Method for on-line real-time prediction of injection rule of fuel injector and evaluation of aging state of injection hole

Abstract

The invention aims to provide a method for online real-time prediction of the injection rule of a fuel injector and evaluation of the aging state of an injection hole, and belongs to the field of engines. The method comprises the following steps: firstly, establishing a non-aging injection characteristic database by collecting excitation current, injector inlet pressure, injection rate and injection quantity data under different working conditions; then, based on the calibrated mathematical model of the fuel injection system, injector states under different injection hole aging degrees are simulated, and an injection characteristic database under the injection hole aging condition is generated; thirdly, constructing an RNN (Recurrent Neural Network)-based injection rate prediction model to realize online prediction of injection characteristics in non-aging and aging states; and finally, on the basis of excitation current and pressure signals which are measured in real time, the injection rate and the injection quantity are predicted online by applying the prediction model, and the aging state of the injection hole is evaluated. According to the invention, real-time prediction of the injection characteristics of the injector working for a long time can be realized, and the aging state of the injection hole can be effectively evaluated online.

Description

Technical Field

[0001] The present invention relates to an engine injection device, in particular to an injection prediction and evaluation method. Background Art

[0002] With the continuous advancement of engine technology and increasingly stringent environmental protection requirements, internal combustion engines are gradually developing towards a more efficient, environmentally friendly and reliable direction. As one of the core technologies in the fuel supply system of modern internal combustion engines, the high-pressure fuel system is widely used in various types of internal combustion engines because it can realize independent control of fuel injection pressure and injection timing. By accurately controlling the injection amount and injection timing of the injector, the combustion efficiency of the engine can be effectively improved, emission pollution can be reduced, and the power performance of the engine can be improved. However, as the fuel injector works under harsh conditions such as high pressure and high temperature for a long time, its spray hole will gradually experience aging phenomena such as wear, carbon deposition, and blockage, which will directly affect the injection law and injection amount of the injector, thereby resulting in reduced engine combustion efficiency, increased emissions and unstable operation. Therefore, during the operation of the engine, real-time monitoring and prediction of the injection law and evaluation of the aging state of the spray hole have become key technical problems to ensure the long-term stable operation of the engine.

[0003] Traditional methods for detecting the status of injectors often rely on physical testing or laboratory measurements after regular shutdowns, which is not only costly and time-consuming, but also difficult to timely reflect the actual working status of the injector. At the same time, due to the complex dynamic characteristics of the injector and the nonlinear characteristics of the injector aging process, it is difficult to accurately predict the changes in the injection law and the impact of nozzle aging using traditional model prediction methods. Summary of the invention

[0004] The purpose of the present invention is to provide a method for online real-time prediction of fuel injector injection law and evaluation of nozzle aging status, which can solve the problems in the prior art of inaccurate prediction of injection law of long-term operating injectors and difficulty in real-time evaluation of nozzle aging status.

[0005] The object of the present invention is achieved in that:

[0006] The present invention provides a method for online real-time prediction of fuel injector injection law and evaluation of nozzle aging state, which is characterized by comprising the following steps:

[0007] (1) Collect the excitation current, injector inlet pressure, injection rate and injection amount of the injector for single injection, pre-main injection and pre-main-post injection under different engine operating conditions to establish an injection characteristic database before aging;

[0008] (2) Construct a mathematical model of the fuel injection system and calibrate the model based on the injection characteristic database. In the model, the injector nozzle diameter is changed to 90%, 80%, 70%, 60%, and 50% of the original nozzle diameter to simulate the injector under different nozzle aging conditions. The modified model is used to calculate the excitation current, injector inlet pressure, injection rate, and injection amount of single injection, pre-main injection, and pre-main-post injection under different working conditions, and establish an injection characteristic database under nozzle aging conditions;

[0009] (3) With the excitation current and the injector inlet pressure as input and the injection rate as output, an RNN-based injection rate prediction model is constructed. The prediction model is trained simultaneously using the injection characteristic database before aging and the injection characteristic database after nozzle aging to achieve the injection characteristic prediction of the injector before and after aging.

[0010] (4) Based on the excitation current signal and the pressure signal at the injector inlet measured in real time by the fuel injection system, the injection rate prediction model established in step (3) is used to perform online real-time prediction of the injection rate, and the real-time injection amount is obtained by integration to evaluate the nozzle aging status.

