JP2525871B2 - Engine controller diagnostic system - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an air-fuel ratio feedback control engine control device, and more particularly to an engine control device diagnostic system suitable for a gasoline engine for automobiles.

[Prior Art] In recent years, various control devices using a microcomputer have been widely adopted, and in response to this, so-called electronic type using the microcomputer also in a gasoline engine for automobiles. Engine control devices (hereinafter referred to as EECs) have been widely adopted.

Therefore, a conventional example of such EEC will be described below.

This conventional example is disclosed in Japanese Patent Laid-Open No. 60-111034, and the conventional example will be described below with reference to FIGS. 2 and 3.

FIG. 2 is a partial cross-sectional view schematically showing the entire control system of the engine. In the drawing, intake air is supplied into the cylinder 8 through the air cleaner 2, the throttle valve 4, the intake pipe 6. . The gas burned in the cylinder 8 is
Through the exhaust pipe 10 and discharged into the atmosphere.

The throttle 4 is provided with an injector 12 for injecting fuel, and the amount of fuel ejected from this injector 12 is atomized in the air passage of the throttle 4 and mixed with intake air. As a result, an air-fuel mixture is formed, and this air-fuel mixture passes through the intake pipe 6 and is supplied to the combustion chamber of the cylinder 8 by opening the intake valve 20.

A throttle valve 14 is provided near the outlet of the injector 12. The throttle valve 14 is configured to be mechanically interlocked with the accelerator pedal and is driven by the driver.

An air passage is provided upstream of the throttle valve 14 of the throttle slot 4.
22 is provided, and a hot-wire type air flow meter, which is an electric heating element, that is, a flow rate sensor 24 is provided in the air passage 22, and an electric signal AF that changes according to the air flow velocity is taken out. Since the flow rate sensor 24 including the heating element (hot wire) is provided in the bypass air passage 22, the flow rate sensor 24 is protected from the high temperature gas generated during the backup from the cylinder 8 and is contaminated by dust in the intake air. It is also protected from things. The outlet of the bypass air passage 22 is opened near the narrowest part of the bench lily, and the inlet is opened on the upstream side of the bench lily.

The pressurized fuel is constantly supplied to the injector 12 from the fuel tank 30 via the fuel pump 32, and when the injection signal from the control circuit 60 is given to the injector 12, the injector 12 is connected to the inside of the suction pipe 6. Is injected with fuel.

The air-fuel mixture sucked from the intake valve 20 is compressed by the piston 50 and burned by the spark of the spark plug (not shown), and this combustion is converted into kinetic energy.
The cylinder 8 is cooled by the cooling water 54. The temperature of the cooling water is measured by the water temperature sensor 56, and this measured value TW is used as the engine temperature.

The exhaust pipe 10 is provided with an O ₂ sensor 142, which detects O ₂ in the exhaust gas.
The presence or absence of ₂ is measured and a measurement value λ is output.

In addition, a crankshaft (not shown) has a fixed angle (for example, 0.5
A crank angle sensor is provided which outputs a reference angle signal and a position signal for each degree.

The output of the crank angle sensor, the output signal TW of the water temperature sensor 56, the output signal λ of the O ₂ sensor 142, and the electric signal AF from the heating element 24 are the control circuit 60 including a microcomputer.
The output of the actuator drives the injector 12 and the ignition coil.

Further, the throttle valve 4 is provided with a bypass 26 that communicates with the intake pipe 6 across the throttle valve 14, and the bypass 26 is provided with a bypass valve 61 that is controlled to open and close.

The bypass valve 61 faces a bypass 26 provided around the throttle valve 14 and is controlled to be opened / closed by a pulse current. The lift amount changes the cross-sectional area of the bypass 26. The drive unit is driven and controlled by the output of 60. That is, the control circuit 60 generates an opening / closing cycle signal for controlling the drive section, and the drive section adjusts the lift amount of the bypass valve 61 by the opening / closing cycle signal.

The EGR control valve 90 controls the passage between the exhaust pipe 10 and the intake pipe 6, and the EGR amount from the exhaust pipe 10 to the intake pipe 6 is controlled.

Therefore, the injector 12 shown in FIG. 2 is controlled to perform the air-fuel ratio (A / F) control and the fuel increase control, and the engine speed control (ISC) at idle is performed by the bipulse valve 61 and the injector 12. In addition, the EGR amount can be controlled.

FIG. 3 is an overall configuration diagram of a control circuit 60 using a microcomputer, which includes a central processing unit 102 (hereinafter C
It comprises a PU), a read only memory 104 (hereinafter referred to as ROM), a random access memory 106 (hereinafter referred to as RAM), and an input / output circuit 108. CPU10 above
Reference numeral 2 calculates the input data from the input / output circuit 108 according to various programs stored in the ROM 104, and returns the calculation result to the input / output circuit 108 again. The RAM 106 is used for intermediate storage required for these calculations. CPU102, ROM104, RAM10
6, data exchange between the input / output circuit 108
This is performed by a bus line 110 composed of a bus, a control bus and an address bus.

The input / output circuit 108 inputs / outputs a first analog digital converter 122 (hereinafter referred to as ADC1), a second analog digital converter 124 (hereinafter referred to as ADC2), an angle signal processing circuit 126, and one bit information. It has an input means with a discrete input / output circuit 128 (hereinafter referred to as DIO) for performing.

The ADC1 includes a battery voltage detection sensor 132 (hereinafter referred to as VBS), a cooling water temperature sensor 56 (hereinafter referred to as TWS), an atmospheric temperature sensor 136 (hereinafter referred to as TAS), and a regulated voltage generator 138 (hereinafter referred to as V).
RS) and slot sensor 140 (hereinafter referred to as OTHS)
And the output of the O ₂ sensor 142 (hereinafter, referred to as O ₂ S) is added to the multiplexer 162 (hereinafter, referred to as MPX), and one of them is selected by the MPX 162 and the analog digital conversion circuit 164 is selected. (Hereinafter referred to as ADC). The digital value output from the ADC 164 is held in the register 166 (hereinafter referred to as REG).

In addition, the flow rate sensor 24 (hereinafter referred to as AFS) is input to the ADC2124, and the analog digital conversion circuit 172 (hereinafter referred to as ADC
Is converted to a register 174 (hereinafter, referred to as REG), and is digitally converted.

