Intelligent optimization method and system for adjustable egr valve exhaust emissions

Abstract

The present application relates to the technical field of engine exhaust emission control, in particular to an intelligent optimization method and system for adjustable EGR valve exhaust emission. The method comprises: obtaining the crank angle and exhaust pressure to identify the pulse peak value and time of each cylinder, correcting the EGR valve response lag based on the integral quantity of the exhaust pressure exceeding the basic pressure and the relative amplitude of the pulse peak value; identifying the NOx response lag and rate according to the EGR rate rising and falling direction, predicting the NOx transient and steady state values corresponding to each candidate opening according to the change direction and amplitude scaling parameters; calculating the memory factor according to the historical opening and dynamically adjusting the three target weights to correct the weighted steady state value and solve the optimal opening by taking the transient peak value as the constraint; taking the predicted arrival time minus the corrected lag time as the advance to issue the control command. The present application can improve the control accuracy and adaptability of the EGR system to nitrogen oxide emission under transient operating conditions.

Description

Technical Field

[0001] This invention relates to the field of engine exhaust emission control technology, and more specifically to an intelligent optimization method and system for exhaust emissions using an adjustable EGR valve. Background Technology

[0002] Adjustable Exhaust Gas Recirculation (EGR) technology is one of the core methods for reducing nitrogen oxide (NOx) emissions from diesel and gasoline engines. Its basic principle is to introduce a portion of the exhaust gas from the engine, after cooling, into the intake manifold, where it mixes with fresh air and enters the cylinder for combustion. The introduced exhaust gas contains a large amount of inert gases (such as water vapor, nitrogen, and carbon dioxide), which have high specific heat capacities and do not participate in combustion and release heat. This reduces the oxygen concentration in the mixture, thereby lowering the peak combustion temperature in the cylinder and effectively suppressing the formation of NOx. The adjustable EGR valve, as a key actuator for precise exhaust gas flow control, directly determines the exhaust gas recirculation rate, thus affecting multiple performance indicators such as NOx emissions, particulate matter emissions, fuel economy, and power output. Therefore, precise control of the exhaust gas recirculation valve is crucial.

[0003] Common exhaust gas recirculation control strategies mainly include open-loop control based on engine steady-state operating condition calibration, closed-loop control based on intake air flow or oxygen concentration, and operating condition-based mapping chart lookup control.

[0004] However, the aforementioned conventional exhaust gas recirculation control strategy faces the following technical problems under transient operating conditions:

[0005] First, the dynamic mismatch arises from neglecting the pulse characteristics of exhaust gas. Under transient conditions, exhaust gas exhibits a significant pulse pattern, with varying pulse frequency and amplitude at different speeds and loads. The exhaust gas recirculation valve experiences pulsating pressure rather than constant pressure. Existing response hysteresis models are mostly based on average exhaust gas flow rates, neglecting the impact of pulses on the valve core's dynamic response, resulting in a significant increase in transient prediction errors.

[0006] Second, the modeling of the asymmetric dynamic characteristics between changes in exhaust gas recirculation rate and nitrogen oxide response is lacking. When the exhaust gas recirculation rate increases, the lag in exhaust gas transport and mixing leads to a significant delay in the decrease of nitrogen oxides; while when the exhaust gas recirculation rate decreases, the memory effect of residual exhaust gas causes the recovery of nitrogen oxides to lag for multiple cycles. Existing emission prediction models fail to characterize this direction-dependent asymmetric response, resulting in insufficient accuracy in transient nitrogen oxide prediction.

[0007] Third, the lack of operating condition memory effect in multi-objective optimization. When switching operating conditions, the historical opening degree of the exhaust gas recirculation valve determines the amount of residual exhaust gas in the intake pipe, which in turn affects the current charging efficiency and exhaust gas recirculation response. Existing methods only determine the optimization weights based on the current operating condition, without considering the impact of previous operating conditions on the current feasible region, resulting in a misalignment between weight allocation and actual emission control requirements. Summary of the Invention

[0008] This invention provides an intelligent optimization method and system for exhaust gas emissions from adjustable EGR valves to solve existing problems.

[0009] The intelligent optimization method for exhaust gas emissions from adjustable EGR valves of the present invention adopts the following technical solution:

[0010] One embodiment of the present invention provides a smart optimization method for exhaust gas emissions from an adjustable EGR valve, the method comprising the following steps:

[0011] Obtain crankshaft angle and exhaust pressure, determine the exhaust phase of each cylinder based on crankshaft angle and cylinder sequence, and extract the pulse peak value and pulse time of exhaust pressure within the exhaust phase of each cylinder;

[0012] Based on the integral of the exhaust pressure exceeding the base pressure within the exhaust phase of each cylinder and the relative amplitude of the pulse peak relative to the base pressure, the effective intensity of the exhaust pulse of each cylinder is determined, and the response hysteresis time of the EGR valve is corrected according to the effective intensity.

[0013] Acquire historical EGR rate change data, and identify the lag time and rate of change of NOx emission response in both directions of EGR rate increase and decrease;

[0014] Calculate the current EGR rate change rate, select the corresponding lag time and change rate according to its direction, scale the selected lag time and change rate according to the current EGR rate change amplitude, and use the scaling result to predict the NOx transient peak value and steady-state value corresponding to the opening degree of each candidate EGR valve.

[0015] The memory factor is calculated based on the historical EGR valve opening, and the weights of the three targets of NOx emissions, fuel economy and combustion stability are adjusted based on the deviation between the memory factor and the target EGR rate.

[0016] Using the opening degree of each candidate EGR valve as the optimization variable, the NOx steady-state value, fuel consumption rate and combustion fluctuation coefficient corresponding to each candidate EGR valve opening degree are weighted by the three corrected weights, and the NOx transient peak value corresponding to each candidate EGR valve opening degree is constrained. The candidate EGR valve opening degree with the smallest weighted sum is then solved as the optimal EGR valve opening degree.

[0017] The predicted arrival time of the next exhaust pulse is obtained, and the advance amount is obtained by subtracting the corrected response lag time of the EGR valve from the predicted arrival time. The control command corresponding to the optimal EGR valve opening is issued in advance according to the advance amount.

[0018] Furthermore, the crankshaft angle and exhaust pressure are obtained, and the exhaust phase of each cylinder is determined based on the crankshaft angle and cylinder sequence. Within each cylinder's exhaust phase, the pulse peak value and pulse moment of the exhaust pressure are extracted, specifically including:

[0019] Obtain the engine's crankshaft angle and exhaust pressure;

[0020] Based on the engine cylinder working sequence, determine the crankshaft angle range corresponding to the exhaust stroke of each cylinder. Within the crankshaft angle range corresponding to the exhaust stroke of each cylinder, extract the local maximum pressure value from the exhaust pressure as the pulse peak value of that cylinder, and record the crankshaft angle time corresponding to the pulse peak value.

[0021] Furthermore, based on the integral of the exhaust pressure exceeding the base pressure within each cylinder's exhaust phase and the relative amplitude of the pulse peak relative to the base pressure, the effective intensity of the exhaust pulse for each cylinder is determined, and the response hysteresis time of the EGR valve is corrected according to the effective intensity, specifically including:

[0022] Calculate the integral of the exhaust pressure exceeding the base pressure within each cylinder's exhaust phase, and calculate the relative amplitude of the pulse peak relative to the base pressure.

[0023] The product of the integral and the relative amplitude is taken as the effective intensity of the exhaust pulse of that cylinder.

[0024] Calculate the deviation rate between the effective action intensity and the preset average intensity. Take the negative value of the product of the correction intensity coefficient and the deviation rate as the exponent of the natural constant to calculate the correction coefficient. Multiply the response hysteresis time of the EGR valve by the correction coefficient to obtain the corrected response hysteresis time of the EGR valve.

[0025] Furthermore, historical EGR rate change data are obtained, and the lag time and rate of change of NOx emission response are identified separately in the two directions of EGR rate increase and decrease, specifically including:

[0026] Acquire historical EGR rate change data, which is a continuous time series of EGR rate changes over time during historical operation.

[0027] From historical EGR rate change data, events in which the EGR rate undergoes a step change are selected as EGR rate change events, and then divided into rising event group and falling event group according to the two directions of EGR rate increase and decrease.

[0028] For each event in the rising event group, the time difference between the start of the change in EGR rate and the start of the change in NOx emission concentration is extracted as the rising lag time, and the average rate of change required for the NOx emission concentration to change from 10% to 90% of the steady-state value before the change is extracted as the rising rate of change.

[0029] For each event in the declining event group, the time difference between the start of the change in EGR rate and the start of the change in NOx emission concentration is extracted as the lag time in the declining direction, and the average rate of change required for the NOx emission concentration to change from 10% to 90% of the steady-state value before the change is extracted as the rate of change in the declining direction.

[0030] Furthermore, the current EGR rate change rate is calculated, and the lag time and rate of change in the corresponding direction are selected based on its direction. The selected lag time and rate of change are scaled according to the current EGR rate change amplitude. The scaling result is used to predict the transient peak and steady-state NOx values ​​corresponding to the opening degree of each candidate EGR valve. Specifically, this includes:

[0031] Calculate the current rate of change of EGR rate, and determine the magnitude of the current EGR rate change based on the difference between the current EGR rate value and the EGR rate value at the moment when the EGR rate began to change;

[0032] The current EGR rate of change is compared with preset positive and negative thresholds. If the current EGR rate of change is greater than the positive threshold, the direction of the EGR rate of change is determined to be upward, and the lag time and rate of change corresponding to the upward direction are selected. If the current EGR rate of change is less than the negative threshold, the direction of the EGR rate of change is determined to be downward, and the lag time and rate of change corresponding to the downward direction are selected. If the current EGR rate of change is between the negative and positive thresholds, the EGR rate of change is determined to be in a quasi-steady state.

[0033] When it is determined that the direction of the EGR rate change is upward or downward, the selected lag time and rate of change are scaled according to the current EGR rate change magnitude.

