Fuel injection control method for diesel generator sets based on adaptive control

Abstract

The present invention relates to the field of diesel generator control technology, and more specifically to a fuel injection control method for a diesel generator set based on adaptive control. The method comprises the following steps: collecting operating data of the diesel generator set; establishing a compression fluctuation model based on the collected compression pressure fluctuation data, wherein the compression fluctuation model identifies in real time the fluctuation amplitude characteristics, frequency characteristics, and peak timing characteristics during the compression process; preliminarily calculating the injection timing and generating a preliminary injection timing control signal; and adjusting the preliminarily calculated injection timing in real time to match the injection timing with the optimal phase of the compression fluctuation, and to optimize the matching between the injection timing and the peak timing characteristics. The present invention effectively improves fuel injection control under different speed and load conditions, thereby increasing the fuel efficiency of the diesel engine, reducing emissions, and extending the engine life.

Description

Technical Field

[0001] The present invention relates to the technical field of diesel generator control, and in particular to a diesel generator set fuel injection control method based on adaptive control. Background Art

[0002] Diesel generator sets are widely used in industrial, commercial and emergency power supply fields. One of their core components is the diesel engine. The performance of the diesel engine directly affects the efficiency, fuel economy and emission level of the generator set. In diesel engines, fuel injection timing control plays a vital role in ensuring efficient combustion of fuel. Precise control of injection timing can not only improve the fuel atomization effect, but also maximize the combustion efficiency in the compression stage. However, traditional fuel injection control systems usually rely on fixed injection models. These models are difficult to adapt to changes in the engine under different operating conditions in real time, especially under high load, low load and rapid operating conditions. The fixed setting of injection timing often leads to incomplete fuel combustion or reduced combustion efficiency, which in turn increases fuel consumption and harmful gas emissions.

[0003] Traditional methods of controlling injection timing rely heavily on basic engine parameters such as speed, piston position, and intake pressure. However, these parameters cannot fully reflect the dynamic characteristics of the compression process within the engine, particularly the amplitude, frequency, and peak characteristics of compression fluctuations. This results in lags and deficiencies in injection timing control during actual operation, making it impossible to accurately respond to changes in compression fluctuations. This leads to a mismatch between fuel injection and the engine's compression state, impacting combustion efficiency. Summary of the Invention

[0004] The present invention provides a fuel injection control method for a diesel generator set based on adaptive control.

[0005] The fuel injection control method of a diesel generator set based on adaptive control includes the following steps:

[0006] S1, operating parameter collection: collects the operating data of the diesel generator set, including engine speed, compression pressure fluctuation, piston movement position and intake pressure;

[0007] S2, Compression Fluctuation Characteristic Analysis and Modeling: Based on the collected compression pressure fluctuation data, a compression fluctuation model is established using a deep learning algorithm based on nonlinear dynamic characteristics. The compression fluctuation model identifies the fluctuation amplitude, frequency, and peak timing characteristics during the compression process in real time. It accurately predicts the potential impact of compression fluctuation on fuel injection timing, especially under high load, low temperature start-up, or engine wear conditions.

[0008] S3, preliminary calculation of injection timing: Based on the traditional fuel injection model, combined with the engine speed, piston position, and intake pressure under the current working conditions, the injection timing is preliminarily calculated and a preliminary injection timing control signal is generated;

[0009] S4, Adaptive Adjustment of Compression Fluctuation: This method comprehensively analyzes the amplitude and frequency characteristics of the compression fluctuation model and the preliminary injection timing calculation results, and adjusts the preliminary calculated injection timing in real time to match the injection timing with the optimal stage of the compression fluctuation. This prevents premature or late fuel injection caused by compression fluctuations. By adjusting the injection start time, injection duration, and injection angle, the injection is ensured to be completed within the optimal time window of the compression process.

[0010] S5, optimization of the matching between injection timing and peak timing characteristics: Based on the real-time monitoring of compression pressure fluctuations, the control system of the diesel generator set aligns the injection timing with the phase of the peak timing characteristics, optimizes the injection time so that the fuel injection timing occurs at the ideal position before the arrival of the peak, so that the fuel is fully atomized in a high-pressure environment, improving combustion efficiency, while reducing pressure shock and reducing wear on engine components.

[0011] Optionally, the operating condition parameter collection in S1 specifically includes:

[0012] S11, real-time acquisition of engine speed: obtain engine speed data from the diesel generator control unit. The unit control unit monitors the crankshaft or flywheel speed in real time through the built-in Hall sensor for calculation of injection timing;

[0013] S12, real-time collection of compression pressure fluctuations: The in-cylinder pressure sensor detects the in-cylinder pressure signal of the diesel generator set, transmits the collected pressure signal to the data processing module, and forms compression pressure fluctuation data P after filtering and amplification. filtered (t), represents the compression pressure fluctuation signal at time t;

[0014] S13, obtaining the piston movement position: obtaining the piston movement position data from the diesel generator set control unit;

[0015] S14, real-time acquisition of intake pressure: obtaining intake pressure data from the diesel generator control unit, and the intake pressure sensor in the intake manifold monitors the intake pressure in real time.

