JP2000170581A - Control device for caseous fuel engine - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve reliability by transforming combustion in a stable condition even when fuel nature is changed, resulting in combustion failure at the time of start. SOLUTION: Whether a combustion failure condition at the time of a start exists or not is judged based on the operation condition of an engine, under a condition for supplying gaseous fuel by feedback control via a fuel injecting means; and when a combustion failure condition at the time of a start is judged, the change of the kind or nature of the gaseous fuel is judged, to correct at least one of an air-fuel ratio correction value in a feedback control used before, and the pressure and ingition timing of the gaseous fuel supplied to the fuel injecting means.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control apparatus for a gaseous fuel engine which supplies gaseous fuel through fuel injection means by feedback control, and more particularly to a case where the kind or property of gaseous fuel changes. The present invention also relates to a device capable of appropriately controlling an engine.

[0002]

2. Description of the Related Art In recent years, environmental protection requirements (eg, exhaust gas regulations, fuel efficiency regulations, etc.) surrounding automobiles have been steadily reinforced. Further, as is well known, current engine fuels are mainly liquid fuels such as gasoline and light oil, and there are concerns about a shortage of fuel supply and a rise in fuel prices due to restrictions on reserves.

[0003] The development of alternative energy is rapidly progressing, and typical examples include electric vehicles, hybrid vehicles composed of liquid fuel and electricity, alcohol and gas (natural gas, propane gas, hydrogen gas, etc.). Gas vehicles, but infrastructure such as fuel supply facilities,
Natural gas vehicles are one step ahead in terms of cost. Since natural gas vehicles use methane gas as the main fuel, exhaust gas emissions can be reduced as compared with conventional liquid fuel vehicles.

Further, by adopting a multipoint injection system (MPI) for injecting fuel independently for each cylinder, an optimum air-fuel ratio control becomes possible, and exhaust gas performance, fuel consumption performance, power performance, drivability, etc. In addition, the conventional engine control system for liquid (gasoline) fuel can be diverted in a large frame, and the production cost and the effect of suppressing the production can be obtained. This satisfies the recent environmental protection requirements (emissions regulations, fuel efficiency regulations, etc.) surrounding automobiles, which are continually being strengthened, and are highly expected in the future.

[0005]

By the way, the gas composition (component) of a commercially available natural gas differs depending on the gas maker that manufactures the gas, and a gas fuel such as a natural gas vehicle is used. In automobiles, the stoichiometric ratio (required air-fuel ratio) for combustion becomes important. FIG. 7 shows a schematic diagram. As shown in FIG. 7, in general, natural gas has a composition centered on a stoichiometric air-fuel ratio of about 17 (16.8: methane 80%) by component adjustment. This is called 13A gas. However, due to variation by the manufacturer,
The required air-fuel ratio indicates a variation width as shown in the figure.
If the methane concentration is 100% and the required air-fuel ratio is the maximum (the right end of the distribution), the frequency distribution spreads at about 17 at the top. The gas having a large variation width is often referred to as 12A gas.

In engine control, there are known feedback control for measuring the oxygen concentration of exhaust gas and correcting the fuel injection amount, and feedback learning control for storing and learning a correction value by the feedback control in a memory (a backup RAM or the like). Thus, it is possible to absorb most of the required air-fuel ratio variations. However, there is also a gas with a composition exceeding the control range (over-lean area A on the left side in the figure), and when this gas is charged, knocking and misfiring due to extreme combustion failure (worst, engine stall, unstartable) There is a risk of occurring.

[0007] Further, the variation in the gas composition also has an octane value, which is a hindrance factor for the optimal ignition timing, and affects power performance and knocking performance. In addition,
A method of setting a methane concentration sensor or the like in the fuel pipe for air-fuel ratio adjustment is also conceivable, but it is disadvantageous in terms of cost, and because the operating environment (particularly temperature, vibration, waterproof environment) of the automobile is severe, It is not practical in terms of reliability.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems, and has as its object to detect that the type and properties (composition) of gaseous fuel have changed, such as an expensive methane concentration sensor. Without using a fuel property detector, it can be reasonably and reliably detected, the effects of the detection can be quickly absorbed, the air-fuel ratio, the ignition timing, etc. can be maintained in the optimum state, the drivability deteriorates, knocking resistance, It is an object of the present invention to provide a control device for a gaseous fuel engine capable of effectively suppressing a decrease in exhaust performance.

[0009]

In order to achieve the above object,
The control device according to the present invention is provided in a gaseous fuel engine configured to supply gaseous fuel through fuel injection means by feedback control, and basically performs combustion at start-up based on an operating state of the engine. It is determined whether or not it is in a defective state, and if it is determined that the combustion is in a poor starting state, it is determined that the type or property of the gaseous fuel has changed, and the feedback control that has been used so far has been determined. , At least one of the air-fuel ratio correction value, the pressure of the gaseous fuel supplied to the fuel injection means, and the ignition timing is corrected.

[0010] In a preferred embodiment of the control device of the present invention, it is determined whether or not the starting combustion is in a poor state, in other words,
It is determined whether the type or properties of the gaseous fuel have changed, whether the state in which the engine speed is equal to or less than a predetermined value at the time of starting has continued for a predetermined period, or the state in which the throttle valve opening is equal to or more than a predetermined value at the time of starting. Is determined based on whether or not has continued for a predetermined period.

In this case, when it is determined that the combustion is in a poor combustion state at the time of starting, preferably, control such as correcting the air-fuel ratio correction value to the increasing side or correcting the pressure of the gaseous fuel to the high pressure side is performed. I do. In another preferred embodiment, a correction value of the air-fuel ratio correction value or a correction value of the pressure of the gaseous fuel at the time of the start-time combustion failure determination is set for each operation region determined based on an operation state of the engine. After the start impossible determination, stopping the fuel injection by the fuel injection means for a predetermined period, and correcting the air-fuel ratio correction value or the pressure of the gaseous fuel after the predetermined period.

According to still another preferred aspect, after the determination of the combustion failure at the time of starting, a total air-fuel ratio correction value is calculated, and a stoichiometric mixture ratio of gaseous fuel is estimated based on the total air-fuel ratio correction value. Estimating the value of the octane number of the gaseous fuel based on the stoichiometric mixture ratio and correcting the preset basic ignition timing with the octane number estimated value, in which case the total air-fuel ratio correction value is determined by A value that takes into account the correction value of the fuel pressure may be added, or a plurality of basic ignition timing maps may be prepared, and these may be switched and used according to the octane number estimated value.