From the angle sensor 146 (hereinafter referred to as ANGLS), a signal indicating a reference crank angle, for example, 180 crank angle (hereinafter referred to as REF) and a minute angle, for example, a signal indicating 1 degree crank angle (hereinafter
And POS) are output and applied to the angle signal processing circuit 126, where the waveform is shaped.

The DIO (128) starts with an idle switch 148 (hereinafter referred to as IDLE-SW) and a top gear switch 150 (hereinafter referred to as TOP-SW) that operate when the throttle valve 14 is returned to the fully closed position.・ Switch 152 (hereinafter referred to as START-SW)
And are entered.

Next, a pulse output circuit and a control target based on a calculation result of the CPU will be described. The eject control circuit 1134 (hereinafter referred to as INJC) is a circuit that converts a digital value of a calculation result into a pulse output. Therefore, a pulse INJ having a pulse width corresponding to the fuel injection amount is generated by the INJC 1134 and applied to the injector 12 via the AND gate 1136.

The ignition pulse generation circuit 1138 (hereinafter referred to as IGNC) includes a register for setting ignition timing (hereinafter referred to as ADV) and a register for setting start time of primary current conduction of the ignition coil (hereinafter referred to as ADV).
DWL) and these data are set by the CPU. A pulse IGN is generated based on the set data, and is supplied to the amplifier 62 for supplying the primary current to the ignition coil.
This pulse IGN is applied via the AND gate 1140.

The opening rate of the bypass valve 61 is the pulse I applied from the control circuit (hereinafter referred to as ISCC) 1142 through the AND gate 1144.
Controlled by SC. ISCC1142 is a register ISCD that sets the pulse width and a register ISCP that sets the pulse period.
And has.

EGR amount control pulse generation circuit 1178 for controlling the EGR control valve 90
(Hereinafter referred to as EGRC) has a register EGRD for setting a value indicating the duty of the pulse and a register EGRP for setting a value indicating the period of the pulse. The output pulse EGR of this EGRC is applied to the transistor 90 via the AND gate 1156.

The 1-bit input / output signal is controlled by the circuit DIO (128). IDLE-SW signal, START-WT as input signal
There are signals and TOP-SW signals. Further, as the output signal, there is a pulse output signal for driving the fuel pump. This DI
O is provided with a register DDR192 for determining whether to use the terminal as an input terminal and a register DOUT194 for latching output data.

The mode register 1160 is a register (hereinafter referred to as MOD) that holds an instruction to instruct various states inside the input / output circuit 108. For example, by setting an instruction set in the mode register 1160, AND gates 1136, 1140, 1144, 1156. To make all of them work or not. By setting the instruction set in the MOD register 1160 in this way, I
It is possible to control stop and start of NJC, IGNC, and ICSS output.

A signal DIO1 for controlling the fuel pump 32 is output to DIO (128).

Therefore, if such an EEC is applied, almost all controls relating to the internal combustion engine, such as control of the air-fuel ratio, can be appropriately performed, and it is possible to sufficiently comply with strict exhaust gas regulations for automobiles.

By the way, the EEC explained in FIGS. 2 and 3
In this case, the fuel injection by the engineer 12 is intermittently performed in synchronization with the rotation of the engine, and the fuel injection amount is controlled by the valve opening time of the injector 12 in one injection operation, that is, the injection time T. It is performed under the control of _i .

Then, in this conventional example, the injection time T _i is basically determined by the following equation.

T _i = α ・ T _P・ (K _l ＋ K _t ＋ K _s ) ・ (1 ＋ ΣK _l ) ＋ T _s ……
(1) T _P = K _const · Q _a / N (2) where α: air-fuel ratio correction coefficient T _p : basic injection time K _l : steady learning coefficient K _t : transient learning coefficient K _s : shift coefficient K _l : Various correction factors T _s : Injector invalid injection time K _const : Injector coefficient Q _a : Intake air flow rate N: Engine speed That is, from the intake air flow rate Q _{A of the} engine and the rotation speed N according to the equation (2,) established a basic fuel injection time T _P, as roughly stoichiometric air fuel ratio (a / F = 14.7) is obtained, O ₂ sensor 1
The 42 signal λ by changing the air-fuel ratio correction coefficient α performs correction of the air-fuel ratio due to fed back, on which is configured to obtain a more accurate stoichiometric air-fuel ratio, further constant learning coefficient K _l
Therefore, it is made to correct the variation and secular change of the characteristics of various actuators and sensors related to the air-fuel ratio control, and this is also made to correct the acceleration and deceleration by the transient learning coefficient K _t . The fuel injection time T _i is determined by subtracting the shift coefficient during sudden deceleration.

Here, the learning coefficient K _l will be described. O ₂ sensor 142
Outputs a binary signal (high or low level voltage) depending on the presence or absence of oxygen in the exhaust gas. It is well known that the air-fuel ratio correction coefficient α is stepwise increased or decreased based on the binary signal, and then gradually increased or decreased to perform the air-fuel ratio control. FIG. 4 shows the state of the air-fuel ratio correction coefficient α, which moves by detecting the air-fuel ratio between rich and lean depending on the output signal λ of the O ₂ sensor.

Here, with the air-fuel ratio correction coefficient α when the O ₂ sensor signal is inverted, the lean-to-lit polar path is α _max , the rich-to-lean extreme value is α _min , and the average value α _ave is calculated by the following equation. To do.

The idea of this average value is well known, but in this conventional example,
When the average value α _ave is outside the range between the upper limit value (TUL) and the lower limit value (TLL), the deviation K _l between the average value α _ave and 1.0 is used as the steady learning correction amount. The calculation of the steady learning correction amount K _l is performed in all areas where the feedback correction is performed by the O ₂ sensor.

FIG. 5 shows a table for writing the steady learning correction amount K _l . In this table, K _l is written at a division point determined by the basic fuel injection time T _P and the engine speed N.
This learning timing is when the division points do not change and the number of extreme values reaches n. The table shown in FIG. 5 is defined as a stationary learning map. In this steady learning map, it is practically unlikely that all division points (64 points in this case) will be filled by learning. Therefore, it is necessary to create the unlearned division points by referring to the learned division points.

 Therefore, next, this creating method will be described.

FIG. 6 shows an example of a buffer map and a comparison map having the same score as the division points of the stationary learning map used for creating the stationary learning map.