[0016] Optionally, the compression fluctuation characteristic analysis and modeling in S2 specifically include:

[0017] S21, data preprocessing: compress the pressure fluctuation data P filtered (t) Perform standardization processing to obtain the standardized compression pressure fluctuation data P norm (t);

[0018] S22, feature extraction: inputting the normalized compression pressure fluctuation data into a deep learning algorithm, wherein the deep learning algorithm performs feature extraction based on a combination of a convolutional neural network (CNN) and a recurrent neural network (RNN), thereby constructing a compression fluctuation model;

[0019] S23, real-time identification of fluctuation characteristics: The real-time input compression pressure fluctuation data is processed using the trained compression fluctuation model. The model output includes fluctuation amplitude, frequency, and peak time characteristics, as follows:

[0020] Fluctuation amplitude characteristics: The local extreme value difference calculated by the compressed fluctuation model reflects the size of the fluctuation amplitude. The fluctuation amplitude characteristics A wave (t) is calculated as:

[0021] A wave (t) = max(P norm (t))-min(P norm (t));

[0022] Frequency characteristics: The main frequency components are extracted from the fluctuation data by fast Fourier transform, and the frequency characteristics f wave It is obtained by the following formula:

[0023] Indicates reaches the maximum value, where represents Fourier transform;

[0024] Peak time feature: Based on the output of the recurrent neural network, the peak time feature t is extracted peak , the peak moment characteristic is expressed as: t peak =arg max(h t ), h t is the hidden state of the recurrent neural network at time step t.

[0025] Optionally, the feature extraction in S22 specifically includes:

[0026] S221, extracting local features using convolutional layers: extracting local features from compression pressure fluctuations through multiple convolutional layers, including the variation pattern of the fluctuation amplitude and the peak time characteristics;

[0027] S222, Recursive layer captures time dependence: Capturing the time dependence and nonlinear dynamic characteristics of compression pressure fluctuation data through recursive neural network.

[0028] Optionally, in S3, the conventional fuel injection model is combined with the engine speed ω and the piston position x piston And the intake pressure P intake, preliminarily calculate the injection time t inj , expressed as:

[0029] t inj =f(ω,x piston ,P intake ), where f(·) represents the traditional fuel injection model function. Based on the relationship between engine speed, piston position and intake pressure, the time point of fuel injection is calculated. The basic time interval of the injection moment is determined according to the currently collected engine speed ω, and the piston position x is combined with the time interval of the injection moment. piston By calculating the relationship between the piston position and the crankshaft angle, the injection timing is corrected and the real-time collected intake pressure P is used. intake , adjust the injection timing and injection duration to adapt to changes in intake pressure and ensure combustion efficiency.

[0030] Optionally, the S4 specifically includes:

[0031] S41, Fluctuation Amplitude Characteristics A wave (t) is too large (i.e. A wave (t) is close to its local or global maximum), delaying the injection timing to ensure that the injection occurs during the high-pressure period, thereby improving the atomization and combustion efficiency of the fuel. The fluctuation amplitude characteristic A wave When (t) is too low, the injection timing is advanced to avoid the injection period overlapping with the low-pressure period, thereby avoiding a decrease in combustion efficiency.

[0032] S42, frequency characteristic f wave When it is too high, the compression fluctuation period is short and the injection time is advanced to ensure that the fuel has enough time to complete injection and combustion within the short compression period; the frequency characteristic f wave When it is too low, the compression cycle becomes longer and the injection time is delayed. According to the compression fluctuation period T wave , calculate the ideal injection timing and adjust the current injection timing to make it in the optimal time period within the compression cycle.

[0033] Optionally, in S41, the actual fluctuation amplitude A wave (t) and the maximum expected fluctuation range A max The relationship between the injection timing is defined as follows:

[0034] Where Δt A is the injection timing offset adjusted based on the fluctuation amplitude, k A is the influence coefficient of the fluctuation amplitude, A max is the maximum expected value or standard value of the fluctuation range, A wave (t) is the real-time fluctuation amplitude, and the injection timing t is adjusted by the fluctuation amplitude characteristics. injdj1 Expressed as:

[0035] t injadj1 =t inj +Δt A , when A wave When (t) is high, the injection timing is appropriately delayed, otherwise, the injection timing is advanced.

[0036] Optionally, the step S42 specifically includes: calculating the current injection time t inj adj1 The position in the current compression cycle is compared with the ideal injection time to obtain the adjustment value Δt f :

[0037] Δt f =t ideal -(t inj adj1 mod T wave ), where t ideal is the ideal injection timing calculated based on the compression cycle, t injadj1 mod T wave is the position of the current injection time in the cycle, mod represents the modulo operation, and the injection time t is adjusted by the frequency characteristic inj adj2 Expressed as:

[0038] t inj adj2 =t inj adj1 +Δt f , when t inj adj1 When the ideal injection timing of the compression cycle is deviated from, it is automatically advanced or retarded to ensure that the injection timing is aligned with the optimal phase of the current compression cycle.

[0039] Optionally, in S5, according to the peak time feature t peak Adjust the injection timing: The ideal injection timing is before the peak moment. Set the injection timing advance amount. According to the peak moment of the compression wave model, calculate the difference between the current injection timing and the peak moment. If the current injection timing is too early or too late, make dynamic adjustments based on the difference to ensure that the injection timing is close to the ideal moment before the peak; if the current injection timing is later than the ideal time before the peak, advance the injection timing; if the current injection timing is earlier than the ideal time before the peak, postpone the injection to ensure that the injection process is aligned with the peak phase.

