Double-spark-plug engine system and control method thereof

Abstract

The invention discloses a double-spark-plug engine system and a control method thereof, belongs to the technical field of engines, and solves the technical problems of long flame propagation distance and long combustion time of a single-spark-plug ignition mode. The system comprises a machine body and an electronic control unit, a piston is arranged in the machine body, cylinder covers are arranged at the top of the machine body, a combustion chamber is formed between the cylinder covers and the piston, the cylinder covers on the two sides of the combustion chamber are provided with a front spark plug and a rear spark plug respectively, and the cylinder cover on one side of the rear spark plug is provided with an oil sprayer; the cylinder cover on one side of the oil sprayer is provided with an air inlet channel communicated with the vortex premixing chamber. And a temperature sensor and a pressure sensor are respectively arranged on the cylinder cover on two sides of the combustion chamber. The front spark plug and the rear spark plug are matched to work, so that the flame propagation distance is shortened, and the combustion time is shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of engines, and more particularly to a dual spark plug engine system and a control method thereof. Background Art

[0002] At present, the conversion of gasoline engine thermal efficiency has always been a difficult problem. The thermal efficiency of most gasoline engines can only reach about 40% at most. Most of the research on engine performance focuses on unilateral influencing factors such as the position of the inlet / outlet port, the structure of the combustion chamber, the position and ignition time of the spark plug installation, the compression ratio and the fuel performance.

[0003] The traditional single spark plug ignition method has a long flame propagation distance and a long combustion time. In traditional direct injection engines, the mixing time of fuel and air is short, and the mixture is unevenly distributed in the combustion chamber, which is prone to local over-rich or over-lean mixing. Traditional engines, due to the lack of real-time monitoring and feedback of the temperature and pressure in the combustion chamber, cannot accurately control the injection amount and ignition timing of the injector and spark plug according to different working conditions and fuel properties. The traditional use of a single camshaft position sensor signal to determine the ignition timing has problems such as large uncertainty, poor accuracy, and unstable ignition timing. All of these will lead to incomplete and unstable combustion of the mixture in the combustion chamber, thereby generating a large amount of unburned gas and carbon soot, resulting in reduced thermal efficiency and increased emissions. Summary of the invention

[0004] The technical problem to be solved by the present invention is to address the above-mentioned deficiencies in the prior art. One purpose of the present invention is to provide a dual spark plug engine system.

[0005] A second objective of the present invention is to provide a control method for a dual spark plug engine system.

[0006] In order to achieve the above-mentioned purpose one, the present invention provides a dual spark plug engine system, including an engine body and an electronic control unit, wherein a piston is arranged in the engine body, a cylinder head is arranged on the top of the engine body, a combustion chamber is formed between the cylinder head and the piston, the cylinder heads on both sides of the combustion chamber are respectively provided with a front spark plug and a rear spark plug, the cylinder head on the side of the rear spark plug is provided with an injector, the cylinder head at the front end of the injector is provided with a vortex premixing chamber, the vortex premixing chamber is provided with a nozzle, and the cylinder head on the side of the injector is provided with an intake duct connected to the vortex premixing chamber; the cylinder heads on both sides of the combustion chamber are also respectively provided with a temperature sensor and a pressure sensor; the electronic control unit is electrically connected to the front spark plug, the rear spark plug, the injector, the temperature sensor, the pressure sensor and the camshaft position sensor of the engine.

[0007] As a further improvement, the center line of the front spark plug, the center line of the rear spark plug, the center line of the nozzle and the center line of the cylinder bore are located in the same plane.

[0008] Furthermore, the included angle between the center line of the front spark plug and the center line of the cylinder bore is 55-65°.

[0009] Furthermore, the angle between the center line of the post-spark plug and the center line of the cylinder bore is 20-30°.

[0010] Furthermore, the included angle between the center line of the nozzle and the center line of the cylinder bore is 65-70°.

[0011] Furthermore, the included angle between the air intake passage and the fuel injector is 20-30°.

[0012] Furthermore, the vortex premixing chamber is a spherical structure, the top of the piston is provided with a groove structure, the groove structure is high in the middle and low at the periphery, and the cylinder cover directly above the piston is provided with a conical groove structure.

