A knock control method for a methanol-diesel dual fuel engine

Abstract

The present application provides a kind of methanol-diesel dual fuel engine knock control method, it is related to engine control technical field, implementation multi-cylinder independent control, realize to each cylinder knock sensor signal individual collection and carry out filtering to each cylinder difference control, utilize fast fourier transform and time domain integral analysis method to extract the knock characteristic intensity of each cylinder;Determine the knock characteristic intensity is located in the response strategy threshold range of which level, execute the response strategy of matching level;Establish adaptive learning mechanism, optimize the threshold of each level response strategy.The present application has significantly improved knock detection accuracy, through adaptive learning mechanism, realize the optimal control boundary under different methanol mixing ratio, adopt differentiation cylinder control strategy, solve the inconsistency of each cylinder under dual fuel working condition, while ensuring the safety of engine, improve the torque output under full load condition, reduce the failure rate of knock related parts, prolong the service life of engine, reduce pollutant emission.

Description

Technical Field

[0001] This invention relates to the field of engine control technology, specifically to a method for controlling knock in a methanol-diesel dual-fuel engine. Background Technology

[0002] With the diversification of energy sources and increasingly stringent emission regulations, methanol is being widely used in diesel engines as a clean alternative fuel. However, methanol's high octane number and low cetane number characteristics result in significant differences in combustion characteristics compared to diesel fuel. In dual-fuel mode, this can easily lead to knocking, seriously affecting engine reliability and service life.

[0003] In existing technologies, traditional knock control strategies are mainly designed for pure gasoline or pure diesel engines. They typically employ simple methods such as ignition timing retarding or reducing boost pressure to address knock. However, these strategies suffer from several drawbacks in dual-fuel operation, including inability to accurately identify the unique knock spectrum characteristics of methanol-diesel mixtures, lack of adaptive control response for different fuel ratios, overly conservative knock handling strategies that sacrifice significant power performance and fuel economy, and failure to consider the differentiated control requirements between different cylinders. Furthermore, application number 202410996310.7, when used to control a methanol-diesel dual-fuel engine, also fails to consider the differentiated control between different cylinders, the extracted knock signals are inaccurate, and it lacks an adaptive learning process, resulting in poor final control performance.

[0004] Therefore, there is an urgent need for an intelligent knock control strategy that targets the characteristics of methanol-diesel dual-fuel mode, in order to improve engine reliability while maximizing performance output. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a knock control method for a methanol-diesel dual-fuel engine, which achieves knock control through high-precision knock detection and a multi-level adaptive control strategy.

[0006] To achieve the above technical objectives, the adopted technical solution is: a method for controlling knock in a methanol-diesel dual-fuel engine, comprising the following steps:

[0007] Step 1: Implement independent control of multiple cylinders to achieve differentiated control for each cylinder;

[0008] Step 2: Collect and filter the knock sensor signal for each cylinder separately, and extract the knock characteristic intensity of each cylinder using fast Fourier transform and time-domain integral analysis.

[0009] Step 3: Determine the response strategy threshold range within which the detonation characteristic intensity falls, and execute the matching level response strategy;

[0010] Step 4: Establish an adaptive learning mechanism and optimize the thresholds of response strategies at each level.

[0011] The method for extracting the knock characteristic intensity of each cylinder using fast Fourier transform and time-domain integral analysis is as follows: A comprehensive knock intensity assessment model is established:

[0012] K comp =W freq ×P(k)+W time ×V knock +K corr

[0013] Among them, K comp W represents the characteristic intensity of detonation. freq W represents the frequency domain feature weighting coefficients. time V represents the time-domain feature weighting coefficients, P(k) represents the target frequency energy calculated by the Fast Fourier Transform, and V represents the target frequency energy. knock K represents the time-domain detonation integral value calculated using the time-domain integral analysis method. corr This is a correction factor for the methanol blending ratio;

[0014] The weighting coefficient is dynamically adjusted based on the methanol blending ratio:

[0015]

[0016] W time =1―W freq

[0017] Correction factor calculation:

[0018]

[0019] Among them, K base R is the baseline correction factor, α is the methanol influence factor, with a value of 0.02-0.05. methanol This indicates the proportion of methanol blended in the mixture.