FIG. 7 is a black diagram showing a routine for creating a stationary learning map. In (1), the steady learning map and the comparison map are all cleared, and the steady learning correction amount is written in the buffer map. However, at this point, double writing is not performed on the buffer map. In (2), when the number of written buffers reaches C, or the contents of the buffer map are transferred to the comparison map, and in (3), all the buffer maps are created by referring to the C contents written in the buffer map. , Transfer the contents to the stationary learning map. In (4), the contents of the comparison map are retransferred to the buffer map. From this point on, the value of K _l is used for the calculation of the fuel injection time. Up to this point, K _{l in} equation (1) is 1.0. In (5), the steady-state learning correction amount is written in both the steady-state learning map and the buffer map, and the air-fuel ratio correction coefficient α
Set 1.0 to 1.0 and compare the contents of the comparison map with the comparison map. When the difference in the compared contents is a certain number,
In (6), the routines (2) to (4) are repeated.

According to this conventional example, the steady-state learning correction K _l stores a deviation from 1.0, so that the air-fuel ratio correction coefficient α
Can be controlled around 1.0, and harmful components of exhaust gas can be reduced.

Further, in the steady learning map shown in FIG. 5, when the basic fuel injection time T _p7 or more and the engine speed N ₇ or more, the map values of the rightmost column and the lowest row are used. It is possible to perform correction so that the power is always optimum.

Next, the eighth example of the learning routine for the stationary learning coefficient K _l will be described.
This will be described with reference to the flow charts of FIGS.

The flow chart was although connexion process after starting the engine, is repeated at a predetermined cycle, firstly, O ₂ at step 300
It is determined whether or not the feedback control is entered, and if the result is Yes, the process proceeds to step 302. If the result is no, go to step 332. In step 302, it is determined whether or not the signal from the O ₂ sensor has crossed λ = 1 (theoretical air-fuel ratio A / F = 14.7). If the result is No, the well-known integration process (not shown) is performed in step 332. Result is Ye
If s, proceed to step 304 and average value α shown in equation (3).
Calculate _ave . At step 306, the average value α _ave is the fourth
Determine whether it is within the upper and lower limits shown in the figure,
If the result is Yes, normal feedback control is being performed, so the counter is cleared at step 326 and the process proceeds to step 332. On the other hand, if the average value α _ave is outside the upper and lower limits, in step 308, the difference between the average value α _ave and 1 is set as the steady learning correction amount K _l . Next, in step 310, the present dividing point determined from the basic fuel injection time T _p and the engine speed N shown in FIG. 5 is calculated, and in step 312, it is compared with the dividing point immediately before this routine. Then, it is determined whether or not the division point has changed. The dividing point has changed (Ye
s), since the division point where the steady learning correction amount K _l is written is not fixed, the process goes to step 326. If the dividing point has not changed, step 314 updates the counter and step 3
At 16, the counter determines whether it has reached n. If the counter value is not n (No), go to step 332. When the counter value reaches n (Yes), the counter is cleared in step 318, and the process proceeds to step 320. In step 320, it is determined whether or not the first preparation of the steady learning map, which is the operation (2) to (4) described in FIG. 7, is performed. If the map has not been created yet, the process proceeds to step 322 and thereafter to perform the operation (1) described with reference to FIG. In step 322, it is determined whether or not the division point has already been written. If already written (Yes), do nothing and step 332
Head to. If the result is No, step 324, step 308.
Write the steady learning correction amount K _l calculated in step ₁ into the division points. If you created the first stationary learning map at step 320 (Y
es), and after step 328, the operations (5) and (6) described with reference to FIG. 7 are performed. At step 328, the steady learning correction amount is applied to the division points of the steady learning map and the buffer map.
Add K _l . Then, in step 330, the air-fuel ratio correction coefficient is set to 1.0.

Therefore, by repeating the processing according to these steps 300 to 332, the operations (1), (5), and (6) described in FIG. 7 are obtained.

Next, the operations of (2), (3) and (4) described in FIG. 7 will be described with reference to the flow chart of FIG.

At step 350, it is determined whether the first stationary learning map has been created. If not yet created (No), step 354
Proceed to and check the number of write buffers. When the number reaches m, proceed to step 35, but m
If not, go to step 370. Step 3
If the first stationary learning map is created at 50 (Yes), at step 352, the difference between the data of the buffer map and the comparison map is checked. If there is a difference of 1 in the content between the comparison map and the buffer map, proceed to step 356,
Create a stationary learning map. If there is no difference in the contents, go to step 370.

At step 356, set the flag that the map is being created,
Prohibit writing learning results. At step 358, the contents of the buffer map are transferred to the comparison map and step 360
Then, the stationary learning map is created using the buffer map. In step 362, the contents of the created buffer map are transferred to the stationary learning map, and in step 364, the contents of the comparison map are transferred to the buffer map. Step 36
In step 6, set the flag that the stationary learning map has been created. This flag is set in steps 350 and 3 of FIG.
Used for judgment at 20. In step 368, in step 356
Reset the map-creating flag that was set in.

Accordance connexion, according to this conventional example, the engine control by the microcomputer, can be performed by a steady learning O ₂ fed back control and air-fuel ratio correction coefficient according to the output of the O ₂ sensor, it is possible to obtain excellent controllability .

[Problems to be Solved by the Invention] By the way, in such a system, as is clear from the above description, various sensors and actuators used therein are not affected by changes in characteristics due to aging or the like. It is inevitable.

And even in such a case, in this conventional technology,
Since it is corrected by learning, there is almost no problem of deterioration of controllability, but no special consideration is given to the abnormality judgment of these sensors and actuators, and there is a problem in maintaining sufficient reliability. .

The object of the present invention is to calculate the characteristics of the intake air flow rate sensor and the fuel injection valve (injector) among various sensors and actuators necessary for engine control by performing inter-element calculation of the air-fuel ratio correction coefficient corresponding to the operating state of the engine. The correction coefficient for is calculated separately, and by using this correction coefficient for evaluation, the self-diagnosis of the intake air flow rate sensor and the fuel injection valve can be performed independently and independently, thereby maintaining high reliability at all times. An object of the present invention is to provide an engine control device diagnosis system that can be performed.

[Means for Solving the Problem] The above-mentioned object is to utilize the fact that the influence of the characteristic change of the intake air flow rate sensor and the fuel injection valve on the air-fuel ratio correction coefficient is different, and The correction coefficient for the characteristics of the intake air flow rate sensor and the fuel injection valve can be separately obtained by the inter-element calculation, and the change in the characteristics of the intake air flow rate sensor and the fuel injection valve can be determined using this correction coefficient. To be achieved.

