.

. ^" _. ^" . DESCRIPTION ^{' ' •} / . ^"

SLIDING ^' MODE CONTROL APPARATUS AND ADJUSTING METHOD .

^' 5. Technical 'Field

The present invention relates to a sliding -mode control apparatus -for se.tting a switching hyperplane at a time of displacing- a controlled object to, which an urging -force-is ^■ . 10 applied by ^" a'n elastic member from, one displacement end to ' the ^■ other displacement end, and controlling the controlled object in ^" such a; manner that a, state quantity of the controlled object 'is converged, on the switching . _v hyperplane, and an adjusting method of the apparatus. ^■ . ^■ is ^■ , ^{; '} ; . ^{• '} . - . . ^{■ ' '}

Background Art . ^■ - , ^{■ • '} _i •

In a control system having ^' a spring mass system such, as,

■an electromagnetic valve _/ ^' it is hard to ^" sufficiently secure 20 robustness with- respect to a disturbance, -a characteristic change of the electromagnetic valve • or the like., in ^■' . ^' ■ accordance with a control method such as a PID control or the like. ^• Accordingly, there has been' considered such a control as to converge ;.a state quantity .of a controlled .25 ^■ object on a switching hyperplane' expressed by a desirably ⁽ ' designed linear function in accordance with a high gain control by applying a sliding mode control to the control system mentioned above, and bind it on ^' the switching hyperplane. In the sliding mod'e control apparatus mentioned 30 above, there has been proposed a technique which can further satisfy a plurality of request elements without contradiction (for example, Japanese Laid-Open Patent Publication No. 2003-202901) .

35 However, ^' the Japanese Laid-Open Patent Publication No.

2003-20290! does not cope with a case that an external force . ^' caused by ^' a cylinder internal pressure of .an internal ^' combustion engine or the like is .applied as a disturbance, or a case 'that a neutral point o'f a spring of an electromagnetic valve is changed due to a change with time, ^{■ ■} ■ and ^' there is a risk that the Operation is destabilized on ' the basis ^' of an operating state of the internal combustion engine, .or ,the 'Change with time. _. The destabilization mentioned .above ^' may be caused by a. machine, error generated at a time _. of manufacturing the controlled object such as the.' ' electromagnetic valve or the like.

Disclosure of the' Invention

An ^' objective of- the present invention is to prevent destabilization in a sliding ^' mode control caused by a ^■ disturbance, a change with trme or a machine .error. ^■

- . ' _^ ?

.To achieve the foregoing objective and in accordance with a first aspect of the present . invention-, a sliding mode control apparatus- is provided. When displacing a controlled object to which an urging force is _. applied by an elastic member from one displacement end to the • other displacement end _/ - the. control apparatus sets a. switching hyperplane and controls the controlled object by using a sliding mode .control in . such a manner that a state quantity of the controlled object is converged on the switching hyperplane. In the sliding mode control, the control apparatus switches an operation mode for controlling the controlled object when ^■ the controlled object passes an operation switching point provided in a displacement region of the controlled object.

The control apparatus includes a disturbance detecting section that detects disturbance in the sliding mode control, and a changing section that changes the operation

switching point in accordance with the disturbance detected ^' by the disturbance detecting section.. Alternatively, ^' the control apparatus may include a change' with time detecting section that detects a 'change .with time of- .the elastic member,' and a changing section that" ^; changes the operation •• switching, point in accordance- with- the change with time detected by the change with, time detecting ^' .section.

In accordance with a second aspect of. the present invention, ' ^' a sliding mode control ^" apparatus is provided.

^■ When ^' displacing a controlled object to which an urging force- is applied, by an elastic member from one- displacement end ^' to the other displacement ■ end, the control apparatus sets a switching hyperplane and controls the controlled object by using a sliding mode - control in such a manner that a ^' state quantity" of the controlled object is converged on the . ^' switching hyperplane. The control apparatus holds the-

^■ controlled Obj ect ά'n- a floating state - at ^' a target floating

• position in the vicinity- ¹ Of at least one of _ the displacement ends . ' ^{' ' '} , . ' _. ^•

The control apparatus includes a disturbance detecting ^■ section that detects disturbance in the -sliding mode control, and a changing. section ^■ that changes the target ^■ floating position in accordance with the disturbance idetected by the disturbance detecting section. Alternatively, the control apparatus may include a change with time ^' detecting section that detects a change with time of the elastic member, and a changing section that changes the target floating position in accordance with the change with time detected by the change with time detecting section.

In accordance with a third aspect of the present invention, an adjusting method for a sliding mode control

apparatus is provided. The . control apparatus obtains, as a detection value of a displacement sensor,, a position of a controlled object to which an urging force is applied by an elastic member, and converts the detection value to a ^■ 5 displacement of the controlled object based on a

■. predetermined ^' relationship". When, displacing the controlled ' object from one displacement end to the other displacement end, the control apparatus set ^' s a switching hyperplane and controls the controlled object by using a sliding _. mode _.

10 ^' control in ^{^} 'such a manner that a state ^' quantity of _. the < controlled object, ^' which state ^' quantity is ^' based on the converted displacement, is converged on the switching hyperplane.. _. In the sliding mode control, the control apparatus switches an operation mode for controlling the,

15 controlled object when the controlled object passes an operation . switching point provided. in a displacement region of the controlled object. , • • • ^■

- . ' _^

. _. ^• , .The method includes, obtaining' a detection value' from

20 the displacement sensor in a state Where the controlled object is located- at. a known displacement, converting a .known displacement state to a detection value of the displacement sensor according to the relationship, ^' - and correcting the relationship based on a difference between

25 • the detection value obtained from the displacement ^' sensor , ^• ,and the detection value obtained from the conversion. Alternatively, the method may include-: obtaining a detection value from the displacement sensor ¹ in a specific displacement state of the controlled object, the specific

30 displacement state being determined based on an equation of motion representing a motional state of the controlled object; converting the specific displacement state to a detection value of the displacement sensor according to the relationship; and correcting the relationship based on a

35 difference between the detection value obtained from the

displacement sensor and the detection ^' value obtained- from ^• - the conversion. ^' , • _{. .} ^••

Brief Description of the Drawings ^' .;

•■ ■ ^' Fig. 1 is an explanatory' view of a structure of an ^• . ^■ electromagnetic valve (in a, closed state) and a control mechanism .thereof in accordance with a first embodiment;

Fig. -2 is an explanatory view, of a neutral state of the ^' electromagnetic valve in accordance with the first

' embodiment; ^• • ^' . . ^{' ■}

Fig. 3 is an explanatory view of an Opening state of' ^' • the electromagnetic valve in accordance with the first . ^• embodiment; ■ _. • ' ^{■ •} _. ^■ Fig. 4 is an explanatory- view of a setting state of a switching hyperplane in a sliding mode control in accordance with the first embodiment; , ^{■ ■ • . ■ Fig. 5 is • a flowchart of a valve -opening control ' .process executed by an ECU in accordance -with the first embodiment; ■ .}

Fig. 6 is an explanatory view pf a structure of a map MAP ^' a ^' for calculating an- attraction current application start displacement A in the- first embodiment; • ^'

Fig.- 7 is a flowchart of a fully .closed position ^■ measuring process executed by- the ECU in accordance with the .first 'embodiment ; ^" . ^■ •

Fig.- 8 is a ^' flowchart of a fuliy open position measuring process;

Fig. 9 is a flowchart of a neutral position measuring process;

Fig. 10 is a timing chart showing an example of each of the fully closed position measuring process, the fully open position measuring process and the neutral position measuring process in accordance with the first embodiment; Fig. 11 is a flowchart of a neutral position deviation

amount calculating process per engine' stop executed by the ECU in accordance with ,the first embodiment;

Fig. 12 is a graph showing the relationship f of a detected voltage V and 'a- displacement x in the first 'embodiment; •, • .

^■ ■ ^' ■ ^' Fig. 13 is a flowchart of a long-period neutral position deviation amount averaging process executed by the ECU in. accordance with the first embodiment;

Fig. .14 is ^' a timing chart showing an example of a process -in." ^' the long-period neutral position deviation amount ^■ averaging process; ^{' '}

Fig. 15 is a timing chart showing an example of a ' ^' . control in accordance with the first embodiment;

' Fig. 16 is a timing chart showing an , example of a control in- accordance with a second ^' embodiment;

Fig. 17 is a flowchart of a valve opening Control process ■ ^' executed by the , ECU \n accordance with the second ^■ embodiment; ■ . ' . ^' . ^■•

• Fig. 18 is an . explanatory view of a structure of a map MAPaI for calculating, a brake current application end displacement Al- in the second embodiment;

Fig. 19-.is an 'explanatory view of a ^' structure of a map. MAPa2' for- calculating an attraction current application start ^" displacement A ^' 2 in the second embodiment; Fig. 20 is. a flowchart of a target floating position ₍ , setting process executed by an ECU in accordance with a third embodiment;

Fig. 21 is an explanatory view of a structure of a map MAPsf for calculating a target floating position Sf in the third embodiment;

Fig. 22 is a flowchart of a shipment correcting process executed by a measuring computer in a fourth embodiment;

Fig. 23 is an explanatory view of a process for preparing a relationship f executed in the fpurth embodiment;

Fig. 24 is a flowchart of an on-board ^■ correcting ^{' '} process executed by the, ECU in the fourth embodiment; ' Fig. 25 is an explanatory view of a process for- updating the relationship f executed. in the fourth' 5- embodiment; ^{' ' • ι 1 ■ ■ • ' Fig. 26 is a flowchart of a cylinder interna! pressure external force _ι calculating process executed by an ECU in accordance- with .a fifth embodiment;}

Fig. '27A is a graph showing a . change of a heat 10 generation 'amount dQ in the fifth' embodiment;

Fig. 27B is a graph showing a change of an estimated cylinder internal pressure Pel . calculated in ^' the fifth ^' . ^'■ embodiment; _t and ^'" . ^■

' Fig. 28 is a timing chart showing an .example of a 15 control in the fifth- embodiment . '

Best Mode for Carrying Out the Invention • . ^' , •

^' • [First embodiment] ■ ^• . . , • ^• \ . 20 ^• . '. ^■ - ^{■ ■} '.

