Cylinder pressure sensing and fusion method based on physical process of four-stroke diesel engine

Abstract

The invention relates to the technical field of diesel engine combustion control, and discloses a cylinder pressure sensing and fusion method based on the physical process of a four-stroke diesel engine. S2, modeling the volume of the air cylinder; s3, combustion turbulence is corrected; s4, updating gaseous parameters; s5, model cylinder pressure calculation; s6, calculating inertia torque; s7, decoupling the torque signal; s8, multi-cylinder torque approximation is calculated according to the characteristics of the four-stroke process; s9, single-cylinder pressure inversion is carried out; s10, initializing a system noise variance matrix, and determining forgetting factors in different stages; and S11, introducing a forgetting factor, and carrying out dynamic weight fusion. Through the physical process characteristics of the four-stroke diesel engine, a model estimation method and an indirect measurement method are optimized respectively, meanwhile, whether turbulence correction and empirical formula selection under different physical processes are considered or not can be freely selected according to the real-time requirement, and expansibility is high.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of diesel engine combustion control, in particular to a cylinder pressure sensing and fusion method based on the physical process of a four-stroke diesel engine. BACKGROUND

[0002] Cylinder pressure is a core physical quantity representing the engine combustion process, which directly affects the combustion efficiency, heat conversion performance and pollutant generation. High-precision cylinder pressure data is crucial for achieving efficient and clean combustion, reducing fuel consumption and meeting stringent emission regulations.

[0003] Current cylinder pressure acquisition techniques mainly fall into three categories, including direct measurement, indirect measurement and model estimation. Among them, the direct measurement method directly measures the pressure in the combustion chamber through a cylinder pressure sensor, which has the advantages of simple operation, high precision and strong real-time performance. However, due to the complex and harsh physical environment in the cylinder, high requirements are placed on the impact resistance and long-term working stability of the sensor. In this regard, the indirect measurement method indirectly measures the cylinder pressure through easily obtained cylinder-out data, which can greatly save costs, but at the expense of certain sensing accuracy. In contrast, the last method, model estimation, is based on mathematical and physical modeling of the engine, which often has higher accuracy than indirect measurement. However, due to the complex mathematical and physical relationships, pure model estimation often has the disadvantages of high modeling difficulty and poor sensing real-time performance. SUMMARY

[0004] To overcome the shortcomings of the prior art, the present application provides a cylinder pressure sensing and fusion method based on the physical process of a four-stroke diesel engine, which solves the problems of long-term running error accumulation and poor real-time performance caused by static weight and fixed noise covariance in existing cylinder pressure multi-source fusion algorithms.

[0005] To achieve the above purpose, the present application is implemented by the following technical scheme: a cylinder pressure sensing and fusion method based on the physical process of a four-stroke diesel engine, comprising the following steps: S1, data preprocessing of CAN bus collected data; S2, establishing a volume calculation module according to engine structure parameters; S3, correcting the turbulent flow model in the four-stroke process of the diesel engine; S4, updating the cylinder gas mass and gas temperature; S5, substituting the ideal gas state equation to update the model estimated cylinder pressure and obtain the model estimated value; S6, calculating the total reciprocating inertia torque according to the crankshaft connecting rod mechanism; S7, decomposing the total instantaneous torque obtained by CAN bus collection into total reciprocating inertia torque and total gas pressure torque; S8. Based on the characteristics of the four-stroke process, a small constant value is used to replace the torque contribution of other cylinders to calculate the gas pressure torque and the corresponding cylinder's gas pressure torque. S9. Obtain the cylinder pressure based on the single cylinder inversion and obtain the cylinder pressure measurement value; S10. Initialize the system noise variance matrix, observe the noise variance matrix, covariance matrix, and initial values ​​of the system state, and determine the forgetting factor at different stages based on the mean square difference between the optimal estimate and the actual value. S11. In the formula for the variance matrix of prediction error, a forgetting factor is introduced to adjust the model estimate and cylinder pressure measurement value according to the characteristics of the four-stroke process. Furthermore, the whole includes an indirect cylinder pressure measurement method based on the instantaneous crankshaft torque signal; and a Kalman filter algorithm that introduces a forgetting factor to achieve process adaptability, which allows dynamic fusion of model prediction and indirect measurement cylinder pressure data to output the optimal cylinder pressure estimate.

[0006] Preferably, in step S1, the data includes crankshaft angle. Intake pressure Fuel injection quantity and total torque Its preprocessing is based on time windows The moving average filter is calculated using the following formula: ; in, The weights for the Hamming window are N=5. The sampling interval is denoted as .

