`Measurement and Control
`
`http://tim.sagepub.com/
`
`
`A microprocessor controlled gearbox for use in electric and hybrid-electric vehicles
`P.W. Masding, J.R. Bumby and N. Herron
` 1988 10: 177Transactions of the Institute of Measurement and Control
`
`DOI: 10.1177/014233128801000402
`
`The online version of this article can be found at:
`
` http://tim.sagepub.com/content/10/4/177
`
`Published by:
`
`http://www.sagepublications.com
`
`
`
`On behalf of:
`
`
`The Institute of Measurement and Control
`
`
`
`Additional services and information for can be found at:Transactions of the Institute of Measurement and Control
`
`
`
`
`
`
`http://tim.sagepub.com/cgi/alertsEmail Alerts:
`
`
`
`
`http://tim.sagepub.com/subscriptionsSubscriptions:
`
`Reprints:
`
`
`http://www.sagepub.com/journalsReprints.nav
`
`
`
`
`http://www.sagepub.com/journalsPermissions.navPermissions:
`
`Citations:
`
`
`http://tim.sagepub.com/content/10/4/177.refs.html
`
`>>
`
`Version of Record
`
`- Sep 1, 1988
`
`What is This?
`
`
`
`Page 1 of 11
`
`Downloaded from
`
`
`
` at WAYNE STATE UNIVERSITY on May 23, 2012tim.sagepub.com
`
`FORD 1228
`
`
`
`
`A microprocessor controlled
`gearbox for use in electric and
`hybrid-electric vehicles
`
`by P. W. Masding. BSc, J. R. Bumby, BSc, PhD, CEng, MIEE and N. Herron
`
`This paper describes the automation and control of a
`‘manual’ synchromesh gearbox in a form suitable for use in
`an electric or hybrid-electric vehicle. Such a computer
`controlled transmission system allows full integration and
`control of the drive train leading to improved vehicle
`performance. The automation process is described, and a
`full description given, of the computer control algorithms
`necessary to ensure smooth and reliable operation of the
`transmission system. It is shown that an upechange can be
`achieved in 1.4 seconds and a down-change in 1.2 seconds.
`These times are shown to represent maximum gear change
`times and details of how they can be reduced substantially
`are discussed.
`
`Keywords: Microprocessor control, vehicle gearbox auto-
`mation. computer control algorithms
`
`1 . Introduction
`
`Automating the operation of a vehicle transmission
`allows the control of the engine and transmission system
`to be integrated. giving substantial benefits in terms of
`vehicle performance, energy efficiency and driveability.
`Although such a statement
`is applicable to internal
`combustion lie) engine vehicles. electric vehicles and
`hybrid-electric vehicles. the details relating to how the
`engine/transmission should be controlled are quite differ—
`ent. The main thrust of this paper is to consider the auto-
`mation and control of a discrete ratio. synchromesh
`transmission for use in an electric or a hybrid-electric
`vehicle. As a hybrid-electric vehicle includes both an
`electric traction motor and an ic engine in its drive system
`it is relevant to outline briefly the benefits to be gained by
`automating the transmission system in both an ic engine
`and an electric vehicle.
`
`Fig 1 shows a typical efficiency map for a 50 kW ic
`engine. Also shown on this diagram is a line corre—
`sponding to the road load seen by the engine when
`operating in a fixed gear. it is only at high loads that the
`engine operates at all efficiently. At low loads the oper-
`ating point is well removed from the high-efficiency (low~
`specific-fuel-consumption) area. At a road load of 10 kW,
`the engine operates at about 3000 rev/min and is rela-
`tively inefficient. By reducing engine speed relative to the
`vehicle speed. through a suitable change in gear ratio, the
`engine operating point can be moved up. along the
`constant power line, towards the high-efficiency region.
`As the operating point moves up this constant power line
`
`School of Engineering and Applied Science, University of Durham.
`South Road, Durham, England 0H1 3LE.
`
`Trans Inst MC Vol 10, No 4, July—September 1988
`
`it would. ultimately, reach the optimum engine operating
`line,
`the locus of which links the maximum engine-
`efficiency points at each speed.
