Exhaust Particle Removing System for an Internal Combustion Engine

This invention relates to a system for removing particles from exhaust produced by internal combustion engines, such as diesel engines.

Exhaust produced by diesel engines has a relatively high content of polluting particles composed of carbon, unburned fuel, and partially burned fuel. Filters disposed in engine exhaust systems are conventionally used to remove the particles from the exhaust. In this case, burners positioned in the exhaust systems upstream of the filters are often employed to burn off the particles deposited on the filters in order to unclog and rejuvenate the filters.

Japanese patent publication No. 56-115809 discloses such a particle removing system. In this system, the pressure across the filter is monitored as an indication of the degree of clogging of the filter. When this pressure rises to a preset level, the burner is activated to unclog the filter. The pressure across the filter, however, inaccurately represents the degree of clogging of the filter for the following reason: this pressure depends on not only the degree of clogging of the filter but also the rate of exhaust flow. This inaccuracy in the recognized degree of clogging of the filter could allow clogging to an unacceptable level or wasteful overuse of the burner.

It is an object of this invention to provide an exhaust particle removing system for an internal combustion engine which accurately determines the degree of clogging of a filter and thereby reliably holds this degree to within an acceptable range.

In accordance with this invention, an exhaust particle removing system includes a filter and a burner. The filter is disposed in an engine exhaust passage to trap particles suspended in exhaust gas. The burner serves to burn off the particles deposited on the filter. The pressures in the exhaust passage at points upstream and downstream of the filter are sensed. A determination is made as to whether or not the sensed upstream pressure is lower than a preset level. When the sensed upstream pressure is not lower than the preset level, the degree of clogging of the filter is deduced on the basis of the sensed upstream and downstream pressures. When the sensed upstream pressure is lower than the preset level, the degree of clogging of the filter is deduced on the basis of a time elapsed since the moment at which the sensed upstream pressure dropped below the preset level and on the basis of the sensed upstream and downstream pressures obtained immediately prior to that moment.

FIG. 1 is a diagram of an exhaust particle removing system according to a first embodiment of this invention.

FIG. 2 is a cross-sectional view of the burner of FIG. 1.

FIG. 3 is a graph of the relationship between the upstream pressure and the pressure difference.

FIG. 4 is a flowchart of a program for controlling the operation of the control unit of FIG. 1.

FIG. 5 is a detailed flowchart of the pressure sampling step of FIG. 4.

FIG. 6 is a diagram of the waveform of the pressure signal appearing in the system of FIG. 1.

FIG. 7 is a detailed flowchart of the burner activation step of FIG. 4.

FIG. 8 is a detailed flowchart of the temperature control step of FIG. 7.

FIG. 9 is a detailed flowchart of the burner deactivation step of FIG. 4.

With reference to FIG. 1, an exhaust passage 21 extends from a diesel engine (not shown) to conduct exhaust gas from the engine. A casing 22 encloses a chamber. The casing 22 has an inlet 22a and an outlet 22b in communication with each other via the chamber. This casing 22 is disposed with respect to the exhaust passage 21 in such a way that the chamber in the casing 22 constitutes a part of the exhaust passage 21 and that exhaust gas enters the chamber via the inlet 22a and exits from the chamber via the outlet 22b.

A filter 23 of a honeycomb structure is disposed within the downstream half of the casing 22 to trap particles suspended in the exhaust gas. A buffer 24 is sandwiched between the filter 23 and the walls of the casing 22 so that the filter 23 is gently supported within the casing 22. This filter 23 has a plurality of parallel holes extending between the upstream and downstream ends of the filter 23. These holes are of two types. The holes of the first type have open upstream ends and closed downstream ends. The holes of the second type have closed upstream ends and open downstream ends. The first holes adjoin the second holes via porous walls of the filter 23 so that the exhaust gas first enters the first holes and then passes through the porous walls into the second holes before exiting via the second holes. As the exhaust gas passes through the porous walls, particles suspended in these gases are trapped by the walls.

A burner 25 disposed within the upstream half of the casing 22 serves to burn off the particles trapped by the filter 23. As shown in FIGS. 1 and 2, the burner 25 includes a cylindrical combustion liner 26, a mixture vaporizing tube 27, a mixture injection tube 28, and a glow plug 29. The combustion liner 26 defines a combustion chamber in a region immediately upstream of the filter 23. The combustion liner 26 has a plurality of apertures 26a which admit exhaust gas into the combustion chamber. After passing through the combustion chamber, the exhaust gas enters the filter 23. The mixture vaporizing tube 27 is disposed within the combustion liner 26. The mixture injection tube 28 extends through the walls of the liner 26 into the mixture vaporizing tube 27. The mixture injection tube 28 serves to discharge a mixture of air and fuel into the mixture vaporizing tube 27. The discharged mixture reverses its flow and then flows out of the mixture vaporizing tube 27 into the combustion chamber via an opening or openings 27a through the walls of the tube 27. The glow plug 29 projects into a section of the combustion chamber near the opening 27a to ignite the mixture entering the combustion chamber. After igniting, the mixture burns in the combustion chamber and thereby the particles trapped by the filter 23 are burned off of the filter 23. Furthermore, the combustion of the mixture heats the mixture vaporizing tube 27, thereby speeding evaporation of the fuel in the tube 27 and thus facilitating ignition of the subsequently supplied mixture.

As shown in FIG. 2, the mixture vaporizing tube 27 and a member 27b fixed to the tube 27 define an ignition chamber 27c in communication with both of the opening 27a and the combustion chamber. The glow plug 29 includes an elongated cylinder 29a and a sleeve cover 29b. The cylinder 29a extends into the ignition chamber 27c and is exposed to the chamber 27c. This exposed section of the cylinder 29a contains a heating element. The sleeve cover 29b surrounds a segment of the cylinder 29a extending in the combustion chamber and the space between the combustion liner 26 and the casing 22 in order to protect this segment of the cylinder 29a from a blaze. The cover 29b is preferably made of a heat resisting alloy.