[0007] Preferably, in step S2, the volume calculation module formula is as follows: ; In the formula, Cylinder diameter, For the itinerary, The compression ratio is... The connecting rod ratio (connecting rod length / crank radius) is 0.25. This refers to the crankshaft rotation angle.

[0008] Preferably, in step S3, the correction mainly involves the combustion process, thermal convection process, and intake / exhaust flow rate correction under turbulent correction, with the turbulent kinetic energy correction coefficient being... The specific formula is as follows: ; ; ; ; In the formula, This is a calibration constant (the turbulence correction for this process can be ignored if it cannot be calibrated). For quality flow, The air intake valve flow area; Because it is a low-calorific-value fuel, This refers to the density of the exhaust gas.

[0009] Preferably, in step S4, the formula for updating the gas mass is as follows: ; In the formula, For the updated gas mass, This represents the change in gas mass. This represents the original mass of the gas inside the cylinder. The formula for updating the gas temperature is as follows: ; ; In the formula, Indicates the temperature of the gas. Indicates the temperature after gas renewal. Indicates the intake air temperature. This represents the volume of the gas at time k. This indicates the volume of the updated gas. This indicates the specific heat ratio of a gas.

[0010] Preferably, in step S5, the ideal gas law formula is: ; In the formula, This indicates the volume of the updated gas. Indicates the temperature after gas renewal. For the updated gas mass, Indicates the pressure of the updated gas. Represents the gas constant; In step S6, the total reciprocating inertial torque is calculated as follows: In the present moment, satisfaction: ; In the formula, Indicates at time The total reciprocating inertial torque, Indicates quality, Indicates acceleration. Represents a certain constant, This represents the angle at time k+1. Indicates radius or distance.

[0011] Preferably, in step S7, the total reciprocating inertial torque and total gas pressure torque ,satisfy: ; In the formula, Indicates total gas pressure torque. This represents the reciprocating inertial torque.

[0012] Preferably, in step S8, when a small constant value is used to replace the torque contribution of other cylinders, the total gas pressure torque... And the gas pressure torque of the nth cylinder Approximately: ; In the formula, For the first Gas pressure torque of each cylinder This represents the total gas pressure torque, from which the gas pressure torque and the corresponding cylinder's gas pressure torque can be calculated.

[0013] Preferably, in step S9, obtaining the cylinder pressure measurement value must satisfy the following: ; In the formula, Where is the piston diameter. For the updated cylinder pressure measurement, For the updated pressure.

[0014] Preferably, in step S11, during the process of adjusting the model estimate and the cylinder pressure measurement, a Kalman filter algorithm is used for filtering, and the specific formula is as follows: ; ; ; ; ; In the formula, The cylinder pressure change rate predicted by the model. For process noise, To reconstruct cylinder pressure for torque, for The measurement transition matrix at time t, For measuring noise.

[0015] This invention provides a cylinder pressure sensing and fusion method based on the physical processes of a four-stroke diesel engine. It has the following beneficial effects: 1. This invention optimizes the model estimation method and the indirect measurement method based on the physical process characteristics of a four-stroke diesel engine. At the same time, it can freely choose whether to consider turbulence correction and empirical formula selection under different physical processes according to real-time requirements, and has strong scalability.

[0016] 2. This invention introduces an adaptive forgetting factor to enable the Kalman filter algorithm to change the weights of the estimated and measured values ​​according to the physical characteristics of the four-stroke process. The forgetting factor can be calibrated in segments by actual measurement fitting, or it can be calibrated in real time by traditional black box models or intelligent algorithms.

[0017] 3. The diesel engine cylinder pressure multi-source fusion estimation method using adaptive Kalman filtering has excellent real-time performance, meets the requirements of engine control cycle, and provides a reliable data foundation for engine optimized control and fault diagnosis. Attached Figure Description

[0018] Figure 1 This is a flowchart of the cylinder pressure multi-source fusion estimation method based on the adaptive Kalman filter algorithm of the present invention; Figure 2 This is a flowchart of the fusion method steps of the present invention; Figure 3 This is a schematic diagram of the crankshaft connecting rod mechanism of the present invention; Figure 4 This is a schematic diagram of the total gas pressure torque and the gas pressure of each cylinder in this invention. Detailed Implementation

[0019] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see the appendix Figure 1 -Appendix Figure 4 This invention provides a cylinder pressure sensing and fusion method based on the physical processes of a four-stroke diesel engine, comprising the following steps: S1. Perform data preprocessing on the CAN bus acquired data.