[0034] Obtain the opening degree of the candidate EGR valve;

[0035] Using each candidate EGR valve opening as the target opening, and utilizing the scaled lag time and rate of change, the dynamic trajectory of NOx emission concentration during the transition from the current EGR valve opening to each candidate EGR valve opening is simulated. The maximum value is extracted from the dynamic trajectory corresponding to each candidate EGR valve opening as the NOx transient peak value corresponding to that candidate EGR valve opening, and the endpoint value is extracted as the NOx steady-state value corresponding to that candidate EGR valve opening.

[0036] Furthermore, the memory factor is calculated based on the historical EGR valve opening, specifically including:

[0037] Obtain the historical EGR valve opening at each moment within the preceding time window;

[0038] An exponentially weighted average of the historical EGR valve openings at each time point within the preceding time window is used as the memory factor. The weight of each time point is negatively correlated with the time elapsed between that time point and the current time point.

[0039] Furthermore, the weights of the three targets—NOx emissions, fuel economy, and combustion stability—are adjusted based on the deviation between the memory factor and the target EGR rate. Specifically, this includes:

[0040] The difference between the memory factor and the target EGR rate is calculated as the residual exhaust gas deviation.

[0041] The weight correction coefficients for NOx emission targets, fuel economy targets, and combustion stability targets are determined based on the ratio of residual exhaust gas deviation to target EGR rate. The weight correction coefficient for each target is an exponential function of the natural constant, and the independent variable of the exponential function is the product of the memory correction sensitivity coefficient of the target and the ratio.

[0042] The base weights of each objective are multiplied by the corresponding weight correction coefficients to obtain the corrected weights of each objective. The three corrected weights are then normalized so that the sum of the three is 1.

[0043] Furthermore, using the opening degree of each candidate EGR valve as the optimization variable, the NOx steady-state value, fuel consumption rate, and combustion fluctuation coefficient corresponding to each candidate EGR valve opening are weighted with three modified weights, and the NOx transient peak value corresponding to each candidate EGR valve opening is used as a constraint. The candidate EGR valve opening with the minimum weighted sum is then calculated as the optimal EGR valve opening. Specifically, this includes:

[0044] The steady-state NOx value, fuel consumption rate, and combustion fluctuation coefficient corresponding to the opening degree of each candidate EGR valve were obtained respectively.

[0045] The comprehensive cost of each candidate EGR valve opening is obtained by summing the products of the corrected NOx emission target weight and the NOx steady-state value, the corrected fuel economy target weight and the fuel consumption rate, and the corrected combustion stability target weight and the combustion fluctuation coefficient.

[0046] The transient peak value of NOx corresponding to each candidate EGR valve opening is compared with the preset transient peak value. If the transient peak value of NOx corresponding to the candidate EGR valve opening is greater than the preset transient peak value, the candidate EGR valve opening is eliminated.

[0047] Select the candidate EGR valve opening with the lowest overall cost from the remaining candidate EGR valve openings as the optimal EGR valve opening.

[0048] Furthermore, the predicted arrival time of the next exhaust pulse is obtained, and the predicted arrival time minus the corrected response lag time of the EGR valve is used as the advance amount. The control command corresponding to the optimal EGR valve opening is then issued in advance according to this advance amount, specifically including:

[0049] Based on the pulse time of the current exhaust pulse and the current engine speed, the interval time between adjacent exhaust pulses is determined. The pulse time of the current exhaust pulse plus the interval time between adjacent exhaust pulses is used as the predicted arrival time of the next exhaust pulse.

[0050] The predicted arrival time of the next exhaust pulse is subtracted from the corrected response lag time of the EGR valve as the advance amount;

[0051] Add a lead time to the current moment as the command issuance time, and issue the control command corresponding to the optimal EGR valve opening at the command issuance time.

[0052] This invention proposes an intelligent optimization system for exhaust gas emissions from an adjustable EGR valve, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When executed by the processor, the computer program implements the steps of the intelligent optimization method for exhaust gas emissions from an adjustable EGR valve as described above.

[0053] The beneficial effects of the technical solution of the present invention are:

[0054] This invention identifies the peak value and timing of exhaust pulses in each cylinder by acquiring crankshaft angle and exhaust pressure. Based on the integral of the exhaust pressure exceeding the baseline pressure and the relative amplitude of the pulse peak, it determines the effective pulse intensity to correct the response lag time of the EGR valve. This ensures precise synchronization between the valve core action and the exhaust pulse phase, avoiding mismatch in exhaust gas introduction caused by valve response lag and effectively suppressing NOx emission spikes under transient conditions. By identifying the lag time and rate of change of NOx emission response in both the rising and falling directions of EGR rate, and adjusting prediction parameters in real time according to the current EGR rate change direction and amplitude, it achieves accurate characterization of NOx response differences in different change directions, improving the accuracy of transient NOx prediction. Furthermore, by calculating a memory factor based on historical EGR valve openings and dynamically adjusting the weights of NOx emissions, fuel economy, and combustion stability based on its deviation from the target EGR rate, the optimization decision reflects the impact of previous conditions on the current state, avoiding weight mismatch. Simultaneously, this invention calculates the lead time based on the predicted arrival time of the exhaust pulse and the corrected response lag time, issuing control commands in advance to further improve transient response speed. This invention effectively improves the control accuracy and adaptability of the EGR system for NOx emissions under transient conditions, without requiring additional hardware costs. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 A flowchart illustrating an intelligent optimization method for exhaust gas emissions from an adjustable EGR valve, provided in one embodiment of the present invention.

[0057] Figure 2 This is a structural diagram of an intelligent optimization system for adjustable EGR valve exhaust gas emissions provided in one embodiment of the present invention. Detailed Implementation

[0058] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the intelligent optimization method for adjustable EGR valve exhaust gas according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0059] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0060] The following description, in conjunction with the accompanying drawings, details the specific scheme of the intelligent optimization method for adjustable EGR valve exhaust gas emissions provided by this invention.

[0061] This invention provides an intelligent optimization method and system for exhaust gas emissions from adjustable EGR valves. Please refer to [link / reference]. Figure 1 The diagram illustrates a flowchart of an intelligent optimization method for exhaust gas emissions from an adjustable EGR valve, according to an embodiment of the present invention. The method includes the following steps:

[0062] S101. Obtain the crankshaft angle and exhaust pressure, determine the exhaust phase of each cylinder based on the crankshaft angle and cylinder sequence, and extract the pulse peak value and pulse time of the exhaust pressure within the exhaust phase of each cylinder.

[0063] In this embodiment, the crankshaft angle and exhaust pressure are obtained, and the exhaust phase of each cylinder is determined based on the crankshaft angle and cylinder sequence. The pulse peak value and pulse moment of the exhaust pressure are extracted within each cylinder's exhaust phase, specifically including:

[0064] Obtain the engine's crankshaft angle and exhaust pressure;

[0065] Based on the engine cylinder working sequence, determine the crankshaft angle range corresponding to the exhaust stroke of each cylinder. Within the crankshaft angle range corresponding to the exhaust stroke of each cylinder, extract the local maximum pressure value from the exhaust pressure as the pulse peak value of that cylinder, and record the crankshaft angle time corresponding to the pulse peak value.

[0066] For example, let's first explain the two key terms involved:

[0067] EGR is an abbreviation for Exhaust Gas Recirculation. In this embodiment, the EGR valve refers to a valve device used to control the flow rate of exhaust gas recirculation. The EGR rate refers to the exhaust gas recirculation rate, which is the proportion of exhaust gas introduced into the intake pipe to the total intake gas volume. The EGR valve and the exhaust gas recirculation valve have the same meaning.

[0068] NOx is the common symbol for nitrogen oxides, where 'x' represents the number of oxygen atoms; it is not the letter 'x' but a subscript variable. Nitrogen oxides are a major pollutant emitted during engine combustion, primarily including nitric oxide and nitrogen dioxide. In this embodiment, NOx refers to nitrogen oxide emissions. Except for formulas and specific annotations, "nitrogen oxides" is used to refer to NOx in this embodiment.

[0069] It should be noted that the above abbreviations are all expressed in full Chinese in the embodiments.

[0070] During engine bench or vehicle operation, various sensor signals from the engine electronic control unit are acquired via the controller area network (CLAN) bus. The acquired signals include, but are not limited to: engine speed, torque, intake manifold pressure, exhaust pressure, exhaust temperature, target opening of the exhaust gas recirculation (EGR) valve, actual opening of the EGR valve, pressure difference across the EGR valve, nitrogen oxide emission concentration, and fresh air flow rate. The sampling frequency is set to 1 kHz to ensure the high-frequency components of the exhaust pulses are captured.

[0071] Among the parameters mentioned above, exhaust pressure is the core input signal for subsequent pulse analysis, used to extract the waveform characteristics of exhaust pulses from each cylinder; the actual opening of the exhaust gas recirculation valve reflects the current true position of the valve and is the basic data for subsequent response lag analysis; nitrogen oxide emission concentration is used as the controlled variable to evaluate the control effect and establish an emission response model; engine speed is used to determine the time interval between adjacent exhaust pulses and is a key parameter for pulse timing analysis.

[0072] Simultaneously, real-time crankshaft angle signals are acquired via a crankshaft position sensor. These crankshaft angle signals and exhaust pressure signals are synchronously acquired on the time axis to precisely locate the crankshaft angle range within which each cylinder's exhaust stroke occurs. The acquisition of the crankshaft angle signal provides a timing reference for the subsequent steps of this invention to achieve "cylinder-by-cylinder exhaust pulse identification." Specifically, the crankshaft angle signal determines which cylinder's exhaust stroke is currently in, thereby attributing the exhaust pressure changes within that time period to a specific cylinder, enabling independent analysis and cylinder-by-cylinder modeling of each cylinder's exhaust pulse.

[0073] In the following embodiments, each step can be implemented in the engine electronic control unit via an embedded program, or completed offline during the bench calibration stage. Intermediate variables involved in each step can be stored in the random access memory or read-only memory of the electronic control unit as needed, for subsequent steps to access. Preset parameters involved in each step (including but not limited to base pressure, preset average strength, correction strength coefficient, positive threshold, negative threshold, preset transient peak limit, etc.) are all predetermined through bench calibration experiments and stored in the electronic control unit.