Fig. 1 shows' an electronic control unit (hereinafter, . refer to as ECU) 2 functioning as a sliding mode control apparatus, and an electromagnetic valve ■ ,4 of an internal combustion ^' engine in which a sliding mode control is 25 executed. The _^ internal combustion engine corresponds to an ^{■ '} .engine for -a vehicle, and is constituted, for ¹ example," by a gasoline ^' engine or a diesel. engine comprising a plurality of cylinders. ^'

30 .The ECU 2 is provided with a cylinder internal pressure sensor 2b and a port pressure sensor 2c per cylinder, detects a cylinder internal pressure Pel within each of the cylinders, and a port pressure Ppt in an exhaust port 20, executes a data communication with respect to the ECU for

35 the internal combustion engine, and executes a data exchange

of various data such as a control data, a detection data ' and the like. The • electromagnetic valve 4 is described as an ■ exhaust valve in the present embodiment, ^* however, can be applied to ^' ah intake valve in .the same manner by arranging . a port pressure sensor in an ^' intake port. ^{•" ■}

^{" The electromagnetic-valve ' ■ 4 is structured such as to be provided with a poppet valve 6, an armature 8, a lower core 10, an upper core 12, a lower spring 14, an .upper spring 16 ' and a lift" sensor ■ 18 (corresponding ' to a displacement sensor ' * constituted by a differential transformer or the like). The- EGU 2- formed mainly by a microcomputer adjusts a current ; amount applied to 'a lower coil 10a provided within the ' lower., core io and an upper coil 12a provided in ,the 'upper core _r 12 via an electromagnetic " drive circuit (hereinafter, refer to • as an EDU) 2a. Accordingly, arriving force ' and a ■ holding force applied to the armature, -8 are adjusted by an ' 'electromagnetic force generated in the Lower, core 10 or the ; upper core 12. ■ . ■ . • • • ■}

In a state in which the armature ^' 8 is attracted by and contacts the upper core- 12 as shown in Fig; 1, an upper . spring 16- is set to a- compressed state between an upper retainer 16a and a cas _. ing 4a. Further, the poppet valve 6 is seated on a _^ seat portion 20a of the exhaust port 20 so ,as , . to be ^" in a valve close state by an urging force applied- from the lower' spring 14 ^' via a lower retainer 14a ^' , and the. exhaust port 20 is closed.

On the other hand, in the case that no electromagnetic force is generated in the lower core 10 and the upper core 12, the armature 8 and the poppet valve 6 stop at a neutral position where the urging forces of the lower spring 14 and the upper spring 16 are balanced as shown in Fig. 2.

In a ^' state in which the armature 8 ^' is -attracted- by and contacts the lower core, 10, the lower spring 14 is set to ^' a compressed state between the lower retainer 14a and a- cylinder head H via the ^! armature rod 8a and a shaft portion ^" 5 6a of the poppet valve 6 as shown in Fig. ^' 3. Further, the ^• poppet valve 6 comes to a position ^■ which is most away from -. the seat portion 20a of the . exhaust port 20, and .the exhaust port 20 comes to _. a fully open state.

10 ^' In the ^' case that the poppet valve 6 is moved from the ' fully closed state- to the fully open state, it is possible to employ a control of fixing the armature 8 in- a floating ^' • , state at .a target ^' floating position slightly floating from the 'lower core 10 mentioned below, in place of setting to

15 the ^' completely fully -open state as shown in Fig. ^' 3. ^'• Alternatively, it is pos _. sible to employ a control of temporarily stopping at the target floating- position and- thereafter setting to the ^' fully open .state as shown in Fig . ^■ • ^; 3 . . - ^{• ' " • ■} • . ' ^{• •}

2 b ' ^{■' ■ ' •} . ^{■ ■ ■ '} '

A description will be given of an operation of ^■ the electromagnetic valve 4 by the ECU 2. In ^' the valve-close state' of the electromagnetic valve 4 as -shown in ^' Fig. 1, a holding current is applied to the upper coil 12a for holding 25 ^■ a state in which the poppet valve -6 is seated on the seat _t .portion 20a.

In the opening timing of the electromagnetic valve 4, the holding current application to the upper coil 12a is 30 stopped. Accordingly, the poppet valve 6 is detached ^' from the seat portion 20a on the basis of the urging force of the upper spring 16. Further, in order to adsorb the armature 8 to the lower core 10, in a displacement region of the poppet valve 6 from the fully closed position to the fully open 35 position, the attraction current application to the lower

coil 10a is started and an amount of the current is ^■ adjusted. If the ^' armature. 8 is- brought into contact with the lower core 10 as shown in Fig.. 3, the holding current is circulated to the lower coil 10a', ^■ and the -valve-open state. • is maintained. ^' \ ^•'

In the closing . timing .of the electromagnetic valve 4, the holding current application to the lower coil 10a is stopped.. ■Accordingly, the poppet valve 6 starts moving toward the ^' seat portion 20a oh the basis of the urging force' ' of the lower spring 14. Further, in order to adsorb the armature 8 to the upper core 12, in a displacement region- of the poppet valve 6 from the fully open position to the fully, closed position, the attraction current application to the upper coil 12a is started and an amount of the current ^" is adjusted-. If the armature 8 ^' is brought into contact with the upper core 12, as shown in Fig. "1, the holding current is ^■ circulated to the -upper coil 12a, and -the valve-close state • is maintained. ■ . . ^■ •• ^{■ • ■'} . . ' ^'

At a time of displacing the poppet valve 6 and ¹ the armature 8 from one ^' displacement end to the other displacement end as mentioned above, the. current application control to' the lower coil 10a and. the .upper coil 12a is ^■ executed by using a sliding mode control as mentioned belςw. . Since' the .control aspects of the respective coils 10a and ■ 12a are the same ^' between the displacement process from ^' the fully closed position to the fully open position, and the displacement process from the fully open position to the fully closed position, a description will be given below by exemplifying the displacement process from the fully closed position to the fully open position.

In the present embodiment, ^• the displacement of the poppet valve 6 (actually including the armature 8)

corresponding to a controlled object comes to a one-' ' ^■ dimensional space (line segment) connecting the fully ^■ closed position and the fully open position, and -a state quantity indicating a dynamic characteristic of the poppet valve 6 comes to a displacement and a displacement speed of the ^• poppet valve 6. Further, a control is executed such that the displacement and the- displacement speed (the state quantity) ^; of the. poppet valve ^' 6 is converged to- a previously set switching hyperplane (line segment) corresponding to a (one-dimerisional) linear partial Space within a two- " dimensional space having the displacement and the displacement speed of the poppet valve _. 6 as degrees, of freedom, . on ..the two-dimensional linear space. Further/ the ^• switching hyperplane _. is set so as- to be variable in correspondence to the displacement of the poppet valve 6, for satisfying a plurality of request elements relating to a control' performanpe of the electromagnetic, valve 4. >

In detail, the- switching hyperplane is set as a- hyperplane which is brought into contact with a previously set reference model about the state quantity of the: poppet valve 6 at _. the corresponding . displacement point. The reference' model is basically set on the -basis ^' of ^' a displacement of the poppet valve 6 (a .locus of the ^■ displacement and the displacement speed of the poppet valχe .6) defined -on the assumption that a disturbance, a damping- element and an electromagnetic force do -not exist. The reference model is set as the hyperplane near the other displacement end. The hyperplane is set such that the displacement speed comes to "0" in the other displacement end. Further, the hyperplane is set such that a change rate of the displacement speed becomes smaller than the reference model set on the basis of the displacement aspect of the poppet valve 6 defined on the assumption that the disturbance, the damping element and the electromagnetic

force do not exist, near the other displacement ^' end. ^•

Accordingly, the switching hyperplane is set on the basis of the reference 'model set' on the basis of the • ^' displacement of the poppet valve 6 defined on the assumption ^■• that ^' the disturbance, the damping, element and the electromagnetic do not exist, in . the other portions ^■ than the portion near the. other displacement end. • In contrast, near the fully -open position, the hyperplane is set - such that, the displacement speed comes to "0 ^' ". at the fully open position.

^■ Fig. 4 shqws a specific procedure for establishing the switching hyperplane in accordance with the present embodiment. In the other displacement regions of. the poppet - valve 6 than a portion near a fully open position Low, the ^' • reference model is structured as follows. In other words, it comes to a locus of the displacement and the displacement 'speed of the poppet valve 6. at a time -when the- poppet valve ^■ 6 is displaced from a state in which the poppet valve 6 exists at a fully closed position Up to the fully open position Low only by the urging for.ce of the lower spring 14 ■and the upper spring..16. ,

_. This locus (the _.. reference model) ■ comes to a quadratic ^• curve calculated from a physical model having an elastic _t ■body constituted by the lower spring ^' 14 and the upper spring 16 and a movable portion (the poppet -valve 6 and the armature 8) coupled to the elastic body as a physical system. In other words, an equation of motion of the physical system is expressed by an expression 1, in which a weight of the movable portion is set to M, an elastic constant of the elastic body constituted by the lower spring 14 and the upper spring 16 is set to K, and a displacement (detected value) of the poppet valve 6 on the basis of a neutral position (position of balance) of the elastic body

is set to x . ^{■ ' •} ■ - ^' . ^■ • ^{' ■}

[Expression I ] ₁ - . ^• *

M - x = -K - x ^{' ■ • ■} . ^{" ■} - ^{■ • •}

^' 5 ^{' ■ • '}

'• ^■ The displacement ^■ x of the poppet valve 6 corresponding to a solution of the expression 1 is determined as a periodic function, ^' and the displacement speed is determined as a periodic function on the basis -of a differential value 10 therefrom. ^{' ■} Further, a relational ¹ expression therebetween is ' determined as a quadratic curve shown in Fig. 4, on the basis- of the displacement and the displacement speed.

' Further,, as shown- in Fig. 4, with respect to the 15 displacement region- of the poppet valve 6 near the fully open position Low, the reference model mentioned above is set to a one-dimensional hypqrplane ^' (line segment) in -which the change rate of 'the displacement speed with respect to : the; displacement is ^■ smaller than. the quadratic curve ^' 20 mentioned above. '. '

It is possible ^' to ^" satisfy a plurality of request elements requested in the control of the. electromagnetic valve ^" 4 by variably- setting the switching hyperplane in

25 accordance with -the displacement x of the poppet valve 6 qn .the basis of the reference model. In other words, in the • other positions than a portion near the fully open position ^' Low, it is possible to displace the poppet valve 6 while effectively utilizing a natural vibration of the physical

30 system, by controlling the actual displacement of the poppet valve 6 in accordance with the displacement of the poppet valve 6 at a time of displacing the poppet valve 6 only by the urging force of the elastic body mentioned above. Therefore, it is possible to reduce a time required for the

35 displacement from the fully closed position Up to the fully

open position Low. ^{" " ■}

On . the contrary, near the fully open position Low, it is possible 'to ^' suppress. the impact at a time when the armatur'e 8 _. is seated on the upper surface ^" of the lower core ^■ 10, by binding the state quantity on the hyperplane in which a change rate of the displacement speed is. small. - •

It. is possible to ^" execute the. control for cushioning. the shock ait a time of being seated while- reducing the

' ^■ displacement time, by controlling the electromagnetic valve ^• 4 _' in such ^' a manner as to bind the 'state, quantity on. the switching hyperplane variably set in correspondence to the displacement of the poppet valve 6. On the contrary, for ^■ : example, in the case- of controlling the electromagnetic valve 4 by using the switching hyperplane shown ^' by a single- ^' dot chain line in, Fig. 4 so as to cushion the shock at a 'time of .being seated, the displacement . time .is' increased.