Returning to FIG. 1, an electromagnetic-type fuel supply control valve 30 has an outlet connected to the mixture injection tube 28 via a fuel feed pipe 31. Fuel is driven from a fuel tank 32 to an inlet of the fuel valve 30 by an electrically-powered fuel pump 33. This fuel may consist of the same petroleum fuel used to run the engine. When the fuel valve 30 is open, fuel is admitted into the mixture injection tube 28 via the mixture feed pipe 31 provided that the fuel pump 33 is activated. When the fuel valve 30 is closed, fuel supply is interrupted. Electrical energization and de-energization of the fuel valve 30 causes the fuel valve 30 to be opened and closed respectively.

An engine-driven air pump 34 has an inlet 34a and an outlet 34b. The inlet 34a leads to the open air via an air cleaner (not shown). The air pump 34 draws air via the inlet 34a and discharges it via the outlet 34b. A pressure-operated three-way valve 35 has a control chamber partially defined by a spring-loaded diaphragm, a first port 35a, a second port 35b, and a third port 35c. The first port 35a is connected to the air pump outlet 34b. The second port 35b leads to the open air via the air cleaner. The third port 35c is connected to the mixture injection tube 28 via an air feed pipe 36. When the control chamber is supplied with atmospheric pressure, the first port 35a is connected to the second port 35b and is disconnected from the third port 35c so that air driven by the air pump 34 is relieved via the valve 35 and thus the mixture injection tube 28 does not receive any air. When the control chamber is supplied with a preset vacuum, the first port 35a is connected to the third port 35c and is disconnected from the second port 35b so that air driven by the air pump 34 enters the mixture injection tube 28 via the valve 35 and the air feed pipe 36.

Adjustment of the pressure in the control chamber of the valve 35 is realized via an electromagnetic three-way valve 37 having a first port 37a, a second port 37b, and a third port 37c. The first port 37a is connected to the control chamber of the valve 35. The second port 37b is connected to a vacuum source, such as an engine-driven vacuum pump. The third port 37c leads to the open air via the air cleaner. When the electromagnetic valve 37 is electrically de-energized, the first port 37a is connected to the third port 37c and is disconnected from the second port 37b so that atmospheric pressure is supplied to the control chamber of the valve 35, thereby interrupting the air supply to the mixture injection tube 28. When the electromagnetic valve 37 is electrically energized, the first port 37a is connected to the second port 37b and is disconnected from the third port 37c so that the preset vacuum is supplied to the control chamber of the valve 35, thereby enabling the air supply to the mixture injection tube 28.

A control unit 50 includes a switch section 51 having switches 51a, 51b, 51c, 51d, and 51e. Each of these switches 51a, 51b, 51c, 51d, and 51e consists mainly of a switching power transistor. The fuel pump 33 is electrically connected across a battery 52 via the switch 51a and an engine key switch 70. Provided that the key switch 70 is closed, the fuel pump 33 is electrically energized and de-energized when the switch 51a is closed and opened respectively. The fuel valve 30 is electrically connected across the battery 52 via the switch 51b and the key switch 70. Provided that the key switch 70 is closed, the fuel valve 30 is electrically energized and de-energized when the switch 51b is closed and opened respectively. The electromagnetic valve 37 is electrically connected across the battery 52 via the switch 51c and the key switch 70. Provided that the key switch 70 is closed, the electromagnetic valve 37 is electrically energized and de-energized when the switch 51c is closed and opened respectively. A relay 53 has a control winding 53a and a switch 53b. Electrical energization and de-energization of the control winding 53a causes the switch 53b to be closed and opened respectively. The glow plug 29 is electrically connected across the battery 52 via the relay switch 53b and the key switch 70. Provided that the key switch 70 is closed, the glow plug 29 is electrically energized and de-energized when the relay switch 53b is closed and opened respectively. The relay winding 53a is electrically connected across the battery 52 via the switch 51d and the key switch 70. Provided that the key switch 70 is closed, the relay winding 53a is electrically energized and de-energized when the switch 51d is closed and opened respectively. Accordingly, provided that the key switch 70 is closed, closing and opening of the switch 51d causes the glow plug 29 to be activated and de-activated respectively. The switches 51a, 51b, 51c, and 51d have control terminals to which electrical signals S1, S2, S3, and S4 are applied respectively to control the switches 51a, 51b, 51c, and 51d. It should be noted that the engine key switch 70 remains closed while the engine is running.

Under conditions where the key switch 70 is closed, when all of the switches 51a, 51b, 51c, and 51d are closed, the fuel pump 33, the fuel valve 30, the electromagnetic valve 37, and the glow plug 29 are all electrically energized. As a result, fuel is driven by the fuel pump 33 into the mixture injection tube 28 via the fuel valve 30 and air is admitted from the air pump 34 into the mixture injection tube 28 via the air valve 35. A mixture of air and fuel thus results, which is discharged into the combustion chamber of the burner 25. The discharged mixture is ignited by the glow plug 29. Thus, in this case, the burner 25 is activated.

In more detail, the fuel valve 30 is energized intermittently at a constant frequency during operation of the burner 25. The rate of fuel supply to the burner 25 is determined by the duty cycle of current pulses flowing through the fuel valve 30.

When all of the switches 51a, 51b, 51c, and 51d are opened, the fuel pump 33, the fuel valve 30, the electromagnetic valve 37, and the glow plug 29 are all electrically de-energized. As a result, the fuel valve 30 interrupts admission of fuel into the mixture injection tube 28 and the air valve 35 interrupts admission of air into the mixture injection tube 28, so that no air-fuel mixture is supplied to the burner 25. Thus, in this case, the burner 25 is de-activated.

An electromagnetic three-way valve 80 has three ports 80a, 80b, and 80c. When this valve 80 is electrically de-energized, the first port 80a is connected to the second port 80b and is disconnected from the third port 80c. When the valve 80 is electrically energized, the first port 80a is connected to the third port 80c and is disconnected from the second port 80b.

A first pressure introduction passage 38 connects the interior of the casing 22 to the second port 80b of the three-way valve 80. The junction of this passage 38 and the interior of the casing 22 is located at a point immediately upstream of the filter 23 so that the second port 80b is supplied with the pressure in the casing 22 at this point. This pressure will be referred to as the upstream pressure P1 hereafter. A second pressure introduction passage 39 connects the interior of the casing 22 to the third port 80c of the three-way valve 80. The junction of this passage 39 and the interior of the casing 22 is located at a point immediately downstream of the filter 23 so that the third port 80c is supplied with the pressure in the casing 22 at this point. This pressure will be referred to as the downstream pressure P2 hereafter.