[Operation] The air-fuel ratio correction coefficient provided as a result of the air-fuel ratio feedback control using the output of the O ₂ sensor reflects the characteristic changes of the intake air flow rate sensor and the fuel injection valve.

For example, when the injector coefficient, which is one of the characteristics of the fuel injection valve, changes to a value larger than an appropriate value,
Due to this, the excessive fuel supply amount appears as a decrease in the value of the air-fuel ratio correction coefficient due to the air-fuel ratio feedback control.

By the way, the air-fuel ratio correction coefficient is calculated and held for each operating region of the engine.

Therefore, looking at how the characteristic changes of the intake air flow rate sensor and the fuel injection valve are reflected in the air-fuel ratio correction coefficient for each operating region, for example,
The change in the injector coefficient affects the air-fuel ratio correction coefficient in each operating region almost uniformly, while the change in the characteristic of the intake air flow rate sensor is on the air flow rate line where the air flow rates are equal. The air-fuel ratio correction coefficient has a particularly large influence, and the change in the invalid injection time of the injector has a particularly large effect on the air-fuel ratio correction coefficient, particularly in a region where the basic injection time is short.

Therefore, by utilizing the difference in the influence on the air-fuel ratio correction coefficient, an index representing the characteristics of each of the intake air flow rate sensor and the fuel injection valve is calculated from a plurality of air-fuel ratio correction coefficients, and based on this, each is calculated. The characteristic correction coefficient for each is separated and calculated independently.

If the abnormality determination is performed by evaluating the characteristic correction coefficient obtained for each of the intake air flow rate sensor and the fuel injection valve, the self-diagnosis can be accurately performed.

[Embodiment] Hereinafter, an engine control device diagnosis system according to the present invention will be described in detail with reference to an illustrated embodiment.

In the following embodiments, the hardware configuration and the contents of engine control by the microcomputer are the same as those of the conventional example described in FIGS. 2 to 9, and therefore the description thereof is omitted.

The sensors and actuators targeted in this embodiment are the intake air flow rate sensor and the injector (fuel injection valve), respectively, as described above, but in the following description of the embodiments, the intake air flow rate sensor will be described. Is simply referred to as a sensor, and the injector is referred to as an actuator.

As described above, in this engine control device, the control constants (injector coefficient K
Control is performed using _const , the invalid injection time T _s , and the intake air flow rate characteristic Q _a ).

Therefore, it is conceivable that this control constant deviates considerably from the physical characteristics of the engine, the sensor, and the actuator. Hereinafter, such a state is referred to as an unmatching state.

Therefore, in this unmatching state, the injector coefficient K _const , the invalid injection time T _s , and the intake air flow rate characteristic Q _a
Assuming that the learning coefficient of equation (1) is in the steady learning state of k _l = 1.0, K _t = 0, K _s = 0, the fuel injection time T _{i in} the case of using Like

T _l = K _const · Q _a / N · COEF · α + T _s (4) The value obtained by matching each of these states K _const *, Q _a *, T _s
Described using *, COEF *, the following formula is obtained.

T _l = K _const * · Q _a * / N · COEF * + T _s * …… (5) From the expressions (4) and (5), the following expression holds.

K _const · Q _a / N · COEF · α + T _s = K _const * · Q _a * / N · COEF * + T _s * (6) where T _P * = K _const * · Q _a * / N Using it, it becomes the following formula.

α (N, T _P *) = E1, E2, E3, E4 (7) E1 = (T _s * −T _s ) / {T _P * / COEF * (N, T _P *)} + 1…
… (8) E2 ＝ K _const ＊ / K _const …… （9） E3 ＝ Q _a ＊ / Q _a … （10） E4 ＝ COEF ＊ (N, T _P ＊) / COEF (N, T _P ) …… According to (11) (7) to (11), T _s ; E1 (mainly a function of T _P *, see Fig. 14) K _const ; E2 (constant) Q _a ; E3 (function of Q _a ) It can be seen that COEF; E4 (function of N, T _P *), etc. are reflected in α as products.

Next, consider the case where N, T _P * is divided so that α (N, T _P *) is arranged so that equal Q _a lines are diagonally arranged as shown in FIG. Here, for simplicity, consider a 4x4 map, and let the learning value of α be the intersection point of the lattice. Next, FIG. 11 shows an error E for each factor when unmatching occurs according to the equation (7). The value of the α map at this time is shown in FIG. Here in equation (11)
COEF is assumed to be almost matched, and it is assumed that E4 = 1. Here, the value of E1 is to the longitudinal axis T _p is a1, a2,
......, and on the diagonal, the unmatching E3 values c1, c2, ... of Q _a are applied, and b1 of the unmatching term E2 of the injector coefficient is present for all map values.

The coefficient of the α map at this time is regarded as a matrix,
Let each element be M _ij as shown in FIG.

Each element of the matrix reflects each unmatching factor in the form shown in Fig. 12. For example, as shown in FIG. 13, a1 to a3 standardized by a4 are obtained as divisions of matrix elements. Therefore, by capturing the characteristic for this T _p, for example 14 was trend (in the basic injection time T _p is a small area, in proportion to T _s Anmatsuchingu amount greatly changes) shown in FIG correct T _s from It can be matched.

Next, in FIG. 15, Q _a is corrected in the same manner as in FIG. At this time, it is standardized with c4.

Considering the above characteristics, the matching coefficient in the unmatched state is set as follows.

K1; K _const correction (scalar; 1 variable) K2 (N, T _p ); K _trm correction (N, T _{p map} ; N division number / T _p division number) K3; T _s correction (scalar; 1 variable) K4 (Q _a ); Q _a correction (vector; Q _a division number) Considering (4) and (5), these coefficients may be respectively set as follows.

K1 ＝ K _const ＊ / K _const …… (12) K2 (N, T _p ) ＝ COEF ＊ (N, T _p ＊) / COEF (N, T _p ) …… (1
3) K3 = T _s * -T _s (14) K4 (Q _a ) = Q _a * / Q _a ...... (15) In addition, the fuel injection is performed by the following formula.

According to Eqs. (16) and (17), from the change of the coefficient appearing in α due to the O ₂ feedback,
The coefficients to be corrected are sorted for each of K _const , T _s, and Q _a . In particular, it is corrected by the product of the correction K1 and Q _a correction K4 (17) basic injection time as shown in equation Injiekuta coefficient (K _const) (Q _a). Further, as shown in the following equation, the fuel injection time T
_i is calculated by multiplying T _p ′ by the product of COEF ′ and α, and is calculated by adding the battery correction voltage T _s ′.