The setting of the current application control to the electromagnetic valve 4 for executing the sliding mode control of binding the state . quantity on the switching hyperplane is executed as follows. Firs.t, there is defined by. an expression 2 a switching hyperplane (a tangent line in ' the displacement x of the quadratic reference model) . , , ■contacted with the quadratic reference model ' in Fig. 4 at the displacement x of the poppet valve 6, and a primary reference ^' model in Fig.. 4.

[Expression 2] x = a-x+b

In this expression 2, coefficients a and b are actually constituted by a function of the displacement x of the poppet valve 6. Further, a switching function σ

corresponding to a linear function defining the switching hyperplane is defined by an expression 3. .

[Expression '3] ^' ■ • \ ^{' ■'} . σ ^' =x—a-x-b ^■

As is known from the expression 3/ the hyperplane in which the _' Switching function σ is/ set to zero corresponds to the switching hyperplane. ^< ,

Next, an actual physical ^" system of the . electromagnetic ^■ valve- 4 is set ,to a system to which a sliding _. resistance ; between the. movable portion and the fixed portion, and an electromagnetic force acting on ^' the armature- 8 are applied, in a structure ih which the lower spring 14 and the upper spring 16 are coupled to the movable portion. An equation of motion of this, system is expressed by an expression 4 by ^' 'using a, damping coefficient C between the movable portion ; and the fixed portion, ^' and a sliding mode input. Ul corresponding to the electromagnetic force ^' applied to the armature 8 in a. sliding mode state,, in addition to the. weight M, the elastic constant K, and the ^'' displacement x mentioned ¹ above. . , ^{' ■}

[Expression 4] ■ ^■

M-x = -K-x-C-x+Ul ^'

In the sliding mode state, the state quantity of the electromagnetic valve 4 is next bound on the switching hyperplane, in other words, on the hyperplane in which the switching function σ comes to zero. Accordingly, the sliding mode input Ul is expressed by an expression 5 by using a fact that a time differential of the equation of motion (the expression 4) and the switching function σ is zero.

[Expression 5] ^' , . ^{■ ■} . ^' . ■ Ul=(C+M-g)-x+K-x

Further, a reaching mode input' ^, (a feedback input) UnI ^■ converging the state quantity on the. switching hyperplane at a time when the state quantity 'is away from the switching hyperplane- is de.fined by an expression ^' 6.

^' [Expression 6] ^■ "

UnI = G-- ^■ - . - . . . ^■ ;

The feedback gain G is set ^' in such a .manner ^' as to satisfy a condition- for reaching the switching hyperplane, - in other words, a reaching condition corresponding to a condition for reaching the sliding mode. The gain G _. • 'satisfying the ^• reac'hing condition is set., by using a Liapunov _: functipn method in the present embodiment. \ In other ^' words, the gain G is set- such that the time differential expressed - by an expression 7 becomes negative,, for example, ^' by setting V = ^' 1/2 x σ x σ ^τ to ^' the Liapunov function.

[Expression 7]- _. . . _ ■ . V = σ ^τ -ά , ^• . ^' _. ^{' ' • ' ' '}

In the expression 7, the switching -function σ is going ^' to converge to zero while using the reaching mode input UnI, by setting positive and negative of the gain G while having a predetermined absolute value, in such a manner that the time differential of the Liapunov function becomes negative.

A description will be given of the valve opening control process of the electromagnetic valve 4 executed by the ECU 2 with reference to Fig. 5. Fig. 5 is a flowchart

• showing a processing procedure of the ^' control . ^• This-process is executed repeatedly at a time ^' cycle . When starting the valve opening qontrdl process, the ECU 2 stops the ^' holding current applied to the ^: upper coil ^' 12a. Accordingly, the ^■ armature 8 starts moving away from the upper core 12, and the poppet valve 6 starts moving away from the seat portion 20a in conjunction therewith. ^'

If the valve opening control process .(Fig.- 5) is started, the ^' attraction current application start ^{' '} ' displacement A is set (SlOO) . ^' The attraction current application start displacement A corresponds to- a threshold value for judging 'the displacement starting the attraction force to the armature 8- by the lower core ,10 -in accordance with the sliding ^' mode control. ^■

. In- ^' the attraction, current application start • displacement A,- a p'roper value on the -sliding mode control :is changed in correspondence to a cylinder -internal pressure external force FcI ^" on. the basis of a pressure difference between an inner side and an outer ,side of a combustion chamber 22, (Fig. I) ^' ^. and a deviation Dmp with time _, of a neutral position Mp shown in Figs. 2 ^' and.4. Accordingly, in the step SlOO, the attraction current 'application start ^■ displacement A is calculated on the basis of the cylinder , ^■ ■internal pressure external force FcI corresponding to the disturbance, and the deviation (the neutral position - ^' deviation) Dmp of the neutral position Mp corresponding to the change with time, in accordance with a map MAPa shown in Fig. 6.

The cylinder internal pressure external force FcI corresponds to an external force generated in the poppet valve 6 on the basis of a difference "Pel - Ppt" between a cylinder internal pressure Pel sequentially detected by the

• cylinder internal pressure sensor 2b and .a port pressure ^' Ppt

-sequentially detected by the port pressure, sensor 2c, and is

■ calculated by a, map 'and a function on the 'basis of the value of the difference "Pel ^• - Ppt". ^' . . ^' .

"' ^• In accordance that the cylinder internal pressure external force FcI generated on the basis of "Pel - ^■ Ppc" becomes larger, it _. is necessary t.o apply the attraction current early against the cylinder internal .pressure external force FcI at a time of the valve ^• opening control. "Accordingly, as shown by a contour line (a broken line) in Fig...6, the attraction current application staxt displacement a is ^' moved to the valve closing side in accordance with an increase of the cylinder internal pressure external force FcI.. Since the attraction current application is delayed in accordance with the increase of the cylinder internal pressure external force FoI at . a- time

'of ^' the valve closing- control, there is' formed a map having

^■ the same tendency as Fig. 6, although the- map value .is different. ■ ■ ^' ' , - . ^' . ^■

The neutral position deviation Dmp employs a value ^■ ■ which ^' has ' been already detected iri accordance ^' with- a neutral position deviation detecting process ('Figs. 7 to 14) mentioned below.- If the neutral position deviation comes to •the valve closing side as shown by the contou'r line (the broken line) in Fig. 6, the. attraction current application start displacement A is moved to the valve closing side so as to be adapted thereto. On the contrary, if the neutral position deviation comes to the valve opening side, the attraction current application start displacement A is moved to the valve opening side so as to be adapted thereto.

The map MAPa is formed by previously determining an optimum parameter value in accordance with an

experimentation and a simulation so as to map. The map MAPa may ^' be formed in accordance with ^' a forecasting expression determined by a statistical method (DoE -or the like) ,• an experimental' formula determined ^' by an ^■ experimentation, or a physically conducted physical formula, in addition to the ^■ •above. The forming method ^' of ^• the. map mentioned above is applied to the other maps mentioned below in the .same manner. ' . ^'

If the ^' attraction current ^' application start

" displacement A is calculated by the step SlOO, it is -judged ^■ whether of not the displacement x of the- poppet valve ,6 detected .by ..the lift sensor 18 is below the attraction current application start displacement A (.S102) If a relationship x ≥'A is established here (no in S102) , the present process is temporarily finished.

If a relationship x < A is established ■ (yes in S102),

'. . .. . )

•the switching function ^' is .set ^' on .the basis of the . displacement x of the poppet valve β (S104) . This is set by previously providing the ECU 2 with a memory function for • storing the reference model shown in Fig.- 4, and calculating _' the tangential ^' line of the reference model in ^' the displacement x of the poppet valve 6. ■ Further, the ECU 2 ^• may be provided -with a memory function for storing ^' a data ,(a .map or the like) relating to the switching function such as the coefficients ^' a and b in the expression 3 ^' in the respective ' displacements x of the poppet valve β within the ECU 2, and execute the switching function setting by the map .

The sliding mode input Ul is calculated on the basis of the switching function set as mentioned above (S106) . In other words, the sliding mode input Ul is calculated by using the expression 5 from the switching function in the

• displacement 'x. Further, the reaching mode input UnI is' calculated on the basis, of the expression ^' 6 from the _. switching function in the displacement x (S108) .

The control input U corresponding to ^' the-

' electromagnetic force applied- to the armature 8 is calculated on the ^' basis of a total of the sliding mode input Ul and, the- reaching mode input ^' UnI calculated as mentioned ^' above (SIlO) . Further, a gap Gp between the armature 8 and the lower core 10 is calculated on the basis ^' of the • . ° displacement x detected by the ^' lift sensor 18 (S112) .. Further, a current application control current amount How ^' to the lower .coil ^' 10a of the electromagnetic valve 4 is calculated in accordance with a calculatiqn F 'using the gap Gp and the control input U (S114). The calculation F executes ^' the calculation of the current application control current ^' amount How by providing the ECU 2 with .a function for storing the physical model- formula defining the _. current •amount, applied to the lower coil- 10a of the electromagnetic valve 4' from the gap Gp and the control input U. Further, a map defining a relationship among the gap Gp, the ^' control • input U and the supply current amount to the lower coil 10a . may be stored in the -ECU 2. If- the control input ^' U ^" is negative,- the current .application control amount is set to ' "0". . ^■

If the current application control current amount How is calculated as mentioned above, the current application control to the lower coil 10a is executed on the basis of the current application control current amount How (S116) . As mentioned above, the sliding mode control at a time of the valve opening drive is applied to the electromagnetic valve 4. Although the value is different, the sliding mode control is basically applied' to the valve closing control time, in the same manner as the valve opening control

process mentioned above. ^{' ' •}

Next, a description will be .given of a detecting process of the neutral position deviation amount Dmp executed by the ECU 2. Figs. 7 and' 8 show a flowchart o-f a ' measuring process of a fully 'closed position Up and a fully open position Low executed for ^' detecting the neutral position deviation amount Dmp. ^' Each of the processes corresponds to a process which is repeatedly executed in .a short time ^"' cycle during the. normal operation of the internal_ ' combustion engine.