[0013] Furthermore, the temperature sensor is located on one side of the front spark plug, and the pressure sensor is located on one side of the rear spark plug.

[0014] In order to achieve the above-mentioned second objective, the present invention provides a control method for a dual spark plug engine system, comprising the following steps:

[0015] Step 1. Establishing a BP neural network, wherein the input parameters of the BP neural network include temperature, pressure and cam phase, and the output parameters of the BP neural network include spark plug usage, front spark plug ignition timing signal, rear spark plug ignition timing signal and fuel injection amount;

[0016] Step 2: Train the BP neural network according to the historical data of the temperature sensor, the pressure sensor and the camshaft position sensor to obtain a control model, and run the control model on the electronic control unit;

[0017] Step 3. The electronic control unit obtains the current temperature, current pressure and current cam phase information in the combustion chamber through the temperature sensor, pressure sensor and camshaft position sensor respectively, and inputs the information into the control model to obtain the best predicted spark plug usage, predicted front spark plug ignition timing signal, predicted rear spark plug ignition timing signal and predicted fuel injection amount;

[0018] Step 4. The electronic control unit controls the operation of the front spark plug, rear spark plug and injector respectively according to the predicted spark plug usage, the predicted front spark plug ignition timing signal, the predicted rear spark plug ignition timing signal and the predicted fuel injection amount.

[0019] As a further improvement, the BP neural network comprises a three-layer structure: an input layer, a hidden layer and an output layer;

[0020] The input layer mainly receives output signals of the temperature sensor, the pressure sensor and the camshaft position sensor after being processed by the electronic control unit;

[0021] The output layer mainly outputs the spark plug usage, the front spark plug ignition timing signal, the rear spark plug ignition timing signal and the fuel injection amount;

[0022] The hidden layer extracts data features according to nonlinear transformation, and the number of nodes in the hidden layer is calculated according to the number of nodes in the input layer and the output layer using an empirical formula:

[0023]

[0024] Where h represents the number of neurons in the hidden layer, m and n represent the number of nodes in the input layer and output layer respectively, and a is a constant ranging from 1 to 10. According to formula (1), the number of nodes in the hidden layer is determined to be 8;

[0025] Signal forward propagation: Network initialization, the weight from input layer to hidden layer is ω ij , the threshold of the hidden layer node is a j , the weight from the hidden layer to the output layer is ω jk , the output layer node threshold is b k , the learning efficiency is η, respectively for ω ij 、a j ,ω jk 、b k Assign a random value in the range of [-1,1], and select the Tanh function as the hidden layer activation function;

[0026]

[0027] Choose any set of input samples X=(x1,x2…x n ), calculate the input I of the jth node in the hidden layer j With output O j :

[0028]

[0029] Output layer output calculation:

[0030]

[0031] Error back propagation: The error back propagation process improves the accuracy of the network by continuously correcting the weights and thresholds. To ensure that the algorithm does not fall into the local minimum, logistic regression is used as the loss function, and the expected output of the output layer is Y k, the error E of the output layer:

[0032]

[0033] In order to minimize the error E, the gradient descent method is used to correct the weights along the direction of the steepest descent of the sum of squared errors. The weights of the hidden layer and the output layer after the update are The weights of the input layer and the hidden layer are

[0034] Weight correction from hidden layer to output layer:

[0035]

[0036] Weight correction from input layer to hidden layer:

[0037]

[0038] The thresholds of the hidden layer and the output layer are updated as follows:

[0039]

[0040]

[0041] Through multiple iterations, the difference between the actual output and the expected output is continuously reduced. When the difference between the errors after two consecutive iterations is less than the specified value, the training process ends.

[0042] Beneficial Effects

[0043] Compared with the prior art, the present invention has the following advantages:

[0044] 1. The dual spark plug design of the present invention can greatly improve the combustion efficiency of fuel within a reasonable ignition energy range. The two spark plugs ignite the mixed gas step by step to form two flame fronts, accelerate the flame propagation efficiency, shorten the combustion time, reduce the misfire rate and knock tendency, achieve more complete combustion, and reduce emissions and fuel consumption. At the same time, the design reduces the demand for ignition energy, ensures the utilization rate of the spark plug under different working conditions, and prolongs its service life.