[0074] Taking a four-stroke engine as an example, in a four-stroke engine, each cylinder completes one exhaust stroke in one working cycle. The cylinders exhaust sequentially according to the firing order, causing the exhaust gas to enter the exhaust pipe in a pulsed manner. For a four-cylinder engine, the interval between adjacent exhaust pulses is 180 crankshaft angles; for a six-cylinder engine, the interval is 120 crankshaft angles. This pulsed flow causes the exhaust gas recirculation valve inlet pressure to fluctuate periodically. Engine speed determines the pulse frequency, and load determines the pulse amplitude. The peak time and interval of the pulse pressure directly determine the rhythm of the transient pressure difference across the valve core, thus affecting the valve core's opening response speed and closing sealing performance.

[0075] Therefore, to quantify pulse load, it is first necessary to accurately pinpoint the peak time and interval of each exhaust pulse. This step utilizes a crankshaft position sensor to obtain the real-time crankshaft angle, combined with the engine cylinder firing order (taking a four-cylinder engine as an example, the firing order is...). This involves determining the crankshaft angle range corresponding to the exhaust stroke of each cylinder. This range is typically the crankshaft angle range from the opening to the closing of the exhaust valve, approximately 200 to 250 degrees.

[0076] Based on this, the local maximum pressure value within the crankshaft angle range corresponding to the exhaust stroke of each cylinder is extracted from the exhaust pressure signal and used as the pulse peak value of that cylinder's exhaust stroke. Simultaneously, the crankshaft angle time corresponding to this pulse peak value is recorded. Through this method, the exhaust pulse peak value and pulse occurrence time generated by each cylinder during operation can be completely identified, forming the true pulse excitation sequence experienced by the intake end of the exhaust gas recirculation valve.

[0077] In this embodiment, the exhaust pressure signal and crankshaft angle signal are synchronously acquired on the time axis. The sampling method adopts constant-angle sampling based on crankshaft angle, that is, one data point is collected at every fixed crankshaft angle interval (such as every two degrees of crankshaft angle), rather than sampling at fixed time intervals. The advantage of constant-angle sampling is that the number of sampling points in each exhaust stroke remains constant when the engine speed changes, avoiding the problem of too few sampling points at high speeds and too many sampling points at low speeds caused by fixed-time sampling. This ensures that the exhaust pulse waveform of each cylinder has a consistent sampling resolution at different speeds, providing a unified data basis for the accurate calculation of the effective intensity of the exhaust pulse of each cylinder.

[0078] S102. Based on the integral of the exhaust pressure exceeding the base pressure within the exhaust phase of each cylinder and the relative amplitude of the pulse peak relative to the base pressure, determine the effective intensity of the exhaust pulse of each cylinder, and correct the response hysteresis time of the EGR valve according to the effective intensity.

[0079] In this embodiment, the effective intensity of the exhaust pulse for each cylinder is determined based on the integral of the exhaust pressure exceeding the baseline pressure within each cylinder's exhaust phase and the relative amplitude of the pulse peak value relative to the baseline pressure. The response hysteresis time of the EGR valve is then corrected according to the effective intensity, specifically including:

[0080] Calculate the integral of the exhaust pressure exceeding the base pressure within each cylinder's exhaust phase, and calculate the relative amplitude of the pulse peak relative to the base pressure.

[0081] The product of the integral and the relative amplitude is taken as the effective intensity of the exhaust pulse of that cylinder.

[0082] Calculate the deviation rate between the effective action intensity and the preset average intensity. Take the negative value of the product of the correction intensity coefficient and the deviation rate as the exponent of the natural constant to calculate the correction coefficient. Multiply the response hysteresis time of the EGR valve by the correction coefficient to obtain the corrected response hysteresis time of the EGR valve.

[0083] For example, after the exhaust pulse is discharged from the cylinder, it propagates along the exhaust manifold to the inlet of the exhaust gas recirculation valve. During this process, the pulse waveform attenuates and deforms. The propagation attenuation of the pulse pressure is mainly affected by factors such as pipe length, pipe diameter, number of bends, and exhaust gas temperature. Simultaneously, there is always a steady-state base pressure in the exhaust manifold determined by the average exhaust flow rate upstream. The pulse peak is superimposed on this base pressure to form a complete pulsating pressure waveform. For the force analysis of the exhaust gas recirculation valve core, the pressure throughout the entire pulse cycle does not contribute equally to the valve core opening. The transient high-pressure impact near the pulse rising edge and peak value has the most significant driving effect on the valve core opening, while the driving force during the pulse falling edge and trough is relatively weak.

[0084] The baseline pressure serves as a benchmark reference value for calculating the effective intensity of the pulse, and its accuracy directly affects the validity of all subsequent calculations. In this embodiment, the baseline pressure is determined using a dynamic fusion method combining physical model feedforward and sensor trough calibration, as detailed below:

[0085] First, using the current engine speed and fuel injection quantity (or intake air quantity) known from the engine electronic control unit, the average mass flow rate in the exhaust pipe is estimated in real time based on the ideal gas law and throttling formula. This is then combined with the exhaust pipe's geometric parameters to convert it into the theoretical value of the base pressure under the current operating conditions. The advantage of this feedforward value is that it has no hysteresis and can reflect the rapid response of the base pressure to transient changes in operating conditions.

[0086] Simultaneously, using the acquired crankshaft angle signal, the crankshaft angle gap range where the exhaust valve of each cylinder is closed but the exhaust stroke of the next cylinder has not yet begun is precisely located. Within this gap range, no new pulse is injected into the exhaust pipe, and the pressure value measured by the exhaust pressure sensor at this time is closest to the current actual physical baseline pressure. The arithmetic mean of the exhaust pressure within this gap range is used as the sensor calibration value for the baseline pressure.

[0087] Finally, the physical feedforward value and the sensor calibration value are fused through complementary filtering to obtain the final base pressure value. The fusion coefficient is adaptively adjusted according to the rate of change of the current operating conditions: under steady-state conditions, the weight of the sensor calibration value is increased to eliminate the cumulative deviation of the physical model caused by carbon buildup or wear; under transient conditions, the weight of the physical feedforward value is increased to avoid the delay in base pressure estimation caused by sensor signal lag.

[0088] To quantify the effective contribution of each exhaust pulse to the driving force of the exhaust gas recirculation valve core, this step, after determining the baseline pressure under the current operating conditions, extracts the pressure waveform between adjacent pulse peaks and calculates the integral of the portion of the exhaust pressure exceeding the baseline pressure over the entire pulse cycle. This integral reflects the total impact energy of the high-pressure exhaust gas on the valve core. Its physical meaning is that the greater the amplitude of the pulse pressure exceeding the baseline pressure and the wider the pulse duration, the larger the integral, indicating a stronger contribution of the pulse to the driving force of the valve core. Simultaneously, the ratio of the pulse peak value to the baseline pressure is calculated as a relative amplitude, reflecting the relative intensity of the pulse relative to the background pressure.

[0089] The product of the integral and the relative amplitude is taken as the effective intensity of the exhaust pulse for that cylinder. This effective intensity is a dimensionless scalar, representing the actual contribution of the current exhaust pulse to the valve core opening drive force. The greater the difference between the pulse peak and the base pressure, the wider the pulse duration, and the slower the fall-edge decay, the higher the effective intensity.

[0090] In terms of calculation timing, the effective intensity is calculated separately for each cylinder's exhaust stroke; that is, one effective intensity value is generated for each cylinder's exhaust stroke. Until the next cylinder's exhaust stroke ends and a new effective intensity is calculated, the system continuously uses the current effective intensity value for subsequent correction calculations. This "cylinder-by-cylinder refresh, interval-based maintenance" update method maintains consistency with the engine's combustion rhythm, avoiding control command oscillations caused by frequent intensity value jumps.

[0091] The response lag time of an exhaust gas recirculation (EGR) valve depends not only on the mechanical inertia of the valve body itself (such as valve core mass, spring stiffness, and friction), but also closely on the transient pressure difference across the valve core. When the peak pulse pressure reaches the EGR valve inlet, the pressure difference across the valve core increases instantaneously, and the opening driving force on the valve core also increases accordingly, thereby accelerating the valve core's opening response and reducing the effective lag time. Conversely, during the pulse trough, the pressure difference decreases, the valve core's opening driving force is insufficient, and the response slows down. Traditional methods using a fixed lag time constant cannot reflect this dynamic characteristic that changes with exhaust pulse fluctuations.

[0092] This step utilizes the effective intensity of the exhaust pulses for each cylinder calculated above to perform pulse-weighted correction on the reference lag time constants for each operating point obtained from the bench step response experiment. Specifically, the step response experiment is conducted at different speeds and loads (e.g., speed range of 1000 to 4000 rpm, step size of 500 rpm; load range of 20% to 100%, step size of 20%) to obtain a mapping table of reference lag time constants for each operating point. In real-time control, the reference lag time constant for the current operating point is obtained by looking up the table based on the current speed and load, and then corrected using the current effective pulse intensity.

[0093] The correction employs an exponential function, specifically: calculating the deviation rate between the current effective pulse intensity and the preset average intensity; taking the negative of the product of the correction intensity coefficient and this deviation rate as the exponent of the natural constant to calculate the correction coefficient; and multiplying the reference lag time constant by this correction coefficient to obtain the corrected response lag time. The preset average intensity is the statistical average of the effective pulse intensity of each cylinder under the current operating conditions. When the effective pulse intensity is higher than the preset average intensity, the correction coefficient is less than one, and the response lag time is shortened, indicating that the pulse pressure boosts the valve core to open. When the effective pulse intensity is lower than the preset average intensity, the correction coefficient is greater than one, and the response lag time is prolonged, indicating that the pulse pressure is insufficient to boost the valve core, and a more favorable opening opportunity must be awaited.