From the above results, it is possible to separate the fuel injection time, which has conventionally been collectively corrected by α, by the generation factor, and particularly correct the basic injection time T _p as shown in equation (17). That is,
Separate learning for each occurrence factor can be realized.

The mating procedure will be examined below based on the above analysis.

First, K _const , T _s, and Q _a are set, O ₂ feed back is performed to realize various operating states, and steady learning of α (N, T _p ) map is performed. At this time, the various correction terms in the equation (1) are operating conditions such that they are 0 except for k _trm which is the air-fuel ratio correction coefficient of the feedback control during steady state, that is,
Stationary learning is performed under the following conditions.

After warming up, the operation is performed and the steady operation (| ΔN | <Δ
Α (n, T _p ) is learned in N _s , | ΔT _p | <ΔT _ps ).

Next, each factor is separated from α. Here, first, the air flow rate is corrected. From the characteristics of the matrix elements shown in Fig. 12, the value normalized by c4, which is the value of E4 at Q ₄ 4 as shown in Fig. 15, is calculated by dividing the matrix elements as shown in the table. I understand. From the table, it can be seen that there are several calculation methods depending on the element. If all the α maps have been learned, it is preferable to perform determination based on the degree of variation in the values and, if the averaging process is effective, perform the averaging process. Further, when the learning number of the α map is small, the necessary minimum value may be acquired and the Q _a correction may be performed.

That is, for K _lcd 4 (Q _a ), first, the following equation is corrected to make the relative error constant at 1 / c4.

K4 # (Q _a ) = ci / c4 …… (19) Q _ai ′ (Q _a ) ＝ K4 # (Q _a ) ・ Q _ai …… (20) K4 (Q _a ) ＝ c4 ・ K4 # (Q _a ) (21) Correct the Q _a table as shown above and, at the same time, calibrate the diagonal elements of the α map as shown in the following equation. This is _a correction to be made because there is no factor affecting the α map by calibrating the Q _a table by the equation (20).

M _ij = M _ij · (c4 / ck) (22) where k = j−i + 4 (23) i, j; 1,2,3,4 In other words, j−i = constant diagonal element Is corrected uniformly. As a result, the term for Q _a of the α map is ci = c4.

 The above is summarized as follows.

First, the Q _a error characteristics are flattened from the learned α (N, T _p ) map, and the Q _a correction table is created. here,
The α map is also corrected for the map corresponding to the Q _a correction.

There are two methods for correcting K _const and T _s . One is to correct T _s by dividing the coefficient of the matrix of α map, and the other is to correct T _s by plotting T _p ′ and T _i .

As shown in FIG. 14, there is Anmatsuchingu the ineffective injection time T _s, T _s * -T _s is "0" if not show the hyperbolic characteristic for T _p, 1 T _p is a large place It becomes the characteristic. Therefore, for example, a1 / a in the low load region of T _p 1 and T _p 2
The optimum value of T _s is found by increasing or decreasing the correction term K _lcd 3 of T _s so that the value of 4, a2 / a4 approaches 1. Here, for example, when T _s is small, a1 in the low load region as shown in FIG.
Since / a4 and a2 / a4 are larger than 1, the operation of increasing K _lcd 3 is performed. When T _s is large, K _lcd 3 is made small by the same method. At this time, if the value of a1 / a4 shows a stable increase / decrease tendency, the increase / decrease amount may be set as in the following equation in order to increase the convergence speed of T _s .

K3 = K3 + (constant) ・ al / a4 …… (24) Calculation of coefficients such as a1 / a4 is shown in Fig. 13.

If the optimum value of T _s falls within the predetermined range in the above procedure, and the value of α map becomes almost constant α s,
When the common term is drawn so that the value of each element is close to “1”, from Eqs. (5) to (10), E1 = 1, E3 = Q _{a in} Eq. (7)
Considering the condition of 4 * / Q _a 4, the following equation holds.

K _const * / K _const c4 = αs (25) K1 = K _const * / K _const (26) From equation (21), K4 (Q _a ) = ci = c4 ・ (ci / c4 ) ＝ c4 ・ K4 # (Q _a ) …… (27) From the equations (26) and (27), the following equation holds.

K1 ・ K4 (Q _a ) ＝ K _const ＊ / K _const・ c4 ・ K4 # (Q _a ) ……
(28) Here, using equation (29) gives the following equation.

K1 ・ K4 (Q _a ) ＝ αs ・ K4 # (Q _a ) …… (29) The above is summarized as follows.

From the revised α map, a characteristic value that depends on T _s unmatching and becomes a function of T _p is calculated, and the T _s correction value is optimized using this characteristic value as a reference value.

Next, the product of the uniform error of Q ₂ and the error rate of K _const is obtained by using the value of the common coefficient of the α map which is almost flattened by the above operation. By the above operation, T _s and K _const can be calibrated.

Next, correction of K _const and T _s by the T _p ′ -T plot method will be shown.

When the correction of the Q _a table is executed, Q _a becomes Q _a ′ ＝ Q _a
* / C4. At this time, fuel injection is expressed by the following equation.

T _i = K _const · k4 # (Q _a ) · Q _a / N · COEF * · (K _const * / K _const · Q _a 4 * / Q _a 4) + T _s * T _p ′ …… (30) here Then, when (T _p ′, T _i ) is obtained and plotted, the first
As shown in Fig. 6, the plotted locus becomes a straight line, the intercept where T _p ′ = 0 becomes T _s *, and the slope of the straight line becomes
The value is as shown in a part of equation (30). It is considered that COEF * = 1 in the case of matching, so if the slope of the straight line is ks, the following equation holds.

K _const * / K _const · Q _a 4 * / Q _a 4 = K1 · c4 = ks (31) With the above, it is possible to match the coefficients, and the summary is as _follows .

In each operating state at regular times, obtain the basic injection time T _p and the injection time T _i calculated using the corrected Q _a table,
From the slope of the straight line formed by (T _p , T _i ), the uniform error of Q _a and K
Calculate the product of _const error and T from the intercept of T _p = 0
_Calculate the correction value for _s . By repeating this arithmetic operation, the optimization is advanced by learning each control constant. In this way, by advancing the optimization of each control constant, the basic injection time becomes accurate, the ignition timing calculated on the basis of this is also optimized, and as a result, the engine that is comprehensively optimized Control will be gained.

Next, a block diagram showing the functions required for the above-mentioned control constant correction operation is as shown in FIG.