[0020] Step two also includes failure protection logic, which switches to the estimation module to calculate the detonation characteristic intensity state when the detonation sensor malfunctions.

[0021] The estimation module is:

[0022]

[0023] Where, k ion For the detonation characteristic intensity based on ion current, I max I is the peak value of the ion current. base y1 and y2 are the baseline values, t1 and t2 are the calibration coefficients, and t1 and t2 are the start and end times of integration.

[0024] The threshold ranges for different levels of response strategies are as follows:

[0025] If V knock_L1 <K comp <V knock_L2 Execute the Level 1 response strategy;

[0026] If V knock_L2 <K comp <V knock_L3 Implement a level-two response strategy;

[0027] If K comp >V knock_L3 Implement a three-level response strategy;

[0028]

[0029] Among them, K comp R represents the characteristic intensity of detonation. methanol V represents the methanol blending ratio. base_time This is the time-domain energy reference value.

[0030] The method for optimizing the thresholds of response strategies at each level is as follows:

[0031] V base_time (n)=V base_time (n―1)×(1―λ time )+V knock ×λ time

[0032] V base_time (n) is the time-domain energy reference value of the nth cycle, V base_time (n-1) is the time-domain energy reference value of the (n-1)th cycle, V knock λ is the time-domain detonation integral value calculated by the time-domain integral analysis method. time The time-domain learning rate is set, and the threshold automatically increases by 1% after 100 consecutive cycles without knocking.

[0033] The response strategy is divided into three levels from low to high:

[0034] Level 1 response strategy: Delay the current cylinder ignition angle by 0.5-2.0°CA and increase the fuel injection quantity;

[0035] Secondary response strategy: all cylinders are delayed by 0.5-2°CA to reduce boost pressure, intake valves are delayed to close, methanol injection is advanced by 5-10°CA to reduce methanol blending ratio by 3-8%;

[0036] Level 3 response strategy: Reduce combustion temperature, limit engine output torque to below 60%, forcibly reduce methanol blending ratio to a safe level, and activate emergency cooling strategy.

[0037] If the internal knocking phenomenon is not completely eliminated after N consecutive cycles, the boost control optimization strategy is activated. This strategy includes switching to a low boost MAP, reducing the upper limit of intake pressure, and reducing boost pressure.

[0038] When the knock rate is >10 times / minute, the following control method is activated:

[0039] (1) Fix the ignition angle to the optimal torque ignition angle minus 8 degrees;

[0040] (2) Activate cylinder pressure soft monitoring;

[0041] (3) The methanol blending ratio is automatically reduced to a safe range;

[0042] (4) Reduce the maximum speed and load.

[0043] It also includes preventive measures when the sensor is not faulty and the cylinder pressure rise rate is greater than the set maximum. The preventive measures include preset control strategies for various operating conditions. The preset control strategies for various operating conditions are updated by the control strategy results obtained in step 3.

[0044] The beneficial effects of this invention are:

[0045] This invention provides a knock control strategy for a methanol-diesel dual-fuel engine, comprising signal preprocessing, knock feature extraction, a graded response strategy, an adaptive learning mechanism, and multi-cylinder coordinated control, forming a complete closed-loop control system. The modules work closely together to achieve real-time monitoring, accurate identification, and intelligent control of knock phenomena in the methanol-diesel dual-fuel engine. This invention significantly improves knock detection accuracy, achieves optimal control boundaries under different methanol blending ratios through an adaptive learning mechanism, addresses cylinder inconsistencies under dual-fuel conditions using a differentiated cylinder control strategy, enhances torque output under full load while ensuring engine safety, reduces the failure rate of knock-related components, extends engine lifespan, and reduces pollutant emissions.

[0046] The detonation feature extraction module integrates frequency domain analysis and time domain integration methods. It employs the sliding window Goertzel algorithm to efficiently calculate energy values ​​in specific frequency bands (6-18kHz) and dynamically adjusts the frequency band weights to address the displacement characteristics of the detonation spectrum under different methanol blending ratios (the spectrum is generally higher by about 1.5-2kHz at high methanol blending ratios). This module overcomes the applicability limitations of traditional detonation detection algorithms in dual-fuel applications, significantly improving the accuracy of detonation identification, especially under conditions of high methanol blending ratios above 25%.