One preferred specific embodiment of the control apparatus for a gaseous fuel engine according to the present invention is an engine operation represented by an intake air amount, an engine speed, a cooling water temperature, a throttle valve opening, an engine load, and the like. Operating state detecting means for detecting the state; fuel state detecting means for detecting the temperature and / or pressure of the gaseous fuel upstream of the fuel injection valve; and an exhaust gas by an oxygen sensor or an air-fuel ratio sensor for detecting the oxygen concentration in the exhaust gas. Exhaust state detecting means for detecting a state, basic injection pulse calculating means for calculating a basic injection pulse of gaseous fuel based on the detected operating state, and an air-fuel ratio correction value based on the detected operating state and fuel state. Air-fuel ratio correction value calculating means for calculating, feedback correction value calculating means for calculating a feedback correction value based on the detected exhaust state, Feedback learning value calculating means for calculating a feedback learning value based on a feedback correction value and storing the feedback learning value in a memory; and an injection pulse for calculating an injection pulse based on the basic injection pulse, the air-fuel ratio correction value, the feedback correction value, and the feedback learning value. Calculating means, an injection drive pulse output means for supplying the calculated injection pulse to the fuel injection valve, and a poor start combustion state determining means for determining whether or not a poor start combustion state based on the operating state; When it is determined that the combustion is in a poor combustion state at the time of starting, the air-fuel ratio correction value at the time of poor combustion at start is set to correct the calculation result of the air-fuel ratio correction value calculating means. A feedback learning value for erasing the feedback learning value stored in the memory when it is determined that the combustion is in a poor combustion state; Removing means, air-fuel ratio correction value returning means for shifting the air-fuel ratio correction value at start-up combustion failure to a temporary air-fuel ratio correction value after setting the air-fuel ratio correction value at start-up combustion failure, and the air-fuel ratio correction value at start-up combustion failure. Feedback learning value calculation re-executing means for permitting the calculation of the feedback learning value after the setting.

Another preferred embodiment is an operating state detecting means for detecting an operating state such as an intake air amount, an engine speed, a cooling water temperature, a throttle valve opening, an engine load, etc., and an upstream of the fuel injection valve. Fuel state detecting means for detecting the temperature and / or pressure of gaseous fuel on the side, exhaust state detecting means for detecting an exhaust state by an oxygen sensor or an air-fuel ratio sensor for detecting oxygen concentration in exhaust gas, and detected operating state A fuel pressure setting means for setting a fuel pressure adjustment value for the fuel pressure regulator based on the fuel pressure regulator; a basic injection pulse calculating means for calculating a basic injection pulse of gaseous fuel based on the detected operating state; An air-fuel ratio correction value calculating means for calculating an air-fuel ratio correction value based on the detected exhaust gas condition; Feedback correction value calculating means, feedback learning value calculating means for calculating a feedback learning value based on the feedback correction value and storing the feedback learning value in a memory, the basic injection pulse, air-fuel ratio correction value, feedback correction value, and feedback learning value. Injection pulse calculating means for calculating an injection pulse according to the following, injection driving pulse output means for supplying the calculated injection pulse to the fuel injection valve, and starting to determine whether or not a combustion failure state at the start based on the operating state And a fuel pressure adjustment value for correcting the fuel pressure adjustment value set by the fuel pressure setting means to set the fuel pressure adjustment value at the time of poor combustion at the start when it is determined that the combustion condition at the start is poor. Means for erasing the feedback learning value stored in the memory when it is determined that the combustion is in a poor start condition. Feedback learning value erasing means, fuel pressure adjustment value return means for temporarily shifting the fuel pressure adjustment value at the time of start-up combustion failure to the fuel pressure adjustment value after setting the fuel pressure adjustment value at the time of start-up combustion failure, and fuel pressure adjustment at the time of start-up combustion failure. Feedback learning value calculation re-executing means for permitting the calculation of the feedback learning value after setting the value.

In this case, it is preferable that the result of the combustion failure determination at the start is stored in the memory as a combustion failure history at the start.
The setting of the air-fuel ratio correction value at the time of poor combustion at the start or the fuel pressure adjustment value at the time of poor combustion at the start is executed based on the combustion failure history at the start, and is stored in the memory when the feedback learning value after the restart of the feedback learning converges within a predetermined range. , The history of poor combustion at startup is erased.

[0016] More preferably, the feedback learning value is composed of a first feedback learning value calculated for each operating region determined according to the operating state of the engine, and a second feedback learning value independent of the operating region. ,
After the setting of the air-fuel ratio correction value at the time of poor combustion at the start or the setting of the fuel pressure adjustment value at the time of poor combustion at the start, only the second feedback learning value is allowed to be operated, and after the air-fuel ratio correction value is restored or the fuel pressure adjustment command value is restored, After completion of the calculation of the second feedback learning value, the first feedback learning value is replaced with the first feedback learning value.
The calculation of the feedback learning value is permitted.
If the feedback correction value from the feedback correction value calculation means is out of the predetermined range, the feedback learning value is deleted, and the feedback learning by the feedback learning value calculation means is started again.

According to a preferred embodiment of the control apparatus for a gaseous fuel engine according to the present invention having the above-described structure,
A change in the type and properties (composition) of the gaseous fuel can be reasonably and reliably detected based on the operating state of the engine without using a fuel property detector such as an expensive methane concentration sensor. Since the air-fuel ratio correction value, fuel pressure, ignition timing, etc. can be set appropriately according to the type and properties of the fuel currently used, the effects of changes in the type and properties (composition) of the gaseous fuel can be reduced. Can be absorbed early,
The air-fuel ratio, ignition timing, etc. can be maintained in optimal conditions,
Deterioration of drivability, anti-knocking characteristics, and reduction of exhaust performance can be effectively suppressed.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the drawings. First, FIG. 1 is a system configuration diagram of an engine control device used in the present invention. The state information of the engine is passed through an air cleaner 3, passed through an intake pipe 4 and taken into the engine 2, an intake air amount sensor 6 for measuring a mass flow rate of intake air, a water temperature sensor 8 for measuring a cooling water temperature of the engine, and an intake air temperature. Detection signals from an intake air temperature sensor 9 to be measured, a crank angle sensor 27 to detect a crank angle, a throttle sensor 21 to detect a throttle valve angle, and the like are input to the control unit 1, and the control unit 1 receives signals from the crank angle sensor 27. The engine speed is calculated based on the detection signal, the engine load is calculated from the throttle opening, the intake air amount, and the engine speed, and the optimum fuel injection pulse (driving pulse width) to be supplied to the engine based on these various information, fuel Injection timing, ignition timing, and the like are calculated, and the injector 18 serving as a fuel injection valve is operated.
Inject fuel.

The ignition system calculates an ignition timing according to the operation state based on various information from the various sensors, and ignites the air-fuel mixture with the ignition plug 7 by energizing / cutting off an ignition coil (not shown). Further, the control unit 1 calculates a feedback correction value based on a signal from the oxygen sensor 20 for detecting the oxygen concentration in the exhaust gas discharged through the exhaust pipe 5 so that the mixture becomes the stoichiometric air-fuel ratio. To correct the fuel injection pulse.