First, the air-fuel ratio feedback means 400, as described above, generates the air-fuel ratio correction coefficient α by O ₂ feedback.

Next, the steady learning means 500 carries out the steady learning processing described in FIGS. 8 and 9 to learn the air-fuel ratio correction coefficient α in the steady state.

Then, using this learned air-fuel ratio correction coefficient α,
The characteristic index calculating means 600 calculates the characteristic index for each of the control constants.

Then, referring to this characteristic index, the control constant correction means
The correction constant of the control constant is executed by 700, and the control constant is optimized.

Here, the characteristic index relating to the control constant is shown in FIG. 13 and FIG.

It defines the values of ai / a4, ci / c4, etc., and is obtained by dividing the learned elements of the air-fuel ratio correction coefficient α.
At this time, since the air-fuel ratio correction coefficient α has a value near 1.0, it can be processed by subtraction instead of the above division.

Next, the above processing will be described in more detail by using a flow chart.

First, FIG. 18 is a schematic flow chart showing steady learning processing 500 (first
The characteristic correction routine 2000 is executed after the steady learning means in FIG. 7). FIG. 19 is a schematic flow of this characteristic correction routine HIMBASE.

First, in process 2010, it is determined whether or not the learning number is equal to or larger than the predetermined value NA, and then the process proceeds to process 2020. When this condition is not satisfied, the characteristic correction process is not performed.

In processing 2020 to processing 2050, whether to execute the detailed logic 2060 or the simple logic is distributed. That is, the detailed logic 2060 is executed only when it is determined that the number QAN of acquisition of the Q _a characteristic is larger than the predetermined value QANS and the acquisition rate NTS of the T _s characteristic is larger than the predetermined value NTSS. In the meantime, the simple logic 2070 is executed.

FIG. 20 is a flow chart showing the processing contents of the simple logic HIMSIMP. When this processing is started, first, the matching status flag calculation processing 2110 is executed. This processing is completed when the value of the learning map obtained by the steady learning processing is within the predetermined range as a whole with respect to the previous value, the matching (correction processing) is completed. age,
On the other hand, when the amount of change exceeds a certain limit, a matching error (correction processing error) is determined. Then, in either case, one of the corresponding flags FHIMC and FHIME is set.

In the judgment processes 2120 and 2130, judgments are made according to these flags, and if the matching is completed, the process ends here. After returning by the completion of the matching, another task not described here is activated in a predetermined relatively long cycle, and the matching process is periodically executed. When a matching error occurs, the matching error processing 2150 is executed. In this embodiment, the content of the matching error processing 2150 is basically such that the correction processing is canceled and only the control by the air-fuel ratio correction coefficient by the steady learning is performed.

Then, if it is determined here that the matching is neither completed nor an error, that is, the result of the process 2130 is
When the result is Yes, the process proceeds to the process 2135 and thereafter.

First, this processing 2135 is a switching processing of i, which has the following contents. That is, in this embodiment, there are two systems of maps, one of which is currently in use, and the other of which is a map for calculation. Therefore, this process 21 is performed so that the map is switched to execute i = 1 or i = 0.
35 are provided.

In the next process 2140, a map value is searched for in an area where T _p in the map is large, which is necessary for correcting K _const . In this region where T _p is increasing,
This is because the effect of T _s is small, and the effect of K _const is dominant here, provided that the variation in Q _a characteristics is small.

Next, the median average calculation process 2160 of the air-fuel ratio correction coefficient α is executed. Here, of the map values α extracted in the process 2140, the process of calculating the average value of the remaining ones except the one showing the maximum value and the minimum value is performed. In addition, when the number of extractions here is two, their average value, and when one, the α
The value of is calculated as it is as the median average value ALPROC.

In process 2170, the average value A is added to KLCD1 which is the correction value of K _const.
Execute the process of substituting LPROC.

Next, a process 2180 is a map condition search process, which is performed in the map.
The map value α is searched for in the region where T _p is small, and the neutral average calculation process is performed in the subsequent process 2190 in the same manner as in the above processes 2140 and 2160.

In the subsequent process 2200, a process of calculating the correction value K3i of T _s by multiplying the gain KKKCD is performed.

Following the above calculation results, the map correction 1 process 2210 corrects the learning map related to K _const and T _s , and then the process 2220 completes the map correction.
The map switching is performed, so that the control by the new coefficient is performed.

The above is the contents of the characteristic correction processing by the simple logic HIMSIMP.

Next, the detailed logic HIMPREC will be described with reference to FIG. In addition, in the flow chart of FIG. 21, Q _a , K
The assistant process when _const and all of T _s are in the unmatching state will be described.

First, the processes 2410, 2420, 2430, 2435, and the processes up to the process 2450 are the matching completion and error determination process similar to the simple logic described in FIG.

In the Q _a characteristic table calculation processing 2440, the characteristic index relating to this Q _a is calculated. At this time, when only a part of the characteristic index cannot be calculated, for the rest, all characteristic indexes are calculated by interpolation calculation. To do so. By this interpolation calculation, the correction process here can be executed even if the learning is not all finished.

In the next T _s characteristic table calculation processing 2460, similarly,
Calculate the characteristic index for T _s . Here, since this T _s characteristic index shows a monotonic characteristic as described in FIG. 14, it can be determined that an error occurs when the calculation result does not exhibit a monotonic characteristic.
SCMPER is set, and as a result, when an error occurs, the processing is ended in the judgment processing 2470.

If no error has occurred, take the next action 2480.
Then, the calculation process of K3, which is the correction value of T _s , is executed.

Subsequently, in map correction processing 2500, the learning map is corrected by K _const , T _a , and Q _a . Then, at the time when the correction of the learning map is completed, like the case described above,
The map is switched so that engine control with a new correction coefficient can be obtained.

 Therefore, the above is the explanation of the detailed logic HINPREC.

Here, the relationship between the control constant and the correction coefficient is summarized as follows.

(1) Control constant (a) Injector (each scalar quantity) K _const : Injector coefficient K _s : Injector invalid injection time (b) Intake air flow sensor (one-dimensional table) Q _a (i) ( ^ex) i = 0 ~ 63): Air flow rate characteristics (2) Correction coefficient K1 ′: Correction coefficient of injector coefficient K _const : K _const0・ K1 ′ …… (32) K3: Correction coefficient of invalid injection time T _s = T _s0 + K3 ′ …… (33) K4 (i): correction factor Q _a of air flow characteristics _{(i) = Q a0 (i} ) · K4 '(i) ...... (34) _{where, K const0 T s0 Q a0 (} i) is These are the initial reference values.