`Such an optimum engine-operating schedule can be
`followed if a continuous variable transmission {CVT ) is
`used (Stubbs, 1981; Stubbs and 1r0nside. 1981: Steig and
`Spencer Worley, 1982; Srinivasan er al. 1982; and Chan
`et a], 1984). Unfortunately, current CVT's have an effi-
`ciency which is lower than comparable discrete ratio
`units. and there is a danger that any efficiency benefits
`accruing from improved engine utilisation can be lost in
`the transmission itself. With this in mind Thring (1981)
`has suggested that a discrete ratio transmission with a
`number of ratios spanning a relatively large range may be
`a more efficient solution. even though the engine does not
`quite follow the optimum engine schedule. With this in
`mind. a number of Workers have investigated the micro-
`processor control of automatic transmissions (Trummel
`and Burke, 1983; Richardson et al. 1983), while others
`seeking greater efficiency and lower costs have investi-
`gated the automation of spur gear transmissions {Busca
`etal. 1979; Main et a1, 1987),
`Whereas the ic engine vehicle meets a variety of
`operating requirements, the electric vehicle is generally
`designed for an urban or sub-urban operating environ—
`ment where, for the advanced electric vehicle,
`traffic
`compatibility over a speed range of typically 0— 100 km/h
`may be required. Such a requirement can be met by an
`electric-traction system with a single gear
`ratio of
`between 521 and 6:1,
`thereby simplifying the trans-
`mission system and reducing weight and cost. Generally,
`such vehicles employ power electronics to give con—
`tinuous control of the motor torque over the full speed
`range of the motor.
`A typical torque/speed characteristic for a DC separ-
`ately excited traction motor is shown in Fig 2 and should
`be compared with the similar diagram, Fig i, for the ic
`engine. The difference in the diagrams, and in particular
`the area of maximum efficiency, is immediately apparent.
`The area of maximum efficiency now tends to be at rela—
`tively high speed, low torque, in contrast to the low-speed
`high-torque requirement of the ic engine. This demands a
`different gear change logic if motor efficiency is to be
`maximiscd.
`
`To operate the motor over its complete speed range,
`two distinct operating schemes are necessary. At about
`2000 rev/min the torque is seen to fall with speed. This
`transition point is termed the break speed‘ and is the
`speed at which the motor will operate with full field
`current and full armature voltage. Below the break speed
`it is necessary to control the armature voltage at full field,
`177
`
`Page 2 of 11
`
`FORD 1228
`
`FORD 1228
`
`
`
`Masdirzg. Bumby and “error:
`
`,0 kW
`
`ZOkW
`
`30 kW
`
`40 kW
`
`Moxirmm engine torque
`
`'p
`
`\
`
`at
`8
`6°
`g
`3 v (-
`c
`E
`
`1.0
`
`20 »
`
`Optimum engine operating
`schema
`
`sex
`‘
`27'“
`
`/
`
`Level road load.
`fixed gear
`
`237.
`
`wr<
`
`0
`
`1000
`
`2000
`
`3000
`
`4000
`
`5000
`
`6000
`
`Fig 1 Road-load and engine-operating curves for an ic engine vehicle
`
`Engine speed. rpm
`
`Skw MW 20M 30M
`
`Laval road load
`
`
`
`
`
`6)
`
`
`
`Motortorque,Nm
`
`20
`
`1.0
`
`60
`
`80
`
`100
`
`I
`0
`
`Motor speed
`rpm
`
`Road $P¢¢d kmlh
`(9:5.65rl)
`
`Fi92 Road~load and motor-operating curves for
`electric vehicle
`
`all—
`
`while speeds above the break speed are achieved by
`reducing the field current but at full armature voltage: 50-
`called field weakening.
`Fig 2 shows that the area of maximum efficiency tends
`to be at speeds just above the break speed in the field
`weakening region. To capitalise on this a number of
`electric vehicle designs have incorporated a transmission
`system with a number of discrete ratios (Bindin, 1986;
`Burba er a], 1986). The use of such a transmission allows
`the motor break speed to appear at a relatively low
`road-wheel speed. therefore extending the range of road
`speeds over which field weakening control can be used,
`By arranging a battery switching system which halves the
`applied armature voltage,
`this range can be exrended
`further. down to below 10 km/h. Armature control
`
`178
`
`ItN
`
`
`
`Tractiveeffort
`
`20
`
`LO
`
`5301001201th0
`so
`Rood speedkmlh
`
`Tractive—effort and road-toad curves fer all-electric
`Fig 3
`vehicle showing the effect of gear ratio
`
`electronics can now be dispensed with. Vehicle movement
`from rest is achieved by the use of starting resistors in the
`armature circuit.