A descriptioh will be given of the fully closed position measuring process (Fig. 7) . If the^ present process is started, a detected voltage V of the lift sensor 18 is first read (S200) . .The lift sensor 18 is structured such as to output a higher voltage in accordance that the poppet valve 6 and the armature 8 come closer to the fully closed

^■ position Up side (the displaced state in Fig. 1), and output a lower' voltage in accordance that ^• they come closer to the fully open position Low side (the displaced state in Fig.

Next, ^' it is judged whether or not the detected voltage - V exists in the- fully closed region (S20.2). In the judgment

■ of the fully closed region, it is determined' that the detected ^' voltage V exists in the fully closed region ,in the case that the detected voltage V is higher than a previously set fully closed reference voltage Vclose.

If the relationship V ≤ Vclose is established (no in S202), a counter nl for a fully closed region is cleared (S204), and the present process is temporarily finished.

If the relationship V > Vclose is established on the

• basis of the drive of the electromagnetic valve 4 (yes in ^' S202>, the detected voltage V exists in the fully closed region, so that it is judged whether or 'not the fully closed region counter ^' nl is smaller than' 100 (S206) . . Since a ^' relationship nl = 0 is established just after the

'•relationship V > Vclose is ^' established ^' from the state of V ≤ Vclose (yes in _ι S206)., a moving ^' average process of the detected voltage. V is next executed in accordance with an expression- 8, and a moving average value tVmax is calculated (S208) ^{■ ■ ■} . ' ^{' ' '}

[Expression 8] -, ^■ • . tVmax _* r .Vmaxold ^• 9/10 + V/10

^' Further, the moving average value tVmax determined by the expression 8 at this time is set as the previous value Vmaxold ■ (S210) . . , ^{" •}

^{' ; -Further, the counter .nl for -the fully closed region is incremented (S212), the present process is temporarily finished. ■ ■}

In the next control ^' cycle and after, as long as the relationship V > Vclose. is established (yes in S202) and the ^• relationship nl ■< 100 is established (yes in S206) , the

■moving average process (S208), the setting of the previous value Vmaxold (S210) and the increment of the fully closed region counter nl (S212) mentioned above are repeated.

If the moving average process (S208) is repeated at 100 times on the basis of continuation of the relationship V > Vclose (yes in S202) , a relationship nl = 100 is established on the basis of the increment of the fully closed region counter nl (S212) . Accordingly, in the next control cycle, since a relationship nl > 100 is established (no in S206)

after the decision outcome in the step S202 is judged to be positive, the moving average value t-Vifiax is set in the fully closed time voltage 'Vmax (S214.) . . '

^' If the relationship V > Vclose> (yes in S202) is

^■ •continued in the later control cycle-, the decision outcome '• in the step S206 is negative. ^' Accordingly, the moving average process .(S208) is not executed, the ^' value of the moving average value tVmax is not changed, and the value of the fully closed time "voltage Vmax is maintained.

As shown in a timing chart in Fig. 10 as mentioned above, the fully closed position is measured as the value of ^' the ' fully closed time voltage Vmax at an early stage (tO.to tl) ^' of each of the period, every time when the electromagnetic valve 4 comes to the fully, closed position.

A description 'will be .given of the fully open position ■measuring process (Fig. 8 ^" ). • The -present process calculates the moving average in a minimum -state of the ^■ detected voltage V, and ^: a basic process is the same as Fig. 7 mentioned above. ^' _. . - >

If the present pr.ocess is started, the detected voltage ^■ V of the lift sensor 18 is first read (S300) . Next, it is, judged whether or not the detected voltage V 'exists in the ^• fully open region (S302) . In the judgment of the fully open region, the fully open region is determined in the case that the detected voltage V is lower than a previously set fully open reference voltage Vopen.

If the relationship V ≥ Vopen is established (no in S302) , a counter n2 for a fully open region is cleared (S304), and the present process is temporarily finished.

If the relationship V < V.open is ^■ established (yes in S302)-, the detected voltage V exists in the fully open region, so that it rs judged whether or -not the fully, open region counter n2 is smaller than 1OQ (S306) . Since a ^■ relationship n2 = 0 is established just after .the

^■ •relationship V < Vopen is established from the state of V ≥ Vopen ^' (yes in S306) ,. a moving average process of _. the detected voltage _. V is next executed in accordance with an expression.9, and a moving average value tVmin is calculated (S308) . ^{' ■ •■ ■} . ^■ . ^■

[Expression 9] ^■ : _λ tVmin _*í .Vminold ^•' 9-/10 + V/10

Further, the moving average value tVmin determined by the expression 9 at this time is set as the previous value Vminold ^■ (S310) . . , ^{' '} _. ■ ^' .

Further, the counter . n2 for .the fully open region is incremented (S312), the present process is temporarily finished. ^• . ^• _

Tn the next control cycle and after, as long ^' as the relationship V < Vopen is established , (yes in S302) and the ^■ relationship n2.< 100 is established (yes in S306) , the

.moving average process (S308), the setting of the previous value Vminold (S310) and the increment of .the fully open region counter n2 (S312) mentioned above are repeated.

If the moving average process (S308) is repeated at 100 times on the basis of continuation of the relationship V < Vopen (yes in S302), a relationship n2 = 100 is established on the basis of the increment of the fully open region counter n2 (S312) .■ Accordingly, in the next control cycle, since a relationship n2 ≥ 100 is established (no in S306)

■ after the decision outcome in the step S302 is- judged to 'be positive, the moving average vaiue tVmin is set in the fully open time voltage VmIn (S314) . .

■ If' the relationship V < Vopen (yes in S302) is

^■ •continued in the later control cycle, the decision outcome ■ in the step -S306 is negative. ^' Accordingly, ^' the moving average process (S308) is not executed, the value of the moving average value tVmin is not changed, . and the value of ^' the fully open time voltage Vmin is maintained.

. As shown in the timing chart in Fig. ^' 10 as- mentioned- ' above, the fully open position is measured as the value of the 'fully open time voltage Vmin at an early ■ s'tage (t2 to t3) of each of the period, every time when the electromagnetic valve 4 comes to the fully open ^' position.

Fig. 9 shows a' flowchart of a neutral position ^■

■measuring process executed for detecting the- neutral ^' position deviation amount Dmp. The 'present process . corresponds to a process executed repeatedly at a short time cycle. _. ^' _, , ^' .

If the present process is started, it is judged whether ^• or not the stop .process of the internal combustion engine _v is ■executed (S400) . In other words, in accordance with the present embodiment, ^" it is judged whether or not an ignition switch is turned off, whereby the stop operation of the internal combustion engine is executed, on the basis of the signal from the ECU for the internal combustion engine.

If the internal combustion engine is under operation, or the stop process of the internal combustion engine is finished as in the present process (no in S400), the present process is temporarily finished.

If the ignition ^' switch is just -after- being turned off, the it is determined that the internal combustion engine is under the ^' stop Operation (yes in ^' S400: t4 in Fig. 10), _, ^' and the detected voltage V of the lift sensor ^' 18 is next read ' (S402) . ^'

Next,- a changing speed SPv of the detected- voltage V is calculated- (S404) . Since the detected voltage V is measured by the ECU ^" 2 at . a fixed cycle, ^: the changing speed SPy

" employs ^' a change dV of the detected voltage V for one- cycle as the changing, speed SPv. ' .

' Next, it is judged whether or not a displacement x , ^' corresponding to 'the detected voltage V exists in the neutral region, and the changing speed SPv exists in the low speed region (S406) .

,- .' ' ^'

^■ . ■ ^■ The neutral region is provided by setting a width δX around a displacement .x = 0 indicating the neutral position. Accordingly, if .an expression 10 is, satisfied, if is •determined that the ^' detected .voltage V exists in the neutral _' . region. ^{• '} • ^{■ '} - ^■ . ^■

^■ [Expression 10J -

-δX < x < +δX ^'

A conversion into the displacement x (corresponding to the displacement data) from the detected voltage V is executed on the basis of a relationship f (Fig. 12) mentioned below.

The low speed region is provided by setting a width δSPv around a state in which a time change (V/s) of the detected voltage V is zero. Accordingly, if an expression

■ 11 is satisfied, it is determined that the change speed SPv exists in the low speed, region. • ^'

[Expression 11] ^{' ■ ' '} -δSPv < SPv < δSPv

ϊn the step S406, it may be judged whether or not the changing speed of the displacement x, that is, the moving speed of the poppet valve 6 exists i-n the low speed region, in place of the changing speed' SPv of the detected voltage ' V. The ^' changing speed of the displacement x is calculated in the step S404. ^{• •} ..

In a state in which the condition of the ^' step S406 is not satisfied yet. just after starting the stop process (no • ^' in S406) , a ■ counter n3 for a neutral position is cleared • (S408), ' ^' and the present . process is temporarily finished.'

If the expression ^' 10.and the expression- 11 are ^■ satisfied (yes in S406: t5 in Fig. 10), the detected -voltage V becomes a value ^■ suitable for measuring the neutral ■position. Accordingly, it is judged whether or not the neutral position counter n3 is smaller than 100 (S410) . If the determination of the step S406- is just after being ' changed to yes, from no, a relationship n3 =0 is established . (yes in S410) . Accordingly, the moving avera'ge process of the detected voltage V is next executed in accordance with an expression 12, and the moving average value tVz is calculated (S412) .

[Expression 12] tVz «- Vzold ^• 9/10 + V/10

A moving average value tVz determined by the expression 12 at this time is set to the previous value Vzold (S414) .

^■ Further, the neutr.al position counter. n3 is incremented (S416) , .and the present process is temporarily finished.

Ih the next control cycle and after, ^' as long as the ^• decision outcome in the step S406 is judged to be positive ^■ and the relationship n3 < 100 Is established (yes in S410) , the moving- average process (41 ^' 2),. the setting of the previous value Vzold (S414) and the -increment of the neutral position' counter n3 ^' (416) are repeated.

• If the moving average process (S412) is repeated at 100 times on ■ the .basis of the continuation of the determination ^• of yes in the step S406, the relationship ,n3- - 100 is established on the basis of the increment of the neutral position counter n3 (S416) . Accordingly, since ^' the relationship n3 ≥, 100 is established (no in S410) after the decision outcome iri the step S406' is. judged to be positive ^{• ■} in the next control cycle, the moving average value tVz is set to the neutral position voltage ^' Vz (S418: t6 in Fig. 10). The neutral position voltage ,Vz ^' is determined' as mentioned above. ^' .. , , . ;

Further, a neutral, position deviation amount ^■ calculating process per engine stop (Fig.. 11) mentioned

■below ^' is executed (S420), and a long-cycle neutral position deviation amount ^' averaging process (Fig. 13) mentioned ^' below is executed (S422) .

Further, a stop process end setting is executed (S424). Accordingly, in the next control cycle, since the decision outcome in the step S400 is negative, a substantial process in accordance with the neutral position measuring process (Fig. 9) is finished.