On the other hand, the gaseous fuel supply system includes a fuel cylinder 10 for filling and storing gaseous fuel in a high pressure state, a fuel pressure regulator 16 for adjusting the fuel pressure of fuel injection, an injector 18 for injecting fuel, the cylinder 10 and the regulator 16. , A low-pressure pipe 17 connecting the regulator 16 and the injector 18, and a high-pressure pipe 1.
Shut-off valves 11 and 12 installed on the cylinder side and regulator 16 side of 5, fuel temperature sensor 13 installed on low-pressure pipe 17 and detecting fuel state (fuel temperature, fuel pressure), fuel pressure sensor 1
4. Although not shown, the high-pressure pipe includes a high-pressure-compatible fuel pressure sensor that detects abnormal fuel pressure in the high-pressure pipe.

A starter signal from the starter 22 is input to the control unit 1 to determine whether the engine is started or not. FIG. 2 shows the regulator 1
6 shows an outline of the internal structure. In this example, a fuel pressure adjustment signal Cf from the control unit 1 is added to the set fuel pressure adjustment function preset in the regulator 16.
This allows variable control of fuel pressure. High pressure piping 15
The high-pressure fuel from is supplied to the mechanically set initial fuel pressure by the spring 23 in the regulator, and the regulated fuel flows to the low-pressure pipe 17 side. When it is desired to change the fuel pressure, the position of the shaft 26 is displaced by applying a fuel pressure adjustment command signal Cf to the coil 26, and the spring constant of the spring 23 is changed, thereby changing the controlled fuel pressure.

The control unit 1 is constructed using a microcomputer. In this case,
It can be considered that the control unit 1 has various function realizing means as shown in a control block diagram of FIG. In the following, the control contents of the control unit 1 will be described with reference to a control block diagram of FIG. 3 and subsequent drawings, a flowchart of FIG. 5 and thereafter showing an example of a program executed by the microcomputer, and a time chart of FIG. It will be described sequentially.

First, a description will be given with reference to the control block diagram of FIG. The basic fuel injection pulse calculation uses a pulse calculation method for gasoline, and the operating state detecting means (6, 8, 21, 27,...) Of the block 28 uses the basic pulse based on the intake air amount, engine speed, and the like. The operation state of the engine for calculating the width is detected, and based on the information, the block 2 is detected.
The basic pulse width calculating means 9 calculates the basic pulse width (the required injection pulse width per cylinder determined from the intake air amount). The fuel state detecting means (13, 14) of the block 30 detects information such as the temperature, pressure, water temperature, throttle angle, etc. of the gaseous fuel, and the air-fuel ratio correction value calculating means of the block 31 detects the air-fuel ratio correction value (for example, 13A gas. An air-fuel ratio adjustment value such as a reference water temperature correction amount or acceleration / deceleration correction amount is calculated. The exhaust gas state detecting means (20) of the block 32 detects the oxygen concentration in the exhaust gas, and the feedback correction value calculating means of the block 33 calculates the feedback correction value. The feedback learning value calculation means of block 34 calculates the feedback learning value based on the calculation result of the feedback correction value calculation means of block 33, and stores (saves) it in a memory such as a backup RAM. In addition,
In the exhaust state detection means of the block 32, accurate information cannot be detected unless the oxygen sensor 20 has risen to the optimum temperature. Therefore, the sensor output has started to operate (the transition from rich to lean or the transition from lean to rich). ), It is determined that the sensor has been activated, and after it is determined that the sensor has been activated, the calculation of the feedback correction value is started by the feedback correction calculation means of the block 33.

Although not shown in the block diagram, the signals from the fuel temperature sensor 13 and the fuel pressure sensor 14 (block 24) described in the fuel supply system are related to the basic fuel temperature (for example, an arithmetic expression: From の (input fuel temperature value) / √ (calculation result of basic fuel temperature value), the relationship between the fuel temperature correction coefficient and the basic pressure (for example, calculation formula: (basic fuel pressure value) / (input fuel pressure value))
The fuel pressure correction coefficient is calculated from the calculation result from the above), and the multiplication result is calculated as a pulse width correction value. (In the present control block diagram, it is included in blocks 30 and 31.) This is the fuel state (fuel pressure and fuel temperature) depending on the operating environment.
Although the air-fuel ratio fluctuates due to a change or a change in the fuel pressure due to the operation state (transient / steady state), this is because the fluctuation is corrected by pulse correction based on the fuel temperature or the fuel pressure.

The basic pulse width, the air-fuel ratio correction value, the feedback correction value, and the feedback learning value calculated in blocks 29, 31, 33, and 34 are finally output from the injector 18 by the injection pulse calculation means in block 35. The injection pulse (pulse width) is calculated as a composite value, and the injection drive pulse output means of the block 36 outputs an injection drive pulse having the injection pulse width to the injector 18 at a predetermined timing to perform fuel injection. On the other hand, the main parts of the device of the present invention are blocks 37 to 41.

An engine speed, a throttle opening, a starter signal, and the like are input to the start-up poor combustion state determination means of the block 37, and the start-up poor combustion state determination means based on the information as described later. Then, it is determined whether or not the engine is in a poor combustion state at the time of starting (that is, a situation where a complete explosion does not continue despite cranking continues). As described above, the used fuel has a composition variation, and if the fuel is extremely deviated, it becomes difficult to start. In the case of natural gas vehicles, it is often impossible to start the engine during fuel filling (such as shutting off the starter power supply or shutting off fuel using a shutoff valve in the fuel pipe), and the opportunity to change the composition by filling with fuel is limited. In the background, there is a case where a fuel of a completely different composition is injected at the time of starting. (For example, from 13A to 1
Filling changed to 2A. )

If it is determined that the combustion is in a poor combustion state at the time of starting, the memory for storing the feedback learning value is erased by the feedback learning value erasing means of the block 38. This is because the feedback learning value cannot be used as it is because the fuel composition is different (it is not known whether it can be used). On the other hand, the air-fuel ratio correction value correction means in block 39 is an air-fuel ratio correction value correction value (for example, an air-fuel ratio correction value based on 12A gas) which is an alternative to the air-fuel ratio correction value by the air-fuel ratio correction value calculation means in block 31. And set
The air-fuel ratio correction value is replaced with an air-fuel ratio correction value correction value. Since it is considered that the air-fuel ratio is excessively lean in the poor combustion state at the time of starting, the air-fuel ratio almost matches the normal air-fuel ratio from the excessive lean state,
A correction value sufficient to match the air-fuel ratio with poor combustion at start-up is defined as an air-fuel ratio correction value correction value.