`
`Apart from a potentially improved drive system effi—
`ciency.
`the use of a stepped transmission can also
`improve the performance of the electric vehicle. This will
`become an even greater requirement in the future when
`electric vehicle range is extended by the use of high-
`energy-density battery systems. such as the sodium!
`
`Trans lnst MC Vol 10, No 4, July-September 1988
`
`Page 3 of 11
`
`FORD 1228
`
`FORD 1228
`
`
`
`(Bindin, 1986). Fig 3 shows the motor
`sulphur cell
`maximum torque profile of Fig 2 converted to an equiva-
`lent traction force at the vehicle wheels for three different
`transmission ratios. Also shown is a typical road load
`requirement at different gradients. With a fixed ratio of
`5.65:1,
`the vehicle can cover the speed range 0—100
`km/h but is incapable of starting on gradients greater than
`15% and travelling at speeds much above 100 km/h. A
`typical performance specification for an ic engine vehicle
`is a start from rest on a 20% gradient and the ability to
`sustain 120 km/h on a 2% gradient Obviously, the elec-
`tric vehicle with fixed gear ratio is a compromise and, for
`a more general-purpose vehicle, it has difficulty meeting
`either requirement. By adding two more gear ratios, both
`low- and high-speed performance is improved, while with
`an integrated motor/transmission control system, drive
`system efficiency can also be maximised.
`As the hybrid ic engine/battery-electric vehicle can
`operate in an ic engine mode, an electric mode or with
`both power sources together providing the propulsion
`power, integrated control of the engine. traction motor
`and transmission is of prime importance (Trummel and
`Burke, 1983: Bumby and Forster. 1987; Forster and
`Bumby, 1988). With these benefits in mind, this paper
`examines the feasibility of automating and controlling a
`four-speed discrete—ratio transmission in a form suitable
`for use in an electric or hybrid-electric vehicle.
`
`2. The hybrid ic engine/battery-electric test
`system
`
`2.1 The test facility
`
`The arrangement of the laboratory test system in which
`the automated transmission is installed is shown in Fig 4
`and is used to examine the control and driveability prob-
`lems relating to a parallel hybrid drive (Fig 5). in this
`drive arrangement a 32 kW ic engine, and a 37 kW separ-
`
`Mas-ding. Bumby and Hemm
`
`ately excited DC traction motor. are connected mech-
`anically in parallel at
`the input
`to the
`standard,
`four-speed, synchromesh transmission by a 1:1 toothed
`belt drive. Both the ic engine and the electric traction
`motor power electronics have been modified for com—
`puter control. In the case of the ic engine, the main butter-
`fly valve is controlled by a stepper motor and is capable
`of moving this valve from a fully closed to fully open
`position in 425 ms. The control system for
`the DC
`motor incorporates both an armature and field chopper
`so that smooth control of the motor over its complete
`speed range, in both motoring and regenerative braking
`modes, is possible.
`Incorporated between the ic engine and the belt drive
`is a one-way clutch, or frccwheel, which allows the drive
`system to be Operated in a pure electric mode with the ic
`engine remaining stationary. When necessary the engine
`can be started, run up to speed and synchronised with the
`rest of the drive train ready to pick up load. All drive
`train control is carried out by a Motorola Mo8000-based
`computer system. This computer system also gathers test
`data, stores it in memory, and, at the end of a test, passes
`this information to the test-bed control computer for
`analysis and graph plotting. Data are also passed during a
`test to generate on-line displays. A fuller description of
`the test facility and its computer control can be found in
`Bumby and Masding (1988}.
`
`2.2 Requirements of an automated transmission
`
`In examining the feasibility of a computer controlled
`gearbox it is important to recognise that if good drive-
`ability is to be achieved without a ‘hot shift', then gear
`change times must be as short as possible. in addition
`when a gear change is completed, drive torque should be
`reapplied quickly and smoothly.
`With the drive arrangement described in section 2.1.
`the permanent connection of the electric motor to the
`
`d.,c. traction motor
`
`Power electronics
`
`‘T' slotted bed plate
`
`Test bod control computer
`DUST-to (RC. )
`
`Satoty trbe.
`lnamunontatlon It
`test-bod control
`
`Toothed bolt drlve
`
`Transmission
`
`"-Q'I-fin'u-nl-Et- n92.
`
`!