[0045] 2. The present invention designs a spherical vortex premixing chamber, and the air inlet is installed at an angle of 20 to 30 degrees with the injector, which ensures that high-concentration fuel will not remain around the air inlet. The air inlet presses the gas into the vortex premixing chamber at high speed to form a high-speed vortex with the fuel injected by the injector, thereby improving the effect of guiding the fuel and air along the spherical wall. The vortex premixing chamber enables the gas and gasoline to achieve the initial full mixing and atomization, reducing the problem of insufficient mixing of the fuel in the combustion chamber.

[0046] 3. In the present invention, the fuel sprayed from the vortex premixing chamber enters the combustion chamber at an angle of 65° to 70° with the center line of the cylinder hole, which is conducive to the preliminary mixed fuel fully entering the combustion chamber, and further secondary mixing and atomization with the air on the ω-shaped piston surface and cylinder wall, and corresponding to the ignition direction of the spark plug, which is more conducive to the formation of disturbances in the entire combustion chamber mixer and full combustion. The combustion efficiency of the fuel is improved, the engine power is increased, and the power, economy and emission of the engine are guaranteed.

[0047] 4. The present invention uses temperature sensors, pressure sensors and camshaft sensors to monitor the temperature, pressure and cam phase information in the combustion chamber in real time, and feeds back to the electronic control unit ECU, which pre-processes this information, including normalization, denoising and feature selection, and then inputs this information into the trained neural network model, so that it can adjust the injection amount of the injector and the ignition timing of the spark plug according to different working conditions and fuel strategies, so that the mixed gas in the combustion chamber is fully burned, reducing fuel waste and pollutant emissions, and improving the safety and reliability of the engine. This multi-sensor data fusion technology has the advantages of good fault tolerance, high precision, fast processing capability and strong complementarity. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a structural schematic diagram of the present invention;

[0049] Figure 2 is a control flow chart of the present invention;

[0050] Figure 3 Schematic diagram of the control model of the present invention.

[0051] Among them: 1-engine body, 2-piston, 3-cylinder head, 4-front spark plug, 5-temperature sensor, 6-rear spark plug, 7-injector, 8-swirl premixing chamber, 9-intake duct, 10-pressure sensor, 11-combustion chamber, 12-nozzle. DETAILED DESCRIPTION

[0052] The present invention will be further described below with reference to the specific embodiments in the accompanying drawings.

[0053] See also Figure 1 to Figure 3A dual spark plug engine system includes a body 1 and an electronic control unit, wherein the electronic control unit is an ECU. A piston 2 is arranged in the body 1, and a cylinder head 3 is arranged on the top of the body 1, and a combustion chamber 11 is formed between the cylinder head 3 and the piston 2. The cylinder heads 3 on both sides of the combustion chamber 11 are respectively provided with a front spark plug 5 and a rear spark plug 6, and the cylinder head 3 on the side of the rear spark plug 6 is provided with an injector 7, and the cylinder head 3 at the front end of the injector 7 is provided with a swirl premixing chamber 8, and the swirl premixing chamber 8 is provided with a nozzle 12, and the cylinder head 3 on the side of the injector 7 is provided with an intake duct 9 connected to the swirl premixing chamber 8. The cylinder heads 3 on both sides of the combustion chamber 11 are also respectively provided with a temperature sensor 4 and a pressure sensor 10. The electronic control unit is electrically connected to the front spark plug 5, the rear spark plug 6, the injector 7, the temperature sensor 4, the pressure sensor 10 and the camshaft position sensor of the engine.

[0054] In this embodiment, the center line of the front spark plug 5, the center line of the rear spark plug 6, the center line of the nozzle 12 and the center line of the cylinder hole are located in the same plane. The angle between the center line of the front spark plug 5 and the center line of the cylinder hole is 55-65°, the angle between the center line of the rear spark plug 6 and the center line of the cylinder hole is 20-30°, the angle between the center line of the nozzle 12 and the center line of the cylinder hole is 65-70°, and the angle between the intake duct 9 and the injector 7 is 20-30°.