Thereafter, the air-fuel ratio correction value correction means of the block 40 gradually returns the air-fuel ratio correction value correction value to the original air-fuel ratio correction value. At the same time, the feedback learning value is re-executed by the feedback learning value calculation re-executing means of the block 41, and a new feedback learning value is calculated by the feedback learning value calculating means of the block 34.
Let me save. This makes it possible to optimally control the fuel injection amount (air-fuel ratio) even when new fuel is charged and the type and properties of the fuel are changed.

However, the blocks 40 and 41 need to be synchronized, so that the blocks 40 and 41
Is triggered by the start of the calculation by the feedback correction value calculation unit (block 33) after the activation determination of the oxygen sensor 20 is completed, and the calculation by the feedback correction value calculation unit (block 33) is not started. Then, the air-fuel ratio correction value in the block 31 is held.

FIG. 5 shows a control flowchart for this.
This flowchart is executed as a fixed-time task. In step 47, it is determined whether or not there is a combustion failure at startup. If NO, the current task processing ends without doing anything. If YES, the process proceeds to step 48, and it is determined whether or not the air-fuel ratio correction value (correction value) at the time of start-up combustion failure has already been set. (That is, it is determined whether or not the YES determination in step 47 is the first time.) If NO (first time in determination of poor combustion at startup), in step 49, the air-fuel ratio correction value for poor startup combustion (correction value) Is set, the feedback learning value is immediately cleared in step 50, and the current task processing ends.

In step 48, if YES (the determination for the second time or later), it is determined whether or not the feedback correction value calculation has been started in step 51. If NO, the current task processing is terminated, and YES If there is, in step 52, a process of returning the air-fuel ratio correction value (correction value) at the time of poor combustion at start-up to the original air-fuel ratio adjustment value by a predetermined amount is performed. The learned value is stored in the memory at step 54.

The timing chart is as shown in FIG. 9, and the air-fuel ratio adjustment value is set on the increasing side (the air-fuel ratio is on the rich side) in order to prevent excessive leaning (as described in FIG. 7).
As means for correcting the air-fuel ratio, there are a method of correcting the air-fuel ratio correction value (injection pulse width) and a method of correcting the gas pressure (fuel pressure) described below. The air-fuel ratio is determined by the mass ratio between air (atmosphere) and fuel (gas). By varying the fuel pressure, the density of the fuel changes, so that the air-fuel ratio can be corrected. When the fuel pressure is a predetermined pressure (for example, 3 atmospheres or more), the gas flow rate from the injector 18 becomes a sonic state and the volume flow rate is constant, but the mass flow rate changes due to the change in the fuel pressure. FIG. 4 shows a case where the adjustment is performed using this. The premise is the same as FIG. 3, and only the differences will be described.

The operating state detecting means of the block 42 detects the operating state, for example, the intake air amount, the engine load, the cooling water temperature, etc., and thereby sets a predetermined fuel pressure (fuel pressure adjustment value) suitable for the operating state to the fuel pressure setting of the block 43. The driving signal corresponding to the fuel pressure adjustment value is supplied to the fuel pressure regulator 16 by the fuel pressure correction signal output means of the block 44.

If it is determined by the starting combustion failure state determining means in block 37 that the combustion is in the starting combustion failure state, the feedback learning value is eliminated by the feedback learning value elimination means in block 38 as in FIG. Block 4
The fuel pressure adjustment value correcting means of No. 5 sets the fuel pressure adjustment value to a predetermined value (13
Corrected value corresponding to the air-fuel ratio difference between A gas and 12A gas)
The fuel pressure setting means of the block 43 sets the fuel pressure adjustment value at the time of poor combustion at the start, and the fuel pressure correction signal output means of the block 44 supplies a drive signal corresponding to the fuel pressure adjustment value at the time of poor combustion at the start to the fuel pressure regulator 16. To supply.

When the calculation of the feedback correction value is started in the block 33, the blocks 46 and 42 are started, and the fuel pressure adjustment value return means in the block 46 gradually reduces the fuel pressure adjustment value at the time of starting combustion failure to the original fuel pressure adjustment value. And return to
Feedback learning value calculation re-executing means 4 of block 41
In step 1, the optimum feedback learning value is calculated in the process of returning the fuel pressure, and is stored in the memory. FIG. 6 shows a control flowchart for this. The assumption is to start from step 47 in FIG. 5, and steps 48 to 52 in FIG. 5 are replaced with steps 55 to 57 (and 51) in FIG.

If YES is determined in step 47,
Proceeding to step 55, it is determined whether the fuel pressure adjustment value at the time of poor combustion at startup has already been set. If NO, the start-up combustion failure fuel pressure adjustment value is set in step 56, and the process proceeds to step 50. If YES, the process proceeds to step 51, and it is determined whether the feedback correction value calculation has been started in step 51. If the determination is NO, the current task process is terminated. If the determination is YES, a process of returning the fuel pressure adjustment value at the time of poor combustion at startup to the original fuel pressure adjustment value by a predetermined amount is performed in step 52, and then the process proceeds to step 52. Move to 53. The timing chart is as shown in FIG. 10, and the fuel pressure adjustment adjustment value is set on the high pressure side (on the side where the gas density is increased) in order to prevent excessive lean (as described in FIG. 7).

Learning by the methods shown in FIGS. 3 (FIGS. 5 and 9) and FIGS. 4 (FIGS. 6 and 10) makes it possible to control the air-fuel ratio optimally according to the composition of the charged fuel. . Also,
As shown in FIG. 11A, the air-fuel ratio correction value at the time of poor combustion at start (correction value) and the fuel pressure adjustment value at the time of poor combustion at start (correction value) are as shown in FIG.
The engine speed and the engine load are set in advance in accordance with each operation region, and even if searched and used in each operation region, they are set in advance according to the cooling water temperature as shown in FIG.
It may be searched and used for each cooling water temperature range, and this enables detailed setting of correction values and adjustment values according to the operating state.

However, while the fuel pressure adjustment value is being operated (changed) and until feedback learning is completed, block 43 is executed.
It is necessary to temporarily prohibit the calculation of the basic fuel pressure and the fuel pressure correction coefficient in and to fix them to a predetermined value (for example, 100%).
This is because the effect of the air-fuel ratio correction with the fuel pressure adjustment value is absorbed (invalidated) by the air-fuel ratio correction value. Note that FIG.
Incombustible state determination at start-up in FIG. 4 (block 37)
Is stored and cleared as a start-up combustion failure history as shown in the flowchart of FIG.