`
`.I'O- rllzl-ln-ulll‘l'n lizl l—lll
`> I
`it
`
`,
`
`armory
`
`M68000
`
`Drive train control a
`Data acquisition
`"
`
`\
`
`Torque transducers
`
`Drlva shalt
`
`Free-whee! unlt
`
`Lc. engine
`
`Fig 4 Test-bed layout
`
`Trans Inst MC Vol 10, No 4, July—September 1988
`
`Page 4 of 11
`
`‘6
` Driver
`
`,
`
`i
`
`.
`
`Flywheel
`
`Dynamomoter
`
`Dynamomoter
`cone-o1 unit
`
`179
`
`FORD 1228
`
`FORD 1228
`
`
`
`Masdt‘ng, Bumby and Herron
`
`free-wheel
`
`Lc. englno
`
`Accelerator
`
`Brake
`
`Motor
`
`
`Controller
`
`
`
`
`
`
`Battery
`
`Fig 5
`
`Parallel hybrid configuration
`
`shaft provides a means of precisely
`gearbox input
`controlling the speed of this shaft. This allows a gear
`change to' be completed without
`the need for a slip-
`ping clutch. In addition, the electric traction system is
`used to move the vehicle away from rest, with the ic
`engine being automatically started and synchronised with
`the moving drive train at speeds above 1000 rev/min. a
`control sequence again not requiring the use of a slipping
`clutch.
`
`l
`
`.
`
`-.
`009an
`53an
`
`Forward
`
`Centre!
`lorward
`Centre!
`back
`
`Back
`
`Longitudinal pneumatic
`cylinder
`
`,Pisbn rod
`
`Fixed base
`
`Bearing wheel
`
`3. Transmission-system hardware
`
`The function of the gear-change mechanism is to move
`the gear lever to any point on the H gate in a similar
`manner to a human operator. As the laboratory test
`system is used extensively to study control problems in
`hybrid electric drives it is advantageous to retain manual
`movement of the transmission lever. Such freedom of
`movement
`is provided by two pneumatic cylinders
`attached to the gear selector as shown in Fig 6. Activating
`the
`longitudinal
`cylinder causes
`a plate
`to move
`backwards and forwards. thus shifting the gear change
`lever between l/3 and 2/4 ends of the H gate. This plate
`is mounted between two sets of bearings rotating in the
`horizontal plane. Each bearing wheel has a grooved edge
`which locates with a bevel on the edge of the plate.
`A second cylinder is mounted on the plate and moves
`with it. This cylinder provides the necessary sideways
`movement of the gear selector, between 1/2 and 3/4 sides
`of the H gate.
`The circuit diagram for the pneumatic system is shown
`in Fig 7. Compressed air is supplied to the system from a
`reservoir which is recharged as necessary by an electric
`pump. During normal operation a regulator valve main—
`tains the working pressure in the system at about 2 bar.
`Each of the two working cylinders is controlled by a
`piston valve. There are five ports on each piston valve.
`180
`
`
`
`
`
`Transverse poslion sensors
`MR'
`Itghl
`
`Lubricated
`sliders
`
`_
`
`_
`
`“““
`
`3
`:
`'
`I
`
`H 9019 /i 1'.
`
`Tmrsversz pmimotic
`cylinder
`
`.4
`Fig 8 The gear-chance mechanism
`
`the flow path between them being controlled by activating
`the appropriate solenoid.
`In the case of the longitudinal cylinder. two additional
`components are used to help stop the piston in the
`central. neutral, position. The first is a fast-acting valve
`which cuts the air supply to the cylinder 12 ms after its
`solenoid has been energised. Second. flow regulators are
`fitted to both supply lines feeding the cylinder. These
`allow air to flow freely into the cylinder but, when air is
`
`Trans inst MC Vol 10, No 4. July-September 1988
`
`Page 5 of 11
`
`FORD 1228
`
`FORD 1228
`
`
`
` Inlet pressure
`
`
`regulator
`
`Minding, Bumby and HerI‘OIl
`
`Solenoid
`(isolates port1>\
`10
`
`
`
`Piston valve
`
`
` Transverse cylinder
`
`Flow regulator __..____,
`Restricted
`tlow
`
`Longitudinal
`cylinder
`
` Port numbering system
`
`Valve ports:
`
`1aMein inlet port
`2I4=Outlat ports
`3I5=Exhcust ports
`11. Connects ports land I. when on
`12 Connects ports land 2 when on
`
`Fig 7 The pneumatic circuit
`
`returning to the exhaust ports, the flow is restricted. As a
`result, the speed of the piston's movement is decreased.