[0047] The graded response strategy module constructs a three-level progressive knock control mechanism: the first-level response addresses single minor knock incidents by implementing localized, gradual intervention, quickly eliminating knock through fine-tuning of ignition angles and optimization of fuel injection parameters; the second-level response addresses continuous knocking trends by adopting global control measures, including retarding ignition angles across all cylinders, adjusting boost pressure, and temporarily reducing the methanol blending ratio; the third-level response serves as a safety protection mechanism, implementing multi-dimensional coordinated control such as forced EGR activation, torque limiting, and emergency cooling in cases of severe knocking. A significant feature of this module is its differentiated control strategy based on the severity of knocking, maximizing engine performance output while ensuring safety.

[0048] The adaptive learning mechanism module improves the detonation prediction capability by dynamically updating the detonation energy benchmark value and detection threshold through real-time monitoring of historical detonation data.

[0049] The failure protection logic establishes a comprehensive sensor failure diagnosis mechanism, assessing sensor health status in real time by monitoring parameters such as signal DC offset, amplitude stability, and signal-to-noise ratio. In case of sensor malfunction, the multi-source information fusion knock estimation model is switched to ensure the reliability of the control system. Simultaneously, a safety backoff mode is designed for excessively high knock rates, ensuring safe engine operation under extreme conditions through measures such as fixing the ignition angle to MBT-8°CA and limiting the methanol blending ratio to below 25%. This module fills the gap in emergency handling mechanisms for dual-fuel engine knock control systems under abnormal operating conditions.

[0050] When the cylinder pressure rise rate exceeds the set maximum, preventive measures are directly adopted, the control response is rapid, and the database is enriched by updating the control strategy corresponding to the operating condition to achieve preventive control.

[0051] The aforementioned modules are tightly integrated to form a highly robust knock control system tailored to the characteristics of methanol-diesel dual-fuel engines. Compared to existing technologies, the knock control strategy proposed in this invention overcomes three key technical challenges: first, it solves the problem of accurately identifying the unique knock spectrum during methanol-diesel mixed combustion; second, it overcomes the challenge of adaptively adjusting knock control parameters under different methanol blending ratios; and third, it overcomes the differentiated control challenges caused by combustion inhomogeneity in multi-cylinder dual-fuel operation. Attached Figure Description

[0052] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0053] The present invention will now be described in detail with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0054] The knock control strategy of this invention is implemented based on a high-performance ECU processor, requiring a processor clock frequency of no less than 1MHz to ensure real-time control performance. The control system consists of a knock sensor, a signal conditioning circuit, an ECU processing unit, and actuators. The knock sensor is mounted on the cylinder block sidewall to collect cylinder block vibration signals; the signal conditioning circuit filters and amplifies the raw signal; the ECU processing unit performs feature extraction and control decisions; and the actuators include fuel injectors, an ignition system, a turbocharger control valve, and an EGR valve.

[0055] like Figure 1 As shown, a knock control method for a methanol-diesel dual-fuel engine implements multi-cylinder coordinated control, employing differentiated control strategies for cylinders prone to knocking. The control strategy for each cylinder includes knock detection and assessment, response strategies, sensor failure diagnosis, and preventative measures. When a knock sensor detects a knock, the control system simultaneously acquires information about the engine's current operating conditions and other control components.

[0056] I. Knock Detection and Assessment

[0057] The knock sensor signal is acquired by the knock sensor and then transmitted to the bandpass filter for filtering. The knock characteristic intensity of each cylinder is extracted using fast Fourier transform and time-domain integral analysis.

[0058] 1. Bandpass filter optimization

[0059] To address the characteristics of methanol-diesel dual-fuel operation, this invention employs an optimized bandpass filter design:

[0060] (1) Frequency range: 5-20kHz, with a focus on the 7-15kHz window.

