Search

CN-121993309-A - Smith-augmentation and reduction ESO common rail pressure control method based on online parameter identification

CN121993309ACN 121993309 ACN121993309 ACN 121993309ACN-121993309-A

Abstract

The invention belongs to the technical field of electronic control of internal combustion engines, and discloses a Smith-augmentation and reduction ESO common rail pressure control method based on online parameter identification, which comprises the following steps of step 1, data acquisition and pretreatment, step 2, modeling and discretizing a high-pressure common rail system to obtain a discretization model of the high-pressure common rail system, and constructing an online parameter identification module solution system pole parameter based on a recursive least square method with forgetting factors Controlling gain parameters Constant bias parameter Step 3, the Smith predictor calculates a time-lag-free virtual predicted rail pressure Step 4, the total disturbance estimated value of the extended state observer observation system with enhanced and reduced order Rate of change estimate for a vehicle Step 5, composite disturbance rejection control law combination 、 、 And Controlling the output control amount Acting on the high-pressure oil pump actuator.

Inventors

  • SONG KANG
  • WANG WEI
  • XIAO WEI
  • LI QIAN
  • WANG XIA
  • GUAN ZHUOWEI
  • SUN YUBO
  • XIE HUI

Assignees

  • 天津大学

Dates

Publication Date
20260508
Application Date
20260123

Claims (10)

  1. 1. The Smith-augmentation and reduction ESO common rail pressure control method based on online parameter identification is characterized by comprising the following steps of: Step1, at each control cycle Collecting rail pressure measured value of high-pressure common rail system Current target rail pressure Control amount of time-lapse alignment , The sampling step number corresponding to the physical time lag of the system; Step 2, modeling and discretizing a high-pressure common rail system to obtain a discretization model of the high-pressure common rail system: , wherein, In order to reflect the damping coefficient of the leakage characteristics of the system, In order to sample the period of time, Is that The rail pressure measurement at the moment in time, In order to control the gain of the gain control, Is that The control amount of the time of day, The constant bias of the system is characterized, Is a discretized time-varying dynamic disturbance term; On-line parameter identification module based on recursive least square method with forgetting factors is constructed and utilized And Resolving system pole parameters Controlling gain parameters Constant bias parameter ; Step 3, the Smith predictor is based on the step 2 、 And Calculating ideal model output without time lag Recombination is carried out Calculating virtual prediction rail pressure without time lag ; Step 4, the augmented reduced-order extended state observer is obtained in the step 3 And the current control amount For input, a reduced order structure is adopted to avoid redundant observation of the system state, and an augmented structure is adopted to expand a disturbance model into a second order so as to directly observe the total disturbance estimated value of the system Rate of change estimate for a vehicle ; Step 5, compounding disturbance rejection control law with differential signal of target rail pressure Differential feed forward to And Constructing model inverse decoupling terms Real-time feedforward compensation and utilization And (3) with Proportional feedback control of tracking error of (c) to output control quantity Acting on the high-pressure oil pump actuator.
  2. 2. The Smith-augmented reduced-order ESO common rail pressure control method based on online parameter identification according to claim 1, wherein in the step 2, the resolving process of the online parameter identification module is as follows: constructing regression vectors containing constant bias terms : Is that Rail pressure at moment; Defining parameter vectors to be identified , Parameters to be identified are updated through the following recurrence formula: Calculating gain matrix : Is that The covariance matrix of the time of day, Is a forgetting factor; Calculating prediction error : Is that A parameter estimation value of the moment; Updating parameter estimates : Updating covariance matrix : Wherein the method comprises the steps of Is that The covariance matrix of the time of day, Is that A gain matrix of the time of day, Is that The prediction error of the moment in time, Is a unitary matrix, and uses updated And (3) calculating physical parameters:
  3. 3. the method for controlling the pressure of the Smith-enhanced reduced-order ESO common rail based on the online parameter identification as claimed in claim 2, wherein in the step 2, when When the system is not excited enough, the parameter update is stopped and the system is kept Wherein Is a preset dead zone threshold.
  4. 4. The method for controlling the pressure of the Smith-enhanced reduced-order ESO common rail based on the on-line parameter identification as claimed in claim 1, wherein in said step 3, the rail pressure is virtually predicted The calculation formula of (2) is as follows: Wherein, the For the actual rail pressure measurement that is acquired, Is: Wherein, the For internal model output with time lag, the length is read by Is obtained from the history model output in the ring buffer of (c), Control amount of time.
  5. 5. The method for controlling the common rail pressure of Smith-augmented reduced-order ESO based on online parameter identification according to claim 1, wherein in the step 4, the augmented reduced-order extended state observer is directed to a time-lag-free first-order object model after Smith pre-estimation compensation: Intermediate variable Intermediate variable ; 、 In order for the gain of the observer to be achieved, As an estimate of the dynamic total disturbance, An estimated value of a first derivative of the dynamic total disturbance; As an intermediate variable Is used as a first derivative of (a), As an intermediate variable Is a first derivative of (a).
  6. 6. The method for controlling the pressure of the Smith-enhanced reduced-order ESO common rail based on the on-line parameter identification as claimed in claim 5, wherein, , Is the observer bandwidth.
  7. 7. The method for controlling the pressure of the Smith-enhanced reduced-order ESO common rail based on the on-line parameter identification as claimed in claim 1, wherein in the step 5, The calculation formula of (2) is as follows: is the controller bandwidth.
  8. 8. The high-voltage common rail controller is characterized by comprising a processor and a memory, wherein the memory can be used for executing a computer program on the processor, and the processor can realize the Smith-augmented and reduced-order ESO common rail pressure control method based on online parameter identification according to any one of claims 1-7 when executing the program.
  9. 9. A computer readable storage medium having stored thereon computer executable instructions which, when executed, are adapted to implement the Smith-augmented reduced order ESO common rail pressure control method based on online parameter identification according to any one of claims 1 to 7.
  10. 10. A diesel engine comprising the high pressure common rail controller of claim 8.