By the way, as described above, the present invention is characterized in that the characteristic correction coefficients of the injector and the intake air flow rate sensor are separated in this way, and the abnormality determination is performed by this. Now, examples of the present invention will be described.

FIG. 1 is a control block diagram showing an embodiment of the present invention. The basic configuration is the same as that of FIG. 17, but the above-mentioned various coefficients K1, K3 calculated by the correction coefficient calculation means 650. , K4
The value of is corrected by the control constant correction means 700 (equation (32), (3
In addition to (3) and (34), the control constant diagnosis means 660 for diagnosing the control constant using these coefficients is provided.

FIG. 22 is a flow chart showing the processing of one embodiment of the control constant diagnosing means 660. First, processing 660002 to processing 66.
0004 is executed to calculate the correction coefficients K1 ′, K3 ′, K4 ′.

Next, in order to perform K _const diagnostic processing, first, processing 66
At step 0010, a process of substituting K1 'for x is executed. Then, the process proceeds to the next process 660012, and the details of this process 660012 are shown in FIG. That is, this process is the result of the process 660100 which determines whether or not the absolute value of the deviation of the value of x, which is now given, from 1.0 exceeds a preset predetermined value XSL (for example, 60%). By this, the data d representing the diagnosis result is set to either "1" indicating the abnormality or "0" indicating that there is no particular abnormality.
2 or performing one of the processes 660104.

This completes the diagnostic processing of the correction coefficient K _const , and stores the result in the diagnostic result RAM memory table D ₁ (K1), D ₂ (K1). In this case, D ₁ is a flag bit indicating whether or not there is an abnormality, and D ₂ is a flag bit indicating a correction coefficient. For this flag bit D ₁ , the value of K ₁ may be used as it is.

The next diagnostic process relates to the correction coefficient T _s .

Regarding this T _s , the value of K3 does not lead to the relative magnification, and therefore x is calculated by performing the calculation shown in the processing 660020. Then, the following processing 66002
The contents of processing in 2,660025 are the same as those in the case of the processing of the correction coefficient K _const (processing 660015).

The final diagnostic process is the process 660030 related to the Q _a table.

In this embodiment, as is clear from the contents of this process 660030, all 64 table values are executed, and a diagnostic table is created for each of them. Further, when diagnosing the Q _a table, a comprehensive evaluation of all 64 pieces (for example, when the sum of Σ of | x−1.0 | exceeds a predetermined value) is performed and representative parameters D ₁ and D ₂ may be formed.

 Another embodiment of this diagnostic processing is shown in FIG.

In this embodiment, with each control constant, for example, K _const and T _s , it is also an effective idea to change the predetermined value XSL for evaluating it, and correspondingly, the processing 660200, 660202, 66
The contents of 0204 are set.

The control constant diagnosis processing is started when each control parameter is updated.

Therefore, according to this embodiment, it is possible to easily execute the diagnosis of the sensor (intake air flow rate sensor) and the actuator (injector) in operation only by comparing the magnitude of the correction coefficient with the predetermined value. Even in the event of an occurrence, early detection is possible and the effect of high reliability is obtained.

 Next, another embodiment of the diagnostic process is shown in FIG.

In this embodiment, the numerical value x in the determination process 660300
₀ (K _i ) represents the correction coefficient when the engine is shipped or immediately after adjustment, and the diagnosis is made by the absolute value of the difference between this and the current correction coefficient x (K _i ). The processes 660302 and 330304 are the same as those in the embodiment shown in FIG.

According to this embodiment, the characteristics peculiar to the device at the time of shipment are set as the primary evaluation value x ₀ (K _i ), and the diagnosis is performed by the difference between this value and the evaluation value at the time of the subsequent diagnosis. Therefore, there is an effect that the determination error due to the difference in the characteristics of the devices can be suppressed.

Next, FIG. 26 shows still another embodiment of the diagnostic processing,
The determination processing 660400 at this time is such that both the determination based on the difference and the deviation from the reference value are used as the evaluation parameters, similar to the embodiment of FIG.

Therefore, according to this embodiment, since both the difference from the initial value and the difference from the reference value are referred to, there is an effect that the objectivity for the determination is increased and a more accurate diagnosis can be obtained.

Next, the storage state of various data in the embodiment of the present invention will be described with reference to FIG.

First, the control constants K _const , Q _a0 (0) to Q _a0 (63) at the time of shipment are stored in the ROM, and the control constants K1, K3, K4 (0) to K4 (63) used by the control constant correcting means are stored. ) Is a correction coefficient for the current control parameters K _const , T _s , Q _a (0) to Q _a (63), and is used on the RAM.

The correction factors for the initial control constants K _const , Q _a (0) to Q _a (63) are K1 ', K3', K4 '(0) to K4' (63), but these are also on the RAM. Used in.

Next, FIG. 28 shows another embodiment of the control constant setting process, which is executed when the control parameter is changed.

First, in the process 7000100, the process of K _const is performed, in the next process 7000110, the process of T _s is performed, and finally, the process regarding Q _a is executed in the process 700120.

According to this embodiment, when the control parameter is changed, the correction calculation is performed each time, so that it is not necessary to perform the correction process every time the control parameter is used.

Further, FIG. 29 and FIG. 30 show, in yet another embodiment of the control constant setting process of the present invention, that the correction calculation is performed every time the control parameter is used.
Processing 7000200 to processing 7000204, processing 7000300, respectively.

According to these embodiments, it is not necessary to hold each data of K _const , T _s , Q _a (0) to Q _a (63) as in the embodiment of FIG. There is an effect that the capacity is small.

Next, FIG. 31 shows an embodiment of the present invention in which the diagnostic processing can be performed externally. In the engine control unit and the external engine diagnostic system, serial communication miniports SCI are provided, respectively. Through the processor of the engine diagnostic system and the RAM of the engine control unit via
The data D ₁ and D ₂ stored in this RAM can be read from the processor. Then, the processing shown in FIG. 32 is executed.

When this processing is executed, first, from the C / U (control unit), the processing 900000 for reading the engine identification code and the processing 900100 for reading the data D ₁ and D ₂ which are pre-assigned to the engine are executed. . Next, processing 900102 of reading the history data of this engine based on the engine identification data read from the C / U from the external storage device, and the past diagnostic result data of another engine similar to this engine. The process 900104 for reading the data is executed. Then, based on these pieces of information, the same processing as the above-mentioned diagnosis processing, and the diagnosis processing 900106 whose main contents are the pattern matching of the above-mentioned history data and diagnosis result data are executed, and the diagnosis result is displayed. The processing 900108 of storing in the external storage device is executed again.