^■ A description will be given of the neutral position' deviation amount, calculating process per engine stop (Fig.

11.) • _. . - _ . . ;

If the 'present process is started, the already calculated fully open time voltage Vmin is first converted ^■ into the fully open displacement Xmin, on the basis of the ^• . relationship f between the detected voltage V of .the lift sensor, 18 .and the displacement x of the poppet valve 6 expressed by the map or the function in Fig. 12 (S500) . The 0 detected voltage V of ¹ the lift sensor 18 does not have a ¹ straight-line relationship with the actual ^' displacement x, but has a deviation from the straight-line relationship as ^' shown by . a curve in Fig. 12. ^• Accordingly, in the case of calculating the neutral position deviation amdunt, since _, an 5 error is generated if the detected voltage V is converted into the displacement x by a coefficient so as to be utilized for the calculation,, the fully open time voltage

^■ Vmin is ^' first converted into the fully open .displacement

• . ^' ■ Xmin on the basis of the- relationship f in Fig. 12. 0 ^■ . ^'

Next, the ■ already calculated fully closed time' voltage Vmax is converted into the fully closed displacement Xmax on the basis- of the relationship f in Fig. ■ _v 12 in ^' the same manner (S502) . ^{' '} , , 5 ^■

Further, the already calculated neutral 'position ^■ voltage Vz is converted into the neutral displacement Xz on the basis of the relationship f in Fig. 12 in the same manner (S504) . 0

Further, a neutral position deviation amount per engine stop dm is calculated on the basis of these displacements Xmin, Xmax and Xz in accordance with an expression 13(S506).

5 [Expression 13]

dm <- |Xz-Xmin| - |Xmax-Xz|

If a relationship the neutral position deviation amount per engine stop ^' dm > 0 ^; is established, the neutral ; position is deviated to the closed side, and- if dπv ^' < 0, the neutral '■position is deviated to the open side.

The present process is ^' finished as mentioned above. The neutra-l position deviation amount calculating .process. per engine ^' stop (Fig. 11) is executed at one time .per the J " internal combustion, engine stop time as mentioned above. Accordingly, the neutral position, deviation amount per engine stop _* dm is ^' calculated per the internal combustion engine stop. The neutral position deviation amount per , engine stop dm is stored within a nόn-yolatile memory within the ECU ^' 2 ■ after the .ignition switch is turned off.

Next, a description will be given of the long-cycle ] neutral position deviation amount averaging process (Fig. 13) for' calculating a . final neutral' position deviation amount Dmp.

Tf the present process is started, ■ ,a long-cycle counter nlong ^" is first incremented (S600). A ^value of a long-cycle ^• counter nlong is stored within the non-volatile memory.

Next', it is ^' judged whether, or not the long-cycle counter nlong is equal to or more than 10000 (S602) . If a relationship the long-cycle counter nlong < 10000 is established (no in S602), the present process is finished. Accordingly, the neutral position deviation amount per engine stop dm is determined in accordance with the neutral position deviation amount calculating process (Fig. 11) just before, however, the final neutral position deviation amount Dmp is not newly calculated.

If the long-cycle .counter nlong is increased to satisfy a relationship ,nlong ≥ 10000 ^' (yes. in S602), during ^' the repeat of the internal -combustion engine stop, a moving average' counter nshort is next incremented (S604) .

Next, the _t moving average process of the neutral position deviati.on amount per engine stop dm calculated in accordance' with the neutral position deviation amount calculating process per engine ^' stop (Fig. 11) at this time ' is next ^' executed in accordance with an expression 14 so as to calculate the moving' average value tDmp (S606) . •

[Expression ^' 14] • , _. tDmp <- Dmpoid ^• 19/20 + dm/20 ^{' '}

Further, the. moving average value tDmp- calculated by

^' the expression ^■ 14 at- this ^' time. is ^' set to the previous value ^' ; Dmpoid (S608) . ^{' ■■} . ^{•' ■} . ' ^{' '} . ^{• ■ '} .

Next, it is judged whether or ,not the moving average ■counter nshort is eqμal to or more than 100 (S610) . If the ■ relationship the moving average- counter ■ nshort < 100 is established (no in S ^' 61.0) , the present process is temporarily ^• finished. IThe -moving average value tDmp, the previous .value ^" Dmpoid and the .moving average counter nshort are stored within the non-volatile memory-.

If the long-cycle neutral position deviation amount averaging process (Fig. 13) is executed in succession to the neutral position deviation amount calculating process per engine stop (Fig. 11) per the internal combustion engine stop, the step S604 and the moving average process (S606 and S608) are repeated.

Further/ if the increment of the moving average' counter nshόrt is repeated, whereby the/ relationship nshort ≥ 100 is established (ye.s in 'S610) , the. long-cycle counter ήlong is cleared (S ^' 61 ^' 2) , and the moving average counter nshort i's 5 cleared (S614) . Further, the moving average value tDmp at 'this time is set as the final ^• neutral position deviation amount Dmp (S616) . . • ^'

Accordingly, the final neutral -position deviation 10 amount Dmp ^" is calculated by. moving averaging the neutral

" position deviation amount per engine stop dm at 100 times as- shown by a timing chart ¹ in Fig. 14' every 10000 engine stops. The final ^' neutral position deviation amount Dmp calculated as mentioned above is. used in the step SlO ₁ O of the valve , 15 opening control process (Fig. 5) mentioned above., whereby it- - ^' is possible to determine the attraction current 'application start displacement A. ^• , • •

^• • ^{'■ ■} Accordingly, as shown by a timing chart in Fig. 15, the 20 attraction current application to the lower coil 10a -is started at the attraction current application start: ■displacement A set on the basis of the .cylinder internal pressure external force FcI and the neutral position deviation- Dmp (til) , _. after the holding current applied to 2 _. 5 ^■ the upper coil 12a is stopped (tlO ¹ and after) . Fig. 15

. shows ^' a state in which the holding current applied to the • lower coil 10a. is controlled in such -a manner as to float to the target floating position Sf at a time of opening the valve. 30

In the structure mentioned above, the correspondence to claims is as follows. A combination of the poppet valve 6 and the armature 8 corresponds to a valve body (a controlled object) . The cylinder internal pressure sensor 2b, the port 35 pressure sensor 2c, and the ECU 2 for calculating the

cylinder internal pressure external force FcI on the basis of the cylinder internal pressure Pel :and the port pressure Ppt detected by thes^e sensors 2b and 2c correspond " to. a disturbance detecting section. The fully closed position 5 measuring process (Fig. 7), the ^' fully open position

^■ •measuring process (Fig. 8), the neutral position measuring • •process (Fig. 9) , the neutral position deviation amount calculating process per engine' stop (Fig. 11) and the long- cycle neutral position deviation amount averaging process 10 (Fig. 13) correspond to processes ¹ performed by a change with ' ^• time detecting section. The steps SlOO and S102 in the valve opening control process (Fig. 5) correspond to ^• ' processes per.formed by .an operation switching point changing- section.- ■ ^' . '

^■ 15 ^' - ^{' ' ;} , ^' ' ^{■ ■ • • '} .

• In ^' accordance with the first ^' embodiment ^' described above, the following advantages can be obtained.

1

(1) Due to influence. of both of the cylinder internal 20 pressure external force FcI corresponding to the disturbance and the change ^: with time of the springs i4 and 16 corresponding to the elastic .member, the attraction current > application start displacement A is deviated ^■ from ^' the proper displacement in the sliding mode control. There is a risk 25 ^■ that the process (Fig. ^' 5: S104 to S116) in the sliding mode .control becomes unstable. ' ^■

Accordingly, in the valve opening control process (Fig. 5) , the proper attraction current application start 30 displacement A is set by changing the attraction current application start displacement A (SlOO) corresponding to the operation switching point in correspondence to both of the cylinder internal pressure external force FcI and the neutral position deviation amount Dmp of the poppet valve 6 35 by springs 14 and 16. Accordingly, it is possible to

prevent the destabilization in the sliding mode control, ^' and it is possible to prevent a deterioration of the seating speed and a logs of 'synchronism. . ^{■ '}

5 (2) The detected value (the detected Voltage V) of the ' lift sensor 18 is not used ^' as ¹ the displacement of the poppet valve 6, but the detected voltage. V is converted .into the actual -displacement x on the basis of the relationship f shown in Fig. 12.- Accordingly, it is possible to _. do, away. 10 with the displacement ' error on ^' the basis of the straight- ^r line relationship between the detected voltage V and the actual displacement x, and it is possible to detect ¹ . _. the change with '.time of the neutral position more accurately.

15 ^' Accordingly^, it' is possible to set the more proper attraction current application. start displacement A, and it i ^' s possible to prevent the de,stabilization in the sliding mode control. ^{■ • ' • ■ '}

20 . [Second' Embodiment] • ^• .

In the present ^' embodiment, as shown by a timing chart in Fig. 16, after the- holding current ■ in the ^" upper coil 12a is - temporarily ^' stopped in accordance with the valve opening 2.5 ' control process * (t20 and after) , a motion of the valve body > constituted by the poppet valve 6 and the armature 8 is temporarily controlled by the brake current. Further, ^' after finishing the brake current application (t21 and after) , the attraction current application start is applied to the lower 30 coil 10a (t22) .

Accordingly, in the present embodiment, the operation switching point is constituted by both of the brake current application end displacement Al and the attraction current 35 application start displacement A2.

Accordingly, in' the present/ embodiment, the valve opening control process is repeated in a time cycle as shown in Fig. 1? in place of the process in- Fig. 5. The othe'r structures are the same as the structures ^" in accordance with ^• the ^' first embodiment. Accordingly, a description will be given ^' with reference to the other drawings than Figs. 5 and 15 of the first embodiment in addition to the new drawings.

^' If the present ^' process, is ^' started, the setting of the ' brake current application end displacement ^" Al is- first executed ^■ (S700) . The brake current application end/ displacement Al corresponds to a threshold value for judging- the ' displacement stopping the brake force .applied to the _, armature 8 which ' is . going to be disconnected from the upper ■ core 12,' on the sliding mode control.

A proper value of the brake current, application end -displacement Al is changed by the cylinder internal pressure external force FcI corresponding to ' the disturbance and the neutral position deviation Dmp corresponding to the ' change .with time. Accordingly, the .brake current application end displacement Al is calculated on the basis of the cylinder internal pressure external force FcI and the neutral ^■ position deviation Dmp in accordance with a map MAPaI show,n , in Fig. 18, ^' in the step S700. Each of the values of the cylinder internal pressure external force FcI and the neutral position deviation Dmp is obtained in the same manner as the first embodiment mentioned above.