[0040] Beneficial effects of the present invention:

[0041] The present invention, by introducing a compression fluctuation model, can dynamically identify the fluctuation amplitude, frequency and peak characteristics during the compression process, make real-time adjustments to the injection timing based on actual operating conditions, and optimize the overall range based on the fluctuation amplitude and frequency characteristics, so that the fuel injection timing can accurately match the optimal stage of the compression fluctuation. Especially under high-load or rapidly changing operating conditions, the injection timing can be delayed or advanced according to the real-time fluctuation amplitude, ensuring that injection occurs during a period of higher pressure, improving the fuel atomization effect, and maximizing combustion efficiency. This adjustment method can adaptively respond to changes in different engine loads and speeds, effectively reducing incomplete combustion, and improving fuel economy.

[0042] The present invention introduces frequency characteristics that can be finely adjusted according to the current compression cycle. When the frequency is high, the injection timing is automatically advanced to adapt to the shorter compression cycle, avoiding the problem of incomplete combustion caused by late fuel injection. When the frequency is low, the injection timing will be appropriately delayed to utilize the longer compression cycle for sufficient combustion preparation. This adjustment mechanism ensures the dynamic synchronization of the injection timing and the compression cycle, effectively improves the fuel injection control under different speed and load conditions, improves the fuel utilization efficiency of the diesel engine, reduces emissions and extends the engine service life.

[0043] The present invention precisely identifies the peak moment characteristics in the compression fluctuation and finely adjusts the injection timing to ensure that fuel injection occurs at the ideal position before the peak arrives. The dynamic optimization mechanism of the injection timing enables the fuel to be injected and atomized when the compression pressure reaches the maximum value, thereby maximizing the fuel combustion efficiency and significantly improving the dynamic response capability of the engine, so that it can maintain efficient and stable combustion performance under various operating conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 Schematic diagram of a method flow in an embodiment of the present invention;

[0046] Figure 2 Schematic diagram of compression fluctuation characteristic analysis and modeling according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It is also noted that, to provide a more detailed description, the following embodiments are best and preferred embodiments, and those skilled in the art may employ alternative methods for implementing certain known technologies. Furthermore, the accompanying drawings are intended only to provide a more detailed description of the embodiments and are not intended to limit the present invention.

[0048] It should be noted that references in the specification to "one embodiment," "an embodiment," "an exemplary embodiment," "some embodiments," etc. indicate that the described embodiments may include specific features, structures, or characteristics, but not every embodiment necessarily includes such specific features, structures, or characteristics. In addition, when specific features, structures, or characteristics are described in conjunction with an embodiment, it is within the knowledge of persons skilled in the relevant art to implement such features, structures, or characteristics in conjunction with other embodiments (whether or not explicitly described).

[0049] In general, terms can be understood, at least in part, from their use in context. For example, depending at least in part on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending at least in part on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0050] like Figure 1-Figure 2 As shown, the fuel injection control method of a diesel generator set based on adaptive control includes the following steps:

[0051] S1, operating parameter collection: collects the operating data of the diesel generator set, including engine speed, compression pressure fluctuation, piston movement position and intake pressure;

[0052] S2, Compression Fluctuation Characteristic Analysis and Modeling: Based on the collected compression pressure fluctuation data, a compression fluctuation model is established using a deep learning algorithm based on nonlinear dynamic characteristics. The compression fluctuation model identifies the fluctuation amplitude, frequency, and peak timing characteristics during the compression process in real time. It accurately predicts the potential impact of compression fluctuation on fuel injection timing, especially under high load, low temperature start-up, or engine wear conditions.

[0053] S3, preliminary calculation of injection timing: Based on the traditional fuel injection model, combined with the engine speed, piston position, and intake pressure under the current working conditions, the injection timing is preliminarily calculated and a preliminary injection timing control signal is generated;

[0054] S4, Adaptive Adjustment of Compression Fluctuation: This method comprehensively analyzes the amplitude and frequency characteristics of the compression fluctuation model and the preliminary injection timing calculation results, and adjusts the preliminary calculated injection timing in real time to match the injection timing with the optimal stage of the compression fluctuation. This prevents premature or late fuel injection caused by compression fluctuations. By adjusting the injection start time, injection duration, and injection angle, the injection is ensured to be completed within the optimal time window of the compression process.

[0055] S5, optimization of the matching between injection timing and peak timing characteristics: Based on the real-time monitoring of compression pressure fluctuations, the control system of the diesel generator set aligns the injection timing with the phase of the peak timing characteristics, optimizes the injection time so that the fuel injection timing occurs at the ideal position before the arrival of the peak, so that the fuel is fully atomized in a high-pressure environment, improving combustion efficiency, while reducing pressure shock and reducing wear on engine components.

[0056] The main purpose of S4 is to make macro adjustments to the preliminarily calculated injection timing based on the fluctuation amplitude and frequency characteristics of the compression fluctuation model, so that the injection timing matches the approximate dynamics of the compression process and ensures that injection does not occur at a time point that is obviously too early or too late. For example, S4 determines the overall injection window by analyzing the fluctuation amplitude and determines the range within which injection should occur based on the fluctuation frequency. If the compression fluctuation is large, the injection timing may need to be postponed; if the fluctuation frequency is high, the injection duration may need to be shortened. S5 is further optimized. Based on the global adjustment of S4, it focuses on aligning the injection timing with the specific phase of the peak moment characteristics. The focus of this step is to fine-tune the injection timing to ensure that it is optimally matched with the compression peak (that is, the moment of highest pressure in the compression process) and to ensure that injection occurs at the ideal position before the peak is about to arrive.