`which reduces the tendency of the piston to overshoot the
`neutral position before the air-cut-off valve can be
`activated.
`
`Both cylinders have magnetic pistons which operate
`sensor switches attached to the outside of the casing {see
`Fig 6). A system of two switches on the transverse
`cylinder and four on the longitudinal cylinder completely
`define the position of the gear selector. The area of the H
`gate covered by each position sensor is shown diagram-
`matically in Fig 8. Each of the position sensors is inter-
`faced to the MoSOOO through appropriate filter and
`isolation circuitry.
`The six sensors alone are sufficient to confirm when a
`
`particular gear is engaged and when the gearbox is in
`neutral. Logical OR of the two centre sensors on the
`longitudinal cylinder gives a single signal which covers the
`whole of the central, neutral, region. However, ambiguous
`positions
`exist;
`for
`example,
`the reading obtained
`between first and neutral is the same as that between
`
`neutral and second. This ambiguity is removed by con-
`necting the centre two sensors on the longitudinal cylin-
`der to a positive edge triggered D-type flip-flop decoder
`Circuit.
`
`4. Software controi
`
`The gear change control software has been designed to
`be readily incorporated into the existing hybrid drive
`control software. The structure adopted for this control
`software, described in Bumby and Masding ( 1988), is that
`of a background program concerned in receiving. and
`transmitting, data to the operator and a level-3 (lowest
`level used) interrupt routine, activated every 20 ms,
`in
`which all time-sensitive control algorithms run.
`in the
`gear change software, gear change requests from the
`
`Trans Inst MC Vol 10, No 4, July—September 1988
`
`0
`
`9
`
`Transverse cylinder
`sensors
`
`Left -——9
`
`Right
`
`6
`
`0
`
`Longitudinal
`
`cylimr sensors forward
`
`Ccntrcl/back
`
`Flip -tlop
`circuit
`
`Logical OR
`of centre
`sensors
`
`Decoder circuit
`Signals
`
`Fig 8 Regions covered by the position sensors
`
`operator can be made in the background routine. but
`during automatic operation the decision as to when to
`change gear, and what gear to change to. is made in the
`r'level-3
`interrupt. The main gear
`change
`control
`algorithms also run at
`this interrupt level. All control
`181
`
`Page 6 of 11
`
`FORD 1228
`
`FORD 1228
`
`
`
`Masding, Bumby and Hermn
`
` “than approprkle sdmold: to mov- to "navel
`
`
`Fur-v4 or 30¢ mob; an pecan! gear
`pins
`
`Lofl 9' rith depudhg a nut gear
`
`
`(hang. up: Front motor cmtockoru ta «cunt:
`
`Change down: Pres“ motor :enlactor: to hrcke
`( Roman: ”when loading tram tho Mozor )
`
`
`
`Englns PM Spted Control
`Uamm: Snack“: Flywhcd-gzwutlu—Ol rpm
`
`Downlhiit: Sczpcflma molanud—Ba rpm
`( Rmoves gearbox loading 1mm lb- Engino ]
`
`
`
`
`Chuck Scam for correct nculrd redoing
`M 10 con-ct rcudhg: on obtain“ mm
`lum air on and "t nautrd lug—TRUE
`
`dqufihm omy
`
`Syn strum" new Spud
`DOWNDMNGE
`hiflofiy apaly broke: mil
`small overspua m
`Men great contacts" to
`water“. my!“ error
`I: zero.
`men uso PH sac-Denna nlgorlthrn
`UPOHANCE
`Us. PM «ad-rotor
`
`i1
`
`maid. tourma-
`Mmln Yuma
`
`So! lynch lag
`ITRUE
`
`
`
`No
`
`Engoqo
`mm
`mud.
`
`Yo:
`
`5.! vow"
`to tune: new gun
`
`for the again:
`Chuck sun-or:
`:11on roofing ooprepduu
`
`
`far me hey gear
`I! 10 some: reading: are
`
`
`catch-l "in not «pogo-TRUE
`
`
`
`
`Vdm cf!