[0061] (2) Filter type: Butterworth 4th order filter

[0062] (3) Attenuation characteristics: -48dB / octave

[0063] Filtering objective: To eliminate piston knocking noise (<4kHz) and electromagnetic interference noise (>25kHz). The filter transfer function expression is:

[0064]

[0065] Where ψ is the gain coefficient, ζ1 is the low-frequency damping ratio, ζ2 is the high-frequency damping ratio, ω1 and ω2 are the low-frequency cutoff angular frequencies and the high-frequency cutoff angular frequencies, respectively, and s is the Laplace variable.

[0066] Comparative experiments revealed that when the methanol blending ratio increased, the knock characteristic frequency was about 1.5-2 kHz higher than that of pure diesel. Therefore, this invention specifically enhanced the detection sensitivity in the 8-16 kHz frequency band.

[0067] 2. Angle detection window calculation

[0068] (1) Sampling precision: 12-bit ADC;

[0069] (2) Sampling frequency: 50kHz;

[0070] (3) Detection window: Synchronize the crankshaft position sensor signal to establish a dynamic detection window.

[0071] Formula for determining the angle detection window:

[0072]

[0073] Where, θ start θ is the starting angle (°CA) of the detonation detection window. end R is the end angle of the detonation detection window (°CA). methanol This indicates the percentage (%) of methanol blending. end -θ start The length of the angle detection window.

[0074] 2. Fast Fourier Transform

[0075] This invention achieves efficient frequency domain feature extraction through the following algorithm:

[0076] (1) FFT analysis: Perform fast Fourier transform in each cycle to identify energy surges in the 6-18kHz frequency band;

[0077] (2) Optimization algorithm: The sliding window Goertzel algorithm is used to calculate the energy value for the target frequency band (12±2kHz);

[0078] The expression for calculating the Discrete Fourier Transform using the Goertzel algorithm:

[0079]

[0080] X(k) is the complex transformation result of the target frequency point k, V signal (θ) represents the filtered knock sensor signal, N represents the length of the angle detection window (number of sampling points), k represents the target frequency point, j represents an imaginary number, and n represents the sampling points.

[0081] Actual algorithm optimization implementation:

[0082] v1(0)=0

[0083] v2(0)=0

[0084] For n = 0 to N-1

[0085]

[0086] v2(n+1) = v1(n)

[0087] v1(n+1)=v0

[0088] End

[0089]

[0090] Where v0, v1, and v2 are intermediate variables of the algorithm, N is the length of the angle detection window, and P(k) is the energy at frequency k.

[0091] 2. Time-domain integration technique

[0092] Integrate the absolute value of the filtered knock sensor signal:

[0093]

[0094] Among them, V knock V is the time-domain detonation integral value, representing the detonation intensity. θ1 and θ2 represent the start and end angles (°CA) of the angle detection window, respectively. signal (θ) represents the filtered knock sensor signal.

[0095] 3. Extraction of detonation characteristic intensity by fusing dual features

[0096] This invention uses the target frequency energy P(k) calculated by Fast Fourier Transform and the time-domain detonation integral value V calculated by the time-domain integral analysis method. knock A fusion judgment strategy is used to establish a comprehensive knock intensity assessment model:

[0097] K comp =W freq ×P(k)+W time ×V knock +K corr

[0098] Among them, K comp W represents the characteristic intensity of detonation. freq W represents the frequency domain feature weighting coefficients. time V represents the time-domain feature weighting coefficients, P(k) represents the target frequency energy calculated by the Fast Fourier Transform, and V represents the target frequency energy. knock K represents the time-domain detonation integral value calculated using the time-domain integral analysis method. corr This is the correction factor for the methanol blending ratio.

[0099] The weighting coefficient is dynamically adjusted based on the methanol blending ratio:

[0100]

[0101] W time =1―W freq

[0102] When the methanol blending ratio is high, the frequency domain feature weight increases; in pure diesel mode, the time domain feature weight increases relatively.

[0103] Correction factor calculation:

[0104]

[0105] Among them, K base is the baseline correction factor (value 1.0), and 'a' is the methanol influence factor (value 0.02-0.05).

[0106] This invention employs a feature extraction method that combines time-domain integration and frequency-domain analysis, thereby improving the reliability of detonation identification.