[0057] The working condition parameter collection in S1 specifically includes:

[0058] S11, real-time acquisition of engine speed: obtain engine speed data from the diesel generator control unit. The unit control unit monitors the crankshaft or flywheel speed in real time through the built-in Hall sensor for calculation of injection timing;

[0059] S12, real-time collection of compression pressure fluctuations: The in-cylinder pressure sensor detects the in-cylinder pressure signal of the diesel generator set, transmits the collected pressure signal to the data processing module, and forms compression pressure fluctuation data P after filtering and amplification. filtered (t), represents the compression pressure fluctuation signal at time t;

[0060] Signal filtering: In order to remove high-frequency noise in the signal, a low-pass filter is used to process the signal. The transfer function H(f) of the low-pass filter is expressed as:

[0061] Among them, f c is the cutoff frequency of the filter, f is the signal frequency, j is the imaginary unit, and the filtered signal V filtered (t) is expressed as:

[0062] in represents the Fourier transform, Represents the inverse Fourier transform. This filter is used to remove high-frequency noise signals and retain only the low-frequency effective compressed fluctuation signal. Through the transfer function of the low-pass filter, the useless noise components in the voltage signal output by the sensor can be filtered out to ensure that the signal transmitted to the control system is clean and noise-free.

[0063] Signal amplification: The filtered signal is amplified by an amplifier with a gain of G. The amplified signal V amp (t) is expressed as:

[0064] V amp (t) = G·V filtered (t), G is the gain constant of the signal amplifier, which is usually greater than 1. Through the amplification process, even small changes in the compression pressure can be captured, and it is ensured that the control system can respond to small changes.

[0065] S13, obtaining the piston movement position: Obtaining the piston movement position data from the diesel generator control unit. The unit control unit monitors the precise position of the piston in real time through the crankshaft displacement sensor and transmits it to the injection timing control module to ensure that the injection timing is accurately matched with the piston movement;

[0066] S14, real-time acquisition of intake pressure: obtaining intake pressure data from the diesel generator control unit, and the intake pressure sensor in the intake manifold monitors the intake pressure in real time.

[0067] By processing the collected compression pressure data, key features of the compression process (fluctuation amplitude, frequency, and peak) are identified, thereby establishing a compression fluctuation model that can reflect the dynamic characteristics of compression pressure fluctuations in real time. The compression fluctuation feature analysis and modeling in S2 specifically include:

[0068] S21, data preprocessing: compress the pressure fluctuation data P filtered (t) Perform standardization processing to obtain the standardized compression pressure fluctuation data P norm (t), so that the data meets the input requirements of the deep learning model. The standardization formula is as follows: Where μ is the mean of the compression pressure data, σ is the standard deviation of the data, and P norm (t) is the normalized data;

[0069] S22, feature extraction: inputting the normalized compression pressure fluctuation data into a deep learning algorithm, which performs feature extraction based on a combination of a convolutional neural network (CNN) and a recurrent neural network (RNN), thereby constructing a compression fluctuation model;

[0070] S23, real-time identification of fluctuation characteristics: The real-time input compression pressure fluctuation data is processed using the trained compression fluctuation model. The model output includes fluctuation amplitude, frequency, and peak time characteristics, as follows:

[0071] Fluctuation amplitude characteristics: The local extreme value difference calculated by the compressed fluctuation model reflects the size of the fluctuation amplitude. The fluctuation amplitude characteristics A wave (t) is calculated as:

[0072] A wave (t) = max(P norm (t))-min(P norm (t)); represents the pressure signal P norm (t) The difference between the maximum and minimum values in time, which is the local extreme value difference, reflects the fluctuation amplitude of the signal;

[0073] Frequency characteristics: The main frequency components are extracted from the fluctuation data by fast Fourier transform, and the frequency characteristics f wave It is obtained by the following formula:

[0074] Indicates reaches the maximum value, where represents Fourier transform;

[0075] Peak time feature: Based on the output of the recurrent neural network, the peak time feature t is extracted peak , the peak moment characteristic is expressed as: t peak =arg max(h t );

[0076] Updating and applying the compression fluctuation model: Through the training results of the deep learning model, the parameters in the compression fluctuation model are updated in real time, so that the model can dynamically adjust under conditions of engine load changes, temperature fluctuations or engine wear, and the identified fluctuation amplitude characteristics, frequency characteristics and peak time characteristics are used for adaptive adjustment of fuel injection timing.

[0077] Fast Fourier Transform (FFT) is an efficient algorithm for Discrete Fourier Transform (DFT) that converts time domain signals into frequency domain signals to analyze the frequency components of the signal. The formula for Fourier transform is: Where x(n) is the signal in the time domain, X(f) is the signal in the frequency domain, N is the length of the signal (number of sampling points), and e -j2πfn / N It is a Fourier basis function used to convert time domain signals into amplitude and phase at different frequencies; by decomposing the time domain signal, the frequency components of the discrete signal can be efficiently calculated.