[0021] Specifically, in this embodiment, step S1 involves data preprocessing to acquire crankshaft rotation angle via the CAN bus. (Unit: degrees), Intake pressure (Unit: kPa), fuel injection quantity (Unit: mg / cycle) and other basic operating parameters and total torque By time window =10ms for moving average filtering.

[0022] ; in, The weights for the Hamming window are N=5 (window length is 11 sampling points). =1ms is the sampling interval.

[0023] S2. Based on the engine structural parameters, establish a volume calculation module.

[0024] Specifically, in this embodiment, step S2 involves establishing a volume calculation module based on the engine structural parameters: ; In the formula, Cylinder diameter, For the itinerary, The compression ratio is... The connecting rod ratio (connecting rod length / crank radius) is 0.25. This refers to the crankshaft rotation angle.

[0025] S3. Correct the turbulence model in the four-stroke process of the diesel engine.

[0026] Specifically, in this embodiment, step S3 is used for turbulence model correction during the four-stroke process of the diesel engine. This mainly involves correction of the combustion process, thermal convection process, and intake / exhaust flow rate under turbulence correction. The turbulence kinetic energy correction coefficient is... : ; ; ; ; In the formula, This is a calibration constant (the turbulence correction for this process can be ignored if it cannot be calibrated). For quality flow, The air intake valve flow area; Because it is a low-calorific-value fuel, This refers to the density of the exhaust gas.

[0027] S4. Update the gas mass and temperature inside the cylinder.

[0028] Specifically, in this embodiment, step S4 mainly addresses the mass of the gas inside the cylinder. renew: ; For the compression and combustion stages, (combustion stage) The values ​​remain unchanged, but changes in gas composition require corrections for R and T); for the intake / exhaust phases, Determined by the isentropic flow equation, these correspond to the supercritical flow state (depending on the ratio of cylinder pressure to intake manifold pressure): ) and subcritical flow state ( ): ; ; In the formula, For the flow area of ​​a single valve, and These represent the intake manifold temperature and pressure, respectively (derived from operating condition data). This is the ratio of the gas's isobaric specific heat capacity to its isochoric specific heat capacity (obtained from a table). This represents the cylinder pressure at time k. For time step, The flow coefficient represents the ratio of the actual to the ideal mass flow rate (obtained from a table). Where is the crankshaft radius; And its cylinder gas temperature The update details are as follows: ; (Method 1) In the formula, Indicates the temperature of the gas. Indicates the temperature after gas renewal. Indicates the intake air temperature. This represents the volume of the gas at time k. This indicates the volume of the updated gas. This indicates the specific heat ratio of a gas.

[0029] (Method 2) It considers heat transfer and friction through the energy equation, and then uses RK4 to solve the differential equation to update T. Heat convection loss under turbulence correction , The convective heat transfer coefficient is obtained through an empirical formula. Sure, This is the equivalent area of ​​the cylinder wall. Taken as the coolant temperature.

[0030] ; in, For combustion to release heat, satisfy ,in Combustion efficiency without turbulence correction. The turbulence correction factor was calibrated experimentally. This refers to the fuel mass in the cylinder.

[0031] (Simplifying assumptions) The differential equations are solved using an improved Runge-Kutta (RK4) solver: ; Equivalent gas constant Based on air and fuel quality, the following content will be dynamically updated: .

[0032] S5. Substitute the ideal gas law into the updated model to obtain the estimated cylinder pressure, and obtain the model estimate value.

[0033] Specifically, in this embodiment, step S5 combines steps S1-S4 by substituting the ideal gas law. The model estimate of cylinder pressure at time k+1 is updated to obtain the result. ; In the formula, This indicates the volume of the updated gas. Indicates the temperature after gas renewal. For the updated gas mass, Indicates the pressure of the updated gas. This represents the gas constant.

[0034] S6. Calculate the total reciprocating inertial torque based on the crankshaft connecting rod mechanism.

[0035] Specifically, in this embodiment, in step S6, we can refer to the crankshaft connecting rod mechanism (see details). Figure 3 ), calculate the total reciprocating inertial torque at time k+1. ,satisfy: ; Among them, the connecting rod crank ratio , For reciprocating inertial mass, This is the piston acceleration.

[0036] S7. Decompose the total instantaneous torque acquired by the CAN bus into the total reciprocating inertial torque and the total gas pressure torque.