A pressure sensor 82 has a sensing port 82a and a reference port 82b. This pressure sensor 82 detects the pressure difference between its ports 82a and 82b. The sensing port 82a is connected to the first port 80a of the three-way valve 80. A diaphragm 84 is disposed in the connection between the sensing port 82a and the first valve port 80a to prevent the transmission of heat and moisture but allow the transmission of pressure between them. In this way, the sensing port 82a is supplied with the pressure developed in the first valve port 80a. The reference port 82b leads to the open air via the air cleaner to be exposed to atmospheric pressure. The pressure sensor 82 is supplied with a constant voltage Vcc from a voltage regulator 60 described in more detail hereafter. The pressure sensor 82 outputs a voltage signal VP which represents the pressure difference between its ports 82a and 82b. A multiplexer 57 in the control unit 50 is electrically connected to the pressure sensor 82 to receive the pressure signal VP.

The pressure sensor 82 may have a piezoelectric element and a gauge resistor provided on a silicon diaphragm whose opposing surfaces are subjected to the pressures in its ports 82a and 82b respectively. In this case, the effective resistance of the gauge resistor varies as a function of the difference in pressure between the ports 82a and 82b. The voltage regulator 60 disposed in the control unit 50 and connected to the battery 52 derives a constant voltage Vcc from the voltage VB across the battery 52. This constant voltage Vcc is applied across a series combination of a reference fixed resistor and the gauge resistor, so that the voltage across the gauge resistor or the voltage across the fixed resistor varies as a function of the resistance of the gauge resistor. Since the resistance of the gauge resistor depends on the pressure difference, the voltage across the gauge resistor or the voltage across the fixed resistor represents that pressure difference. This voltage is outputted by the sensor 41 as a pressure signal VP.

The three-way valve 80 is electrically connected across the battery 52 via the switch 51e in the control unit 50 and via the key switch 70. Provided that the key switch 70 is closed, the three-way valve 80 is electrically energized and de-energized when the switch 51e is closed and opened respectively. A control signal S5 applied to a control terminal of the switch 51e selectably determines the operating position of the switch 51e.

Under conditions where the key switch 70 is closed, when the switch 51e is opened, the three-way valve 80 is de-energized so that the first valve port 80a is connected to the second valve port 80b and is disconnected from the third valve port 80c. As a result, the upstream pressure P1 is supplied to the sensor sensing port 82a so that the pressure sensor 82 measures this upstream pressure P1 with respect to atmospheric pressure. Under the same conditions, when the switch 51e is closed, the three-way valve 80 is energized so that the first valve port 80a is connected to the third valve port 80c and is disconnected from the second valve port 80b. As a result, the downstream pressure P2 is supplied to the sensor sensing port 82a so that the pressure sensor 82 measures this downstream pressure P2 with respect to atmospheric pressure.

A temperature sensor 44 is disposed within the casing 22 at a point directly upstream of the center of the end face of the filter 23. The temperature sensor 44 outputs a voltage VT representing the temperature at that point. This temperature sensor 44 may consist basically of a thermocouple.

An engine speed sensor 45 monitors the rotational speed of the engine. This speed sensor 45 includes a crank angle sensor generating pulses corresponding to evenly spaced angular positions of the engine crankshaft. The frequency of these pulses is thus proportional to the engine speed. The speed sensor 45 outputs the resulting pulse signal Rev, the frequency of which indicates the engine speed. A frequency-to-voltage (F/V) converter 56 in the control unit 50 is connected to the engine speed sensor 45 to receive the pulse signal Rev. This circuit 56 converts the pulse signal Rev to a voltage VR which varies as a function of the engine speed.

An engine load sensor 46 monitors the load on the engine. This load sensor 46 includes a potentiometer 46a, the adjustment shaft of which is linked to the control lever 47a of a fuel injection pump 47. The angular position of the control lever 47a determines the rate of fuel injection into the engine. This control lever 47a is connected to an accelerator (not shown) so that the angular position of the control lever 47a reflects the power output required of the engine, that is, the load on the engine. The constant voltage Vcc outputted by the voltage regulator 60 is applied across the resistor of the potentiometer 46a. As a result, the potentiometer 46a outputs a voltage VL which varies as a function of the engine load.

The control unit 50 includes a digital central processing unit (CPU) 54, a read-only memory (ROM) 55, a multiplexer (MPX) 57, an analog-to-digital (A/D) converter 58, and a peripheral input/output (PIO) circuit 59.

The multiplexer 57 is connected to the sensors 82, 44 and 46 and the F/V converter 56 to receive the pressure signal VP, the temperature signal VT, the engine load signal VL, and the engine speed signal VR. The multiplexer 57 selects one of these signals in accordance with a channel selection signal CH issued by the PIO circuit 59 and passes it to the A/D converter 58. This signal CH has four different states corresponding to the four different selections. After receiving a start signal START from the PIO circuit 59, the A/D converter 58 commences converting the selected signal to a corresponding digital signal DATA. Upon completion of the conversion, the A/D converter 58 outputs an end-of-conversion signal EOC to the PIO circuit 59 and then the digital signal DATA is transmitted to the PIO circuit 59.

The PIO circuit 59 also outputs the control signals S1, S2, S3, S4, and S5 to the switches 51a, 51b, 51c, 51d, and 51e of the section 51 via connections between the circuit 59 and the switches 51a, 51b, 51c, 51d, and 51e.

The CPU 54 is connected to the PIO circuit 59 and the ROM 55 holding a program and fixed data. The CPU 54 has an internal random-access memory (RAM). The CPU 54, the ROM 55, and the PIO circuit 59 are connected to the voltage regulator 60 so as to be powered by the constant voltage Vcc. It should be noted that the other circuits 51, 56, 57, and 58 are also powered by this constant voltage Vcc.