[0078] To extract frequency features from compression pressure fluctuation data to find the main frequency related to the compression process of the diesel engine, assume that the normalized compression pressure fluctuation data P has been obtained from the sensor. norm (t), the following are the specific steps:

[0079] a. Data acquisition and sampling: The compression pressure fluctuation data obtained from the sensor is P norm (t), which is a time domain signal, indicating the change of compression pressure over time, with a frequency of f s The signal is sampled, and the number of sampling points is N, then the time domain signal is P norm (t), where t = 0, 1, 2, ..., N-1 corresponds to each sampling point.

[0080] b. Apply fast Fourier transform: transform the time domain signal P norm (t) Perform fast Fourier transform and convert it into frequency domain signal P fft (f), represents the amplitude of each frequency component, and the Fourier transform is: Among them, P fft (f) is the frequency domain signal, which represents the amplitude and phase information of each frequency component.

[0081] c. Calculate the amplitude of the spectrum: The output of the Fourier transform is in complex form, which contains the amplitude and phase of the frequency components. To analyze the main frequency, we only need to care about the amplitude of the spectrum, that is, the modulus of each frequency component: Among them, θ(P fft (f)) and δ(P fft (f)) are the real and imaginary parts of the frequency components, respectively.

[0082] d. Spectrum analysis and main frequency extraction: After obtaining the amplitude of the frequency domain signal |P fft (f)|, we can analyze the spectrum and find the main frequency, that is, the main frequency component corresponding to the compression fluctuation, the main frequency f wave is the frequency component with the largest amplitude, which can be expressed as:

[0083] f wave =argmax(|P fft(f)|), which means finding the frequency component f with the largest amplitude, i.e., the main frequency. The main frequency is the main frequency component in the compression process, which usually corresponds to the periodicity of the compression fluctuation in the cylinder.

[0084] The feature extraction of S22 specifically includes:

[0085] S221, Convolution layer extracts local features: Multiple convolution layers are used to extract local features in the compression pressure fluctuation, including the change pattern of the fluctuation amplitude and the peak time feature. The output of the convolution layer is expressed as: F conv (t) = f(W conv ·P norm (t)+b conv ), where W conv and b conv are the weight matrix and bias of the convolutional layer, f(·) is the activation function, F conv (t) is the local feature output by the convolutional layer;

[0086] S222, recursive layer captures time dependency: The recursive neural network is used to capture the time dependency and nonlinear dynamic characteristics of the compression pressure fluctuation data. The state update formula of the recursive layer is:

[0087] h t =f(W h ·h t-1 +W x ·F conv (t)+b h ), where h t is the hidden state of the recurrent neural network at time step t, W h and W x are the weight matrices of the recursive layers, b h is the bias term, h t-1 is the hidden state of the recurrent neural network at the previous time step t-1.

[0088] In S3, combined with the traditional fuel injection model, according to the engine speed ω, piston position x piston And the intake pressure P intake , preliminarily calculate the injection time t inj , expressed as:

[0089] t inj =f(ω,x piston ,P intake ), where f(·) represents the traditional fuel injection model function. Based on the relationship between engine speed, piston position and intake pressure, the time point of fuel injection is calculated. The basic time interval of the injection moment is determined according to the currently collected engine speed ω, and the piston position x is combined with the time interval of the injection moment. pistonBy calculating the relationship between the piston position and the crankshaft angle, the injection timing is corrected and the real-time collected intake pressure P is used. intake , adjust the injection timing and duration to adapt to changes in intake pressure and ensure combustion efficiency;

[0090] Influence of engine speed ω: Determine the basic time interval Δt of the injection timing based on the currently collected engine speed speed , the formula is: Time interval Δt speed Define the time distribution within each combustion cycle. The higher the engine speed, the earlier the injection timing.

[0091] Piston position x piston Correction: Combined with the actual position of the piston, the injection timing is corrected by calculating the relationship between the piston position and the crankshaft angle. The correction formula is:

[0092] t inj =t inj +Δt piston , where Δt piston It is a correction value based on the piston motion phase, ensuring that the injection timing matches the moment when the piston reaches the optimal compression point;

[0093] Inlet pressure P intake Impact on fuel injection quantity: According to the real-time collected intake pressure data, the injection timing and injection duration are adjusted to adapt to the changes in intake pressure and ensure combustion efficiency. The initial injection timing t inj Adjust according to the change of intake pressure:

[0094] t inj =t inj -k p ·(P intake -P ref ), where k p is the sensitivity coefficient of intake pressure to injection timing, P ref is the reference intake pressure.

[0095] Generate preliminary injection timing control signal: Based on the above calculation and correction results, generate preliminary injection timing control signal and transmit this signal to the fuel injection control module to ensure that the injection timing initially meets the combustion requirements under the current working conditions. The formula is expressed as:

[0096] t inj control =t inj ·G control , where G control It is the injection timing control gain, which is used to further adjust the sensitivity of the control signal.

[0097] Δt pistonIndicates the correction value of the injection timing according to the piston position, which is calculated by the dynamic change of the physical position of the piston movement relative to the crankshaft rotation angle, Δt piston It can be calculated by the geometric relationship between the piston and the crankshaft. For a four-stroke engine, the position of the piston and the rotation angle of the crankshaft (θ crank ) is related to the piston displacement x piston The formula is:

[0098] Where r is the crankshaft radius, l is the connecting rod length, θ crank is the crankshaft angle; in a real engine, θ crank The relationship with time t is monitored in real time by the sensor to determine the current position of the piston. The correction value Δt of the injection time is calculated according to the time when the piston reaches the compression top dead center. piston .