In the map MAPaI (Fig. 18), since the brake force is likely to be applied by the cylinder internal pressure external force FcI at a time of opening the valve in accordance with the increase of the cylinder internal pressure external force FcI, it is necessary to finish the

brake current in an early stage. Accordingly, the brake ^' current application end displacement. Al is moved to the valve closing side in accordance with the increase of the cylinder internal pressure external force FcI. Further, if ^' 5 the neutral position deviation comes close to the valve

' closing side, the brake current application end displacement Al is moved to _ι the valve closing side so as to adapt thereto, and if .the neutral ^' position deviation comes close to the valve opening side, the brake current application .end 0 displacement Al is moved to. the valve opening side so as to ' adapt thereto. ^{' '} ■

In this. case, since the brake force is hard to be applied at a time of the valve closing control due to the 5 cylinder internal pressure external force FcI, it is ^' necessary to delay the end of the brake current ^' . Accordingly, a map having the. same tendency as that of the map MAPaI in Fig. 18 is employed, however, the ^' map value is ^■ different . ^• ■ ■ ^{• ■} . 0 ^' ■ ^{' ■}

Next, the- attraction current application start displacement A2 is ^' set (S702) . In this case, a relationship A2 < Al i's established. The attraction' current application start displacement A2.corresponds to a threshold value for 5' judging the displacement starting 'the attraction force

. applied to- the armature 8 which is going to be disconnected from the ^' upper core 12, on the sliding mode, which has the same tendency as that of the attraction current application start displacement A in the first embodiment. 0

Accordingly, a map MAPa2 shown in Fig. 19 having the same tendency as that of the map MAPa shown in Fig. 6 is employed. In other words, in the map MAPa2 (Fig. 19) , since it is necessary to supply the attraction current in an early 5 stage against the cylinder internal pressure external force

FcI at a time ^; of opening the valve in ^' accordance with the increase of the cylinder internal pressure, external, force ^' FcI, the. attraction 'current application 'start displacement A2 is moved to ^' the valve closing ^' side in accordance with the increase of the cylinder internal pressure ^' external force ^{' 1} FcI. Further, if the neutral ^■ position deviation comes close to the valve closing, side, the /attraction current, application start displacement ^' A2 _. is moved to the valve closing side so as to adapt thereto,- and if .the neutral ^' position ¹ deviation comes close ^' to' the valve opening side,

" the attraction current application start displacement A2 is • moved, to the valve opening side so ¹ as to adapt thereto.

' In this case, at a time of the valve .clos'ing control, a map having the same tendency as that of the map MAPa2 (Fig. 19) is employed, however, a relationship A2 > Al is established. . ,. ^■

1

If the brake current .application end displacement Al and the ' attraction current application start displacement A2 are set, it is judged whether or not the displacement x of the poppet valve 6 detected by the lift sensor 18 is below the brake current application end displacement Al (S704). In, this case, if it is, just after the valve opening control, ^■ and a relationship x ≥ Al is established (no in S704), the _v .current application start to the upper coil 12a in accordance with the sliding mode control is executed (S706) . Accordingly, the current application for the braking force is executed to the upper coil 12a from the timing t20 in Fig. 16.

In other words, in this timing, the calculation of the control current Iup applied to the upper coil 12a is executed by providing the ECU 2 with the function for storing the physical model formula defining the current

amount applied to the upper coil 12a on the basis of- the ^' gap _. Gp and the control input U in the ECU.2, in accordance with the sliding mode control described in the first embodiment mentioned ^' above. Further, the map defining the ^' relationship among the gap Gp, the control input > U and the • supply current ^• •amount to the upper coil 12a may be stored in the ECU 2. In this case, if the ^' control input U is negative, the current application control amount is set. to "0".

Thereafter, as far as the ^' relationship x ≥ Al is

' established (no in S704), the braking force on the basis of ^■ the electromagnetic force of the upper core 12 is applied ^" to the armature .8 in the sliding mode control.

^' Further, if _. ^' the- relationship x < Al is established on the basi ^' s of the movement of the armature 8 (yes in S704), the current application to the upper coil ^' 12a is ■ finished ^• .(S798: timing t21 in- Fig. ^' 16) . ^{■ •} .

^■ Next, it is judged whether or not the ^' displacement x is below the attraction current application start displacement ■A2 (S710) . In this ^' case, if .a relationship x > .A2- is established (no in S710) , the present . process ^' is temporarily finished. Thereafter,, as far as the relationship x > A2 is established (no -in S710), the armature 8 and the poppet

.valve ^' 6 are moved to the valve opening side o ^' n the basis of the urging force of the upper spring 16 without current application to the upper coil 12a and the lower coil 10a.

Further, if the relationship x < A2 is established (yes in S710) , the current application start to the lower coil 10a is executed in accordance with the sliding mode control (S712: timing t22 in Fig. 16). The contents correspond to the description of the first embodiment (Fig. 5: S104 to S116) .

•In the structure mentioned above, the correspondence to claims is as follows. A combination of 'the poppet ' valve 6 and the armature 8 corresponds to a valve body ( ^' a controlled object)'. The cylinder internal pressure sensor 2b, the port ^■ pressure sensor 2c, and the ECU 2 for calculating the cylinder internal ^' pressure external force FcI on .the basis of the , cylinder internal pressure Pel and the port pressure Pbt detected by these sensors 2b and 2c correspond to a ^' disturbance detecting section. ^' The fully closed position ' measuring process (Fig. 7), the fully open position measuring process (Fig.- 8), the neutral position measuring ¹ process (Fig. 9), ^' the neutral position deviation amount ^' calculating process per engine stop (Fig. ,11) 'and the long- cycle neutral position deviation amount averaging process

(Fig. 13 ^' ) correspond to processes performed by a change with ^' time detecting section. The- .steps S700 to S704 and S710- in the valve opening control process (Fig. 17) correspond to ^' ■processes performed by ^' an .operation switching point changing ^' section: • \ •

In accordance with the second embodiment described above, the following advantages- are obtained. ^' • ^'

^■ In addition to the effects (1) and (2) of the first .embodiment mentioned above, the following advantages are generated'. . ^' •

In the sliding mode control applying the brake current together with the attraction current, the deviation from the proper displacement is generated about the brake current application end displacement Al due to both of the cylinder internal pressure external force FcI and the change with time of the springs 14 and 16.

Accordingly, in the valve opening control process (Fig. 17) , the proper brake current application end displacement Al _. and the attraction current application start displacement

A2 are set by changing ^; the brake' current application end displacement Al together with the attraction current

^■ application start displacement A2 in correspondence to both of the cylinder internal pressure external force .FcI and the neutral position, deviation amount Dmp caused by- the springs 14 and 16.- Accordingly, it is possible to. prevent the ^' destabilization in the sliding ^' mode control.

[Third Embodiment] ^' _ •

' in the sliding mode control in accordance with the . first and second '.embodiments mentioned above, as shown in each of the. Figs. 15 and 16, the holding current applied to the lower coil 10a is controlled in such a . manner as _. to "float the armature '8- to a target floating position Sf in ■place of bringing the armature 8..into contact with the lower core 10 ' at a time of the valve opening control.

In' the target floating position Sf, the proper value is changed by the cylinder internal pressurβ .external force FcI and the neutral position deviation Dmp. In the present embodiment, the -target floating position. Sf is changed in , •correspondence to the cylinder internal pressure external ■ force FcI' and the neutral position deviation ^' Dmp in accordance with a target floating position setting process shown in Fig. 20. The other structures are the same as those of the first or second embodiment. Accordingly, a description will be given with reference to the drawings in each of the embodiment in addition to the new drawings.

A description will be given of the target floating position setting process (Fig. 20) . The present process

corresponds to a process executed just before the. valve ^• •. opening control process, (Fig. 5 or 17) ^' . In the present process,, the target 'floating position Sf is calculated on the basis ^' 'of the cylinder ^' internal pressure external force FcI corresponding to the disturbance and the neutral

^■ position deviation Dmp corresponding to the change with ^' • time, ^' in accordance .with ■ a map 'MAPsf (S800) ^' . In this case, each of the values of the cylinder internal pressure external force FcI and the ^' neutral position deviation Dmp ^' are obtained in the same manner as the first embodiment ' mentioned above .

The, map .MAPsf is shown in Fig. 21. In this map MAPsf, • since the armature 8 is hard to- come ^' into collision with the lower core 10 in ' accordance with the increase of the ^■ cylinder'- internal pressure external force FcI, the target floating position Sf is moved, close to the valve opening- side . ^" • ^' • ' ^'' ■ . ^•• . . ^■

If' the neutral position deviation -comes to- the valve closing side, the target floating position Sf is moved to .the valve closing side so as _, to adapt ^' thereto, and if the . , neutral position deviation Dmp comes to . the valve 'opening side, the- target floating position Sf .is moved to the valve ^• opening side so.as to adapt thereto.

Accordingly', in the valve opening control process- (Fig. 5 or 17), the target floating position Sf is changed in correspondence to the cylinder internal pressure external force FcI and the neutral position deviation Dmp in the sliding mode control applied to the lower coil 10a executed for holding to the opening state.

In the structure mentioned above, the correspondence ^" to claims is as follows. A combination of the poppet valve 6

and the armature 8 corresponds to a valve body (a controlled object) . The cylinder internal ^" pressure sensor 2b, the port, pressure sensor 2c, 'and the ECU 2. for calculating the cylinder infernal pressure external force FcI on the basis- of the cylinder internal pressure Pel and ^' the port pressure ^■ Pbt detected by these sensors- 2b and.2c correspond to a disturbance detecting section. ^' The fully closed .position measuring proces.s (Fig. 7) ^' , ^' the fully open position measuring process (Fig. 8), the neutral position measuring ^' process (Fig. 9), the ' neutral position deviation amount

" calculating process per engine stop (Fig..11) and the^ long-. ^• pycle neutral p.osition deviation amount averaging process ^' (Fig. 13)- correspond to processes performed by a change with ^' time detecting section; The target floating ^■ position setting process _. (Fig: 20) corresponds to a process performed ^* by_ a target floating position changing section.

In accordance ^' with the present third embodiment

^■ described above, the following advantages- are obtained. (1) Since the target floating position Sf is set to the proper position' in correspondence t.o the disturbance and the change with time as ^' well as the effects of the first or second embodiment mentioned above are - generated, it is . possible to prevent the destabilization in the sliding mode ^• control and it, is possible to- improve the seating speed in, .■the valve opening side.