Since exhaust gas generally exhibits laminar flow in the filter 23, the pressure P1 in the exhaust passage 21 at a point upstream of the filter 23, the pressure P2 in the exhaust passage 21 at a point downstream of the filter 23, and the pressure difference .DELTA.P across the fitler 23 are all approximately proportional to the rate of exhaust gas flow through the filter 23 provided that the resistance of the filter 23, that is, the degree of clogging of the filter 23 is constant. The pressure difference .DELTA.P equals the upstream pressure P1 minus the downstream pressure P2. At a fixed degree of clogging of the filter 23, the ratio between the pressures P1 and P2 as well as the ratio between the upstream pressure P1 and the pressure difference .DELTA.P thus remain at approximately constant levels independent of the rate of exhaust gas flow.

As the degree of clogging of the filter 23 increases, the ratio .DELTA.P/P1 also increases. In addition, the ratio .DELTA.P/P1 is independent of the rate of exhaust gas flow. Accordingly, this ratio .DELTA.P/P1 represents the degree of clogging of the filter 23.

FIG. 3 shows the typical relationship between the upstream pressure P1 and the pressure difference .DELTA.P. In this graph, the lower line A represents a minimum level of clogging of the filter 23, while the upper line B represents a reference level separating acceptable and unacceptable ranges of clogging of the filter 23. When the ratio of the upstream pressure P1 to the difference pressure .DELTA.P reaches this reference level, the burner 25 should be activated to unclog the filter 23.

At low engine speeds and small engine loads near values realized under engine idling conditions, the upstream pressure P1 is so low that the ratio .DELTA.P/P1 critically or sensitively depends on fluctuations in exhaust pressures and that the accuracy of calculation of the ratio .DELTA.P/P1 is considerably affected by the accuracy of pressure sensing. As a result, errors in deducing the degree of clogging of the filter 23 directly from the ratio .DELTA.P/P1 may be significant in this engine speed and load range. This invention solves this problem as will be apparent from the following description.

The control unit 50 operates in accordance with a program stored in the ROM 55. FIG. 4 is a flowchart of this program.

In the initial step 101 of this program, the current value of the engine speed derived from the engine speed signal VR is stored in the RAM of the CPU 54. In this flowchart, the variable N represents this engine speed value. After the step 101, the program advances to a step 102.

In the step 102, the CPU 54 determines whether or not the engine has started, specifically whether or not the engine speed value N exceeds a preset level Nref1 preferably chosen to be 500 rpm. If the engine has not yet started, that is, if the engine speed value N does not exceed the preset level Nerf1, the program returns from the step 102 to the initial step 101 via a step 119 in which the burner 25 is deactivated. If the engine has started, that is, if the engine speed value N exceeds the preset level Nref1, the program advances from the step 102 to a step 103.

in the step 103, the current value of the engine load derived from the engine load signal VL is stored in the RAM of the CPU 54. In this flowchart, the variable Q represents this engine load value. After the step 103, the program advances to a step 104.

In the step 104, the CPU 54 determines whether or not the burner 25 is active. If the burner 25 is active, the program advances via a step 117 to a step 118 in which activation of the burner 25 is maintained. If the burner 25 is inactive, the program advances to a step 105.

In the step 105, the current values of the upstream and downstream pressures P1 and P2 are sampled and stored as described in more detail hereafter. In this flowchart, the variables VP1 and VP2 represent the upstream and downstream pressure values respectively. As will be made clear later, this step 105 branches either back to the initial step 101 or onward to a step 106.

In the step 106, the CPU 54 determines whether or not the upstream pressure value VP1 is equal to or greater than a preset reference value VP1ref. This reference value VP1ref is chosen so as to represent the lower limit of an acceptable range of the upstream pressure values within which derivation of the degree of clogging of the filter 23 directly from the ratio .DELTA.P/P1 is accurate and reliable. If the upstream pressure value VP1 is equal to or greater than the reference value VP1ref, that is, if the degree of clogging of the filter 23 can be derived directly from the ratio .DELTA.P/P1 with acceptable accuracy and reliability, the program advances to a step 107. If the upstream pressure value VP1 is smaller than the reference value VP1ref, that is, if the above derivation can be neither accurate nor reliable, the program advances to a step 111.

In the step 107, the CPU 54 determines the ratio of the upstream pressure value VP1 to the pressure difference value V.DELTA.P, where V.DELTA.P=(VP1-VP2) and VP2 is the downstream pressure value. In this flowchart, this ratio is represented by the variable K0. Specifically, "K0=(VP1-VP2)/VP1" is executed. It should be noted that (VP1-VP2)/VP1=1-(VP2/VP1).

In a step 108 following the step 107, the CPU 54 determines whether or not this is the first cycle of execution of the program since the last unclogging of the filter 23 was completed. Specifically, a determination is made as to whether or not the value of the variable K1 held in the RAM of the CPU 54 is zero. In general, this variable K1 represents the preceding value of the variable K as will be made clear with reference to a step 116. In any case, the variable K1 will be zero during the first cycle of execution of the program following the completion of an unclogging operation. If the variable K1 is zero, that is, if this is the first cycle of execution, the program proceeds to a step 110. If the variable K1 is not zero, that is, if this is not the first cycle of execution, the program proceeds to a step 109.

It should be noted that since the CPU 54 is continuously powered by the constant voltage Vcc from the voltage regulator 60 independent of the position of the engine key switch 70 as understood from FIG. 1, the RAM of the CPU 54 holds the data even while the engine remains at rest.

In the step 110, the ratio value K0 is stored in a location in the RAM of the CPU 54 which is designated by the variable K in this flowchart. In other words, "K=K0" is executed. After the step 110, the program advances to a step 115.

In the step 109, "K=(K0+7K1)/8" is executed. In this equation, K1 is the variable representing the preceding value of the variable K as will be made clear with reference to a step 116 described hereafter. Accordingly, this step 109 calculates the weighted mean value of the current ratio value K0 and the preceding value of the variable K having components of the previous ratio values K0. This current mean value of the variable K is used as a final indication of the degree of clogging of the filter 23. The weighted mean prevents erroneous determination of the degree of clogging of the filter 23 resulting from questionable abrupt changes in the ratio value K0. The weighting factor of this averaging operation may differ from that used in this embodiment. After the step 109, the program advances to the step 115.