[0099] Intake pressure correction coefficient k p It is the sensitivity coefficient of the influence of intake pressure on injection timing, which indicates the offset of injection timing due to changes in intake pressure. Changes in intake pressure will affect combustion conditions, so injection timing needs to be adjusted according to intake pressure; k p Obtained through experimental calibration. By testing at different intake pressures, the relationship between injection timing and combustion efficiency is recorded. In the experiment, at different intake pressures P intake Adjust the injection timing under certain conditions, record the engine's combustion efficiency and emission data, and find out the impact of intake pressure changes on injection timing.

[0100] S4 specifically includes:

[0101] S41, fluctuation amplitude A wave (t) represents the amplitude of the pressure change in the compression fluctuation. A larger fluctuation amplitude means a stronger compression process inside the cylinder, providing a higher compression pressure, which is suitable for full atomization and efficient combustion of the fuel. Therefore, the injection timing should be matched with the larger stage of the fluctuation amplitude to utilize the higher compression pressure, so that the fuel can be better atomized and fully burned during injection. Therefore, the fluctuation amplitude characteristic A wave (t) is too large (i.e. A wave (t) is close to its local or global maximum), delaying the injection timing to ensure that the injection occurs during the high-pressure period, thereby improving the atomization and combustion efficiency of the fuel. The fluctuation amplitude characteristic A wave When (t) is too low, the injection timing is advanced to avoid the injection period overlapping with the low-pressure period, thereby avoiding a decrease in combustion efficiency.

[0102] Fluctuation A wave The size of (t) is defined based on the historical data of the engine at different loads and speeds, and a reference value A is set.ref , which represents the standard amplitude of compression fluctuation under ideal or full-load conditions. Based on this benchmark value, the range of "too large" and "too small" is divided.

[0103] Excessive fluctuation: A wave (t)≥0.8·A ref When the fluctuation amplitude is greater than or close to the reference value (reaching or exceeding 80% of the reference value), it is considered that the compression fluctuation is strong and it is appropriate to delay the injection timing to fully utilize the higher compression pressure;

[0104] Smaller fluctuation range: A wave (t)≤0.5·A ref When the fluctuation amplitude is less than 50% of the reference value, the compression fluctuation is considered weak. At this time, the injection timing needs to be advanced to avoid injection in a lower compression pressure range and ensure that the fuel can have better combustion conditions when injected.

[0105] S42, frequency characteristic f wave Indicates the main frequency of compression fluctuation, that is, the compression / expansion frequency of each compression cycle, the frequency characteristic f wave When it is too high, the compression fluctuation period is short and the injection time is advanced to ensure that the fuel has enough time to complete injection and combustion within the short compression period; the frequency characteristic f wave When it is too low, the compression cycle becomes longer and the injection time is delayed. According to the compression fluctuation cycle T wave , calculates the ideal injection timing and adjusts the current injection timing to make it in the optimal time period within the compression cycle to ensure that the injection can still match the compression fluctuation during a longer compression process.

[0106] The frequency characteristics are also set with a reference frequency based on historical data. When the compression fluctuation frequency is higher than 120% of the reference value, it means that the engine is in a high-speed or high-load state. At this time, the compression cycle is shortened and the injection timing needs to be advanced to ensure that the fuel is injected and burned within a shorter compression cycle. When the compression fluctuation frequency is lower than 80% of the reference value, it means that the engine is in a low-speed or low-load state. At this time, the compression cycle is longer and the injection timing can be appropriately delayed to match the injection with the longer compression cycle.

[0107] In S41, according to the actual fluctuation range A wave (t) and the maximum expected fluctuation range A max The relationship between the injection timing is defined as follows:

[0108] Where Δt A is the injection timing offset adjusted based on the fluctuation amplitude, k A is the influence coefficient of the fluctuation amplitude, A max is the maximum expected value or standard value of the fluctuation range, Awave (t) is the real-time fluctuation amplitude, and the injection timing t is adjusted by the fluctuation amplitude characteristics. injdj1 Expressed as:

[0109] t injadj1 =t inj +Δt A , when A wave When (t) is high, the injection timing is appropriately delayed, otherwise, the injection timing is advanced.

[0110] S42 specifically includes: the injection timing should be consistent with the compression fluctuation period Matching ensures that fuel injection occurs at the optimal time period within the compression cycle (usually injection is completed in the first half of compression); therefore, according to the frequency characteristics of the compression fluctuations, the injection timing is advanced or delayed to make the injection timing reasonably adapt to the current compression cycle.

[0111] Define T wave is the period of compression fluctuation, then the ideal injection time in each compression cycle is approximately the first α part of the cycle (for example, the first 25% or 30%), that is: ideal =α·T wave , calculate the current injection time t inj adj1 The position in the current compression cycle is compared with the ideal injection time to obtain the adjustment value Δt f :

[0112] Δt f =t ideal -(t inj adj1 mod T wave ), where t ideal is the ideal injection timing calculated based on the compression cycle, t injadj1 mod T wave is the position of the current injection time in the cycle, mod represents the modulo operation, that is, the remainder operation, which calculates the remainder of one number divided by another number and calculates the injection time t inj adj1 In the current compression cycle T wave The position in the, the returned value is t inj adj1 Relative to T wave The remainder of the time represents the relative time of the moment in the cycle, and the injection time t is adjusted by the frequency characteristic. inj adj2 Expressed as:

[0113] t inj adj2 =t inj adj1 +Δt f , when t inj adj1 When the ideal injection timing of the compression cycle is deviated from, it is automatically advanced or retarded to ensure that the injection timing is aligned with the optimal phase of the current compression cycle.