[Fourth Embodiment]

The present embodiment is different from the first to third embodiments mentioned above in a point that a process of preparing and correcting the relationship between, the displacement x and the lift sensor detected voltage V expressed by the map shown in Fig. 12, the function or the like is executed, and the other structures are the same as

any one of the first to ^' third embodiments." Accordingly, 'a . description will.be given with reference to the drawings in the first to third embodiments in, addition to the new- drawings . ^{' ' •' ' !} \ ^{' ■ ■} ,

^■ • ^■' ■ ^' There are executed a shipment correcting process as shown ^' in Fig. 22 as the preparing process of the relationship f, and an on-board correcting • process as shown in Fig. 24- as the correcting proces.s- of the relationship, f. ^' - ^■ _ ^{' ■ ' ■ ' '}

A description will be given of the shipment correcting process (Fig ^' . 22) . The present process is executed' _. by a - ^' me'asuring computer at a time of finishing the assembly of the ^' electromagnetic valve 4. Alternatively, .it is executed by the measuring ^' computer at a time of installing the electromagnetic valve 4 in the internal combustion engine. Ih addition, it may be _. executβd by the ECU _. 2 for control- at ^' a time of installing- the electromagnetic, valve 4 in the 'internal combustion engine. In the case of being corrected by the measuring computer, the corrected value is written ^' as a corrected value ^• in the non-volatile memory of the 'ECU 2. ■In the present embodiment, the process is ^' executed by the measuring ^• computer provided in a measuring instrument in which the- electromagnetic valve 4 is set after the assembly ^■ of the electromagnetic valve 4 is finished.

In this case, it is assumed that- an adjustment of ^' setting it at a proper neutral position with respect to the electromagnetic valve 4 as a hardware is finished in the final step of the assembling stage before being set to the measuring instrument.

If the shipment correcting process (Fig. 22) is executed, the measuring computer first applies the current to the lower coil 10a of the electromagnetic valve 4 by the

drive apparatus provided in the measuring instrument-, and . moves the poppet valve ,6 and the/ armature .8 to the fully ^' open position (the displacement Xmin) (S900) . Further, the detected voltage V of the lift sensor 18 in the ^' fully open state is stored in the fully open time measuring voltage Va ■ ^■ (S90 ^' 2) . ^{' • •} - . . - ^{' ■ ■}

Next,:- the current application to the lower coil l ^' Oa is stopped, the current is applied to the upper coil _, 12a, and the poppet ^" valve 6 and the armature 8 are moved to the ^' fully ' closed position (the displacement Xmax) (S90.4). Further, the detected _. voltage V of the lift sensor 18 in the. fully ^' closed state .is stored -as the fully closed time measured voltage Vb (S906) . ■ ' :

Next, the poppet valve 6 and the armature 8 are moved to the neutral position (the .displacement Xz) by stopping the current application to the upper coil 12a, ^• that is, ■stopping the current application .to both the- coils 10a and 12a (S908) . Further, .the detected voltage ^' V of the lift sensor 18 in the neutral position state is stored as the neutral time measured voltage Vc (S910) . ^' In this neutral position measurement (S908, S910), the neutral time measured voltage Vc may ^• be ,determined as the moving average value by ^■ executing the steps S400 to S418 of the neutral position , , .measuring process (Fig. 9) executed in the first embodiment mentioned' above . ^' Further, the neutral time measured voltage Vc may be determined by averaging the detected voltage V of the lift sensor 18 in the neutral position state obtained by executing both of the movement from the fully closed position to the neutral position and the movement from the fully open position to the neutral position at plural times, or moving averaging them.

Next, a fully open time voltage initial relationship

value Vamap is calculated on the basis of the ^■ value of the fully open position (the displacement Xmin) determined on design or obtained by actual measurement from an initial relationship' g ^' expressing a relationship between the ^{' '} displacement x and the detected voltage V previously stored ^• -within the measuring computer- as an initial map or an initial function and shown in Fig. 23 (S912) .

A fully closed time voltage initial relationship value Vbmap is calculated on the basis of the value of the fully ^" closed position (the displacement Xmax) from the initial ^' relationship . g in the same manner (S914)-. ^' Further, -.a neutral time voltage initial relationship value Vcmap is calculated on the basis of the value of the neutral position (the displacement Xz) from the relationship g (S91 _. 6) .

Further, a deviation of ,the voltage between the _. ■ ^{' •} actually measured Value and the initial relationship value •is calculated so as to ^' be .averaged. as shown _^ by an expression 15 at the fully open position, the fully closed position and the neutral position, and is calculated as an average deviation amount δVs _. (S918) . , .

[Expression 15] . . ■ ^• δVs <- { (Va-Vamap) + (Vb-Vbmap) + (Vc-Vcmap) } /3

Next, the .value of the initial relationship g previously stored within the measuring computer and shown in Fig. 23 is corrected by the average deviation amount δVs, that is, the initial relationship g is offset at the average deviation amount δVs, whereby the relationship f is prepared and is stored in the memory for the ECU 2 (S920) .

The relationship f prepared as mentioned above is used as the relationship f described by Fig. 12 in the first

embodiment mentioned above. ^{' • •}

A flowchart in 'Fig. 24 shows, an on- ^í -board correcting process executed by the. ECU 2 for correcting the relationship f with respect to the electromagnetic .valve 4 ' after being installed ^' in the internal combustion engine. The present process -corresponds to a process which is repeatedly executed in a short ^' time cycle.

If the present process, is .started, it is judged whether; ' or not the current' application to both the coils 10a and 12a' of the electromagnetic valve 4 is stopped for stopping the ^' operation ^• of the internal combustion engine (SlOOO). In the case that the electric current application is 'not stopped : (no in SlOOO), the present process ^" is temporarily finished. ^•

If the current application is /stopped, (yes in SlOOO), the current displacement x is calculated in accordance with ^■ the eqμation of motion ^' of • the poppet valve θ and the ^' armature 8 corresponding to the controlled ^' object (S1002).

In the sliding mode control executed by the ECU 2 described in the fir.st embodiment mentioned above, the > expression 4 can be expressed by an expression 16 in a state in> which the electromagnetic force is not applied. ^• _ . * ^{" ■ • •'} ,

■ [Expression 16]

After calculating the current displacement x from the equation of motion, it is judged whether or not the calculated displacement x corresponds to the neutral position (S1004). If it does not correspond to the neutral position (no in S1004) , the present process is temporarily finished. The process in the step S1002 is repeated until

the displacement x calculated from the equation of motion • mentioned above comes to the neutral position (no in.S1004).

If the ^' displacement x comes' to thee neutral position .. 5 (yes in' S1004), the detected voltage V of ^" the lift sensor 18 'in this timing is set to a ^' calculation neutral position voltage Vzcal ,(S1006) . • . .

Next,- a relationship neutral position. voltage Vzmap at 0 the neutral position is calculated on the basis of the " relationship f stored in the memory of the ECU 2 (S1O08) .

Further,, a calculation map deviation amount δVcm is calculated in accordance with an expression 17 ^' (SlOlG) . , 5 _. ^{' • '} : ^•

[Expression 17] _t ■

δVcm <- Vzcal. - Vzmap , . ^■

^{■ Further, the relationship f 'is updated by the, ' 0 calculation map deviation amount δVcm (S1012) . In ' other words, a new relationship f is set by executing ah offset •moving the original " relationship f. in- the- voltage direction ■ • as shown in Fig. 25 at an amount of the ■ calculation ' map deviation- amount δVcm.. . Accordingly, the relationship f is ,5 ■ changed in correspondence to the case that the change with ..time is generated in the electromagnetic valve 4 after being installed ' in the internal combustion engine. '}

In the structure mentioned above, the correspondence to 0 claims is as follows. In addition to the relations mentioned in each of the embodiments, each of the shipment correcting process (Fig. 22) and the on-board correcting process (Fig. 24) corresponds to an adjusting method of the sliding mode control apparatus. 5

In accordance with the fourth embodiment mentioned above, the following ^' advantages are obtained.

(1) It ^' is possible to obtain ^' the advantages of any one 5 of the ^' first to third embodiments to which the present

^■ embodiment is applied. ^"

(2) I-n the .shipment correcting process (Fig. 22) , the measured voltages Va, Vb and Vc are 'obtained by the lift .

10 sensor 18 in the state in which, the poppet valve 6 is, positioned at the known displacement (the fully open state, ^• the fully closed state and the neutral position) (S900 to ^' S910) . Further, the initial relationship values .Vamap, Vbmap and Vcmap are determined by- converting- tfhe ^' respective .

15 known displacements ^■ by the initial relationship g (S912 to - S916) . ^{" •} , ^' . ^' . - ^■

.' ' ^' Further, the average deviation amount δVs' is _. calculated

: as the corrected value on -the basis of the ^' difference 20 between' the measured voltages Va, Vb and Vc and the initial relationship values Vamp, Vbmap and _. Vcmap (S918), and the ■relationship f used ^' for the actual, sliding mode .control is ' determined by correcting the initial relationship g by the average deviation amount δVs (S920) . 2.5 - _.

It is possible to set the relationship f expressing the detected voltage of the lift sensor 18 and the displacement of the poppet valve 6 accurately by determining two detected voltages on the basis of the respective known displacements 30 and correcting the initial relationship g on the basis of the differences.

Accordingly, it is possible to prevent the destabilization in the sliding mode control in accordance 35 with the machine error of the electromagnetic valve 4.

(3) Since the on-b.oard correcting process (Fig. 24) exists, it is possible to obtain the detected voltage Vzcal in the specific displacement state (the neutral position in this case)- of the poppet valve 6 in ' ^■ accordance with the

^■ •equation of motion in the ECU 2 without using the measuring instrument, even after the electromagnetic valve 4 is installed in the. internal combustion engine (S1002, S1004 and S1006) ^■ . Further, it is possible to obtain the. detected voltage Vzmap of the lift sensor 18 by converting the

' specific displacement in accordance with the relationship f mentioned above, (S1008) . It is possible to update the relationship, f in such a manner as to accurately express the detected voitage of the lift sensor 18 and. the ^' displacement of the poppet valve _: 6 by correcting the relationship ^' £ on the basis of the difference AV ₁ cm (SlOlO) between two detected voltages .determined as mentioned above .(S1012) . ^" Accordingly, it is possible to prevent the dest ^' abilization •in the sliding mode control in accordance with the machine error of the electromagnetic valve 4.

Therefore, even . if the change with, time is generated and the actual relationship between the detected voltage of the lift sensor 18 and the displacement of the poppet valve 6 is changed, it- is possible to return the relationship f to the accurate state by correcting as mentioned ^' above. Accordingly, it is possible to prevent the destabilization in the sliding mode control in accordance with the change with time of the electromagnetic valve 4.