[0107] II. Tiered Response Strategy

[0108] 1. Establish a V-based system knock The three-level response threshold system:

[0109] Level 1 response threshold:

[0110]

[0111] Secondary response threshold:

[0112]

[0113] Level 3 response threshold:

[0114]

[0115] Among them, V base_time Using the time-domain energy baseline, the boundaries of the three levels of response strategies are updated and optimized through adaptive learning:

[0116] V base_time (n)=V base_time (n―1)×(1―λ time )+V knock ×λ time

[0117] V base_time (n) is the time-domain energy reference value of the nth cycle, V base_time (n-1) is the time-domain energy reference value for the (n-1)th cycle, V knockλ is the time-domain detonation integral value calculated by the time-domain integral analysis method. time The temporal learning rate (ranging from 0.005 to 0.02) needs to be defined, and the temporal energy baseline value for the first cycle needs to be defined. After 100 consecutive cycles without detonation, the threshold automatically increases by 1%.

[0118] The knock response level is determined according to the following rules:

[0119] If V knock_L1 <K comp <V knock_L2 Execute the first-level response strategy, V knock_L1 and V knock_L2 This defines the boundary of the Level 1 response strategy.

[0120] If V knock_L2 <K comp <V knock_L3 Execute a two-level response strategy, V knock_L2 and V knock_L3 This defines the boundary of the secondary response strategy.

[0121] If K comp >V knock_L3 Implement a three-level response strategy, V knock_L3 This defines the boundaries of the three-level response strategy.

[0122] 2. This invention employs a three-level response strategy, implementing different levels of intervention measures based on the severity of the detonation, such as... Figure 1 As shown:

[0123] 2.1. Level 1 Response (Single Detection of Detonation)

[0124] (1) Ignition angle control: The ignition angle of the current cylinder is delayed by 0.5-2.0°CA, and the ignition of the current cylinder is delayed by 0.5°CA / cycle step (maximum cumulative delay of 8°CA).

[0125] (2) Fuel injection control: Increase fuel injection by 2-5% to reduce combustion temperature.

[0126] Ignition angle adjustment formula:

[0127] S a (n)=S a (n-1)-min[0.5·K i (n),2]

[0128] Among them, S a (n) represents the current cycle ignition advance angle (°CA), S a (n-1) is the ignition advance angle (°CA) of the previous cycle, K i is the current cycle detonation intensity, and n is the current cycle.

[0129] The characteristics of a Level 1 response are local, gradual, and recoverable, with the aim of suppressing minor knocking with minimal impact on performance.

[0130] 2.2. Level 2 response (three consecutive detections of detonation)

[0131] (1) Global ignition angle delay: all cylinders are delayed by 0.5-2°CA, reducing the methanol blending ratio by 3-8%.

[0132] (2) Pressure control: Reduce the pressure by 5 kPa through the exhaust valve.

[0133] (3) Intake and exhaust phase adjustment: The intake valve closes 5°CA later to reduce the effective compression ratio.

[0134] (4) Advance methanol injection time: Advance methanol injection by 5-10°CA to enhance the uniformity of the mixture.

[0135] The secondary response acts on global parameters, controlling the detonation trend through comprehensive measures. Simultaneously, the system records detonation counts and triggers OBD diagnostic code storage.

[0136] 2.3. Level 3 Response (Safety Protection)

[0137] Activated when the knock intensity exceeds the material's safety threshold (e.g., >4.5V peak):

[0138] (1) Force open the EGR valve to the maximum position to reduce the combustion temperature.

[0139] (2) The engine output torque is limited to below 60%.

[0140] (3) The methanol blending ratio is forcibly reduced to a safe level (usually less than 25%).

[0141] (4) Activate emergency cooling strategy (increase water pump speed and increase radiator fan speed).

[0142] Torque limit calculation:

[0143] T limit =T max -K t K i_max

[0144] Among them, T limit To limit the rear torque (Nm), T max For maximum torque (Nm), K i_max For the maximum detonation intensity, K t This is the torque adjustment coefficient.

[0145] The innovation of the three-level response lies in expanding the traditional single-parameter adjustment into multi-dimensional collaborative control, especially for the special operating conditions under dual fuels, by designing a differentiated methanol blending ratio control strategy.