Therefore, according to this embodiment, the data in the engine C / U is transferred to the external engine diagnosis system, and the diagnosis is performed by referring to the history data specific to the engine and the case of other engine. As a result, it is possible to make an accurate diagnosis that is highly objective.
In this embodiment, the diagnosis result may be written in the memory in the C / U. Further, in this embodiment, the engine may be operated and the data during driving may be taken in for diagnosis.

EFFECTS OF THE INVENTION According to the present invention, it is possible to arbitrarily diagnose the characteristics of the intake air flow rate sensor and the fuel injection valve among the sensors and actuators in the engine control device. It is possible to reliably grasp, online exhaust gas management and self-diagnosis become possible, and rational vehicle operation and maintenance can be easily obtained.

[Brief description of drawings]

FIG. 1 is a block diagram for explaining an embodiment of an engine control device diagnosis system according to the present invention, FIG. 2 is a configuration diagram of an engine system to which an embodiment of the present invention is applied, and FIG. 3 is a block of a control circuit. 4 and 5 are explanatory diagrams of the air-fuel ratio correction coefficient, FIG. 5 is an explanatory diagram of a learning map, FIG. 6 is a configuration diagram of the map, FIG. 7 is an explanatory diagram of a map changing procedure, and FIG. 8 is a learning procedure. FIG. 9 is a flow chart showing the map making / changing procedure, FIG. 10 is an explanatory diagram of map contents, FIG. 11 is an explanatory diagram of characteristic coefficients, FIG. 12 is an explanatory diagram of learning values, and FIG. Fig. 14, Fig. 14 and Fig. 15 are explanatory diagrams of characteristic indexes, Fig. 16 is an explanatory diagram of the relationship of characteristic coefficients, Fig. 17 is a block diagram for explaining operation functions, and Fig. 18 is a schematic flow of characteristic correction routine. Chart, Fig. 19 is a flow chart of correction logic, and Fig. 20 is a simple logic. FIG. 21, FIG. 21 is a detailed logic flow chart, and FIG. 22 is a diagnostic processing flow chart in one embodiment of the present invention.
24, 25, and 26 are flow charts of the judgment process with different judgment conditions, and FIG. 27 is an explanatory diagram of the memory and FIG.
FIG. 28 is a flow chart of control constant calculation processing, and FIG. 29 is a flow chart showing another example of control constant calculation processing.
FIG. 31 is a flow chart showing still another example of the control constant calculation processing, FIG. 31 is a block diagram showing still another embodiment of the present invention, and FIG. 32 is a flow chart for explaining the operation. 12 …… Injector (fuel injection valve), 24 …… Intake air flow rate sensor, 56 …… Cooling water temperature sensor, 142 …… O ₂ sensor.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (56) Reference JP-A-57-173569 (JP, A) JP-A-57-171046 (JP, A) JP-A-56-47805 (JP, A) JP-A-55- 140626 (JP, A)

Claims (9)

(57) [Claims] 1. An air-fuel ratio sensor and an intake air flow rate sensor, at least two sensors being provided, and a predetermined air-fuel ratio control target value set in advance according to a plurality of predetermined engine operating states, and a plurality of corresponding air-fuel ratio control target values. The deviation from the actual air-fuel ratio detected value by the air-fuel ratio sensor under the operating condition of the engine is calculated and held as a plurality of predetermined air-fuel ratio correction coefficients, and the fuel injection valve is controlled using these air-fuel ratio correction coefficients. As a result, in the engine control device diagnosis system of the method of performing feedback control of the air-fuel ratio, the characteristic correction coefficient for correcting the flow rate detection characteristic of the intake air flow rate sensor is calculated by the plurality of calculated air-fuel ratio correction coefficients. Of these, a calculation processing unit that performs calculation based on a change in the air-fuel ratio correction coefficient that is calculated and held when the detection values of the intake air flow rate sensor are equal; A characteristic correction coefficient for correcting the injector characteristic of the fuel injection valve is calculated based on the average change of the calculated plurality of air-fuel ratio correction coefficients, and the characteristic correction coefficient for correcting the invalid injection time of the fuel injection valve is calculated. The calculation coefficient is calculated based on the change in the air-fuel ratio correction coefficient calculated and held when the basic fuel injection amount given to the fuel injection valve is small among the calculated and held air-fuel ratio correction coefficients. An engine control device diagnostic system is characterized in that it is configured to independently perform abnormality determination of the intake air flow rate sensor and the fuel injection valve based on the values of these characteristic correction coefficients. 2. The invention according to claim 1, wherein the comparison value level for determining the abnormality is a memory retention of a characteristic correction coefficient at the start of use of the intake air flow rate sensor and the fuel injection valve and at the time of predetermined maintenance. An engine control device diagnostic system characterized in that it is configured to be at least one of the values. 3. The invention according to claim 1, wherein the abnormality determination is made based on whether or not at least one of the absolute value and the rate of change of the characteristic correction coefficient exceeds a predetermined comparison value level. An engine control device diagnosis system characterized by the above. 4. The invention according to any one of claims 1 to 3, further comprising storage means for holding the characteristic correction coefficient and communication means coupled to the storage means, wherein the characteristic correction coefficient is read from the outside. An engine control device diagnostic system characterized by being configured so that it can be output. 5. The invention according to claim 4, wherein the characteristic correction coefficient is externally read, and abnormality determination of the intake air flow rate sensor and the fuel injection valve is performed externally based on the externally read characteristic correction coefficient. An engine control device diagnostic system characterized by being configured. 6. The invention according to claim 5, wherein identification data unique to each engine for which the characteristic correction coefficient is to be read is provided, and the characteristic correction coefficient read out is classified and stored externally based on the identification data. An engine control device diagnosis system, characterized in that, each time an abnormality judgment is repeated, the judgment processing is performed by referring to the characteristic correction coefficient at the time of the previous diagnosis. 7. The engine control according to claim 6, wherein the determination processing is configured to be performed using an inference mechanism of knowledge processing with reference to a diagnosis result of another case. Device diagnostic system. 8. The engine control device diagnosis system according to claim 7, wherein the determination process is performed by also referring to a past determination result. 9. The engine control device diagnosis system according to claim 8, wherein the determination process is performed by referring to other information as well.