[0055] The vortex premixing chamber 8 is a spherical structure, and a groove structure is provided on the top of the piston 2. The groove structure is high in the middle and low at the periphery, that is, the cross-section of the groove structure is ω-shaped. The cylinder head 3 directly above the piston 2 is provided with a conical groove structure, which is conducive to further secondary mixing and atomization of the flow guide and air.

[0056] The temperature sensor 4 is located at one side of the front spark plug 5 , and the pressure sensor 10 is located at one side of the rear spark plug 6 .

[0057] The working process of the system is:

[0058] 1) The fuel injector 7 injects fuel into the vortex premixing chamber 8, and the air intake duct 9 injects gas into the vortex premixing chamber 8, thereby achieving the initial mixing and atomization of gasoline and air in the vortex premixing chamber 8.

[0059] 2) The fuel after the primary mixing and atomization is sprayed into the combustion chamber 11 through the nozzle 12 of the vortex premixing chamber 8, and at the same time, the air introduced through the intake valve is mixed with the fuel after the primary mixing and atomization for secondary mixing.

[0060] 3) The temperature sensor 4, the high pressure sensor 10 and the camshaft position sensor measure the current temperature, pressure and cam phase information in the combustion chamber, and feed the measured signals back to the electronic control unit.

[0061] 4) The electronic control unit processes the collected signals, and determines the spark plug energy required for ignition, the number of spark plugs and the time for the spark plugs to be ignited in steps through the data obtained after signal conversion and fusion. The front spark plug 5 and the rear spark plug 6 are ignited in steps to fully burn the fuel in the combustion chamber, that is, the front spark plug 5 ignites first and the rear spark plug 6 ignites later.

[0062] In order to effectively improve the combustion efficiency of fuel, the present invention also provides a control method for a dual spark plug engine system, comprising the following steps:

[0063] Step 1. Establishing a BP neural network, the input parameters of the BP neural network include temperature, pressure and cam phase, and the output parameters of the BP neural network include spark plug usage, front spark plug ignition timing signal, rear spark plug ignition timing signal and fuel injection amount;

[0064] Step 2: Train the BP neural network according to the historical data of the temperature sensor 4, the pressure sensor 10 and the camshaft position sensor to obtain a control model, and run the control model on the electronic control unit;

[0065] Step 3. The electronic control unit obtains the current temperature, current pressure and current cam phase information in the combustion chamber 11 through the temperature sensor 4, the pressure sensor 10 and the camshaft position sensor respectively, and inputs the information into the control model to obtain the best predicted spark plug usage, the predicted front spark plug ignition timing signal, the predicted rear spark plug ignition timing signal and the predicted fuel injection amount;

[0066] Step 4. The electronic control unit controls the operation of the front spark plug 5, the rear spark plug 6 and the injector 7 respectively according to the predicted spark plug usage, the predicted front spark plug ignition timing signal, the predicted rear spark plug ignition timing signal and the predicted fuel injection amount.

[0067] The BP neural network consists of three layers: input layer, hidden layer and output layer.

[0068] The input layer mainly receives the output signals of the temperature sensor 4, the pressure sensor 10 and the camshaft position sensor after being processed by the electronic control unit.

[0069] The output layer mainly outputs the spark plug usage, front spark plug ignition timing signal, rear spark plug ignition timing signal and fuel injection amount.

[0070] The hidden layer extracts data features based on nonlinear transformation. The number of nodes in the hidden layer is calculated based on the number of nodes in the input layer and the output layer using an empirical formula:

[0071]

[0072] Where h represents the number of neurons in the hidden layer, m and n represent the number of nodes in the input layer and output layer respectively, and a is a constant ranging from 1 to 10. According to formula (1), the number of nodes in the hidden layer is determined to be 8. The network topology is shown in Figure 3 shown.