First, a start-up combustion failure history indicating whether or not the start-up combustion failure state has been determined in step 58 is checked. If there is no history, a start-up state is confirmed in step 59, and in step 60 If it is determined that the start-up combustion is poor, a start-up combustion failure history is set in step 61. The start-up combustion failure history may be set (stored) in a memory such as a backup RAM in order to prevent a delay in the determination of the start-up combustion failure state due to the repeated ON / OFF of the ignition key. If YES is determined in step 58, that is, if it is already determined that the combustion at start is poor, the convergence state of the feedback learning value is confirmed in step 62, and if NO, the convergence state is determined (if not converged). The task processing is ended, and if YES (if the convergence is achieved), the combustion failure history at start is cleared. This is because it is determined that the control of the air-fuel ratio has absorbed the variation in the gas composition due to the convergence of the feedback learning value, and it is no longer necessary to save the combustion failure history at startup.

By the way, as shown in FIG. 15, the convergence determination of the feedback learning value is such that the change amount of the calculation result of the feedback learning executed in the fixed time task cycle falls within a predetermined range (time T1), and the state is maintained for a certain period (time T1). Td) When the feedback learning value has continued (at time T2), it is determined that the feedback learning value has converged. Here, the target value is an ideal feedback learning value for absorbing the gas composition, and is not preset as data.

Next, a description will be given of the determination of combustion failure at the start in the above description. Embodiments include a method shown by the flowchart in FIG. 12 and a method shown by the flowchart in FIG. First, a description will be given from the flowchart of FIG. FIG. 12 detects a state in which the engine rotation does not increase in spite of the start state, that is, does not shift to a complete explosion. If NO is determined in the determination as to whether or not there is a combustion failure history at the start in step 58 of FIG. 8, a signal from the starter (a signal indicating whether or not the starter is cranking) in step 65 of FIG. See step 66
To determine if the engine is starting. Here, if the engine is not being started in the determination of NO, the combustion failure timer at start described later is cleared and the routine proceeds to step 71.

If it is determined that the engine is being started, the process proceeds to step 67 to determine whether the engine speed is equal to or less than a predetermined value. If NO, the process goes to step 69 without doing anything. If YES, the combustion failure timer at start is incremented (added), and it is determined in step 69 whether the low rotation state has continued for a predetermined period or more. If NO, the process proceeds to step 61, and if YES, the process proceeds to step 61. It is determined that the combustion is in a poor combustion state at the time of starting, and the routine proceeds to step 61.

FIG. 13 shows the use of the habit of the driver trying to start the engine by scavenging the combustion chamber by fully opening the accelerator pedal in a state where starting is difficult. The driver is mechanically or electrically connected to the accelerator pedal. The starting combustion failure is determined based on the throttle valve opening. The processing of steps 65 and 66 is performed in the same manner as in FIG.
If the ES determination is made, it is determined in step 72 whether the throttle opening is equal to or more than a predetermined value.

If NO, the process proceeds to a step 61,
In the case of YES, the process shifts to step 68, where the processes from step 68 to step 70 are executed as in FIG. FIGS. 12 and 13 show that, even if the determination results alone are used to determine the start-up combustion failure (OR determination), if both are satisfied, the start-up combustion failure is determined (AND determination).
Judgment) may be performed.

FIG. 14 shows another embodiment as an application related to feedback learning. FIG.
FIG. 5 is a flowchart showing the feedback learning performed in two stages separately from FIG. 5. The first feedback learning means learns for each operation region determined according to the operation state of the engine, and the learning is performed independently of the operation region. And second feedback learning means. (The same applies to the adjustment of the fuel pressure adjustment value, but the description is omitted here.)

If NO is determined in step 48, the air-fuel ratio correction value at the time of poor combustion at start is set in step 73, the first feedback learning value is cleared in step 74, and the second feedback learning value is cleared in step 75. In step 76, the first feedback learning control is prohibited. If YES is determined in the step 48, it is determined whether or not the feedback correction value calculating means is started in a step 51. If NO, the process is ended. If YES, the process proceeds to a step 52, where the air-fuel ratio correction value is determined. Is temporarily changed, the second feedback learning value is read in step 77. In step 78, the convergence of the second feedback learning value is determined.
At step 9, the second feedback learning value is further calculated and stored at step 80 in the memory.

If the determination is YES, the first determination is made in step 81.
The feedback learning value calculation is restarted, and in step 82, the calculation result of the first feedback learning value is stored in a predetermined memory allocated according to the operation region. This allows
It is possible to absorb variations in the type and properties (gas composition) of the gaseous fuel in a short period of time, and to absorb variations in the air-fuel ratio in accordance with the operating state, thereby enabling precise learning control.

The normal fuel filling operation is performed while the engine is stopped. The interlock mechanism which stops the operation of the engine (makes starting impossible) during the charging is operated. In such a case, there is a risk that the fuel may be charged even during operation, and if the fuel is mixed in an uneven composition, the operating state (combustion) may be disturbed. Such a state also needs to be detected, and this state is implemented by monitoring the feedback learning value as shown in FIG.

The normal feedback learning value is almost constant within the control range. However, if the feedback learning value is large during operation, the feedback learning value is stuck to the upper limit side in an attempt to correct to the rich side. In this state, there is a great risk that combustion will be hindered. Therefore, a state in which the fuel is out of the control range is detected, and by monitoring the duration thereof, it is detected that the fuel with a different composition has been charged.

Although the detailed description is omitted here, the fuel temperature may be measured, and if the fuel temperature changes rapidly, it may be determined that the fuel has been charged during operation. (However, in this case, since a change in the composition cannot be detected, it is better to make a composite determination with the feedback learning value monitor.) This flowchart is shown in FIG.
And is arranged between step 47 and step 48 in FIG. After the NO determination in step 47 (and during engine rotation), the flow shifts to step 83 to determine whether the feedback correction value is an appropriate value. It is assumed that the determination is made based on whether the state is within the range and the duration of the state.

If NO, the process is terminated, but if YES, the process proceeds to a step 48, and the subsequent processes are executed in the same manner as the process at the time of starting combustion failure. This makes it possible to cope with a composition change due to filling during operation. Here, in a gasoline-powered vehicle, there is a risk of being directly connected to "defective combustion at start-up" = "spark plug smolder", and control for stopping the fuel injection at the time of the operation of fully opening the accelerator at the time of start-up and promoting scavenging is known. I have. As described above, the basic system of a natural gas vehicle uses gasoline, but from the viewpoint of promoting scavenging, it is better to perform the same processing on a natural gas vehicle.

As an example, FIG. 18 shows a flowchart of stopping fuel injection at the time of starting combustion failure. In step 47, it is determined whether or not the combustion at startup is poor. If YES, the process proceeds to step 84, and it is determined whether or not the fuel injection is stopped due to the combustion failure at startup. If the fuel injection is stopped, the process proceeds to step 87. If not, the fuel injection timer is cleared and the fuel injection is stopped.

Thereafter, at step 87, it is determined whether or not the injection stop period has exceeded a predetermined value. If NO, the process is terminated (injection stop is continued), and if YES, fuel injection is permitted. Thereafter, the process proceeds to step 48 in FIG. 5, and the same processing as described above is executed. However, in this case, it is meaningless to start the adjustment value setting means during the fuel stop, and it is better to start the adjustment value setting means after the fuel stop processing is completed.