[0094] Because the exhaust gas recirculation valve is propelled by upstream pulse pressure when opening, but mainly relies on spring restoring force to act against the exhaust gas flow pressure when closing, the pulse conditions they experience are different. Therefore, the opening and closing response states should be corrected separately, and separate opening and closing response lag time mapping tables should be established. This allows the dynamic response lag time of the exhaust gas recirculation valve to be calculated at any given time based on the current exhaust pulse conditions, serving as a time reference correction parameter for subsequent emission response prediction models.

[0095] S103. Obtain historical EGR rate change data, and identify the lag time and rate of change of NOx emission response in both directions of EGR rate increase and decrease.

[0096] In this embodiment, historical EGR rate change data is acquired, and the lag time and rate of change of NOx emission response are identified in both the rising and falling directions of EGR rate. Specifically, this includes:

[0097] Acquire historical EGR rate change data, which is a continuous time series of EGR rate changes over time during historical operation.

[0098] From historical EGR rate change data, events in which the EGR rate undergoes a step change are selected as EGR rate change events, and then divided into rising event group and falling event group according to the two directions of EGR rate increase and decrease.

[0099] For each event in the rising event group, the time difference between the start of the change in EGR rate and the start of the change in NOx emission concentration is extracted as the rising lag time, and the average rate of change required for the NOx emission concentration to change from 10% to 90% of the steady-state value before the change is extracted as the rising rate of change.

[0100] For each event in the declining event group, the time difference between the start of the change in EGR rate and the start of the change in NOx emission concentration is extracted as the lag time in the declining direction, and the average rate of change required for the NOx emission concentration to change from 10% to 90% of the steady-state value before the change is extracted as the rate of change in the declining direction.

[0101] For example, in actual engine operation, changes in the exhaust gas recirculation (EGR) rate alter the combustion temperature by affecting the specific heat capacity and oxygen concentration of the in-cylinder working fluid, thereby influencing the amount of nitrogen oxides (NOx) generated. When the EGR rate increases, the actual increase in the EGR rate in the intake manifold, from the change in the EGR valve opening, requires the exhaust gas to flow out of the exhaust manifold, pass through the EGR cooler, flow through the EGR valve, mix with fresh air, enter the intake manifold, and finally be drawn into the cylinder for combustion. This process involves multiple stages, including pneumatic transmission, mixing and diffusion, and air exchange, resulting in a significant lag in the reduction of NOx emissions. Conversely, when the EGR rate decreases, the residual exhaust gas in the intake manifold and intake passage is not immediately and completely removed. The residual exhaust gas can still affect the formation of NOx in subsequent combustion cycles, exhibiting a decay memory characteristic. Therefore, the NOx emission response characteristics in the directions of increasing and decreasing EGR rate differ significantly in both lag time and rate of change, exhibiting typical asymmetry.

[0102] To quantify this asymmetry, this step first obtains continuous time series data of the exhaust gas recirculation rate over time recorded in the historical operating database. Then, from this continuous time series, events in which the exhaust gas recirculation rate undergoes a step change are selected as exhaust gas recirculation rate change events.

[0103] The specific screening method is as follows: Calculate the rate of change of the exhaust gas recirculation rate at each moment in the historical exhaust gas recirculation rate change data, that is, the amount of change in the exhaust gas recirculation rate per unit time. When the absolute value of the rate of change of the exhaust gas recirculation rate exceeds a set threshold, it is determined as the starting point of a step change. From this starting point, the change process of the exhaust gas recirculation rate is continuously monitored until the absolute value of the rate of change of the exhaust gas recirculation rate falls back below the set threshold and the exhaust gas recirculation rate tends to a new stable value. A complete process from the starting point to the point of stabilization is recorded as an exhaust gas recirculation rate change event. At the same time, based on whether the exhaust gas recirculation rate increases or decreases in this event, it is divided into an increasing event group and a decreasing event group.

[0104] For each selected event showing a change in exhaust gas recirculation rate, two key parameters are extracted from the synchronously recorded nitrogen oxide emission concentration response curve, using the moment when the exhaust gas recirculation rate begins to change as a baseline:

[0105] The first parameter is the lag time, which is the time difference between the baseline moment when the exhaust gas recirculation rate begins to change and the moment when a detectable change in the nitrogen oxide emission concentration begins to appear. This parameter reflects the physical delay in exhaust gas transport and mixing.

[0106] The second parameter is the rate of change, which is the average rate of change required for the nitrogen oxide emission concentration to increase from 10% to 90% of its steady-state value. This parameter reflects the drastic response of nitrogen oxides to changes in the exhaust gas recirculation rate.

[0107] Due to the volumetric inertia of the exhaust gas recirculation path, the response lag time of nitrogen oxides in the upward direction (i.e., the direction of increasing exhaust gas recirculation rate) is usually greater than that in the downward direction (i.e., the direction of decreasing exhaust gas recirculation rate), because the transmission delay of exhaust gas entry is longer than the purging delay; while the rate of change in the upward direction is usually less than that in the downward direction, because the downward process is limited by the mixing process, and the upward process is limited by the combustion recovery rate. This asymmetry is the core characteristic of transient emission dynamics.

[0108] Using the above method, complete characteristic parameters of the nitrogen oxide response lag time and rate of change in both the upward and downward directions can be obtained. These parameters are stored separately according to engine speed and load conditions, forming a lag time mapping table, a rate of change mapping table, a lag time mapping table, and a rate of change mapping table in the downward direction, which can be used for subsequent real-time control steps to look up and call the tables according to the current operating conditions.

[0109] S104. Calculate the current EGR rate change rate, select the corresponding lag time and change rate according to its direction, scale the selected lag time and change rate according to the current EGR rate change amplitude, and use the scaling result to predict the NOx transient peak value and steady-state value corresponding to the opening degree of each candidate EGR valve.

[0110] In this embodiment, the current EGR rate change rate is calculated, and the lag time and change rate in the corresponding direction are selected based on its direction. The selected lag time and change rate are scaled according to the current EGR rate change amplitude. The scaling result is used to predict the transient peak value and steady-state value of NOx corresponding to the opening degree of each candidate EGR valve. Specifically, this includes:

[0111] Calculate the current rate of change of EGR rate, and determine the magnitude of the current EGR rate change based on the difference between the current EGR rate value and the EGR rate value at the moment when the EGR rate began to change;

[0112] The current EGR rate of change is compared with preset positive and negative thresholds. If the current EGR rate of change is greater than the positive threshold, the direction of the EGR rate of change is determined to be upward, and the lag time and rate of change corresponding to the upward direction are selected. If the current EGR rate of change is less than the negative threshold, the direction of the EGR rate of change is determined to be downward, and the lag time and rate of change corresponding to the downward direction are selected. If the current EGR rate of change is between the negative and positive thresholds, the EGR rate of change is determined to be in a quasi-steady state.

[0113] When it is determined that the direction of the EGR rate change is upward or downward, the selected lag time and rate of change are scaled according to the current EGR rate change magnitude.

[0114] Obtain the opening degree of the candidate EGR valve;

[0115] Using each candidate EGR valve opening as the target opening, and utilizing the scaled lag time and rate of change, the dynamic trajectory of NOx emission concentration during the transition from the current EGR valve opening to each candidate EGR valve opening is simulated. The maximum value is extracted from the dynamic trajectory corresponding to each candidate EGR valve opening as the NOx transient peak value corresponding to that candidate EGR valve opening, and the endpoint value is extracted as the NOx steady-state value corresponding to that candidate EGR valve opening.

[0116] For example, the lag time and rate of change of nitrogen oxide emissions in both the rising and falling directions of the exhaust gas recirculation rate have been obtained through the above steps. However, in real-time prediction, it is necessary to dynamically select appropriate response parameters based on the current direction of change of the exhaust gas recirculation rate. Furthermore, the magnitude of the change in the exhaust gas recirculation rate also affects the response characteristics. Small changes in the exhaust gas recirculation rate may only disturb combustion in a localized area, while large changes may alter the combustion mode, leading to nonlinear scaling of the response characteristics. Therefore, based on the direction determination, it is also necessary to appropriately adjust the response parameters according to the magnitude of the change.

[0117] At each control moment, the current rate of change of the exhaust gas recirculation rate is first calculated, which is the change in the exhaust gas recirculation rate per unit time. This rate of change is calculated by dividing the change in the actual opening value of the exhaust gas recirculation rate by the corresponding time interval.

[0118] The calculated rate of change of the current exhaust gas recirculation rate is compared with preset positive and negative thresholds. If the rate of change of the current exhaust gas recirculation rate is greater than the positive threshold, it is determined that the current exhaust gas recirculation rate is in an upward state, and the lag time and rate of change in the upward direction are selected as the basic response parameters. If the rate of change of the current exhaust gas recirculation rate is less than the negative threshold, it is determined that the current exhaust gas recirculation rate is in a downward state, and the lag time and rate of change in the downward direction are selected as the basic response parameters. If the absolute value of the rate of change of the current exhaust gas recirculation rate is between the positive and negative thresholds, it is determined that the current state is quasi-steady, and the nitrogen oxide emission value is directly determined using the steady-state mapping relationship without the need for transient prediction.

[0119] The threshold for the rate of change is determined as follows: Statistical analysis is performed on all exhaust gas recirculation rate change events in the historical operating database. The distribution characteristics of the exhaust gas recirculation rate change rate in each event are calculated. Using the standard deviation of the rate of change distribution as a benchmark, a positive threshold is set to 0.5 times the positive standard deviation, and a negative threshold is set to 0.5 times the negative standard deviation. This threshold setting method allows direction discrimination to adapt to the characteristic differences of different engines, avoiding misjudgments or omissions caused by using a fixed threshold.

[0120] When it is determined that the current exhaust gas recirculation rate is in an upward or downward state, it is necessary to further scale and correct the selected lag time and rate of change based on the magnitude of the change in the current exhaust gas recirculation rate.