[0073] 1) Signal forward propagation:

[0074] Network initialization, the weight from input layer to hidden layer is ω ij , the threshold of the hidden layer node is a j , the weight from the hidden layer to the output layer is ω jk , the output layer node threshold is b k , the learning efficiency is η, respectively for ω ij 、a j ,ω jk 、b k Assign a random value in the range of [-1,1]. The hidden layer activation function selects the Tanh function;

[0075]

[0076] Choose any set of input samples X=(x1,x2…x n ), calculate the input I of the jth node in the hidden layer j With output O j .

[0077]

[0078] Output layer output calculation:

[0079]

[0080] 2) Error back propagation:

[0081] The error back propagation process improves the accuracy of the network by continuously correcting the weights and thresholds. To ensure that the algorithm does not fall into the local minimum, logistic regression is used as the loss function, and the expected output of the output layer is Y k , the error e of the output layer:

[0082]

[0083] In order to minimize the error e, the gradient descent method is used to correct the weights along the direction of the steepest descent of the sum of squared errors. The weights of the hidden layer and the output layer after the update are The weights of the input layer and the hidden layer are

[0084] Weight correction from hidden layer to output layer:

[0085]

[0086] Weight correction from input layer to hidden layer:

[0087]

[0088] The thresholds of the hidden layer and the output layer are updated as follows:

[0089]

[0090] Through multiple iterations, the difference between the actual output and the expected output is continuously reduced. When the difference between the errors after two consecutive iterations is less than the specified value, the training process ends.

[0091] The front spark plug 5, the rear spark plug 6, and the nozzle 12 of the present invention point to the center line of the cylinder respectively. According to the judgment of the electronic control unit, the ignition time of the front spark plug 5 and the rear spark plug 6 is determined, which is conducive to the ignition of the fuel in the center of the combustion chamber and reduces the flame propagation distance as much as possible. The principle of this design is to optimize the mixed gas distribution and flame propagation path in the combustion chamber by using the installation position and angle of the spark plug and the injector, so that the ignition direction of the spark plug corresponds to the injection direction of the injector, thereby forming a central ignition area, and at the same time, the flame front is diffused from the two spark plugs to the surroundings at the same time, shortening the flame propagation distance, reducing the combustion time, so that a larger compression ratio can be used, improving combustion efficiency and engine performance, and reducing emissions and fuel consumption.

[0092] The above are only preferred embodiments of the present invention. It should be pointed out that for those skilled in the art, several modifications and improvements can be made without departing from the structure of the present invention, which will not affect the effect of the implementation of the present invention and the practicality of the patent.

Claims

1. A dual spark plug engine system, comprising an engine body (1) and an electronic control unit, wherein a piston (2) is arranged in the engine body (1), a cylinder head (3) is arranged on the top of the engine body (1), and a combustion chamber (11) is formed between the cylinder head (3) and the piston (2), wherein: The cylinder heads (3) on both sides of the combustion chamber (11) are respectively provided with a front spark plug (5) and a rear spark plug (6); the cylinder head (3) on one side of the rear spark plug (6) is provided with a fuel injector (7); the cylinder head (3) at the front end of the fuel injector (7) is provided with a vortex premixing chamber (8); the vortex premixing chamber (8) is provided with a nozzle (12); the cylinder head (3) on one side of the fuel injector (7) is provided with an intake passage (9) connected to the vortex premixing chamber (8); the cylinder heads (3) on both sides of the combustion chamber (11) are also respectively provided with a temperature sensor (4) and a pressure sensor (10); the electronic control unit is electrically connected to the front spark plug (5), the rear spark plug (6), the fuel injector (7), the temperature sensor (4), the pressure sensor (10) and the camshaft position sensor of the engine.

2. A dual spark plug engine system according to claim 1, characterized in that: The center line of the front spark plug (5), the center line of the rear spark plug (6), the center line of the nozzle (12) and the center line of the cylinder hole are located in the same plane.

3. A dual spark plug engine system according to claim 1, characterized in that: The included angle between the center line of the front spark plug (5) and the center line of the cylinder hole is 55-65 degrees.

4. A dual spark plug engine system according to claim 1, characterized in that: The included angle between the center line of the post-spark plug (6) and the center line of the cylinder hole is 20-30 degrees.

