λtrend estimator' is applied, after the estimated parameter reaches its mean value.

If (p) is the estimated parameter (in this case (L) or (G) ) , then the parameter model can be written as: p=mt+n (47)

Intheaboveequation (t) isthetime, (m) and (n) areparameters. In the ideal case the parameter (m) equals zero and the parameter (n) equals the estimated parameter.

Fig. 28 shows the principle of the trend estimator. The parameters (m) and (n) of the trend estimator are estimated recursively according to the gradient method, which gives: m_k=m_k__λ+γj_k(p_k-m_k__xt_k-n_k_-ύ (48) n_k=n_k__λ+γ_n(p_k-m_k__xt_k-n_k_ύ (49) where (γ_m) and ( γ_n ) are positive constants.

Fig. 29 shows the recursive trend estimator section 57 according to Equations (48) and (49) .

Using this approach, the estimated parameter becomes the (n) parameterofthetrendestimator. Moreover, theestimatedparameter (m) is usedtomonitorthetransient states oftheestimatedparameter (p) . The estimated value (n) can be considered valid (for a given armature position) if the estimated parameter (m) was in a

well-defined range ( -ε < m < ε ) in the last period of time, denoted by (Δt) . In the ideal case, time-invariant parameters (m) equals zero.

Fig. 30 shows the normal and abnormal position indicator section 52, where the related algorithm is described in pseudo-programming language.

In Fig. 30 (i) is the measured current, and (i_HτH ) and (i_RTH ) are current threshold levels during armature hold and after armature release. Moreover, (n) denotes the estimatedparameterby the trend estimator. The parameters, (pHmin) and (pHmax) defines the normal parameter range during armature hold, and (pRmin) and (pRmax) defines the normal parameter range after armature release - these parameters are defined a priori by the user.

Fig. 31 shows an example, the estimated inductivity for different armature positions.

Fifth Embodiment

Armature position estimation using reference model based switching frequency estimation

Another approach to estimate the armature position is to estimate the switching frequency of the current, which is under hysteresis control. The principle of this approach is shown in Fig. 24 .

Hereinafter, inthe fifth embodiment ofthe present invention, reference is made to armature position estimation using reference model based switching frequency estimation, when the output of the referencemodel is processed inparallel withthe real systemoutput, shown in Fig. 27.

In Fig. 27 the demodulated error signal between the switching frequency of the real signal and the switching frequency of the reference model provides the armature position dependent parameter, which is used for armature position estimation. Precision enhancement is achievedin a similarway, byusing a ^Λtrendestimator' section 57, which provides the inputs for the position indicator section 52.

This extends the application field to elevator brake systems, inwhichmagnetic core is usually saturatedandthe air-gap is usually less than 1 mm.

Using Laplace transform representation the reference model in Fig. 27 is given by:

*->⁾⁼zb; ⁽⁵⁰⁾ where L_n and R_n are the nominal parameters, corresponding to the normal operation mode, during armature hold and armature release.

Moreover, in Fig. 27 the low-pass filter (LPF) as well as the high-pass filter (HPF) can be implemented in a similar way as in Equation (45) as well as in Equation (46) . The block (ABS) in Fig. 27 means that the absolute value of the signal is taken.

InFig.27 an error signal 66 is computedas adifferencebetween a signal 64 anda signal 65. The signal 64 is obtainedafterhigh-pass filtering (ahigh-pass filtering section 58) ofthemeasuredcurrent, provided by the current detector section 13 and after rectifying, providedby a rectifying section 59. The signal 65 is obtained after taking the output of the reference model (a section 24) , whose input is provided by the drive circuit 9 or the voltage detector 14, which is high-pass filtered by a high-pass filtering section 61 and rectified by a rectifying section 62.

In Fig. 27 the error signal 66 is demodulated using a low pass filter section 63, whose output is armature position dependent, which provides the input for the trend estimator section 57.

Thearmaturepositionestimationaswellaspositionindication is performed in an identical manner as described in the fourth embodiment of the present invention.

Claims

1. An armature movement detection apparatus for an elevator brake, the elevator brake comprising: a brake rotor; a brake shoe for frictionally braking rotation of the brake rotor; a spring for urging the brake shoe to be pressed against the brake rotor; and a brake release section for releasing the brake shoe away from thebrake rotor, thebrake release sectionincludinganelectromagnet comprising a brake coil, and an armature that is attracted to the electromagnet against a resilient force ofthe springuponenergizing the electromagnet, the armature movement detection apparatus detecting movement of the armature of the elevator brake and comprising: a current detector fordetectinganelectric current that flows through the brake coil; a voltage detector for detecting a voltage appliedto the brake coil; a voltage variation detector for detecting an abnormal voltage drop occurring in a constant voltage source for energizing the electromagnet; and a movement detector for detecting movement of the armature relative to the electromagnet by comparing information obtained fromthe current detector andthe voltage detectorwith set threshold levels and by judging whether or not the abnormal voltage drop has been detected by the voltage variation detector.