`momma. varieties
`
`
`
`
`RETURN
`Ruth. "hum WE
`
`
`an! m. goat a...” x. mm
`
`
`
`
`Fig 9 Gear-change algorithm
`flowchart
`
`182
`
`Trans Inst MC Vol 10, No 4, July-September 1988
`
`Page 7 of 11
`
`FORD 1228
`
`FORD 1228
`
`
`
`algorithms are written in ‘C’ with small assembler
`routines being used to interface with input/output
`devices.
`
`The flowchart of the ‘C' interrupt routine which pro-
`vides the gear change logic is shown in Fig 9.
`
`4.1 Stage 1 - Gear change
`
`At the start of any gear change, the loading on the gear‘
`box input shaft is removed and the gear selector moved
`into neutral. On its first pass,
`the control software
`activates the necessary solenoids to move the gear selec-
`tor out of gear and into neutral by sending a bit pattern to
`the output register of the appropriate versatile interface
`adaptor (VlA). A two-dimensional array provides the
`program with all the bit patterns necessary, based on the
`current and next gear. These codes activate the longi-
`tudinal cylinder as appropriate to move out of the current
`gear and, at
`the same time, activate the transverse
`cylinder to move left or right, ready for the next gear.
`When the gear selector moves into the region covered by
`the centre sensor - ie, the neutral zone. a high-level inter-
`rupt signal
`(level 4)
`is generated to ensure prompt
`removal of the air supply to the longitudinal valve. Con-
`trol now passes back to the level-3 interrupt routine
`which waits until the correct neutral position has been
`reached (ie, left or right centre on the 'H' gate) before
`cutting off the air supply to the transverse cylinder.
`Checking the position consists of reading the VIA input
`register ten times and only accepting that the gear selector
`is in neutral if all ten give the same correct reading.
`immediately after setting the appropriate V [A output
`to activate the cylinders, the loading on the gearbox is
`removed. Unfortunately,
`if the ic engine is producing
`power,
`it is not simply a matter of setting the throttle
`demand to zero and closing the butterfly valve. Compli-
`cations in reducing the engine loading arise from the fact
`that the engine must pick up load smoothly once a gear
`change is completed. if the engine throttle is simply set to
`zero, its speed will fall rapidly behind the rest of the drive
`train during the gear change. Consequently, when the
`throttle is returned to its previous setting, the engine will
`accelerate rapidly until it ‘hits’ the rest of the drive train,
`causing an unacceptable jolt. To overcome this problem
`either a controlled synchronisation procedure can be
`used or. as is the case here, closed-loop speed control of
`the engine is provided to maintain the engine speed about
`88 rev/min behind the motor during the gear change.
`With such a small speed difference the engine does not
`'hit‘ the drive train with any significant force when it
`comes back on load, with the result that drive train power
`can be restored quickly and smoothly.
`Setpoints for the engine speed controller depend on
`whether the gear change is up or down. With an up—
`change, because of the presence of the one-way clutch,
`the engine can immediately start
`to slow towards the
`speed it will need when the new gear is engaged.
`in
`contrast it
`is not physically possible to accelerate the
`engine to the speed that will be necessary after a down-
`change until the gearbox is in neutral. Two setpoints are
`used to overcome this problem. For a down-change the
`engine follows 88 rev/min behind the motor as it acceler—
`ates, and for an up-change the engine is sent immediately
`towards the speed for the next gear.
`When the electric motor has been supplying the power
`particular attention must be paid to the accelerator and
`
`Trans Inst MC Vol 10, No 4, July—September 1988
`
`illrtsdtrig, Bumby and Herron
`
`is simply
`brake signals in any situation in which it
`required to rotate the motor without loading the drive
`shaft. If both brake and accelerator signals are set to zero,
`then the logic within the power electronic controller is
`undefined. Application of a small signal to either over-
`comes this problem and the controller switches positively
`into one or other mode. Accordingly, the gear-change
`routine applies a small signal
`to the accelerator
`in
`preparation for a down-change and a brake signal for an
`upochange.
`
`4.2 Stage 2 - Speed matching
`
`the speed of the
`Once neutral has been obtained.
`gearbox input shaft is synchronised to that of the output
`by sensing the speed of the flywheel (road-wheel speed)
`and controlling the speed of the electric motor to a set-
`point based on the flywheel speed times the new gear
`ratio. The action of
`the synchronisation algorithm
`depends on the size and sign of the error so that the time
`taken to match speeds is minimised. lntuitively mixed use
`of the motoring and regenerative braking modes, depend—
`ing on the sign of the demand from, say, a P+I algorithm,
`would seem an obvious form of control. However,
`because of the delays that arise due to the way in which
`the power electronics switch between motoring and
`regenerative braking, this is not possible. These delays.