[0114] The peak moment indicates the point in time when the compression fluctuation in the cylinder reaches the maximum pressure. This is the highest pressure stage of the compression process and the moment with the best combustion efficiency, but the injection timing should be slightly advanced to ensure that the fuel has enough time to atomize and complete the injection when the peak arrives.

[0115] The peak amplitude reflects the maximum compression pressure in the cylinder. A higher peak amplitude means greater compression pressure, which is suitable for more precise injection control; a lower peak amplitude means incomplete compression under the current working conditions, and the injection timing needs to be appropriately adjusted to match the actual situation.

[0116] In S5, according to the peak time feature t peak Adjust the injection timing. Here, the injection timing is the injection timing t after adjustment based on the frequency characteristics. inj adj2 :The ideal position of the injection time is before the peak time. The injection time advance amount is set. The advance amount depends on the injection time, the time of the fuel atomization process and the actual working conditions in the cylinder. The injection is appropriately advanced before the peak arrives so that the fuel injection can be completed at the peak time to ensure sufficient combustion. According to the peak time of the compression fluctuation model, the difference between the current injection time and the peak time is calculated. If the current injection time is too early or too late, it is dynamically adjusted according to the difference to ensure that the injection time is close to the ideal time before the peak; the injection time needs to be advanced by a reasonable time window of the peak time. This window ensures that the compression pressure in the cylinder is close to the maximum value when the injection is completed. This advance amount is dynamically determined according to the engine working conditions (such as speed, load) and fuel injection characteristics. If the current injection time is later than the ideal time before the peak, the injection time is advanced; if the current injection time is earlier than the ideal time before the peak, the injection is delayed to ensure that the injection process is aligned with the peak phase;

[0117] Peak time characteristic t peak Indicates the time point when the compression fluctuation reaches the maximum pressure. In order to ensure that the fuel is injected before the peak moment and has enough time to atomize, calculate the current injection time t inj and peak time t peak the gap between;

[0118] Adjustment formula: Δt peak =t inj -t peak , Δt peak Indicates the time difference between the current injection moment and the peak moment. When Δt peak When Δt > 0, it means the injection time is later than the peak and the injection should be advanced. peak When <0, it means that the injection timing is earlier than the peak and the injection should be delayed;

[0119] Based on the peak time, adjust the injection time t inj finalAdvance the peak time t peak , that is, the advance amount Δt adv :t inj_final =t peak -Δt adv , where Δt adv The amount by which the injection timing should be advanced to the peak depends on the engine operating conditions and the time required for fuel atomization. For example, at higher speeds, the fuel atomization time is shorter, and Δt adv Can be smaller, and under low speed conditions, Δt adv Needs to be larger.

[0120] The present invention encompasses any alternatives, modifications, equivalents, and solutions that fall within the spirit and scope of the present invention. To provide a thorough understanding of the present invention, specific details are described in detail below in connection with the preferred embodiments of the present invention, but those skilled in the art will be able to fully understand the present invention without these detailed descriptions. Furthermore, to avoid unnecessary confusion regarding the essence of the present invention, well-known methods, processes, procedures, components, and circuits have not been described in detail.