2. The armature movement detection apparatus for an elevator brake according to claim 1, wherein the movement detector estimates an induced electromotive force induced into the brake coil by using the electric current detectedbythe current detector andthe voltage detectedby the voltage detector, and compares the estimated induced electromotive force with threshold levels.

3. The armature movement detection apparatus for an elevator brake according to claim 2, wherein, during pull-up of the armature: if a signal of the induced electromotive force is less than a preset threshold level THl, the movement detector judges that the induced electromotive force did not increase and the armature did not move; ifthe signal ofthe inducedelectromotive forcebecomes bigger than the threshold level THl and after a while becomes smaller than a preset threshold level TH2, and if no abnormal voltage drop has been detected by the voltage variation detector, the movement detector judges that the induced electromotive force has changed due to movement of the armature; and ifthe signal ofthe inducedelectromotive forcebecomes bigger thanthethresholdlevel THl anddoes not decreaseunderthe threshold level TH2, the movement detector judges that the induced electromotive force has changed due to a voltage increase and not due to movement of the armature.

4. The armature movement detection apparatus for an elevator brake according to claim 2, wherein, during release of the armature: if a signal of the induced electromotive force is bigger than a preset threshold level TH3, the movement detector judges that the induced electromotive force has not been induced andthe armature did not move; if the signal of the induced electromotive force becomes smaller than the threshold level TH3 and after a while becomes bigger than a preset threshold level TH4, and if no abnormal voltage drop has been detected by the voltage variation detector, the movement detector judges that the electromotive force has changed due to movement of the armature; and if the signal of the induced electromotive force becomes smaller than the threshold level TH3 and does not increase over the threshold level TH4, the movement detector judges that the electromotive force has changed due to a voltage drop and not due to movement of the armature.

5. The armature movement detection apparatus for an elevator brake according to claim 1, wherein the movement detector estimates an instantaneous electromagnetic power stored into an electromagnetic field of the electromagnet based on information from the current detector and the voltage variation detector, and compares the estimated instantaneous electromagnetic power with threshold levels.

6. The armature movement detection apparatus for an elevator brake according to claim 1, wherein when an armature current is under control during pull-up and hold of the armature, the movement detector detects an applied voltage or control signal on the brake coil, and compares the applied voltage or control signal with threshold levels.

7. An armature position estimation apparatus for an elevator brake, the elevator brake comprising: a brake rotor; a brake shoe for frictionally braking rotation of the brake rotor; a spring for urging the brake shoe to be pressed against the brake rotor; and a brake release section for releasing the brake shoe away from thebrake rotor, thebrake release sectionincludingan electromagnet comprising a brake coil, and an armature that is attracted to the electromagnet against aresilient force ofthe springuponenergizing the electromagnet, the armature position estimation apparatus estimating a position of the armature of the elevator brake and comprising: a current detector fordetectingan electric current that flows through the brake coil; a voltage detector for detecting a voltage appliedto the brake coil; anarmaturepositionestimationsectionforestimatingat least one of the position of the armature and a parameter dependent on the position of the armature based on information obtained from the current detector and the voltage detector; and a position indicator section for judging whether or not the positionofthe armature isnormalbasedonanoutput fromthe armature position estimation section, at least one of a preset range of the position of the armature and a preset parameter, and information obtained from the current detector.

8. The armaturepositionestimation apparatus for anelevator brake according to claim 7, wherein the armature position estimation section comprises a recursive parameter estimator for estimating the parameter according to a gradient method based on information obtained from the current detector and the voltage detector.

9. The armatureposition estimationapparatus foranelevator brake according to claim 7, wherein the armature position estimation sectionobtains theparameter fromademodulatederror signalbetween a switching frequency of a real signal and a switching frequency of a reference model.

10. The armature position estimation apparatus for an elevator brake according to claim 8 or 9, wherein: the armature position estimation portion further comprises a trend estimator section for judging a validity of the parameter; and the position indicator section judges whether or not the position of the armature is normal based on information provided from the trend estimator section, a preset parameter range, and information obtained from the current detector.