In the step 115, the CPU 54 determines whether or not the value of the variable K is equal to or greater than a preset value Kmax. This reference value Kmax is chosen to correspond to the boundary between acceptable and unacceptable ranges of clogging of the filter 23. If the value K is smaller than the reference value Kmax, that is, if the degree of clogging of the filter 23 is acceptable, the program proceeds to the step 116. If the value K is equal to or greater than the reference value Kmax, that is, if the degree of clogging of the filter 23 is unacceptable, the program proceeds to a step 117.

In the step 116, the value of the variable K is stored in a location in the RAM of the CPU 54 which is designated by the variable K1 in this flowchart. After the step 116, the program returns to the initial step 101.

In the step 117, the value of the variable K1 is set to zero. Specifically, "K1=0" is executed. After the step 117, the program proceeds to a step 118.

In the step 118, the control unit 50 activates the burner 25 to unclog the filter 23. As a result of cooperation of the steps 115 and 118, when the degree of clogging of the filter 23 reaches an unacceptable level, the burner 25 is automatically activated to unclog the filter 23. This step 118 branches to either the initial step 101 or the burner deactivation step 119. After the step 119, the program returns to the initial step 101.

In the step 111, the CPU 54 determines whether or not the time elapsed since the moment at which the upstream pressure value VP1 drops below the reference value VP1ref exceeds a preset interval Tref1. As will be made clear hereinafter, this interval Tref1 determines a timing at which the value of the variable K is to be corrected. If this elapsed time does not exceed the preset interval Tref1, the program returns to the initial step 101. If this elapsed time exceeds the preset interval Tref1, the program advances to a step 112. It should be noted that as shown in FIG. 1, the CPU 54 includes a combination of a fixed frequency clock 54a and counters 54b for measuring this elapsed time and other times described hereafter. As a result of cooperation of the steps 106 and 111, the program advances to the step 112 provided that the upstream pressure value VP1 remains below the preset value VP1ref for more than the preset interval Tref1.

In the step 112, the CPU 54 determines whether or not this is the first cycle of execution of the program since the last unclogging of the filter 23 was completed. Specifically, a determination is made as to whether or not the value of the variable K1 held in the RAM of the CPU 54 is zero. If the value of the variable K1 is not zero, that is, if this is not first execution cycle of the program, the program proceeds to a step 113. If the value of the variable K1 is zero, that is, if this is the first execution cycle of the program, the program advances to a step 114.

In the step 113, the CPU deduces the increase in the value of the variable K which could be expected over a preset length of time equal to the preset interval Tref1 on the basis of the engine speed value N and the engine load value Q sampled in the steps 101 and 103. In this flowchart, the variable .DELTA.K represents this increase value. The increase value .DELTA.K is chosen to correspond to the increase in the degree of clogging of the filter 23 over the interval Tref1. It should be noted that the rate of increase in the degree of clogging of the filter 23 generally depends on the engine speed and the engine load. Specifically, the ROM 55 holds a table in which a set of values of the increase value .DELTA.K are plotted as a function of the engine speed and the engine load. By referring to this table, the CPU 54 determines the increase value .DELTA.K. After determining the increase value .DELTA.K, the CPU 54 executes "K=K1+.DELTA.K". In other words, the CPU 54 calculates the sum of the values of the variables K1 and .DELTA.K and then writes the result in the variable K. In this way, the value of the variable K, serving as a final indication of the degree of clogging of the filter 23, is corrected in cases where the upstream pressure value VP1 drops below the reference value VP1ref and then remains lower than the value VP1ref for longer than the preset interval Tref1. From this step 113, the program advances to the step 115.

The increase value .DELTA.K may be equal to a predetermined constant independent of the current engine speed and load. In this case, the increase value .DELTA.K is chosen on the basis of typical values of the engine speed and load occurring under conditions where the upstream pressure value VP1 is lower than the reference value VP1ref.

In the step 114, the CPU 54 sets the variable K to a preset value Kmin. This preset value Kmin is chosen to correspond to the value (VP1=VP2)/VP1 prevailing when the degree of clogging of the filter 23 is zero or minimal. After the step 114, the program advances to the step 115.

As a result of return of the program to the initial step 101, this program is reiterated so that the engine speed value N, the engine load value Q, the upstream pressure value VP1, the downstream pressure value VP2, the ratio value K0, and the value of the variable K are all continuously updated.

In a re-initialization step (not shown) between the steps 111 and 112, the CPU 54 resets an internal counter for defining a variable representing the time elapsed since the moment at which the upstream pressure value VP1 drops below the preset value VP1ref. Therefore, as long as the upstream pressure value VP1 continues to be lower than the preset value VP1ref, one of the step 113 or the step 114 will be executed at regular intervals equal to the preset interval Tref1 used in the step 111. It should be noted that a similar re-initialization step resides between the steps 106 and 107.

FIG. 5 is a detailed flowchart of the pressure sampling step 105. In a step 121 following the step 104 (see FIG. 4), the CPU 54 determines whether or not the engine is running normally, specifically whether or not the engine speed value N exceeds a preset level Nref2 preferably chosen to be 500 rpm. If the engine has stopped, that is, if the engine speed value N does not exceed the preset level Nref2, the program returns to the initial step 101 (see FIG. 4) via a step 130 in which the pressure sampling operation is suspended and reinitialized. If the engine is running normally, that is, if the engine speed value N exceeds the preset level Nref2, the program advances to a step 122.

In the step 122, the CPU 54 determines whether or not an upstream pressure value VP1 derived from the pressure signal VP has been stored in the RAM of the CPU 54 since the last execution of an intialization step 129 described hereafter. If no upstream pressure value VP1 has yet been stored, the program proceeds to a step 123. If an upstream pressure value VP1 has already been stored, the program advances to a step 126.

In the step 123, the electromagnetic valve 80 (see FIG. 1) is de-energized or is held in the de-energized state. Specifically, the control signal S5 designed to open the switch 51e (see FIG. 1) is outputted to the control terminal of the switch 51e. As described previously, opening the switch 51e causes de-energization of the electromagnetic valve 80, thereby allowing the pressure sensor 82 (see FIG. 1) to measure the upstream pressure P1 relative to atmospheric pressure.