[0121] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A fuel injection control method for a diesel generator set based on adaptive control, characterized in that: The following steps are involved: S1, operating parameter collection: Collect the operating data of the diesel generator set, including engine speed, compression pressure fluctuation, piston movement position and intake pressure; specifically including: S11, real-time acquisition of engine speed: obtaining engine speed data from the diesel generator set control unit; S12, real-time collection of compression pressure fluctuations: The in-cylinder pressure sensor detects the in-cylinder pressure signal of the diesel generator set, transmits the collected pressure signal to the data processing module, and forms compression pressure fluctuation data P after filtering and amplification. filtered (t), represents the compression pressure fluctuation signal at time t; S13, obtaining the piston movement position: obtaining the piston movement position data from the diesel generator set control unit; S14, real-time acquisition of intake pressure: obtaining intake pressure data from the diesel generator control unit, and the intake pressure sensor in the intake manifold monitors the intake pressure in real time; S2, Compression Fluctuation Characteristic Analysis and Modeling: Based on the collected compression pressure fluctuation data, a compression fluctuation model is established using a deep learning algorithm based on nonlinear dynamic characteristics. The compression fluctuation model identifies the fluctuation amplitude characteristics, frequency characteristics, and peak time characteristics during the compression process in real time. Specifically, it includes: S21, data preprocessing: compress the pressure fluctuation data P filtered (t) Perform standardization processing to obtain the standardized compression pressure fluctuation data P norm (t); S22, feature extraction: inputting the normalized compression pressure fluctuation data into a deep learning algorithm, wherein the deep learning algorithm performs feature extraction based on a combination of a convolutional neural network and a recurrent neural network; S23, real-time identification of fluctuation characteristics: The real-time input compression pressure fluctuation data is processed using the trained compression fluctuation model. The model output includes fluctuation amplitude, frequency, and peak time characteristics, as follows: Fluctuation amplitude characteristics: The local extreme value difference calculated by the compressed fluctuation model reflects the size of the fluctuation amplitude. The fluctuation amplitude characteristics A wave (t) is calculated as: A wave (t)=max(P norm (t))-min(P norm (t)); Frequency characteristics: The main frequency components are extracted from the fluctuation data by fast Fourier transform, and the frequency characteristics f wave It is obtained by the following formula: Indicates reaches the maximum value, where represents Fourier transform; Peak time feature: Based on the output of the recurrent neural network, the peak time feature t is extracted peak , the peak moment characteristic is expressed as: t peak =arg max(h t ), where h t is the hidden state of the recurrent neural network at time step t; S3, preliminary calculation of injection timing: Based on the traditional fuel injection model, combined with the engine speed, piston position, and intake pressure under the current working conditions, the injection timing is preliminarily calculated and a preliminary injection timing control signal is generated; S4, Adaptive Adjustment of Compression Fluctuation: This step comprehensively analyzes the amplitude and frequency characteristics of the compression fluctuation model with the preliminary injection timing calculation results, and adjusts the preliminary calculated injection timing in real time to match the injection timing with the optimal phase of the compression fluctuation, thus avoiding premature or late fuel injection caused by compression fluctuation. Specifically, it includes: S41, Fluctuation Amplitude Characteristics A wave When (t) is too large, the injection timing is delayed and the fluctuation amplitude characteristic A wave If (t) is too low, the injection timing is advanced; S42, frequency characteristic f wave When it is too high, the compression fluctuation period is short and the injection time is advanced; the frequency characteristic f wave When it is too low, the compression cycle becomes longer and the injection time is delayed. According to the compression fluctuation period T wave , calculate the ideal injection time and adjust the current injection time to make it in the best time period within the compression cycle; S5, optimization of the matching between injection timing and peak timing characteristics: Based on the real-time monitoring of compression pressure fluctuations, the control system of the diesel generator set aligns the injection timing with the phase of the peak timing characteristics, and optimizes the injection time so that the fuel injection timing occurs at the ideal position before the peak arrives.

2. The diesel generator fuel injection control method based on adaptive control according to claim 1, characterized in that: The feature extraction of S22 specifically includes: S221, extracting local features using convolutional layers: extracting local features from compression pressure fluctuations through multiple convolutional layers, including the variation pattern of the fluctuation amplitude and the peak time characteristics; S222, Recursive layer captures time dependence: Capturing the time dependence and nonlinear dynamic characteristics of compression pressure fluctuation data through recursive neural network.

3. The diesel generator fuel injection control method based on adaptive control according to claim 1, characterized in that: In the above S3, the conventional fuel injection model is combined with the engine speed ω and the piston position x piston And the intake pressure P intake , preliminarily calculate the injection time t inj , expressed as: t inj =f(ω,x piston ,P intake ), where f(·) represents the traditional fuel injection model function. Based on the relationship between engine speed, piston position and intake pressure, the time point of fuel injection is calculated. The basic time interval of the injection moment is determined according to the currently collected engine speed ω, and the piston position x is combined with the time interval of the injection moment. piston By calculating the relationship between the piston position and the crankshaft angle, the injection timing is corrected and the real-time collected intake pressure P is used. intake , adjust the injection timing and injection duration to adapt to changes in intake pressure and ensure combustion efficiency.

4. The diesel generator fuel injection control method based on adaptive control according to claim 1, characterized in that: In S41, according to the actual fluctuation amplitude A wave (t) and the maximum expected fluctuation range A max The relationship between the injection timing is defined as follows: Where Δt A is the injection timing offset adjusted based on the fluctuation amplitude, k A is the influence coefficient of the fluctuation amplitude, A max is the maximum expected value or standard value of the fluctuation range, A wave (t) is the real-time fluctuation amplitude, and the injection timing t is adjusted by the fluctuation amplitude characteristics. injdj1 Expressed as: t injadj1 =t inj +Δt A 。 5. The diesel generator fuel injection control method based on adaptive control according to claim 4, characterized in that: The step S42 also includes calculating the current injection time t injadj1 The position in the current compression cycle is compared with the ideal injection time to obtain the adjustment value Δt f : Δt f =t ideal -(t injadj1 mod T wave ), where t ideal is the ideal injection timing calculated based on the compression cycle, t injadj1 mod T wave is the position of the current injection time in the cycle, mod represents the modulo operation, and the injection time t after adjustment of the frequency characteristics injadj Expressed as: t injadj2 =t injadj1 +Δt f 。 6. The diesel generator fuel injection control method based on adaptive control according to claim 5, characterized in that: In said S5, according to the peak time feature t peak Adjust the injection timing: The ideal injection timing is before the peak moment. Set the injection timing advance amount. According to the peak moment of the compression wave model, calculate the difference between the current injection timing and the peak moment. If the current injection timing is too early or too late, make dynamic adjustments based on the difference to ensure that the injection timing is close to the ideal moment before the peak; if the current injection timing is later than the ideal time before the peak, advance the injection timing; if the current injection timing is earlier than the ideal time before the peak, postpone the injection to ensure that the injection process is aligned with the peak phase.

Images