[0037] Specifically, in this embodiment, step S7 involves obtaining the total instantaneous torque from the sensor in step S1. Decomposed into total reciprocating inertial torque and total gas pressure torque ,satisfy: ; In the formula, Indicates total gas pressure torque. Indicates reciprocating inertial torque; And it can be written as an explicit expression for cylinder pressure: ; Although the torque represents the combined individual gas pressure of all cylinders, typically only one cylinder's gas pressure torque dominates during a given engine cycle. Other cylinders are either in the exhaust or intake stroke at the same time, producing relatively small torques. Therefore, in calculations, a small constant value can be used to represent the torque contribution of other cylinders. In this case, the gas torque value of the cylinder at top dead center can be used as the gas torque value of the remaining cylinders, thus separating the gas torque contribution of each cylinder.

[0038] S8. Based on the characteristics of the four-stroke process, a small constant value is used to replace the torque contribution of other cylinders to calculate the gas pressure torque and the corresponding cylinder's gas pressure torque.

[0039] Specifically, in this embodiment, step S8 is based on step S5, at this time Based on the characteristics of a four-stroke engine cycle, typically only one cylinder's gas pressure torque dominates during a given engine cycle. Other cylinders are either in the exhaust or intake stroke at the same time, resulting in relatively smaller torque contributions. Therefore, in calculations, a small constant value can be used to replace the torque contribution of other cylinders, such as... Figure 2 As shown. At this time, the total gas pressure torque And the gas pressure torque of the nth cylinder Approximately: ; In the formula, For the first Gas pressure torque of each cylinder This represents the total gas pressure torque, from which the gas pressure torque and the corresponding cylinder's gas pressure torque can be calculated.

[0040] S9. Obtain the cylinder pressure based on the single cylinder inversion and obtain the cylinder pressure measurement value.

[0041] Specifically, in this embodiment, step S9 is based on a single cylinder. The cylinder pressure is obtained by inversion. ,satisfy: ; In the formula, Where is the piston diameter. For the updated cylinder pressure measurement, For the updated pressure.

[0042] S10. Initialize the system noise variance matrix, observe the noise variance matrix, covariance matrix, and initial values ​​of the system state, and determine the forgetting factor at different stages based on the mean square difference between the optimal estimate and the actual value.

[0043] Specifically, in this embodiment, in step S10, when using the adaptive Kalman filter algorithm with forgetting factor for estimation, it is necessary to initialize the system noise variance matrix. Observation noise variance matrix Covariance Matrix and initial values ​​of system state The initial values ​​of the system state. The initial in-cylinder pressure for each operating condition was selected. Subsequently, the forgetting factor for different stages was determined based on the mean square error between the optimal estimate and the actual value. Generally, the model predicts cylinder pressure with high accuracy during the non-combustion phase, but during the combustion phase, due to the extensive use of empirical corrections for turbulence during combustion, it often contains significant errors, resulting in a decrease in accuracy. Conversely, reconstructing cylinder pressure has high accuracy during the combustion phase, but accuracy decreases during the compression and expansion phases due to the significant influence of the other five cylinders. Therefore, a larger forgetting factor is used during the combustion phase. During the non-combustion phase, the reduction The value of is a better choice. The value can be determined by minimizing the mean square error between the optimal estimated value of cylinder pressure after fusion and the actual cylinder pressure, or it can be calibrated by combining other intelligent algorithms or black box models, such as particle swarm optimization.

[0044] S11. A forgetting factor is introduced into the formula for the prediction error variance matrix to adjust the model estimate and cylinder pressure measurement based on the characteristics of the four-stroke process.

[0045] Specifically, in this embodiment, step S11 introduces a forgetting factor into the formula for the prediction error variance matrix. Used to adjust according to the characteristics of the four-stroke process and The weights are then determined, and the Kalman filter algorithm becomes: ; ; ; ; ; In the formula, The cylinder pressure change rate predicted by the model. For process noise, To reconstruct cylinder pressure for torque, for The measurement transition matrix at time t, For measuring noise.

[0046] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for cylinder pressure sensing and fusion based on the physical processes of a four-stroke diesel engine, characterized in that, Includes the following steps: S1. Perform data preprocessing on the CAN bus acquired data; S2. Establish a volume calculation module based on the engine structural parameters; S3. Correct the turbulence model during the four-stroke process of the diesel engine; S4. Update the gas mass and gas temperature in the cylinder; S5. Substitute the ideal gas law into the updated model to obtain the estimated cylinder pressure, and obtain the model estimate value; S6. Calculate the total reciprocating inertial torque based on the crankshaft connecting rod mechanism; S7. Decompose the total instantaneous torque acquired by the CAN bus into the total reciprocating inertial torque and the total gas pressure torque; S8. Based on the characteristics of the four-stroke process, a small constant value is used to replace the torque contribution of other cylinders to calculate the gas pressure torque and the corresponding cylinder's gas pressure torque. S9. Obtain the cylinder pressure based on the single cylinder inversion and obtain the cylinder pressure measurement value; S10. Initialize the system noise variance matrix, observe the noise variance matrix, covariance matrix, and initial values ​​of the system state, and determine the forgetting factor at different stages based on the mean square difference between the optimal estimate and the actual value. S11. A forgetting factor is introduced into the formula for the prediction error variance matrix to adjust the model estimate and cylinder pressure measurement based on the characteristics of the four-stroke process.