Finally, a method of absorbing octane number variation, which is another variation in gas composition, will be described.
As is well known, the octane number is related to the burning speed (ignitability), and as the octane number is higher, knocking is less likely to occur,
Although it depends on the operating state and the engine, when the best torque point (ignition timing) of the engine is on the advanced side of the knock occurrence point (ignition timing), the ignition timing can be made closer to the best torque point.

In the case of natural gas, the main component is methane gas, but it also contains a small amount of ethane, propane, butane gas, etc., and the octane number changes with the stoichiometric air-fuel ratio depending on this content. There is a degree of variation width. It is very difficult to analyze the components strictly and detect the air-fuel ratio and octane number based on the analysis. However, it is presumed that if the stoichiometric air-fuel ratio (air / fuel ratio) decreases, the octane number decreases.

FIG. 19 shows blocks 29 and 31 and FIG.
33 to 36, 39, and 40 (the explanation is omitted), the main points of the invention of the ignition timing correction (blocks 90 to 97)
Is added. Blocks 90 to 92 are an outline of the ignition timing control which is a premise of the present invention. The operating state detecting means (6, 8, 21, 27,...) Of the block 90 detects the operating state of the engine from various sensors. Find the engine load and engine speed.

The basic ignition timing setting means of the block 91 sets an optimum ignition timing based on the operation state. For example, the ignition timing according to the engine load and the engine speed is set in advance in the memory (set as an ignition timing map),
The ignition timing map in the memory is searched from the engine load and the engine speed obtained from various sensors to calculate the ignition timing. (If there is information such as cooling water temperature and intake air temperature, the ignition timing after search may be corrected.)
Thereafter, the process proceeds to the ignition timing output means of the block 92, and the air-fuel mixture is ignited by operating the ignition plug 7 at the ignition timing calculated from the crank angle information of the crank angle sensor and the cylinder discrimination information.

The current supply time to the ignition coil is calculated separately from the above, but will not be described here. In FIG. 2, the injection pulse calculation is performed from blocks 29, 31, 33, and 34. The total air-fuel ratio correction value for the basic injection pulse (total correction value) is calculated by the total air-fuel ratio correction value calculation means in block 93. ) Is calculated. The stoichiometric air-fuel ratio calculating means in block 94 calculates the displacement of the gas composition (the deviation of the stoichiometric air-fuel ratio) from the sum of the correction values of the air-fuel ratio (total air-fuel ratio correction value).
Ask for. Since the basic pulse width calculated in the block 29 is calculated based on the reference gas composition (for example, the stoichiometric air-fuel ratio of 16.8), the displacement of the gas composition (the stoichiometric air-fuel ratio) is obtained by calculating the sum of the air-fuel ratio correction values. Deviation) can be obtained.

For example, if the sum of the correction values of the air-fuel ratio is 1.3 times the basic pulse width, since air mass / fuel mass = air-fuel ratio, 16.8 / 1.3 = 12.0. It becomes 9. Therefore, the ratio of 1/1.
It can be easily calculated as a value that is tripled. However, if the increase correction is intentionally performed during warm-up or the like, this correction value should be excluded from the summation calculation in the block 93. The octane number estimating means of block 95 estimates the octane number of the gaseous fuel used (gas filled in the cylinder) from the calculated stoichiometric air-fuel ratio.

As described above, from the relationship between the stoichiometric air-fuel ratio and the octane number, the estimated octane value is calculated as the stoichiometric air-fuel ratio (required air-fuel ratio).
20 (a), and the octane number estimating means in the block 95 estimates the octane number. In the ignition timing correction means of block 96, an ignition timing correction value corresponding to the obtained octane number estimation value is set in advance as shown in FIG. 20 (b), and the ignition timing correction value is searched based on the octane number estimation value.

The ignition timing calculating means in block 97 corrects the basic ignition timing from block 91 with the searched ignition timing correction value to calculate the final ignition timing. Ignition is performed with the corrected final ignition timing.

As described above, by offsetting the ignition timing by the ignition timing correction value obtained from the estimated octane value, the ignition timing can be set in accordance with the gas composition. Note that the octane number estimation here is a simple one and does not accurately analyze the gas composition. Therefore, a strict error occurs between the octane number estimation value and the actual octane number value. Is very small, and this estimated value is used as the octane number estimated value.

Next, FIG. 21 will be described. FIG.
FIG. 4 is an excerpt from blocks 29, 31, 33, 3;
4 to 36, 43, and 44 are the same as those in FIG. 4, and the blocks 90 to 92 and 94 to 97 are also the same as those in FIG. The feature here is block 9
8 is in the method of calculating the sum of the air-fuel ratio correction values corresponding to the fuel pressure adjustment values affected by the gas composition variation in FIG.

The sum of the air-fuel ratio correction values in blocks 29, 31, 33, 34 to 36 is the same as in FIG. 19, but the sum of the air-fuel ratio correction values here is adjusted by the fuel pressure adjustment in block 43. The air-fuel ratio correction value corresponding to the value is further added and the air-fuel ratio correction value is summed (total air-fuel ratio correction value air-fuel ratio correction value)
Is calculated. In the term of the relational expression with the basic fuel pressure, it has been described that the fuel pressure correction coefficient = the basic fuel pressure value / the input fuel pressure value (actual fuel pressure value), but in this case, the actual fuel pressure value is replaced with the fuel pressure adjustment value. , A correction value corresponding to the fuel pressure correction coefficient is calculated. For example, if the basic fuel pressure value is 0.686MP
a, when the fuel pressure adjustment value is 0.784 MPa, the conversion result to the air-fuel ratio correction value is 1 / (0.686 MPa / 0.7
84 MPa) = 1.14.

The fuel pressure adjustment value used in this case is a fuel pressure set value which is a target value of the fuel pressure adjustment, but may be a fuel pressure value actually measured by a sensor. (Block 31
It is indispensable to fix the fuel pressure correction coefficient at. Although the above method enables simple ignition timing correction based on octane value, ignition timing correction for each individual region over the details of the operating state is also essentially required. Therefore, the following method is proposed.
The description will be made with reference to FIGS. In this method, the basic ignition timing is also changed by capturing the octane number estimated value stepwise. In FIG. 22, the octane number estimated value is identified in two regions by a predetermined value.

For example, if the octane number estimated value is equal to or larger than a predetermined value, it is determined that the region is a high octane number region, and a high octane number ignition timing A (ignition timing map A) is selected as the basic ignition timing. If the octane number estimated value is less than the predetermined value, it is determined that the region is in the low octane number range, and the ignition timing B for low octane number (ignition timing map B) is selected as the basic ignition timing. (If it is divided into more subdivided areas, the ignition timing adjustment accuracy is improved, but it is not advisable in terms of processing load and memory capacity in the control unit. In this description, two areas are used.)