In a step 124 subsequent to the step 123, the CPU 54 determines whether or not the time elapsed since the moment of de-energization of the electromagnetic valve 42 exceeds a preset interval Tref2. If this elapsed time does not exceed the preset interval Tref2, the program returns to the initial step 101. As a result, the electromagnetic valve 80 continues to be de-energized until the elapsed time reaches the preset interval Tref2. If the elapsed time exceeds the preset interval Tref2, the program advances to a step 125.

In the step 125, the current upstream pressure value VP1 derived from the pressure signal VP is stored in the RAM of the CPU 54. After the step 125, the program proceeds to the step 126.

In the step 126 following one of the steps 122 and 125, the electromagnetic valve 80 is energized. Specifically, the control signal S5 designed to close the switch 51e is outputted to the control terminal of the switch 51e. As described previously, closing the switch 51e results in energization of the electromagnetic valve 80, thereby allowing the pressure sensor 82 to measure the downstream pressure P2 relative to atmospheric pressure.

In a step 127 subsequent to the step 126, the CPU 54 determines whether or not the time elapsed since the moment of energization of the electromagnetic valve 80 exceeds a preset interval Tref3. If this elapsed time does not exceed the preset interval Tref3, the program returns to the initial step 101. As a result, the electromagnetic valve 80 continues to be energized until the elapsed time reaches the preset interval Tref3. If the elapsed time exceeds the preset interval Tref3, the program advances to a step 128.

In the step 128, the current downstream pressure value VP2 derived from the pressure signal VP is stored in the RAM of the CPU 54. After the step 128, the program advances to the step 106 (see FIG. 4) via the step 129 in which the pressure derivation operation is initialized, specifically in which the variables relating to the pressure derivation operation are cleared and initialized.

The upstream and downstream pressure values VP1 and VP2 are sampled in each cycle of execution of the main flow of this pressure derivation step 105, so that these values VP1 and VP2 are periodically sampled and updated.

FIG. 6 shows a typical waveform of the pressure signal VP. The electromagnetic valve 80 switches from the de-energized state to the energized state at moments denoted by the letter C. The electromagnetic valve 80 switches from the energized state to the de-energized state at moments denoted by the letter D. During the time intervals Tref2 determined in the step 124 (see FIG. 5), the electromagnetic valve 80 remains de-energized. During the time intervals Tref3 determined in the step 127 (see FIG. 5), the electromagnetic valve 80 remains energized. After a certain time lag following each change in the state of the electromagnetic valve 80, the pressure signal VP reflects the true pressure value. The time intervals Tref2 and Tref3 are chosen to be longer than this time lag. The time interval Tref2 is preferably longer than the time interval Tref3 to decrease the frequency of changes in the states of the electromagnetic valve 80 in order to maximize the service life of the valve 80. The shorter time interval Tref3 is preferably 0.2 seconds. The upstream pressure VP1 is sampled immediately prior to the moments C at which the electromagnetic valve 80 is energized. The downstream pressure VP2 is sampled immediately prior to the moments D at which the electromagnetic valve 80 is de-energized. The interval between the sampling of the paired values VP1 and VP2 in that order is essentially equal to the preset interval Tref3. Since this interval Tref3 is short, the paired values VP1 amd VP2 reflect pressures prevailing at essentially the same moment. This arrangement ensures the accuracy of these values VP1 and VP2 even under rapidly changing operating conditions of the engine.

FIG. 7 is a detailed flowchart of the burner activation step 118. In the step 141 following the step 117 (see FIG. 4), the CPU 54 determines whether or not the engine is running normally, specifically whether or not the engine speed value N exceeds a preset level Nref3 preferably chosen to be 500 rpm. If the engine has stopped, that is, if the engine speed value N does not exceed the preset level Nref3, the program advances to the burner deactivation step 119 (see FIG. 4). If the engine is running normally, that is, if the engine speed value N exceeds the preset level Nref3, the program advances to a step 142.

In the step 142, the CPU 54 determines whether or not ignition of the air-fuel mixture has been achieved on the basis of the current temperature value derived from the temperature signal VT. Specifically, a check is made for whether or not a preset increase in the temperature value occurs within a predetermined interval after commencement of air and fuel supply, that is, whether or not the temperature value reaches a preset level within the predetermined interval. If ignition has been achieved, that is, if the temperature value exceeds the preset level, the program advances to a step 150. If ignition has not been achieved, that is, if the temperature value does not exceed the preset level, the program advances to a step 143. Generally, ignition is not achieved immediately, so that the program will repeatedly advance from the step 142 to the step 143 at early stages.

In the step 143, the control unit 50 activates the glow plug 29. Specifically, the control signal S4 designed to close the switch 51d is outputted to the control terminal of the switch 51d. As described previously, closing the switch 51d results in activation of the glow plug 29. After the step 143, the program proceeds to a step 144.

In the step 144, the CPU 54 determines whether or not the elapsed time since the commencement of activation of the glow plug 29 exceeds a preset interval Tref4. This reference interval Tref4 is chosen so that the sustained activation of the glow plug 29 over this period enables the temperature around the glow plug 29 to assume a level high enough to ignite the air-fuel mixture. The preset interval Tref4 is preferably 50 seconds. If elapsed time from commencement of activation of glow plug 29 does not exceed the preset interval Tref4, the program returns to the initial step 101 (see FIG. 4). As a result, activation of the glow plug 29 is sustained for the preset interval provided that the engine keeps running. When this elapsed time exceeds the preset interval Tref4, the program proceeds to a step 145.

In the step 145, the CPU 54 determines whether or not a temperature value derived from the temperature signal VT immediately prior to ignition of the air-fuel mixture has already been stored in its internal RAM. If a temperature value has not yet been stored, the program advances to a step 152 in which the current value of temperature derived from the temperature signal VT is stored in the RAM of the CPU 54. In this flowchart, the variable VT0 represents this temperature value. As will be made clear, the temperature value VT0 is sampled immediately prior to ignition of the air-fuel mixture. After the step 152, the program advances to a step 146. If a temperature value has already been stored, the program advances from the step 145 directly to the step 146.