Particularly, just after stopping the operation of the internal combustion engine, there is employed the equation of motion in the case that both the coils 10a and 12a of the electromagnetic valve 4 do not generate the electromagnetic force, but the poppet valve 6 and the armature 8 are

operated oniy by the springs 14 and 16.- Under the circumstance mentioned above, there are not much factors causing the disturbance at a time- of obtaining the data for correcting the relationship f,- and it is possible to correct more accurately. ^l

(4) Even if the on-board correcting process ^■ (Fig. 24) does not exist in the ECU 2, it is possible to cope even with the- change with time by the shipment correcting process (Fig. 22), ^' in .the case that • the electromagnetic valve 4 is detached from the internal combustion engine- and is set to the measuring instrument.

In other words, .it is possible to return the relationship f to the accurate state by detaching the electromagnetic valve 4. from the internal combustion engine and attaching to the measuring instrument as mentioned above, so as to measure, and correcting the map data within the ECU ^• 2 so as to rewrite. Accordingly,- it is possible to prevent the destabilization in the sliding mode control in accordance with the change with time of the electromagnetic valve 4. . • ^' • . ^• .

[Fifth Embodiment] , ^■

The cylinder internal pressure external force FcI in each of the embodiments is calculated by the pressure difference between the cylinder internal pressure Pel detected by the cylinder internal pressure sensor 2b, and the port pressure Ppt detected by the port pressure sensor 2c. In place thereof, as in the present embodiment, the cylinder internal pressure external force FcI is calculated in accordance with the physical equation without the cylinder internal pressure sensor 2b and the port pressure sensor 2c, thereby being used for the sliding mode control

sueh as in each of the first to fourth embodiments.

The cylinder internal pressure external force calculating process _, is shown in Fig. 26. The present process ^' corresponds to a process which is repeatedly . ^' executed in a rotation cycle of a fixed crank angle (for example, a fixed crank angle between 3 degrees and 6 degrees) of the internal combustion engine. If the present process is- started, the cylinder internal pressure Pel is- first calculated ^' (SlIOO) . . . ■

. ' The cylinder internal pressure Pel can be, estimated in accordance .with expressions 18 to 23 obtained by taking into consideration a heat insulation and a fuel- heat generation model in the present embodiment. A cooling loss, a time loss and a pump loss- are excluded, however, it is possible to calculate by taking each of the losses into consideration. ^{■ ' ■} . ₍

[Expression 18] ^■ . ^" ' ^' .

- Vd _n + AQ _n r PrCIl _n — -

2VcI _n - VcI _n _ _λ

[Expression 19 ] dQ . Q

[Expression 20 ]

[Expression 21 ]

1

• Vπ l - ⁷ ^ ^{v m} f = V (Alll

AFR 100

[Expression 22]

a dx = ^-(m+ϊ)- y ^M - expi-a -y" ¹ *)

Aa _c

[Expression 23 ]

Aa,.

PcIn 'corresponds to an estimated cylinder internal pressure [Pa] at the present sample 'time, and PcIn-I- . ^• corresponds to an estimated- cylinder internal pressure [Pa] - at the previous sample time. Vein corresponds to a cylinder ^' volumetric capacity at the present sample time, [cubic meter] , VcIn-I corresponds to a cylinder volumetric capacity at the previous sample time [cubic meter] , ^■ VaIl corresponds to a total cylinder volumetric capacity [L], δQn corresponds ^' to a fuel heat generation amount at the present sample time [J] , and Q corresponds to a total fuel heat generation ^'

^' amount [J] . dQ corresponds to a heat generation amount by a ^• vibe model [J] , dx corresponds to' a heat generation rate obtained from the vibe model [1/deg] , Hu corresponds to a low calorific power of the fuel [J/kg: about 44 MJ/kg in case of gasoline], mf corresponds to a fuel mass [kg], v v- corresponds to a volumetric efficiency, ^' and AFR corresponds to -an. air-fuel ratio (about 14.5) . a and m correspond to a shape parameter Of the vibe, ά corresponds to a crank angle, •αO corresponds to a combustion start, crank .angle, and δαc corresponds to a combustion period. The ^' shape parameters a and m of the vibe, the combustion start crank angle αO and the combustion period δαc are calculated by the map in accordance with the internal combustion engine rotating speed and the internal combustion engine load rate (or further including the air-fuel ratio) .

The cylinder volumetric capacity VcI (Vein and VcIn-I) can be calculated by an expression 24.

[ Expression 24 ] . ^■ _

VcI = Abr - ((l + r) - costøp) - r ^■ cos(a + ψ) -l - SQRT(I - (r / / • sin(α + ψ) - e 11) ² )) + v ₀

r ^' corresponds to a crank radius, 1 corresponds to a connecting rod length, e corresponds to a crank pin offset, ^' Abr corresponds to a bore area, vθ corresponds to a combustion chamber volumetric capacity, and ψ = arcsin(e/(r

+ D ) • ^{• " • ' ■ s ■ •} • - ^■ . ^{■ ""} _. . _. .

In accordance with the calculation mentioned above, the ^' cylinder internal pressure Pel is calculated in ^' . correspondence to the crank angle ( ¹ CA) as shown in Fig. 27Bl Fig. 27A shows .a ^■ crank angle change -of the heat generation amount dQ ^' in accordance with the vibe model.

Next, the port pressure -Ppt [Pa] is calculated (S1102). In the case of -the exhaust port, it is calculated as shown :by an expression 25 from- the physical equation. In the case .of the intake port, it is as shown by an expression 26.

[Expression 25] . ■ .

P _J pf = (100+^^-._V _β )/1000 - . ^■ 7000 . . _. . .

■ [Expression 26]

P>f = 77 _v /lOOθ ^'

Ne corresponds to an internal combustion engine rotating speed [rpm] .

The cylinder internal pressure external force Ft is sequentially calculated as shown by an expression 27 on the basis of the cylinder internal pressure Pel calculated as mentioned above (S1104) .

[Expression 27] ^' .

^'■ d _v corresponds to a diameter. of the poppet valve 6.

Next,- it is. judged whether or not the poppet valve 6 is in the operation start timing (S110-6) . The operation start corresponds to a timing for .starting the valve opening- ^' operation under the valve closing state, and- a timing for starting the valve closing operation under the-, opening state. • . •

If the valve operation start timing is not established (no in SllOβ) , the present proqess is temporarily finished. Accordingly, in the state in which the valve operation start timing is not established (no in S1106) , the' pfoce.sses of the steps SIlOO to S1104 'are repeated, and the cylinder internal pressure external force Ft is repeatedly calculated • sequentially.

If the valve operation start timing ¹ for opening the valve , or closing the valve is established (yes in SllOβ) . ^■ The sequential .cylinder internal pressure external force Ft calculated in the immediately preceding steps SIlOO to S1104 is set as ^' the cylinder internal pressure external force FcI (S1108) .

Accordingly, in each of the first to fourth embodiments mentioned above, the cylinder internal pressure external force FcI is determined, and the attraction current application start displacements A and A2, the brake current application end displacement Al and the target floating position Sf are determined by using it together with the

neutral position deviation. ^' . ^•

In the valve opening control . of the exhaust valve, for example, as 'shown in a timing chart in Fig. 28, the cylinder internal pressure external force- FcI in the initial stage ^■ (t30) of the valve opening timing is known. Accordingly, it is possible to thereafter execute the proper sliding mode control. - ^'

In the structure mentioned. above, the correspondence to _. claims is as follows. In the relations mentioned in each of' the embodiments, the relationship is different from the other embodiments in a point that the cylinder internal pressure external force calculating process (Fig. 26) corresponds to a /process performed by a disturbance detecting section.

In accordance with the fifth -embodiment- mentioned above, the following advantages are obtained;

(1) It is possible to obtain the advantages of any one of the first to fourth embodiments .mentioned above to which ' the present embodiment is applied.

(2) Since the cylinder internal pressure external force FcI can be detected without using the cylinder internal pressure sensor 2b and the port pressure sensor 2c, it is possible to simplify the entire system with respect to the electromagnetic valve 4.

[Other Embodiments]

(a) In each of the embodiments mentioned above, the maps MAPa, MAPa2 and MAPaI for calculating the attraction current application start displacements A and A2 and the

brake current application end -displacement Al, and the map MAPsf for calculating the target ^' floating position S.f ^' are calculated by the cylinder internal pressure external force FcI and the 'neutral position deviation Dmp. In addition, it is possible to form' a map for calculating the attraction ^■ •current application start displacements. A and A2 and the brake ^' current application end displacement Al only by the cylinder internal pressure external force FcI, or a map for calculating the target floating position Sf. Further, it. is possible to ^" form a map for calculating the attraction

^■ current application start displacements A and A2 and the brake current application end displacement Al .only by the ^' neutral position- deviation Dmp, or a map for calculating the ^" target floating position Sf. . ^{' ' ■}

(b) ^' In the second embodiment, -the brake .current application end displacement Al and the attraction current application start displacement A2 'are respectively •determined from the maps -MAPaI and MAPa2, however, the structure ■ may be made such that any one of them is set to a fixed threshold value and only the' other is calculated by the map. In accordance with this structure, it is also possible to prevent the destabilization in the sliding mode control . ^■ _, -

In the third embodiment mentioned above, ^' the structure- may be made such that the attraction current application start displacement A or the brake current application end displacement Al and the attraction current application start displacement A2 is set to a fixed threshold value, and only the target floating position Sf is calculated by the map MAPsf. In accordance with this structure, it is also possible to prevent the destabilization in the sliding mode control .

(c) In each of all the embodiments, the application ^' of the maps MAPa, MAPaI, MAPa2 and MAPsfmay be provided ^' in each of the electromagnetic valves, or only one map for an average electromagnetic valve is provided so as to be applied to the control of the electromagnetic valve.

^" (d) In the fourth embodiment, the average deviation amount, δVs- is calculated at ^' three, points comprising the fully open- position, the fully closed position and the neutral position in the shipment correcting process _. (Fig. ^• 22), however, the average deviation amount δVs may be calculated by measuring only the neutral point- Alternatively, the average deviation amount δVs may be calculated by measuring four points or more. 'Further, a , new ■ relationship f may be formed by the measured value itself by- measuring ten or more points.

In the on-board correcting process ^' (Fig. 24), it is ■possible to set the detected voltage V obtained after a time delay corresponding to the response ^' of the lift sensor 18 from the timing -.at which the displacement x calculated from the equation of motion comes .to the neutral position, to the ¹ calculation neutral position voltage Vzcal.

^• Further, the relationship f may be calculated on the , , basis of an ^' approximation in accordance with a least square method in place of offsetting the relations g and f by the average deviation amount δVs or the calculation map deviation amount δVcm.

(e) In the fifth embodiment mentioned above, there is employed the cylinder internal pressure external force FcI in the early stage of the valve opening timing, however, it is possible to utilize for the sliding mode control by determining the cylinder internal pressure external force

FcI with respect to the other timings than the early ^' stage.