`caused by contactors closing and the field current
`reversing. would introduce an unacceptable 500 ms
`delay. To avoid this problem. and to achieve synchronisa-
`tion in about 700 ms,
`the controller takes a different
`course of action depending on whether an up~change or a
`down-change is requested. On a down‘change, having
`preset the contactors to accelerate prior to making the
`neutral move, speed control
`is achieved by a P+l
`algorithm using the accelerator alone. if the system over-
`shoots. natural slowing of the motor is adequate without
`recourse to regenerative braking. On an up-change the
`system has been primed for braking and a constant brake
`value is applied for as long as the overspeed error is
`greater
`than 200 rev/min. Below this error natural
`slowing of the motor is used to achieve synchronisation.
`Simultaneously, the controller outputs a small accelerator
`signal to switch the power electronic controller into the
`motoring mode ready for subsequent use. Once speed
`matching to within 100 rev/min has been achieved. a P+I
`control algorithm takes over to maintain this speed until
`the gear is successfully engaged.
`
`4.3 Stage 3 — Engaging the new gear
`
`When the speed error is less than 100 rev/min. the
`longitudinal solenoid is activated to engage the new gear.
`Any residual speed errors are partly corrected by the
`synchromesh in the gearbox and partly by the control
`aigorithm described above. Once the engage move has
`been made. the program waits for the sensors to confirm
`that the position for the new gear has been reached. The
`program now returns ‘false' to the calling routine. signify-
`ing that the gear change has been successfully completed.
`thereby allowing power to be restored on the drive shaft.
`
`4.4 Location on power up
`
`On initial power up. a gear change initialisation routine
`is activated that locates the gear selector in the neutral left
`position ready for the first gear change. This is done
`183
`
`Page 8 of 11
`
`FORD 1228
`
`FORD 1228
`
`
`
`Masding, Bumby and Herron
`
`without use of the decoder circuit since the output from
`the flip—flop might be erroneous until the neutral position
`is crossed for the first
`time. A simple series of logical
`movements based on the sensor ~readings make up the
`structure of the location routine.
`
`5. Results
`
`Typical results for both a down-change and an up
`change are shown in Figs 10a and 10b, respectively. for
`the drive operating in a hybrid mode. Before the gear
`change, the ic engine is providing all the traction power. A
`controlled gear change then takes place, with power being
`quickly and smoothly re-applied by the engine once the
`gear change is complete. Figs 11a and 11b show the
`speed synchronisation phase in more detail and also show
`how the motor-speed setpoint varies during the gear
`change. In Fig 10, the response of the torque transducer
`monitoring the ic engine torque is also recorded as this
`
`gives some indication as to the smoothness of the gear
`change.
`speed
`the gear change the flywheel
`Throughout
`(vehicle road speed) remains substantially constant. Once
`neutral has been reached. speed synchronisation takes
`place using the electric traction motor to control the gear-
`box input shaft speed. The motor setpoint is now defined
`and depends on the gear being selected. At all times it is
`related to the flywheel speed through the appropriate
`gear ratio. Clearly shown in Fig 10 is the way in which the
`engine speed is immediately allowed to ‘drop away’ in an
`up-change towards the final engine speed while during a
`down—change it is tightly linked to motor speed as soon as
`neutral has been achieved.
`
`6. Discussion
`
`the gear change system as
`The requirements of
`described in section 2 were to minimise gear change time
`
`
`
`Electric motor speed
`
`......... ic engine speed
`
`.
`
`ic engine torque
`
`Fig 10a Torque and speed profiles
`during down-change from third to
`second
`
`
`
`ic engine speed
`
`......... electric motor speed
`
`.
`
`ic engine torque
`
`electric motor speed
`.
`.
`LC. engine speed
`
`LC. engine torque
`
`3000
`
`2500
`
`, 2000
`E.D.
`
`‘-; 1500
`‘U
`
`8 8
`
`. 1000
`
`500
`
`
`
`
`0.5
`
`1.0
`
`1.5
`
`20
`
`2.5
`
`3.0
`
`Time (secs)
`
`electric motor speed
`
`to engine torque
`.
`,“._.-'
`\_I
`
`a
`
`'
`
`r‘
`\,1 ._.