2. The cylinder pressure sensing and fusion method based on the physical process of a four-stroke diesel engine according to claim 1, characterized in that, In step S1, the data includes crankshaft angle. Intake pressure Fuel injection quantity and total torque Its preprocessing is based on time windows The moving average filter is calculated using the following formula: ； in, The weights for the Hamming window are N=5. The sampling interval is denoted as .

3. The cylinder pressure sensing and fusion method based on the physical process of a four-stroke diesel engine according to claim 1, characterized in that, In step S2, the volume calculation module formula is as follows: ； In the formula, Cylinder diameter, For the itinerary, The compression ratio is... The connecting rod ratio (connecting rod length / crank radius) is 0.

25. This refers to the crankshaft rotation angle.

4. The cylinder pressure sensing and fusion method based on the physical process of a four-stroke diesel engine according to claim 1, characterized in that, In step S3, the correction mainly involves the combustion process, thermal convection process, and intake / exhaust flow rate correction under turbulence correction, with the turbulence kinetic energy correction coefficient being... The specific formula is as follows: ； ； ； ； In the formula, This is a calibration constant (the turbulence correction for this process can be ignored if it cannot be calibrated). For quality flow, The air intake valve flow area; Because it is a low-calorific-value fuel, This refers to the density of the exhaust gas.

5. The cylinder pressure sensing and fusion method based on the physical process of a four-stroke diesel engine according to claim 1, characterized in that, In step S4, the formula for updating the gas mass is as follows: ； In the formula, For the updated gas mass, This represents the change in gas mass. This represents the original mass of the gas inside the cylinder. The formula for updating the gas temperature is as follows: ； ； In the formula, Indicates the temperature of the gas. Indicates the temperature after gas renewal. Indicates the intake air temperature. This represents the volume of the gas at time k. This indicates the volume of the updated gas. This indicates the specific heat ratio of a gas.

6. The cylinder pressure sensing and fusion method based on the physical processes of a four-stroke diesel engine according to claim 1, characterized in that, In step S5, the ideal gas law is as follows: ； In the formula, This indicates the volume of the updated gas. Indicates the temperature after gas renewal. For the updated gas mass, Indicates the pressure of the updated gas. Represents the gas constant; In step S6, the total reciprocating inertial torque is calculated as follows: In the present moment, satisfaction: ； In the formula, Indicates at time The total reciprocating inertial torque, Indicates quality, Indicates acceleration. Represents a certain constant, This represents the angle at time k+1. Indicates radius or distance.

7. The cylinder pressure sensing and fusion method based on the physical process of a four-stroke diesel engine according to claim 1, characterized in that, In step S7, the total reciprocating inertial torque and total gas pressure torque ,satisfy: ； In the formula, Indicates total gas pressure torque. This represents the reciprocating inertial torque.

8. The cylinder pressure sensing and fusion method based on the physical process of a four-stroke diesel engine according to claim 1, characterized in that, In step S8, when a small constant value is used to replace the torque contribution of other cylinders, the total gas pressure torque... And the gas pressure torque of the nth cylinder Approximately: ； In the formula, For the first Gas pressure torque of each cylinder This represents the total gas pressure torque, from which the gas pressure torque and the corresponding cylinder's gas pressure torque can be calculated.

9. The cylinder pressure sensing and fusion method based on the physical process of a four-stroke diesel engine according to claim 1, characterized in that, In step S9, the cylinder pressure measurement must meet the following requirements: ； In the formula, Where is the piston diameter. For the updated cylinder pressure measurement, For the updated pressure.

10. The cylinder pressure sensing and fusion method based on the physical processes of a four-stroke diesel engine according to claim 1, characterized in that, In step S11, during the adjustment of the model estimate and cylinder pressure measurement, a Kalman filter algorithm is used for filtering. The specific formula is as follows: ； ； ； ； ； In the formula, The cylinder pressure change rate predicted by the model. For process noise, To reconstruct cylinder pressure for torque, for The measurement transition matrix at time t, For measuring noise.

Images