[0121] In this embodiment, the change range of the exhaust gas recirculation rate is determined by the difference between the current exhaust gas recirculation rate value and the exhaust gas recirculation rate value at the moment the change begins. Scaling is applied based on this difference: when the change range is large, the pressure difference generated by the large change is stronger, the equivalent lag time is shortened, and the equivalent change rate is accelerated; when the change range is small, the equivalent lag time is prolonged, and the equivalent change rate is slowed down. Specifically, the scaling method is as follows: the larger the change range of the current exhaust gas recirculation rate, the shorter the lag time is shortened, and the faster the change rate is amplified, i.e., the faster the response; the smaller the change range of the current exhaust gas recirculation rate, the longer the lag time is shortened, and the slower the change rate is amplified, i.e., the slower the response.

[0122] The specific scaling methods can be as follows:

[0123] The moment when the exhaust gas recirculation rate begins to change is taken as the starting moment of the change, and the exhaust gas recirculation rate value at that moment is recorded as the initial value. At the current moment, the initial value is subtracted from the current exhaust gas recirculation rate value, and the absolute value is taken as the current change range of the exhaust gas recirculation rate.

[0124] For different engine speeds and load conditions, a baseline variation range is pre-determined at each operating point through bench calibration. This baseline variation range represents a reference value for the typical step-like change in exhaust gas recirculation rate under that operating condition.

[0125] The scaling factor is obtained by dividing the current change in exhaust gas recirculation rate by the baseline change at the current operating point.

[0126] Divide the selected lag time by the scaling factor to obtain the scaled lag time; multiply the selected rate of change by the scaling factor to obtain the scaled rate of change.

[0127] When the scaling factor is greater than 1 (i.e., the current change is greater than the baseline change), the lag time is shortened and the rate of change is accelerated, reflecting the strong pressure difference driven by the large change, which accelerates the response; when the scaling factor is less than 1 (i.e., the current change is less than the baseline change), the lag time is extended and the rate of change is slowed, reflecting the weak driving force of the small change, which slows the response; when the scaling factor is equal to 1 (i.e., the current change is equal to the baseline change), the lag time and the rate of change remain unchanged, and the baseline response parameters are directly used for subsequent prediction.

[0128] To prevent scaling results from exceeding physically reasonable ranges under extreme operating conditions, upper and lower limits are set for the scaling factor. When the scaling factor exceeds the upper limit threshold, it is truncated according to the upper limit threshold; when the scaling factor is below the lower limit threshold, it is truncated according to the lower limit threshold. These upper and lower limits are predetermined through bench calibration to ensure that the hysteresis time and rate of change after scaling are always within the physically achievable range.

[0129] During each optimization decision, the system generates multiple candidate exhaust gas recirculation valve openings at preset intervals within the adjustable opening range of the exhaust gas recirculation valve, serving as alternative solutions to be evaluated. After completing direction selection and amplitude scaling, a first-order exponential decay form is used to describe the dynamic process of nitrogen oxide emissions transitioning from the current value to the steady-state value. The decay rate of this first-order exponential decay is jointly determined by the direction-dependent lag time and the amplitude correction factor, accurately characterizing the asymmetric characteristics of different response rates in the rising and falling directions.

[0130] The specific prediction method is as follows: For each candidate exhaust gas recirculation (EGR) valve opening, the dynamic trajectory of nitrogen oxide (NOx) emission concentration is simulated during the transition from the current EGR valve opening to that candidate EGR valve opening using scaled lag time and rate of change. This simulation process employs a constant-angle discretization method based on crankshaft angle, using the crankshaft angle as the discretization reference (e.g., every two degrees of crankshaft angle is one step) to recursively calculate the emission response process. For a four-cylinder engine, a complete working cycle is 720 degrees of crankshaft angle, totaling 360 discrete steps; for a six-cylinder engine, the same 720 degrees of crankshaft angle and 360 discrete steps are used. Constant-angle discretization can solve the problem of inconsistent sampling accuracy at different speeds with a fixed time step: at lower speeds, every two degrees of crankshaft rotation corresponds to approximately 0.33 milliseconds, which can fully capture every detail of the pulse cycle; at higher speeds, every two degrees of crankshaft rotation corresponds to approximately 0.083 milliseconds, which still maintains the same number of sampling points within each pulse cycle, ensuring complete capture of the dynamic response of nitrogen oxides excited by exhaust pulses. Each discretization step calculates the nitrogen oxide emission concentration at the next moment according to a first-order exponential decay recursive formula, until the nitrogen oxide concentration approaches the steady-state value corresponding to the candidate opening degree.

[0131] After obtaining the complete dynamic response trajectory of nitrogen oxides corresponding to each candidate opening degree, the maximum value is extracted from each trajectory as the transient peak value of nitrogen oxides corresponding to that candidate opening degree, and the endpoint value is extracted as the steady-state value of nitrogen oxides corresponding to that candidate opening degree.

[0132] Simultaneously, for the change in the opening direction of the decreasing exhaust gas recirculation rate, the recovery delay time of nitrogen oxide emissions from the suppressed state to the new steady-state value is also calculated. The "suppressed state" is defined as the period from the current moment until the nitrogen oxide emission concentration first reaches 90% of the steady-state predicted value corresponding to the target opening; the "new steady-state value" is the steady-state emission mapping value under the target opening, which is obtained through emission calibration experiments at different steady-state operating points (different speed and load combinations). The recovery delay time is calculated as follows: starting from the beginning of the suppressed state, the change in nitrogen oxide emission concentration near the steady-state predicted value corresponding to the target opening is recorded. The steady-state arrival time is determined using the sliding window averaging method. When the deviation between the sliding average value of nitrogen oxide emission concentration and the steady-state predicted value is continuously less than ±2% of the steady-state predicted value within a continuous 50-millisecond window, steady-state is determined, and the beginning of this window is taken as the steady-state arrival time. The time difference between the beginning of the suppressed state and the steady-state arrival time is the recovery delay time. The longer the recovery delay time, the more persistent the impact of residual exhaust gas, and the more complex the corresponding emission control. This parameter can be used as an auxiliary indicator to evaluate the control complexity of candidate opening schemes in subsequent multi-objective optimization.

[0133] It should be noted that the aforementioned equal-angle discretization recursive calculation and subsequent optimization solution process employs a combination of offline calibration and online table lookup in actual engineering implementation. During the calibration phase or on a cloud-based high-performance computing platform, typical operating condition combinations (including dimensions such as engine speed, load, direction of change of exhaust gas recirculation rate, change range, and pulse intensity) are pre-traversed. The aforementioned asymmetric dynamic response prediction model is used to perform equal-angle discretization recursive calculations, and the corresponding transient peak and steady-state values ​​of nitrogen oxides are calculated for each candidate exhaust gas recirculation valve opening at each calibration operating point, forming transient peak constraint mapping tables and steady-state value mapping tables. The solution results are stored in the engine electronic control unit (ECU) in the form of an optimal opening mapping table. During actual operation, the ECU only needs to look up the tables based on the current operating condition parameters to obtain the optimal opening, transient peak constraint, and steady-state target value for the current operating condition, and then output control commands after feedforward compensation, without requiring online iterative calculations. By combining offline and online engineering approaches, the ability to fully capture and accurately predict the dynamic response of nitrogen oxides excited by exhaust pulses is preserved, while also meeting the real-time constraints of the vehicle's electronic control unit.

[0134] By comparing the transient peak emissions and steady-state emissions under different candidate opening degrees, candidate schemes that meet steady-state emissions standards but exceed transient peak emissions limits can be identified. These can then be eliminated or penalized in subsequent multi-objective optimization. This yields the transient peak and steady-state values ​​of nitrogen oxide emissions corresponding to all candidate opening degrees within the adjustable range of the exhaust gas recirculation valve, providing emission constraint boundaries for subsequent multi-objective optimization. The candidate exhaust gas recirculation valve opening degrees are multiple alternative opening degree schemes set by the control system during the optimization process. The transient peak and steady-state values ​​corresponding to each candidate opening degree are obtained through the above recursive calculation, allowing subsequent optimization steps to select the optimal one.

[0135] S105. Calculate the memory factor based on the historical EGR valve opening, and adjust the weights of the three targets of NOx emissions, fuel economy and combustion stability based on the deviation between the memory factor and the target EGR rate.

[0136] In this embodiment, the memory factor is calculated based on the historical EGR valve opening, specifically including:

[0137] Obtain the historical EGR valve opening at each moment within the preceding time window;

[0138] An exponentially weighted average of the historical EGR valve openings at each time point within the preceding time window is used as the memory factor. The weight of each time point is negatively correlated with the time elapsed between that time point and the current time point.

[0139] The weights of the three targets—NOx emissions, fuel economy, and combustion stability—are adjusted based on the deviation between the memory factor and the target EGR rate. Specifically, this includes:

[0140] The difference between the memory factor and the target EGR rate is calculated as the residual exhaust gas deviation.

[0141] The weight correction coefficients for NOx emission targets, fuel economy targets, and combustion stability targets are determined based on the ratio of residual exhaust gas deviation to target EGR rate. The weight correction coefficient for each target is an exponential function of the natural constant, and the independent variable of the exponential function is the product of the memory correction sensitivity coefficient of the target and the ratio.

[0142] The base weights of each objective are multiplied by the corresponding weight correction coefficients to obtain the corrected weights of each objective. The three corrected weights are then normalized so that the sum of the three is 1.

[0143] For example, the engine intake manifold (including the intake duct and intake passage) has a certain volume, equivalent to a pneumatic low-pass filter. When the exhaust gas recirculation (EGR) valve opening changes, the exhaust gas flow rate into the intake manifold changes accordingly, but the exhaust gas already stored in the manifold is not immediately replaced, but gradually renewed at a certain rate. This "volume inertia" effect means that the amount of residual exhaust gas in the intake manifold depends not only on the current EGR valve opening, but also on the history of the EGR valve opening over a previous period. When the engine quickly switches from a high EGR rate condition to a low EGR rate condition, although the EGR valve has closed slightly, a large amount of exhaust gas remains in the intake manifold, continuing to dilute fresh air and affecting combustion and nitrogen oxide emissions. This influence of previous operating conditions on the current state is referred to as the "condition memory effect" in this embodiment.