[0146] 2.4 Boost Control Optimization Strategy

[0147] If the implosion phenomenon is not completely eliminated after N consecutive cycles, the boost control optimization strategy is activated. Here, N is set to 5.

[0148] Boost control optimization: Switch to low boost MAP to reduce peak combustion temperature by limiting the upper limit of intake pressure, while avoiding excessive boost that could exacerbate knocking.

[0149] Boost pressure adjustment calculation:

[0150] p(n) = p(n-1) - K p K i_sum

[0151] Where p(n) is the boost pressure of the nth cycle (kPa), p(n-1) is the boost pressure of the (n-1)th cycle (kPa), and K p K is the pressure adjustment proportional coefficient. i_sum It is the sum of the intensities of three consecutive detonations.

[0152] 3. Safety rollback mode

[0153] The following strategy is activated when the knock rate is >10 times / minute:

[0154] (1) Fix the ignition angle to MBT-8°CA (optimal torque ignition angle minus 8 degrees);

[0155] (2) Activate cylinder pressure soft monitoring (via ion current detection);

[0156] (3) The methanol blending ratio is automatically reduced to a safe range (usually ≤25%);

[0157] (4) Limit the maximum speed and load.

[0158] III. Sensor Failure Diagnosis

[0159] This diagnostic is performed before knock detection and assessment, which assumes the sensor is in good working order when knock detection and assessment are performed.

[0160] 1. Sensor Failure Diagnosis

[0161] (1) Monitor the DC offset of the signal (normally it should be <0.5V). If the offset is continuously >2V, it is determined that the sensor is abnormal.

[0162] (2) If the signal amplitude is consistently below the threshold (<0.1V) and there is no fluctuation, it is determined that the sensor is open-circuited;

[0163] (3) If the signal noise is too high (signal-to-noise ratio < 5dB), it is determined that the sensor has poor contact.

[0164] Signal-to-noise ratio calculation:

[0165]

[0166] Where SNR is the signal-to-noise ratio (dB), S rms N represents the effective value of the signal (V). rms The effective value of the noise is (V).

[0167] The failure protection logic of this invention specifically considers the extreme case of complete sensor system failure. Through a knock prediction algorithm based on a physical model, it ensures that the engine can still operate within the safety boundary even in the event of sensor failure.

[0168] 2. When the detonation sensor malfunctions, the system automatically switches to model estimation mode and uses the ion current detonation estimation model to calculate the detonation intensity.

[0169] Ion current detonation estimation model:

[0170]

[0171] Where, k ion I is the detonation intensity index based on ion current. max The peak ion current (μA), I base y1 and y2 are the baseline value (μA), y1 and y2 are the calibration coefficients, and t1 and t2 are the integration start and end times (s).

[0172] IV. Preventive Measures

[0173] The conditions under which this preventive measure can be implemented include two: (1) the sensor is in normal condition and (2) the maximum pressure rise rate of the cylinder is greater than the set maximum pressure rise rate. In this embodiment, the set maximum pressure rise rate is 5 bar / °CA, and the maximum pressure rise rate is derived from the real-time combustion analysis of the in-cylinder pressure sensor.

[0174] By introducing real-time combustion analysis technology based on in-cylinder pressure sensors, knock prevention control is achieved by monitoring the maximum pressure rise rate of each cylinder.

[0175] (1) Real-time calculation of the maximum pressure rise rate of each cylinder:

[0176]

[0177] Where dp / dθ max(i) represents the maximum pressure rise rate of the i-th cylinder (bar / °CA), p(θ,i) represents the cylinder pressure (bar) of the i-th cylinder at crankshaft angle θ, and Δθ represents the sampling interval (°CA).

[0178] A preset five-dimensional lookup table (speed × load × intake air temperature × methanol blending ratio × coolant temperature) is used. This table is a collection of various operating conditions, specifying the preset control strategies for different conditions. When preventative measures are implemented, knock detection and evaluation are no longer performed; instead, the five-dimensional lookup table is updated based on the operating conditions corresponding to each control strategy implemented during knock detection and evaluation.