In the step 146, the control unit 50 performs air supply and fuel supply to the burner 25. Specifically, the control signal S3 designed to close the switch 51c (see FIG. 1) is outputted to the control terminal of the switch 51c. As described previously, closing the switch 51c results in air supply to the burner 25. The control signal S1 designed to close the switch 51a (see FIG. 1) is outputted to the control terminal of the switch 51a. As described previously, closing the switch 51a results in activation of the fuel pump 33 (see FIG. 1). The control signal S2 designed to close the switch 51b (see FIG. 1) intermittently at a fixed frequency is outputted to the control terminal of the switch 51b. As described previously, this periodic closing of the switch 51b results in the corresponding periodic opening of the fuel valve 30 (see FIG. 1). Accordingly, fuel is supplied to the burner 25. The rate of fuel supply depends on the duty cycle of the control signal S2.

In more detail, in the fuel supply section of the step 146, the CPU 54 determines a desired basic value of the pulse width of the driving pulse current to the fuel valve 30 on the basis of the engine speed value N given in the step 101 and the engine load value Q given in the step 103. Specifically, the ROM 55 holds a table in which a set of desired basic pulse-width values are plotted as a function of the engine speed value N and the engine load value Q. The determination of the basic pulse-width value is carried out by referring to this table. It should be noted that the pulse-widths of the current pulses driving the fuel valve 30 correspond to the interval of time for which the fuel valve 30 remains open. The current pulses driving the fuel valve 30 have a fixed frequency, preferably chosen to be 25 Hz. The desired basic pulse-width values are preferably in the range of 0 to 40 milliseconds. Then, the CPU 54 determines a desired final pulse-width value of the current pulses driving the fuel valve 30 on the basis of the desired basic pulse-width value. Specifically, the desired final pulse-width value is equal to the desired basic pulse-width value multiplied by a factor of 2.0 so as to facilitate ignition. Finally, the control signal S2 in the form of a pulse train having a constant frequency equal to the fixed value and having a pulse-width equal to the desired final pulse-width value is outputted to the control terminal of the switch 51b so that the switch 51b is closed intermittently at the same fixed frequency and the duration of each closing of the switch 51b is maintained at a value equal to the desired final pulse-width value. This periodic closing of the switch 51b actuates a correspondingly periodic opening of the fuel valve 30. This opening of the fuel valve 30 allows fuel supply to the burner 25 at a rate determined by the desired final pulse-width value.

In a step 147 subsequent to the step 146, the current value of the temperature is derived from the temperature signal VT. In this flowchart, the variable VT1 represents this temperature value. The temperature value VT1 is sampled immediately following ignition of the air-fuel mixture. Then, the CPU 54 determines the difference between the temperature value VT1 and the temperature value VT0 given in the step 152. Specifically, ".DELTA.VT=VT1-VT0" is executed, where .DELTA.VT is a variable representing the temperature difference. Since the variables VT0 and VT1 represent the values of temperature at moments immediately preceding and following ignition of the air-fuel mixture, the variable .DELTA.VT represents the increase in the temperature caused by ignition. Finally, the CPU 54 determines whether or not the temperature difference .DELTA.VT exceeds a preset level Href1 preferably chosen to be 100.degree. C. If the temperature difference .DELTA.VT exceeds the preset level Href1, the program advances to a step 148. If the temperature difference .DELTA.VT does not exceed the preset level Href1, the program advances to a step 153.

In the step 153, the CPU 54 determines whether or not the time elapsed since commencement of air-fuel mixture supply exceeds a preset interval Tref5 equal to the longest value necessary for the temperature difference .DELTA.VT to reach the preset level Href1. If this elapsed time does not exceed the preset interval Tref5, the program returns to the initial step 101 (see FIG. 4). If the elapsed time exceeds the preset interval Tref5, the program advances to the burner deactivation step 119 (see FIG. 4) and then returns to the initial step 101. The reference interval Tref5 is preferably 10 seconds.

In the step 148, the CPU 54 determines whether or not the temperature value VT1 exceeds a preset level Href2 at which self-sustaining combustion of the air-fuel mixture will occur. This preset level Href2 is preferably 500.degree. C. If the temperature value VT1 exceeds the preset level Href2, the program advances to a step 149. If the temperature value VT1 does not exceed the preset level Href2, the program advances to a step 154.

In the step 154, the CPU 54 determines whether or not the time elapsed since ignition of the air-fuel mixture exceeds a preset interval Tref6 preferably chosen to be 40 seconds. If this elapsed time does not exceed the preset interval Tref6, the program returns to the initial step 101 (see FIG. 4). If the elapsed time exceeds the preset interval Tref6, the program advances to the burner deactivation step 119 (see FIG. 4) and then returns to the initial step 101. Accordingly, in the case where self-sustaining combustion is not attained within the fixed length of time defined by the preset interval Tref6, the burner 25 is deactivated.

In the step 149, the control unit 50 deactivates the glow plug 29. Specifically, the control signal S4 designed to open the switch 51d is outputted to the control terminal of the switch 51d. As described previously, opening the switch 51d results in deactivation of the glow plug 29. Accordingly, in the case where self-sustaining combustion is attained, the glow plug 29 is deactivated. After the step 149, the program proceeds to the step 150.

In the step 150 following one of the steps 142 and 149, the control unit 50 performs temperature control designed to ensure adequate unclogging of the filter 23. As will be made clear, the program then branches to either a step 151 or the burner deactivation step 119 (see FIG. 4). After the step 119, the program returns to the initial step 101 as illustrated in FIG. 4.

In the step 151, the CPU 54 determines whether or not the time elapsed since the commencement of unclogging of the filter 23 exceeds a preset interval Tref7 preferably chosen to be 3 minutes. If this elapsed time does not exceed a preset interval Tref7, the program returns to the initial step 101 (see FIG. 4). If the elapsed time exceeds the preset interval Tref7, the program advances to the burner deactivation step 119 (see FIG. 4) and then returns to the initial step 101.

FIG. 8 is a detailed flowchart of the temperature control step 150. In a step 161 following one of the steps 142 and 149 (see FIG. 7), the CPU 54 determines a desired basic value of the pulse-width of the current pulses driving the fuel valve 30 on the basis of the engine speed value N and the engine load value Q in the same way as in the step 146 (see FIG. 7). After the step 161, the program proceeds to a step 162.