[0144] This step constructs a cumulative estimation model for the residual exhaust gas volume in the intake manifold based on historical time-series data. The intake manifold is considered a single-cavity structure, and the dynamic change in the proportion of exhaust gas within it (i.e., the exhaust gas recirculation rate) satisfies a first-order inertial plus pure delay model. The pure delay time parameter is derived from the obtained nitrogen oxide emission response lag time constant, which covers the complete transmission delay of exhaust gas from the exhaust manifold through the exhaust gas recirculation path to the cylinder, ensuring the consistency of time parameters throughout the entire process.

[0145] The current value of the intake manifold exhaust gas recirculation (EGR) rate is obtained by weighted averaging of the EGR valve openings at previous times. Specifically, historical EGR valve opening values ​​are obtained for each time window preceding the current time; these historical opening values ​​are then subjected to an exponentially weighted average, which is used as a memory factor. In the exponentially weighted average, the weight of each time point decreases exponentially with the time elapsed since the current time; that is, historical opening values ​​closer to the current time have a larger weight, and historical opening values ​​farther away have a smaller weight. The time constant of this weighted average depends on the intake manifold volume, engine displacement, and current engine speed, reflecting the intake manifold displacement rate. The larger the volume, the smaller the displacement, and the lower the engine speed, the slower the displacement and the larger the time constant; conversely, the faster the displacement and the smaller the time constant.

[0146] The essence of this calculation process is to weight and sum the exhaust gas recirculation valve openings at each preceding time step according to an exponential decay law. Weight normalization ensures that the dimensions of the memory factor are consistent with the exhaust gas recirculation valve opening (i.e., an estimated exhaust gas recirculation rate in percentage form). The memory factor represents the estimated amount of residual exhaust gas in the intake manifold at the current time. When the memory factor is significantly greater than the current target exhaust gas recirculation rate, it indicates that there is excessive residual exhaust gas in the intake manifold; when the memory factor is significantly less than the current target exhaust gas recirculation rate, it indicates that the amount of exhaust gas in the intake manifold is insufficient.

[0147] At this point, the working condition memory factor can be calculated at any time based on the trajectory of the preceding working conditions.

[0148] After obtaining the operating condition memory factor, this step compares it with the current target exhaust gas recirculation rate and calculates the residual exhaust gas deviation, which is the memory factor minus the current target exhaust gas recirculation rate.

[0149] The sign of the residual exhaust gas deviation determines the direction of adjustment for the weights of each target. When the residual exhaust gas deviation is greater than zero, it indicates that there is more residual exhaust gas in the intake manifold than the target value. On the one hand, residual exhaust gas itself helps to suppress nitrogen oxide (NOx) formation, so the weight of the NOx emission target can be appropriately reduced. On the other hand, excessive residual exhaust gas may worsen combustion stability, and the replacement of some fresh air by residual exhaust gas may lead to a decrease in torque and an increase in fuel consumption rate. Therefore, the weights of the combustion stability and fuel economy targets need to be appropriately increased. When the residual exhaust gas deviation is less than zero, it indicates that there is less residual exhaust gas in the intake manifold than the target value. In this case, NOx emissions may increase due to the lack of dilution and cooling effect from exhaust gas, so the weight of the NOx emission target needs to be increased. At the same time, combustion stability and charging efficiency are relatively good, so the weights of these two targets can be appropriately reduced.

[0150] The baseline weights of each objective differ under different operating conditions: under low-speed, high-torque conditions, combustion stability is the primary constraint; under medium- to high-speed steady-state conditions, nitrogen oxide emissions have a higher weight; and under high-load conditions, fuel economy should have a higher weight. This step first determines the basic weight coefficients of each objective by referring to a table based on the current engine speed and load.

[0151] Then, the basic weights are corrected using the operating condition memory factor. Specifically, the correction method is as follows: calculate the ratio of residual exhaust gas deviation to the target exhaust gas recirculation rate; for each of the nitrogen oxide emission target, fuel economy target, and combustion stability target, multiply this ratio by the preset memory correction sensitivity coefficient corresponding to that target, using it as the independent variable of an exponential function of the natural constant; calculate the value of this exponential function as the weight correction coefficient for each target; and multiply the basic weight of each target by the corresponding weight correction coefficient to obtain the corrected weight for each target.

[0152] The memory correction sensitivity coefficients for each target are preset parameters, determined in advance through bench calibration, and stored in the electronic control unit. The sensitivity coefficient for the nitrogen oxide emission target is negative; when the residual exhaust gas deviation is greater than zero, its correction coefficient is less than 1, and its weight is reduced. The sensitivity coefficients for the fuel economy and combustion stability targets are positive; when the residual exhaust gas deviation is greater than zero, their correction coefficients are greater than 1, and their weights are increased. After correction, the three weights need to be normalized so that their sum equals 1.

[0153] The aforementioned correction method achieves non-linear correction of the weights based on the residual exhaust gas deviation using an exponential function. The greater the deviation of the residual exhaust gas from the target value, the larger the weight adjustment; when the residual exhaust gas equals the target value, the correction coefficient is 1, and the weight remains unchanged from the base weight. This correction method ensures that the optimization direction can be adaptively adjusted according to the actual exhaust gas state within the intake manifold. Thus, the dynamic weight coefficients of each target after operating condition memory correction can be calculated in real time at any given moment.

[0154] S106. Using the opening degree of each candidate EGR valve as the optimization variable, the NOx steady-state value, fuel consumption rate and combustion fluctuation coefficient corresponding to each candidate EGR valve opening degree are weighted by the three corrected weights, and the NOx transient peak value corresponding to each candidate EGR valve opening degree is constrained to find the candidate EGR valve opening degree with the smallest weighted sum as the optimal EGR valve opening degree.

[0155] In this embodiment, the opening degree of each candidate EGR valve is used as the optimization variable. The NOx steady-state value, fuel consumption rate, and combustion fluctuation coefficient corresponding to each candidate EGR valve opening are weighted using three modified weights. The transient peak value of NOx corresponding to each candidate EGR valve opening is used as a constraint. The candidate EGR valve opening with the smallest weighted sum is then calculated as the optimal EGR valve opening. Specifically, this includes:

[0156] The steady-state NOx value, fuel consumption rate, and combustion fluctuation coefficient corresponding to the opening degree of each candidate EGR valve were obtained respectively.

[0157] The comprehensive cost of each candidate EGR valve opening is obtained by summing the products of the corrected NOx emission target weight and the NOx steady-state value, the corrected fuel economy target weight and the fuel consumption rate, and the corrected combustion stability target weight and the combustion fluctuation coefficient.

[0158] The transient peak value of NOx corresponding to each candidate EGR valve opening is compared with the preset transient peak value. If the transient peak value of NOx corresponding to the candidate EGR valve opening is greater than the preset transient peak value, the candidate EGR valve opening is eliminated.

[0159] Select the candidate EGR valve opening with the lowest overall cost from the remaining candidate EGR valve openings as the optimal EGR valve opening.

[0160] For example, after obtaining the dynamic weighting coefficients corrected by operating condition memory, this step constructs a comprehensive cost function that comprehensively considers three factors: steady-state emission target, transient peak constraint, and dynamic response cost.

[0161] The comprehensive cost function consists of five terms:

[0162] The first item is the nitrogen oxide emission target, calculated by multiplying the corrected nitrogen oxide emission target weight by the ratio of the predicted steady-state nitrogen oxide value to the reference value. The predicted steady-state nitrogen oxide value corresponds to the steady-state nitrogen oxide value for the current candidate operating degree; the reference value is the ideal target value for nitrogen oxide emissions at this operating point, pre-determined through bench calibration. A smaller value indicates better performance of the candidate operating degree in terms of steady-state emissions.

[0163] The second item is the fuel economy target, which is calculated by multiplying the corrected fuel economy target weight by the ratio of the estimated fuel consumption rate to the reference value. The estimated fuel consumption rate is based on the charging efficiency and exhaust gas recirculation rate under the current candidate opening degree and is estimated through a fuel consumption model; the reference value is the ideal target value of fuel consumption rate under this operating condition, which is predetermined through bench calibration.

[0164] The third item is the combustion stability target, which is calculated by multiplying the corrected combustion stability target weight by the ratio of the estimated cycle fluctuation coefficient to the reference value. The cycle fluctuation coefficient characterizes the degree of variation in combustion state between operating cycles; a higher value indicates greater combustion instability. The cycle fluctuation coefficient is estimated using a combustion stability model based on the exhaust gas recirculation rate and intake air volume at the current candidate opening degree. The reference value is the allowable upper limit of the cycle fluctuation coefficient at this operating point, predetermined through bench calibration.

[0165] The fourth item is a secondary penalty for exceeding peak emission limits. The transient peak value of nitrogen oxides corresponding to the current candidate emission level is compared with the preset transient peak value limit. If the transient peak value exceeds the limit, a secondary penalty is applied. The penalty amount is the square of the portion of the transient peak value exceeding the limit multiplied by a preset penalty coefficient; the greater the exceedance, the larger the penalty. If the transient peak value does not exceed the limit, this item has a value of zero. This penalty identifies candidate schemes that meet steady-state emission standards but exceed transient peak values, allowing them to be excluded or penalized during optimization.

[0166] The fifth item is a smoothing penalty for the rate of change of valve opening. This penalty is the square of the difference between the current candidate opening and the actual opening in the previous control cycle, multiplied by a preset smoothing coefficient. It is used to suppress drastic jumps in the opening command. Too rapid a change rate may lead to new transient disturbances or increased mechanical wear of the valve core.

[0167] The overall cost function is the sum of the above five terms. The weights of the first three terms have been obtained through operating condition memory correction, which can reflect the impact of the current residual exhaust gas volume in the intake manifold on the priority of the optimization target; the fourth term ensures that transient nitrogen oxide emissions do not exceed the limit; and the fifth term ensures the smoothness of control actions.