Claims

1. A method for controlling knock in a methanol-diesel dual-fuel engine, characterized in that, Includes the following steps: Step 1: Implement independent control of multiple cylinders to achieve differentiated control for each cylinder; Step 2: Collect and filter the knock sensor signal for each cylinder separately, and extract the knock characteristic intensity of each cylinder using fast Fourier transform and time-domain integral analysis. Step 3: Determine the response strategy threshold range within which the detonation characteristic intensity falls, and execute the matching level response strategy; Step 4: Establish an adaptive learning mechanism and optimize the thresholds of response strategies at each level; The method for extracting the knock characteristic intensity of each cylinder using fast Fourier transform and time-domain integral analysis is as follows: Establish a comprehensive knock shock intensity assessment model: in, K comp The characteristic intensity of detonation. W freq These are the frequency domain feature weighting coefficients. W time These are the time-domain feature weighting coefficients. P ( k The target frequency energy is calculated using the Fast Fourier Transform. V knock The time-domain detonation integral value is calculated using the time-domain integral analysis method. K corr This is a correction factor for the methanol blending ratio; The weighting coefficient is dynamically adjusted based on the methanol blending ratio: Correction factor calculation: in, K base As the benchmark correction factor, a The methanol influence coefficient ranges from 0.02 to 0.

05. R methanol Indicates the methanol blending ratio; The threshold ranges for different levels of response strategies are as follows: like V knock_L1 < K comp < V knock_L2 Execute the Level 1 response strategy; like V knock_L2 < K comp < V knock_L3 Implement a level-two response strategy; like K comp > V knock_L3 Implement a three-level response strategy; in, K comp Indicates the characteristic intensity of detonation. R methanol Indicates the methanol blending ratio. This serves as the time-domain energy reference value. The method for optimizing the thresholds of response strategies at each level is as follows: V base_time ( n ) is the first n Cyclic time-domain energy reference value, V base_time ( n -1) is the first n -1 Cyclic time-domain energy reference value, V knock The time-domain detonation integral value is calculated using the time-domain integral analysis method. λ time The time-domain learning rate is set such that the threshold automatically increases by 1% after 100 consecutive cycles without knock-up. The response strategy is divided into three levels from low to high: Level 1 response strategy: Delay the current cylinder ignition angle by 0.5-2.0°CA and increase the fuel injection quantity; Secondary response strategy: all cylinders are delayed by 0.5-2°CA to reduce boost pressure, intake valves are delayed in closing, methanol injection is advanced by 5-10°CA to reduce methanol blending ratio by 3-8%; Three-level response strategy: reduce combustion temperature, limit engine output torque to below 60%, forcibly reduce methanol blending ratio to a safe level, and activate emergency cooling strategy.

2. The method for controlling knock in a methanol-diesel dual-fuel engine as described in claim 1, characterized in that: Step 2 also includes failure protection logic, which switches to the estimation module to calculate the detonation characteristic intensity state when the detonation sensor malfunctions.

3. The method for controlling knock in a methanol-diesel dual-fuel engine as described in claim 2, characterized in that: The estimation module is: in, k ion The detonation characteristic intensity is based on ion current. I max This represents the peak value of the ion current. I base Baseline value, y 1 and y 2 is the calibration coefficient. t 1 and t 2 represents the start and end times of the integration.

4. The method for controlling knock in a methanol-diesel dual-fuel engine as described in claim 1, characterized in that: If the internal knocking phenomenon is not completely eliminated after N consecutive cycles, the boost control optimization strategy is activated. This strategy includes switching to a low boost MAP, reducing the upper limit of intake pressure, and reducing boost pressure.

5. The method for controlling knock in a methanol-diesel dual-fuel engine as described in claim 1, characterized in that: When the knock rate is >10 times / minute, the following control method is activated: (1) Fix the ignition angle to the optimal torque ignition angle minus 8 degrees; (2) Activate cylinder pressure soft monitoring; (3) The methanol blending ratio automatically decreases to a safe range; (4) Reduce the maximum speed and load.

6. The method for controlling knock in a methanol-diesel dual-fuel engine as described in claim 1, characterized in that: It also includes preventive measures when the sensor is not faulty and the cylinder pressure rise rate is greater than the set maximum. The preventive measures include preset control strategies for various operating conditions. The preset control strategies for various operating conditions are updated by the control strategy results obtained in step 3.

Images