Here, the selected ignition timing (ignition timing map) is previously set in the memory as an ignition timing map as shown in FIG. 23, and the engine load and the engine information, which are information from various sensors for detecting the operating state, are stored. The basic ignition timing on the ignition timing map is retrieved from the engine speed. Although each ignition timing map value is a value adapted to each of the predetermined octane numbers, it is a representative point. Therefore, the representative octane number point needs to be corrected by the above-described ignition timing adjustment value. FIG. 24 shows a setting example of the ignition timing adjustment value. Ma and Mb on the octane number estimated value axis are representative octane points of the ignition timing maps A and B.

According to this method, the required value for adjusting the ignition timing based on the octane number can be compatible with the required value for setting the ignition timing for each operation region, so that the ignition timing can be corrected more precisely, and the processing load on the microcomputer can be further improved. Can also be reduced. However, when this ignition timing map switching is performed during engine operation, it is necessary to add a damper process (a process of gradually changing) to the ignition timing switching amount in order to eliminate a step in the ignition timing map switching. .

In the control apparatus for a gaseous fuel engine according to the present embodiment configured as described above, the change in the kind and the property (composition) of the gaseous fuel is determined based on the operating state of the engine. It can be reasonably and reliably detected without using a fuel property detector such as a sensor, so that the air-fuel ratio correction value, fuel pressure, ignition timing, etc. according to the type and property of the fuel currently used can be detected. Can be set appropriately, so that the effects of changes in the type and properties (composition) of the gaseous fuel can be absorbed at an early stage, and the air-fuel ratio, ignition timing, etc. can be maintained in the optimal state, and drivability deteriorates.
It is possible to effectively suppress a reduction in knocking resistance and exhaust performance. In addition, in the embodiment in the text, the gaseous fuel is described as natural gas, but the present invention is applicable even if the gaseous fuel is another gaseous fuel.

[0070]

As will be understood from the above description, according to the control apparatus for a gaseous fuel engine according to the present invention, the change in the kind and the property (composition) of the gaseous fuel is determined based on the operating state of the engine. Therefore, without using a fuel property detector such as an expensive methane concentration sensor, it can be reasonably and reliably detected, whereby the air-fuel ratio correction value according to the type and property of the fuel currently used, the fuel Since the pressure and ignition timing can be set appropriately, the effects of changes in the type and properties (composition) of the gaseous fuel can be absorbed at an early stage, and the air-fuel ratio, ignition timing, etc. can be maintained in optimal conditions, and operability can be maintained. Deterioration, knocking resistance, and reduction in exhaust performance can be effectively suppressed.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram showing an embodiment of a control device for a gaseous fuel engine according to the present invention.

FIG. 2 is a configuration diagram showing a main part of a fuel pressure adjustment system of the gaseous fuel engine of FIG. 1;

FIG. 3 is a control block diagram showing an embodiment of a control device according to the present invention.

FIG. 4 is a control block diagram showing another embodiment of the control device according to the present invention.

FIG. 5 is a flowchart showing control contents of an embodiment of the control device according to the present invention.

FIG. 6 is a flowchart showing control contents of another embodiment of the control device according to the present invention.

FIG. 7 is a diagram showing a relationship between a required air-fuel ratio and combustion stability.

FIG. 8 is a flowchart showing control contents of an embodiment of the control device according to the present invention.

FIG. 9 is a timing chart for explaining one embodiment according to the present invention.

FIG. 10 is a timing chart for explaining another embodiment of the present invention.

FIG. 11 is a diagram showing an example of adjustment value setting according to the embodiment of the present invention.

FIG. 12 is a flowchart showing another control content of the embodiment of the control device according to the present invention.

FIG. 13 is a flowchart showing another control content of the embodiment of the control device according to the present invention.

FIG. 14 is a flowchart showing another control content of another embodiment of the control device according to the present invention.

FIG. 15 is a timing chart for explaining one embodiment of the present invention.

FIG. 16 is a timing chart for explaining one embodiment of the present invention.

FIG. 17 is a flowchart showing another control content of the embodiment of the control device according to the present invention.

FIG. 18 is a flowchart showing another control content of one embodiment of the control device according to the present invention.

FIG. 19 is a control block diagram showing another embodiment of the control device according to the present invention.

FIG. 20 is a diagram showing an example of setting an ignition timing correction value according to the embodiment of the present invention.

FIG. 21 is a control block diagram showing another embodiment of the control device according to the present invention.

FIG. 22 is a view showing a relationship between an octane number and an ignition timing in the control device according to the present invention.

FIG. 23 is a diagram provided for describing an example of setting an ignition timing for each octane number region.

FIG. 24 is a view showing a relationship between an octane number and an ignition timing adjustment value in the control device according to the present invention.

[Explanation of symbols]

 1. Control unit 2. Engine 8. Water temperature sensor 9. Intake air temperature sensor 10. Cylinder 11,12. 12. Fuel cutoff valve 13. Fuel temperature sensor Fuel pressure sensor 16, regulator 18. Injector 20. Exhaust sensor 22. Starter

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Kenji Hashimoto 2477 Takaba, Hitachinaka-shi, Ibaraki F-term in Hitachi Car Engineering Co., Ltd. 3G084 AA05 BA09 BA13 BA14 BA17 CA01 DA27 DA34 DA38 EB06 EB11 EB17 EB22 FA00 FA02 FA07 FA10 FA14 FA16 FA20 FA29 FA33 FA38 3G301 HA22 JA22 JA23 JA31 JB09 KA01 LA00 LB06 MA24 NB02 NC04 ND01 ND07 ND21 ND33 NE01 NE23 PA01Z PA10Z PA11Z PB01Z PB02Z PB08Z PD03A PD03Z PD04A PD04ZPE01Z

Claims (15)