[0168] Then, using the candidate opening degree as the optimization variable, a numerical optimization algorithm is employed to find the minimum value of the comprehensive cost function within the opening degree range of 0% to 100%. In this embodiment, a sequential quadratic programming algorithm is used for the solution. The basic idea of ​​this algorithm is as follows: in each iteration, a quadratic programming subproblem of the original problem is constructed near the current candidate solution, and the search direction is solved. The Hessian matrix approximation is updated using the BFGS method, gradually approaching the optimal solution. When the convergence condition is met (such as the change in the cost function value between two adjacent iterations being less than a preset convergence threshold, or the number of iterations reaching a preset upper limit), the opening degree corresponding to the minimum value of the comprehensive cost function is obtained as the optimal exhaust gas recirculation valve opening degree.

[0169] It should be noted that the above optimization solution process adopts a combination of offline and online methods in actual engineering implementation: During the calibration phase or on a cloud-based high-performance computing platform, typical operating condition combinations (including dimensions such as speed, load, direction of change of exhaust gas recirculation rate, change range, and pulse intensity) are pre-traversed, and the above-mentioned sequential quadratic programming algorithm is used for offline iterative solution. The solution results are stored in the engine electronic control unit in the form of an optimal opening degree mapping table. During actual operation, the electronic control unit obtains the optimal opening degree by looking up the table according to the current operating condition parameters, and then outputs control commands after feedforward compensation, without the need for online iterative calculation. Through this engineering division of labor, both the accuracy of the optimization solution and the real-time constraints of the on-board electronic control unit are guaranteed.

[0170] S107. Obtain the predicted arrival time of the next exhaust pulse, and use the predicted arrival time minus the corrected response lag time of the EGR valve as the advance amount, and issue the control command corresponding to the optimal EGR valve opening in advance according to the advance amount.

[0171] In this embodiment, the predicted arrival time of the next exhaust pulse is obtained, and the predicted arrival time minus the corrected response lag time of the EGR valve is used as the advance amount. The control command corresponding to the optimal EGR valve opening is then issued in advance according to this advance amount, specifically including:

[0172] Based on the pulse time of the current exhaust pulse and the current engine speed, the interval time between adjacent exhaust pulses is determined. The pulse time of the current exhaust pulse plus the interval time between adjacent exhaust pulses is used as the predicted arrival time of the next exhaust pulse.

[0173] The predicted arrival time of the next exhaust pulse is subtracted from the corrected response lag time of the EGR valve as the advance amount;

[0174] Add a lead time to the current moment as the command issuance time, and issue the control command corresponding to the optimal EGR valve opening at the command issuance time.

[0175] For example, the optimal exhaust gas recirculation valve opening is used as the desired opening of the feedforward compensation module. After feedforward compensation, the final control command is generated and output to the exhaust gas recirculation valve actuator. The specific calculation method of feedforward compensation is as follows: the interval time between adjacent exhaust pulses is determined based on the pulse time of the current exhaust pulse and the current engine speed. The pulse time of the current exhaust pulse plus the interval time between adjacent exhaust pulses is used as the predicted arrival time of the next exhaust pulse. The predicted arrival time minus the corrected exhaust gas recirculation valve response lag time is used as the advance amount. The advance amount is added to the current time as the command issuance time. At the command issuance time, the control command corresponding to the optimal exhaust gas recirculation valve opening is issued, so that the valve core action can be precisely synchronized with the arrival time of the exhaust pulse.

[0176] The above steps have completed the dynamic response coefficient correction of the exhaust gas recirculation valve based on exhaust pulses, the construction of the asymmetric emission dynamic response prediction model, and the multi-objective dynamic weight correction based on the operating condition memory effect. Based on the coordinated operation of these steps, during actual operation, the system performs real-time optimized control in each control cycle according to the following process:

[0177] In each control cycle (typically 10 to 50 milliseconds), the system acquires real-time operating parameters (including engine speed, load, exhaust pressure, etc.) and crankshaft angle signals. Based on the crankshaft angle signal, the peak intensity and occurrence time of the most recent exhaust pulse are identified. Combining the extracted pulse peak value and pulse time with the determined baseline pressure, the effective pulse intensity at the current moment is calculated. Based on this, the response lag time of the exhaust gas recirculation valve is corrected to obtain the pulse-corrected dynamic response lag time at the current moment.

[0178] The exhaust gas recirculation valve opening sequence at multiple recent moments is input into the memory factor calculation formula. An exponentially weighted average of the historical opening values ​​at each moment within the preceding time window is then applied to obtain the estimated residual exhaust gas volume in the intake manifold at the current moment (i.e., the memory factor). The memory factor is compared with the current target exhaust gas recirculation rate to calculate the residual exhaust gas deviation. Based on this deviation, the weights of the nitrogen oxide emission target, fuel economy target, and combustion stability target are adjusted to obtain the dynamic weights of each target after operating condition memory correction.

[0179] Based on the current direction of change in the exhaust gas recirculation rate (increasing or decreasing), the lag time and rate of change in the corresponding direction are selected from the obtained response parameters, and the selected parameters are scaled according to the magnitude of the change in the current exhaust gas recirculation rate. Using the scaled response parameters, the dynamic response trajectory of nitrogen oxide emissions is simulated for the opening degrees of multiple candidate exhaust gas recirculation valves, and the steady-state value and transient peak value of nitrogen oxides corresponding to each candidate opening degree are extracted.

[0180] Substitute the obtained pulse-corrected dynamic response lag time, the dynamic weights of each target, and the steady-state and transient peak values ​​of nitrogen oxides corresponding to each candidate opening degree into the comprehensive cost function. Using the candidate opening degree as the optimization variable, solve for the opening degree corresponding to the minimum value of the comprehensive cost function, which is taken as the optimal exhaust gas recirculation valve opening degree for the current control cycle.

[0181] The optimal exhaust gas recirculation valve opening is input into the feedforward compensation module. The specific feedforward compensation method is as follows: The interval between adjacent exhaust pulses is determined based on the pulse time of the current exhaust pulse and the current engine speed. The pulse time of the current exhaust pulse plus the interval between adjacent exhaust pulses is used as the predicted arrival time of the next exhaust pulse. The predicted arrival time minus the corrected response lag time is used as the advance amount. The current time plus this advance amount is used as the command issuance time. At this command issuance time, a control command corresponding to the optimal exhaust gas recirculation valve opening is issued, ensuring precise synchronization between the valve core movement and the arrival time of the exhaust pulse.

[0182] The current operating parameters, control commands, and execution results are recorded in the historical database. This recorded data can be used for parameter identification and online model updates in subsequent steps, enabling continuous optimization of the control model.

[0183] In summary, in this embodiment of the invention, by acquiring crankshaft angle and exhaust pressure to identify the peak value and timing of exhaust pulses in each cylinder, and determining the effective action intensity based on the integral of exhaust pressure exceeding the baseline pressure and the relative amplitude of the pulse peak to correct the response lag time of the exhaust gas recirculation valve, the valve core action is precisely synchronized with the exhaust pulse phase. By identifying the lag time and rate of change of nitrogen oxide emission response in both directions of increase and decrease of exhaust gas recirculation rate, and scaling the response parameters in real time according to the current direction and amplitude of change to predict the transient peak and steady-state values ​​of nitrogen oxides corresponding to each candidate opening, the asymmetric dynamic characteristics are accurately characterized. By calculating the memory factor through historical exhaust gas recirculation valve opening and dynamically adjusting the weights of each target according to its deviation from the target exhaust gas recirculation rate, the multi-objective optimization can reflect the memory effect of previous operating conditions. By predicting the arrival time of the next exhaust pulse and issuing control commands in advance by subtracting the corrected response lag time from that time, the transient response speed is further improved. Thus, the overall exhaust gas recirculation system significantly improves the control accuracy and adaptability of nitrogen oxide emissions under transient operating conditions.

[0184] This invention also proposes an intelligent optimization system for exhaust emissions from adjustable EGR valves; please refer to [link / reference]. Figure 2The diagram shows a structural diagram of an intelligent optimization system for adjustable EGR valve exhaust emissions provided by an embodiment of the present invention. The system includes: a data acquisition module 101, an impulse response evaluation module 102, an asymmetric emission prediction module 103, a memory correction optimization module 104, and a feedforward compensation control module 105.

[0185] The data acquisition module 101 is used to acquire crankshaft angle and exhaust pressure, determine the exhaust phase of each cylinder based on crankshaft angle and cylinder sequence, and extract the pulse peak value and pulse time of exhaust pressure within the exhaust phase of each cylinder.

[0186] The pulse response evaluation module 102 is used to determine the effective intensity of the exhaust pulse of each cylinder based on the integral of the exhaust pressure exceeding the base pressure in the exhaust phase of each cylinder and the relative amplitude of the pulse peak value relative to the base pressure, and to correct the response hysteresis time of the EGR valve according to the effective intensity.

[0187] The asymmetric emission prediction module 103 is used to acquire historical EGR rate change data, identify the lag time and rate of change of NOx emission response in both directions of EGR rate increase and decrease; calculate the current EGR rate change rate, select the lag time and rate of change in the corresponding direction according to its direction, scale the selected lag time and rate of change according to the current EGR rate change magnitude, and use the scaling result to predict the transient peak value and steady-state value of NOx corresponding to the opening degree of each candidate EGR valve.

[0188] The memory correction optimization module 104 is used to calculate the memory factor based on the historical EGR valve opening, and correct the weights of the three targets of NOx emission, fuel economy and combustion stability based on the deviation between the memory factor and the target EGR rate. Using the opening of each candidate EGR valve as the optimization variable, the corrected three weights are used to weight the steady-state value of NOx, fuel consumption rate and combustion fluctuation coefficient corresponding to each candidate EGR valve opening, and the transient peak value of NOx corresponding to each candidate EGR valve opening is used as the constraint to solve for the candidate EGR valve opening with the smallest weighted sum as the optimal EGR valve opening.

[0189] The feedforward compensation control module 105 is used to obtain the predicted arrival time of the next exhaust pulse, and use the predicted arrival time minus the corrected response lag time of the EGR valve as the advance amount, and issue the control command corresponding to the optimal EGR valve opening in advance according to the advance amount.

[0190] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the intelligent optimization system for adjustable EGR valve exhaust gas emissions provided in the above embodiments and the intelligent optimization method for adjustable EGR valve exhaust gas emissions belong to the same concept. The specific implementation process is detailed in the method embodiments and will not be repeated here.