In the step 162, the current value of the temperature is derived from the temperature signal VT. In this flowchart, a variable VT2 represents this temperature value. Then, the CPU 54 determines whether or not the temperature value VT2 exceeds a preset level Href3 corresponding to a minimum temperature for reliable unclogging of the filter 23. This reference level Href3 is preferably 550.degree. C. If the temperature value VT2 does not exceed the preset level Href3, the program proceeds to a step 172. If the temperature value VT2 exceeds the preset level Href3, the program proceeds to a step 163.

In the step 172, the control unit 50 activates the glow plug 29 as in the step 143 (see FIG. 7). This activation of the glow plug 29 is to facilitate combustion of the air-fuel mixture.

In the step 173 subsequent to the step 172, the CPU 54 determines a desired final value of the pulse-width of the current pulses driving the fuel valve 30 on the basis of the desired basic pulse-width value given in the step 161. Specifically, the desired final pulse-width value is equal to the corresponding basic value multiplied by a factor of 1.6. After the step 173, the program advances to a step 174.

In the step 174, the CPU 54 determines whether or not the time elapsed since the moment at which the temperature value VT2 first dropped below the reference level Href3 in the step 162 or since the moment at which the temperature value VT1 first rose above the reference level Href2 in the step 148 (see FIG. 7) exceeds a preset interval Tref8. This reference interval Tref8 is preferably 15 seconds. If this elapsed time exceeds the preset interval Tref8, the program advances to the burner deactivation step 119 (see FIG. 4) and then returns to the initial step 101 (see FIG. 4). If the elapsed time does not exceed the preset interval Tref8, the program advances to a step 171.

In the step 163, the control unit 50 deactivates the glow plug 29 as in the step 149 (see FIG. 7). After the step 163, the program advances to a step 164.

In the step 164, the CPU 54 determines whether or not the temperature value VT2 exceeds a preset level Href4 preferably chosen to be 580.degree. C. If the temperature value VT2 does not exceed the preset level Href4, the program advances to a step 165. If the temperature value VT2 exceeds the preset level Href4, the program advances to a step 166.

In the step 165, the CPU 54 determines a desired final value of the pulse-width of the current pulses driving the fuel valve 30 on the basis of the desired basic pulse-width value given in the step 161. Specifically, the desired final pulse-width value is equal to the corresponding basic value multiplied by a factor of 1.4. After the step 165, the program proceeds to the step 171.

In the step 166, the CPU 54 determines whether or not the temperature value VT2 exceeds a preset level Href5 preferably chosen to be 600.degree. C. If the temperature value VT2 does not exceed the preset level Href5, the program advances to a step 167. If the temperature value VT2 exceeds the preset level Href5, the program advances to a step 168.

In the step 167, the CPU 54 determines a desired final value of the pulse-width of the current pulses driving the fuel valve 30 on the basis of the desired basic pulse-width value given in the step 161. Specifically, the desired final pulse-width value is equal to the corresponding basic value multiplied by a factor of 1.2. After the step 167, the program proceeds to the step 171.

In the step 168, the CPU 54 determines whether or not the temperature value VT2 exceeds a preset level Href6 preferably chosen to be 620.degree. C. If the temperature value VT2 does not exceed the preset level Href6, the program advances to a step 169. If the temperature value Href6 exceeds the preset level Href6, the program advances to a step 170.

In the step 169, the CPU 54 determines a desired final value of the pulse-width of the current pulses driving the fuel valve 30 on the basis of the desired basic pulse-width value given in the step 161. Specifically, the desired final pulse-width value is equal to the corresponding basic value multiplied by a factor of 1.0. After the step 169, the program advances to the step 171.

In the step 170, the CPU 54 determines a desired final value of the pulse-width of the current pulses driving the fuel valve 30 on the basis of the desired basic pulse-width value given in the step 161. Specifically, the desired final pulse-width value is equal to the corresponding basic value multiplied by a factor of 0.8. After the step 170, the program advances to the step 171.

In the step 171, the control unit 50 energizes the fuel valve 30 intermittently at the fixed frequency while maintaining the duration of each energization cycle of the fuel valve 30 at a value equal to the desired final pulse-width value determined in the steps 165, 167, 169, 170, or 173. Specifically, the control signal S2 in the form of a pulse train having a constant frequency equal to the fixed value and having a pulse-width equal to the desired final pulse-width value determined in the steps 165, 167, 169, 170, or 173 is outputted to the control terminal of the switch 51b so that the switch 51b is closed intermittently at the same frequency and the duration of each cycle of closure of the switch 51b matching the desired final pulse-width value. This periodic closing of the switch 51b actuates a correspondingly periodic opening of the fuel valve 30. Opening the fuel valve 30 allows fuel supply to the burner 25. It should be noted that air supply to the burner 25 is allowed by the step 146 (see FIG. 7). After the step 171, the program advances to the step 151 (see FIG. 7).

The rate of fuel supply to the burner 25 determined by the desired final pulse-width value varies as a function of the engine speed and load. Furthermore, as a result of execution of the steps 164, 165, 166, 167, 168, 169, and 170, the rate of fuel supply to the burner 25 decreases with increases in the temperature value VT2. Accordingly, the temperature in the burner 25 is maintained within a range, preferably of 600.degree.-620.degree. C., which is suitable for reliable unclogging of the filter 23.

FIG. 9 is a detailed flowchart of the burner deactivation step 119 of FIG. 4. In a step 181 following one of the steps 102 and 118 (see FIG. 4), the control unit 50 de-energizes the fuel pump 33. Specifically, the control signal S1 designed to open the switch 51a is outputted to the control terminal of the switch 51a. Opening the switch 51a results in de-energization of the fuel pump 33.

In a step 182 subsequent to the step 181, the control unit 50 holds the fuel valve 30 closed. Specifically, the control signal S2 designed to open the switch 51b is continuously applied to the control terminal of the switch 51b. Opening the switch 51b results in closure of the fuel valve 30. Accordingly, fuel supply to the burner 25 is interrupted.

In a step 183 subsequent to the step 182, the control unit 50 de-energizes the air valve 37 (see FIG. 1). Specifically, the control signal S3 designed to open the switch 51c is outputted to the control terminal of the switch 51c. Opening the switch 51c results in de-energization of the air valve 37. This interrupts air supply to the burner 25. After the step 153, the program returns to the initial step 101 (see FIG. 4).