`
`‘
`
`i.c. engine speed
`
`0.5
`
`to
`
`1.5
`
`2.0
`
`2.5
`
`3.0
`
`3.5
`
`Time (secs)
`
`Fig 10b Torque and speed profiles
`during an tip-change from second
`to third
`
`184
`
`Trans inst MC Vol 10, No 4, July-September 1988
`
`Page 9 of 11
`
`FORD 1228
`
`FORD 1228
`
`
`
`Muscling, Bumby and Herron
`
` elmiric motor set point
`
`electric motor speed
`
`
`
` re- connected
`speed synchronisation
`
`
`.20
`
`.40
`
`.60
`
`.80
`
`1.0
`
`1.2
`
`1.4
`
`1.6
`
`1.8
`
`Time (secs)
`Fig 11a Diagram showing 'stage' timings during a down<change from third to second
`
`electric motor speed
`
`/
`
`motortbroking
`./
`
`-
`eiectrtc motor
`set point We!
`
`.___/.. motor :noturol decoy
`
`"‘ ' "‘ .'.v-.: ~.--~.......-
`,.-I'
`\i.c.enghe speed
`
`\ flywheel (vehicle)
`speed
`
`speed synchronisation
`
`engage move
`
`Le. engine
`roe-connected
`
`.20
`
`1.0
`
`.60
`
`.80
`
`1.0
`
`l.2
`
`1.4
`
`1.6
`
`1.8
`
`Fig 11b Diagram showing ’stage' timings during an up—change from second to third
`
`Time (secs)
`
`and, when drive power is restored, to do this quickly and
`smoothly to give good driveability. With the gear change
`system described here the time taken to gear engagement
`for an tip-change is, typically, 1.4 s and 1.2 s for a dorm-
`change. Smaller adiustments in speed are necessary if the
`gear change is between sequential gears. and typically this
`reduces the total gear change time by 200 ms.
`The way in which the gear change time is distributed
`between the three distinct stages of a gear change is
`shown in Fig ll. Typically 300*400 ms are used engag-
`ing neutral. If the H gate does not have to be crossed, the
`lower time is appropriate. Speed matching then takes
`place in, typically, 600—800 ms. During a down-change,
`when an increase in motor speed is demanded, the PM
`controller has been designed so as to give a good, fast
`
`Trans inst MC Vol 10, No 4, July—September 1988
`
`is
`response with little or no overshoot. This control
`clearly demonstrated in Fig 11b. The third stage of the
`gear change, that of gear engagement, takes, typically.
`100-300 ms depending on the accuracy of speed match-
`ing in the synchronisation phase.
`Analysis of the time breakdowns described above
`suggests that. with modifications to the system hardware.
`gear change times can be cut by up to 50%. For example,
`modification of the electric-motor controller to allow fast,
`and easy, movement between motoring and regenerative
`braking modes, coupled with modifications to the speed-
`synchronisation algorithms,
`could reduce
`the
`time
`required for speed matching to under 500 ms. Further, if
`modifications were made to the shift mechanism, so that
`it acted more directly on the gear selector inside the gear-
`
`185
`
`Page100f11
`
`FORD 1228
`
`FORD 1228
`
`
`
`Muscling, Bumby and Herron
`
`box, then gear movements (ie, gear to neutral, neutral to
`gear) could be reduced to 100—200 ms. With such modi-
`fications a complete gear change could be achieved in
`500-700 ms.
`
`Perhaps one of the greatest benefits of the transmission
`control system described is the smooth way in which
`torque is reapplied to the drive system following a gear
`change (see Fig 10). This smooth response is due entirely
`to the speed-synchronisation approach adopted in the
`gear change strategy and gives some indication of the
`driveability to be expected in an actual vehicle. An alter-
`native to using speed—matching algorithm in the control
`is to incorporate a slipping clutch at the input to the trans-
`mission and control
`the re-engagement of this clutch
`following a gear change (Main, 1987; Walzer and
`Miersch, 1987). However, it is difficult with such systems
`to avoid torque oscillations in the propshaft following re-
`engagement. In addition, the clutch-control system must
`be able to account for clutch wear. Both of these prob-
`lems are avoided if speed synchronisation is used, as the
`need for clutch control during a gear change can be dis
`pensed with.
`
`7. Conclusion
`
`This paper has demonstrated the feasibility of auto-
`m