5. A dual spark plug engine system according to claim 1, characterized in that: The included angle between the center line of the nozzle (12) and the center line of the cylinder hole is 65-70 degrees.

6. A dual spark plug engine system according to claim 1, characterized in that: The included angle between the air intake passage (9) and the fuel injector (7) is 20 to 30 degrees.

7. A dual spark plug engine system according to claim 1, characterized in that: The vortex premixing chamber (8) is a spherical structure, the top of the piston (2) is provided with a groove structure, the groove structure is high in the middle and low at the periphery, and the cylinder cover (3) directly above the piston (2) is provided with a conical groove structure.

8. A dual spark plug engine system according to claim 1, characterized in that: The temperature sensor (4) is located on one side of the front spark plug (5), and the pressure sensor (10) is located on one side of the rear spark plug (6).

9. A control method for a dual spark plug engine system, characterized in that: The following steps are involved: Step 1. Establishing a BP neural network, wherein the input parameters of the BP neural network include temperature, pressure and cam phase, and the output parameters of the BP neural network include spark plug usage, front spark plug ignition timing signal, rear spark plug ignition timing signal and fuel injection amount; Step 2. training the BP neural network according to the historical data of the temperature sensor (4), the pressure sensor (10) and the camshaft position sensor to obtain a control model, and running the control model on the electronic control unit; Step 3. The electronic control unit obtains the current temperature, current pressure and current cam phase information in the combustion chamber (11) through the temperature sensor (4), the pressure sensor (10) and the camshaft position sensor respectively, and inputs the information into the control model to obtain the best predicted spark plug usage, the predicted front spark plug ignition timing signal, the predicted rear spark plug ignition timing signal and the predicted fuel injection amount; Step 4. The electronic control unit controls the operation of the front spark plug (5), the rear spark plug (6) and the injector (7) respectively according to the predicted spark plug usage, the predicted front spark plug ignition timing signal, the predicted rear spark plug ignition timing signal and the predicted fuel injection amount.

10. The control method of a dual spark plug engine system according to claim 9, characterized in that: The BP neural network comprises a three-layer structure: an input layer, a hidden layer and an output layer; The input layer mainly receives output signals of the temperature sensor (4), the pressure sensor (10) and the camshaft position sensor after being processed by the electronic control unit; The output layer mainly outputs the spark plug usage, the front spark plug ignition timing signal, the rear spark plug ignition timing signal and the fuel injection amount; The hidden layer extracts data features according to nonlinear transformation, and the number of nodes in the hidden layer is calculated according to the number of nodes in the input layer and the output layer using an empirical formula: Where h represents the number of neurons in the hidden layer, m and n represent the number of nodes in the input layer and output layer respectively, and a is a constant ranging from 1 to 10. According to formula (1), the number of nodes in the hidden layer is determined to be 8; Signal forward propagation: Network initialization, the weight from input layer to hidden layer is ω ij , the threshold of the hidden layer node is a j , the weight from the hidden layer to the output layer is ω jk , the output layer node threshold is b k , the learning efficiency is η, respectively for ω ij 、a j ,ω jk 、b k Assign a random value in the range of [-1, 1], and select the Tanh function as the hidden layer activation function; Choose any set of input samples X = (x1, x2...x n ), calculate the input I of the jth node in the hidden layer j With output O j : Output layer output calculation: Error back propagation: The error back propagation process improves the accuracy of the network by continuously correcting the weights and thresholds. To ensure that the algorithm does not fall into the local minimum, logistic regression is used as the loss function, and the expected output of the output layer is Y k , the error E of the output layer: In order to minimize the error E, the gradient descent method is used to correct the weights along the direction of the steepest descent of the sum of squared errors. The weights of the hidden layer and the output layer after the update are The weights of the input layer and the hidden layer are Weight correction from hidden layer to output layer: Weight correction from input layer to hidden layer: The thresholds of the hidden layer and the output layer are updated as follows: Through multiple iterations, the difference between the actual output and the expected output is continuously reduced. When the difference between the errors after two consecutive iterations is less than the specified value, the training process ends.

Images