Description

Smith-augmentation and reduction ESO common rail pressure control method based on online parameter identification Technical Field The invention relates to the technical field of electronic control of internal combustion engines, in particular to a Smith-enhanced and reduced ESO common rail pressure control method based on online parameter identification, which aims at the pressure control method of a high-pressure common rail diesel engine fuel injection system with large time lag, nonlinearity and parameter time-varying characteristics. Background The high pressure common rail fuel injection system (HPCRS) acts as the "heart" of modern advanced diesel engines, and its rail pressure control accuracy directly determines the combustion efficiency, fuel economy and emissions compliance of the engine. With the evolution of emission regulations to more stringent standards, the injection pressure of a common rail system breaks through 250MPa, and the trend of multi-point injection and high-frequency response is presented, which puts extremely high demands on the robustness and rapidity of a control system. However, in practical engineering application, rail pressure control always faces three major core challenges of 'large time lag, parameter time varying and strong nonlinear disturbance', and the prior art is not yet solved perfectly. First, the contradiction between physical time lags and system stability is increasingly prominent. The fuel is pumped from the high-pressure oil pump to the common rail via the pipeline to build up pressure, with a considerable delay in delivery. Recent researches show that rail pressure fluctuation of a high-pressure common rail system is not only influenced by oil injection quantity, but also is highly coupled with hardware parameters such as plunger motion, valve body spring stiffness and the like, and the complex fluid dynamics characteristics enable a traditional fixed gain control strategy to be difficult to keep stable （Jiang H, Li Z, Jiang F, et al. Analysis of Rail Pressure Stability in an Electronically Controlled High-Pressure Common Rail Fuel Injection System via GT-Suite Simulation[J]. Energies (19961073), 2025, 18(3)）. under all working conditions, and the existence of time lag can lead to phase lag of feedback control, so that system oscillation is extremely easy to induce, and the improvement of closed loop bandwidth is limited. Second, the time-varying nature of the model parameters exacerbates the difficulty of controller tuning. Because the bulk modulus of fuel varies strongly with pressure and temperature, and the volumetric efficiency of an actuator (e.g., a high pressure oil pump) decreases with operational wear, a controller based on a fixed model is difficult to accommodate for full life cycle variations. In response to this problem, a variety of adaptive control strategies have recently been proposed in the academia. For example, zhang et al propose a self-adaptive nonsingular terminal sliding mode control method for a high-voltage common rail system, and although the problem of uncertainty of a part of parameters is solved through a self-adaptive law, the method has large calculation load, is sensitive to high-frequency measurement noise, and still has challenges in engineering landing （Zhang J, Yu Y, Wu S, et al. Adaptive Continuous Non-Singular Terminal Sliding Mode Control for High-Pressure Common Rail Systems: Design and Experimental Validation[J]. Processes, 2025, 13(8): 2410）. Again, active Disturbance Rejection Control (ADRC) has limitations in handling complex dynamic disturbances. While ADRC performs well in terms of immunity using Extended State Observers (ESO), in large time-lapse systems, standard linear ESO can result in distortion of disturbance estimates due to timing mismatch of observed inputs and actual outputs. In addition, conventional reduced-order ESO typically assumes that the disturbance is constant or slowly varying signal #) And the fuel injection disturbance of the common rail system under the rapid acceleration working condition shows a severe slope-like change. Recent studies by Cui et al (2025) indicate that the steady state error is difficult to completely eliminate in the face of fast varying disturbances of the common rail system, often requiring the introduction of complex algorithms such as neural networks (RBFs) to compensate, which further increases the computational burden of the ECU （Cui Y, Li W, Huo T, et al. Adaptive Control of Rail Pressure in High-Pressure Common Rail System Based on RBF-LADRC[C]//2025 4th International Conference on Energy and Electrical Power Systems (ICEEPS). IEEE, 2025: 806-813）. In summary, the existing advanced control strategy is still insufficient in coping with the problem of large time lag and slope disturbance coupling of the high-voltage common rail system. Therefore, there is an urgent need to develop a novel control method that integrates on-line parameter learning (solving model m