[Claims] In a control apparatus for a gaseous fuel engine configured to supply gaseous fuel via fuel injection means by feedback control, it is determined whether or not a start-up combustion failure state is based on an operation state of the engine. If it is determined that the combustion is in a poor combustion state at the time of starting, it is determined that the type or property of the gaseous fuel has changed, and the air-fuel ratio correction value in the feedback control that has been used so far, the fuel injection A control apparatus for a gaseous fuel engine, wherein at least one of a pressure of a gaseous fuel supplied to the means and an ignition timing is corrected. 2. The control of the gaseous fuel engine according to claim 1, wherein when the state in which the engine speed is equal to or less than a predetermined value at the time of starting continues for a predetermined period, it is determined that the combustion is in a poor starting state. apparatus. 3. The gaseous fuel engine according to claim 1, wherein when the state in which the throttle valve opening is equal to or larger than a predetermined value at the time of starting continues for a predetermined period, it is determined that the combustion is in a poor starting state. Control device. 4. The gaseous fuel engine according to claim 1, wherein, when it is determined that the combustion is in a poor starting state, the air-fuel ratio correction value is corrected to an increased amount. Control device. 5. The gaseous fuel engine according to claim 1, wherein the pressure of the gaseous fuel is corrected to a high pressure side when it is determined that the combustion is in a poor start state. Control device. 6. A correction value of the air-fuel ratio correction value or the correction value of the pressure of the gaseous fuel at the time of the start-time combustion failure determination
The control apparatus for a gaseous fuel engine according to any one of claims 1 to 5, wherein the control unit is set for each operation region determined based on an operation state of the engine. 7. The method according to claim 1, wherein after the start-up combustion failure determination, the fuel injection by the fuel injection means is stopped for a predetermined period, and the air-fuel ratio correction value or the pressure of the gaseous fuel is corrected after the predetermined period. The control device for a gaseous fuel engine according to any one of claims 1 to 6, wherein: 8. After the start-up combustion failure determination, a total air-fuel ratio correction value is calculated, a stoichiometric mixture ratio of gaseous fuel is estimated based on the total air-fuel ratio correction value, and the stoichiometric mixture ratio is estimated based on the estimated stoichiometric mixture ratio. 8. The control device for a gaseous fuel engine according to claim 1, wherein an octane value of the gaseous fuel is estimated, and a preset basic ignition timing is corrected by the octane number estimated value. 9. The gas fuel engine control device according to claim 8, wherein the total air-fuel ratio correction value is a value that takes into account a correction value of the pressure of the gas fuel. 10. The gaseous fuel engine according to claim 8, wherein a plurality of basic ignition timing maps are provided, and the plurality of basic ignition timing maps are switched and used according to the octane number estimated value. Control device. 11. An operating state detecting means for detecting an operating state of the engine represented by an intake air amount, an engine speed, a cooling water temperature, a throttle valve opening, an engine load, etc., and a gaseous fuel upstream of the fuel injection valve. Fuel state detecting means for detecting the temperature and / or pressure of the fuel, exhaust state detecting means for detecting the exhaust state by an oxygen sensor or an air-fuel ratio sensor for detecting the oxygen concentration in the exhaust gas, and gas detection based on the detected operating state. Basic injection pulse calculating means for calculating a basic injection pulse of fuel; air-fuel ratio correction value calculating means for calculating an air-fuel ratio correction value based on the detected operating state and fuel state; and feedback based on the detected exhaust state. A feedback correction value calculating means for calculating a correction value, and a feedback learning value calculated based on the feedback correction value and stored in a memory. Feedback learning value calculating means for storing; an injection pulse calculating means for calculating an injection pulse based on the basic injection pulse, the air-fuel ratio correction value, the feedback correction value, and the feedback learning value; and the calculated injection pulse for the fuel injection valve. Injection drive pulse output means for supplying a fuel injection, a start-up combustion failure state determination means for determining whether or not a start-up combustion failure state based on the operating state, and an air-fuel ratio when the start-up combustion failure state is determined. An air-fuel ratio correction value at start-up combustion failure to set an air-fuel ratio correction value at start-up combustion failure so as to correct the calculation result of the correction value calculation means; Feedback learning value erasing means for erasing the stored feedback learning value; and starting combustion failure after setting the start combustion failure air-fuel ratio correction value. Air-fuel ratio correction value returning means for temporarily shifting the air-fuel ratio correction value to the air-fuel ratio correction value, and feedback learning value calculation re-executing means for permitting the calculation of the feedback learning value after the start-up combustion failure time air-fuel ratio correction value is set. Control device for gaseous fueled engine. 12. An operating state detecting means for detecting an operating state such as an intake air amount, an engine speed, a cooling water temperature, a throttle valve opening, an engine load, etc., and a temperature and / or pressure of the gaseous fuel upstream of the fuel injection valve. State detecting means for detecting the exhaust gas state, exhaust state detecting means for detecting the exhaust state by an oxygen sensor or an air-fuel ratio sensor for detecting the oxygen concentration in the exhaust gas, and a fuel pressure adjustment value for the fuel pressure regulator based on the detected operating state. , A basic injection pulse calculating means for calculating a basic injection pulse of gaseous fuel based on the detected operating state, and calculating an air-fuel ratio correction value based on the detected operating state and fuel state. Air-fuel ratio correction value calculating means, feedback correction value calculating means for calculating a feedback correction value based on the detected exhaust state, Feedback learning value calculating means for calculating a feedback learning value based on the feedback correction value and storing the feedback learning value in a memory; and an injection for calculating an injection pulse based on the basic injection pulse, the air-fuel ratio correction value, the feedback correction value, and the feedback learning value. Pulse calculation means, injection drive pulse output means for supplying the calculated injection pulse to the fuel injection valve, and poor start combustion state determination means for determining whether or not a poor start combustion state based on the operating state; A fuel pressure adjustment value correction unit that corrects the fuel pressure adjustment value set by the fuel pressure setting unit to set the fuel pressure adjustment value at the time of start-up combustion failure if it is determined that the start-up combustion failure state occurs; Feedback learning value erasing means for erasing a feedback learning value stored in the memory when the state is determined to be in a state A fuel pressure adjustment value return means for shifting the fuel pressure adjustment value at the start of combustion failure to the temporary fuel pressure adjustment value after setting the fuel pressure adjustment value at the start of combustion failure, and a feedback learning value after the fuel pressure adjustment value at the start of combustion failure is set. And a feedback learning value calculation re-executing means for permitting the calculation of (i). 13. The start-up combustion failure judgment result is stored in a memory as a start-up combustion failure history, and the start-up combustion failure air-fuel ratio correction value or the start-up combustion failure fuel pressure adjustment value is set in the start-up combustion failure history. 13. The gaseous fuel engine according to claim 11, wherein when the feedback learning value after restarting the feedback learning converges within a predetermined range, the combustion failure history at the time of starting in the memory is deleted. Control device. 14. The feedback learning value includes a first feedback learning value calculated for each operation region determined according to an operation state of the engine, and a second feedback learning value independent of the operation region. After the setting of the air-fuel ratio correction value at the time of poor combustion or the setting of the fuel pressure adjustment value at the time of poor combustion at startup, only the second feedback learning value is allowed to be operated, and after the air-fuel ratio correction value is restored or the fuel pressure adjustment command value is restored, and After the calculation of the second feedback learning value,
14. The control device for a gaseous fuel engine according to claim 11, wherein calculation of the first feedback learning value is permitted instead of the second feedback learning value. 15. When the feedback correction value from the feedback correction value calculating means is out of the predetermined range, the feedback learning value is deleted and feedback learning by the feedback learning value calculating means is started again. 15. The gaseous fuel injection device according to any one of 11 to 14.