[0191] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0192] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0193] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart optimization method for exhaust gas emissions from an adjustable EGR valve, characterized in that, include: Obtain crankshaft angle and exhaust pressure, determine the exhaust phase of each cylinder based on crankshaft angle and cylinder sequence, and extract the pulse peak value and pulse time of exhaust pressure within the exhaust phase of each cylinder; Based on the integral of the exhaust pressure exceeding the base pressure within the exhaust phase of each cylinder and the relative amplitude of the pulse peak relative to the base pressure, the effective intensity of the exhaust pulse of each cylinder is determined, and the response hysteresis time of the EGR valve is corrected according to the effective intensity. Acquire historical EGR rate change data, and identify the lag time and rate of change of NOx emission response in both directions of EGR rate increase and decrease; Calculate the current EGR rate change rate, select the corresponding lag time and change rate according to its direction, scale the selected lag time and change rate according to the current EGR rate change amplitude, and use the scaling result to predict the NOx transient peak value and steady-state value corresponding to the opening degree of each candidate EGR valve. The memory factor is calculated based on the historical EGR valve opening, and the weights of the three targets of NOx emissions, fuel economy and combustion stability are adjusted based on the deviation between the memory factor and the target EGR rate. Using the opening degree of each candidate EGR valve as the optimization variable, the NOx steady-state value, fuel consumption rate and combustion fluctuation coefficient corresponding to each candidate EGR valve opening degree are weighted by the three corrected weights, and the NOx transient peak value corresponding to each candidate EGR valve opening degree is constrained. The candidate EGR valve opening degree with the smallest weighted sum is then solved as the optimal EGR valve opening degree. The predicted arrival time of the next exhaust pulse is obtained, and the advance amount is obtained by subtracting the corrected response lag time of the EGR valve from the predicted arrival time. The control command corresponding to the optimal EGR valve opening is issued in advance according to the advance amount.

2. The intelligent optimization method for exhaust gas emissions of an adjustable EGR valve according to claim 1, characterized in that, The process of acquiring crankshaft angle and exhaust pressure, determining the exhaust phase of each cylinder based on crankshaft angle and cylinder sequence, and extracting the pulse peak value and pulse moment of exhaust pressure within each cylinder's exhaust phase specifically includes: Obtain the engine's crankshaft angle and exhaust pressure; Based on the engine cylinder working sequence, determine the crankshaft angle range corresponding to the exhaust stroke of each cylinder. Within the crankshaft angle range corresponding to the exhaust stroke of each cylinder, extract the local maximum pressure value from the exhaust pressure as the pulse peak value of that cylinder, and record the crankshaft angle time corresponding to the pulse peak value.

3. The intelligent optimization method for exhaust gas emissions of an adjustable EGR valve according to claim 1, characterized in that, The effective intensity of the exhaust pulse for each cylinder is determined based on the integral of the exhaust pressure exceeding the baseline pressure within each cylinder's exhaust phase and the relative amplitude of the pulse peak relative to the baseline pressure. The response hysteresis time of the EGR valve is then corrected according to the effective intensity. Specifically, this includes: Calculate the integral of the exhaust pressure exceeding the base pressure within each cylinder's exhaust phase, and calculate the relative amplitude of the pulse peak relative to the base pressure. The product of the integral and the relative amplitude is taken as the effective intensity of the exhaust pulse of that cylinder. Calculate the deviation rate between the effective action intensity and the preset average intensity. Take the negative value of the product of the correction intensity coefficient and the deviation rate as the exponent of the natural constant to calculate the correction coefficient. Multiply the response hysteresis time of the EGR valve by the correction coefficient to obtain the corrected response hysteresis time of the EGR valve.

4. The intelligent optimization method for exhaust gas emissions of an adjustable EGR valve according to claim 1, characterized in that, The acquisition of historical EGR rate change data, and the identification of the lag time and rate of change of NOx emission response in both directions of EGR rate increase and decrease, specifically includes: Acquire historical EGR rate change data, which is a continuous time series of EGR rate changes over time during historical operation. From historical EGR rate change data, events in which the EGR rate undergoes a step change are selected as EGR rate change events, and then divided into rising event group and falling event group according to the two directions of EGR rate increase and decrease. For each event in the rising event group, the time difference between the start of the change in EGR rate and the start of the change in NOx emission concentration is extracted as the rising lag time, and the average rate of change required for the NOx emission concentration to change from 10% to 90% of the steady-state value before the change is extracted as the rising rate of change. For each event in the declining event group, the time difference between the start of the change in EGR rate and the start of the change in NOx emission concentration is extracted as the lag time in the declining direction, and the average rate of change required for the NOx emission concentration to change from 10% to 90% of the steady-state value before the change is extracted as the rate of change in the declining direction.

5. The intelligent optimization method for exhaust gas emissions of an adjustable EGR valve according to claim 1, characterized in that, The calculation of the current EGR rate change rate, selecting the corresponding lag time and change rate according to its direction, scaling the selected lag time and change rate based on the current EGR rate change amplitude, and using the scaling result to predict the transient peak and steady-state NOx values ​​corresponding to the opening degree of each candidate EGR valve, specifically includes: Calculate the current rate of change of EGR rate, and determine the magnitude of the current EGR rate change based on the difference between the current EGR rate value and the EGR rate value at the moment when the EGR rate began to change; The current EGR rate of change is compared with preset positive and negative thresholds. If the current EGR rate of change is greater than the positive threshold, the direction of the EGR rate of change is determined to be upward, and the lag time and rate of change corresponding to the upward direction are selected. If the current EGR rate of change is less than the negative threshold, the direction of the EGR rate of change is determined to be downward, and the lag time and rate of change corresponding to the downward direction are selected. If the current EGR rate of change is between the negative and positive thresholds, the EGR rate of change is determined to be in a quasi-steady state. When it is determined that the direction of the EGR rate change is upward or downward, the selected lag time and rate of change are scaled according to the current EGR rate change magnitude. Obtain the opening degree of the candidate EGR valve; Using each candidate EGR valve opening as the target opening, and utilizing the scaled lag time and rate of change, the dynamic trajectory of NOx emission concentration during the transition from the current EGR valve opening to each candidate EGR valve opening is simulated. The maximum value is extracted from the dynamic trajectory corresponding to each candidate EGR valve opening as the NOx transient peak value corresponding to that candidate EGR valve opening, and the endpoint value is extracted as the NOx steady-state value corresponding to that candidate EGR valve opening.

6. The intelligent optimization method for exhaust gas emissions of an adjustable EGR valve according to claim 1, characterized in that, The calculation of the memory factor based on historical EGR valve opening specifically includes: Obtain the historical EGR valve opening at each moment within the preceding time window; An exponentially weighted average of the historical EGR valve openings at each time point within the preceding time window is used as the memory factor. The weight of each time point is negatively correlated with the time elapsed between that time point and the current time point.

7. The intelligent optimization method for exhaust gas emissions of an adjustable EGR valve according to claim 1, characterized in that, The method of adjusting the weights of the three targets—NOx emissions, fuel economy, and combustion stability—based on the deviation between the memory factor and the target EGR rate specifically includes: The difference between the memory factor and the target EGR rate is calculated as the residual exhaust gas deviation. The weight correction coefficients for NOx emission targets, fuel economy targets, and combustion stability targets are determined based on the ratio of residual exhaust gas deviation to target EGR rate. The weight correction coefficient for each target is an exponential function of the natural constant, and the independent variable of the exponential function is the product of the memory correction sensitivity coefficient of the target and the ratio. The base weights of each objective are multiplied by the corresponding weight correction coefficients to obtain the corrected weights of each objective. The three corrected weights are then normalized so that the sum of the three is 1.

8. The intelligent optimization method for exhaust gas emissions of an adjustable EGR valve according to claim 1, characterized in that, The process involves using the opening degree of each candidate EGR valve as the optimization variable, weighting the steady-state NOx value, fuel consumption rate, and combustion fluctuation coefficient corresponding to each candidate EGR valve opening with three corrected weights, and using the transient peak value of NOx corresponding to each candidate EGR valve opening as a constraint. The goal is to find the candidate EGR valve opening with the minimum weighted sum, which is then considered the optimal EGR valve opening. Specifically, this includes: The steady-state NOx value, fuel consumption rate, and combustion fluctuation coefficient corresponding to the opening degree of each candidate EGR valve were obtained respectively. The comprehensive cost of each candidate EGR valve opening is obtained by summing the products of the corrected NOx emission target weight and the NOx steady-state value, the corrected fuel economy target weight and the fuel consumption rate, and the corrected combustion stability target weight and the combustion fluctuation coefficient. The transient peak value of NOx corresponding to each candidate EGR valve opening is compared with the preset transient peak value. If the transient peak value of NOx corresponding to the candidate EGR valve opening is greater than the preset transient peak value, the candidate EGR valve opening is eliminated. Select the candidate EGR valve opening with the lowest overall cost from the remaining candidate EGR valve openings as the optimal EGR valve opening.

9. The intelligent optimization method for exhaust gas emissions of an adjustable EGR valve according to claim 1, characterized in that, The process of obtaining the predicted arrival time of the next exhaust pulse, subtracting the corrected response lag time of the EGR valve from the predicted arrival time as an advance amount, and issuing the control command corresponding to the optimal EGR valve opening in advance according to this advance amount, specifically includes: Based on the pulse time of the current exhaust pulse and the current engine speed, the interval time between adjacent exhaust pulses is determined. The pulse time of the current exhaust pulse plus the interval time between adjacent exhaust pulses is used as the predicted arrival time of the next exhaust pulse. The predicted arrival time of the next exhaust pulse is subtracted from the corrected response lag time of the EGR valve as the advance amount; Add a lead time to the current moment as the command issuance time, and issue the control command corresponding to the optimal EGR valve opening at the command issuance time.

10. An intelligent optimization system for exhaust emissions from an adjustable EGR valve, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by the processor, it implements the steps of the intelligent optimization method for exhaust gas emissions of an adjustable EGR valve as described in any one of claims 1-9.

Images