[0011] The present invention may also include:

[0012] 1. The different working conditions in step (1) are specifically:

[0013] In the single injection data set, the operating conditions include pressure conditions and injection pulse width conditions. The pressure conditions include all pressure conditions with an interval of 10 MPa from the idle pressure to the rated pressure, and the injection pulse width conditions include all injection pulse width conditions with an interval of 0.2 ms from the minimum start-up pulse width to the rated pulse width.

[0014] The pre-main injection working conditions include pressure conditions, pre-injection pulse width conditions, pre-main injection interval conditions and main injection pulse width conditions. The pre-main-post injection working conditions include pressure conditions, pre-injection pulse width conditions, pre-main injection interval conditions, main injection pulse width conditions, main-post injection interval conditions and post injection pulse width conditions. The pressure conditions include all pressure conditions between the idle pressure and the rated pressure with an interval of 20MPa. The pre-injection pulse width conditions include all pre-injection pulse widths between the minimum start-injection pulse width and the maximum pre-injection pulse width with an interval of 0.05ms. Conditions, main spray pulse width conditions include all main spray pulse width conditions with an interval of 0.2ms between the minimum start-injection pulse width and the rated pulse width, post-injection pulse width conditions include all post-injection pulse width conditions with an interval of 0.05ms between the minimum start-injection pulse width and the maximum post-injection pulse width, pre-main spray interval conditions include all interval conditions with an interval of 1ms between the minimum pre-main spray interval and the maximum pre-main spray interval, main-post-injection interval conditions include all interval conditions with an interval of 1ms between the minimum main-post-injection interval and the maximum main-post-injection interval.

[0015] 2. The method for constructing the injection rate prediction model in step (3) specifically includes:

[0016] (31) The excitation current and injector inlet pressure data in the database are used as input data sets, and the injection rate is used as output data sets, and batch normalization is performed:

[0017]

[0018] In the formula, x * is the processed data; represents the mean of unnormalized data; represents the standard deviation of the data; ε is the margin used to prevent division by zero;

[0019] (32) Constructing the framework of injection rate prediction model:

[0020] The input layer is established, and the input X(t) is the combination of the excitation current and the injector inlet pressure:

[0021]

[0022] In the formula, I drive (t) is the excitation current of the injector at time t, P in (t) is the value of the injector inlet pressure at time t;

[0023] The hidden state of each time step of the model is established through a loop structure:

[0024] h t =σ(W h ·h t-1 +W x ·X(t)+b h )

[0025] In the formula, h t is the hidden state at time t, W h is the weight matrix between different hidden states, W x is the weight matrix input to the hidden state, b h is the bias vector of the hidden state, σ is the tanh activation function;

[0026] Establish the output layer and define the output Q inj (t) Determined by the current implicit state:

[0027] Q inj (t) = W y ·h t +b y

[0028] Where Wy is the weight matrix from hidden state to output layer, b y is the bias vector of the output layer;

[0029] (33) A loss function is established, which is defined as the mean square error between the injection rate predicted by the model and the true value:

[0030]

[0031] In the formula, N is the number of samples in the data set, T is the time series length of each sample, For time t i The predicted injection rate, For time t i The actual injection rate.

[0032] 3. The method for evaluating the nozzle aging state in step (4) specifically includes:

[0033] The maximum injection rate will be predicted and total injection volume The maximum injection rate before aging and total injection volume As a ratio, we use the mean of the two quantities η total Comprehensive evaluation of the aging degree of the injector nozzle:

[0034]

[0035] When 90%<η total When ≤100%, the aging state is excellent; when 80%<η total When ≤90%, the aging state is good; when 70%<η total When ≤80%, the aging state is medium; when 60%<η total When ≤70%, the aging state is poor; when η total When ≤60%, the aging state is extremely poor.

[0036] The advantages of the present invention are:

[0037] (1) The present invention combines the injection characteristic database under unaged and aged conditions, so that the injection law can be predicted online in real time under different aged conditions;

[0038] (2) The present invention can evaluate the aging status of the nozzle hole online without stopping the engine for inspection, thereby effectively determining the performance status of the injector, reducing maintenance costs and improving engine reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0040] The present invention is described in more detail below with reference to the accompanying drawings:

[0041] Combination Figure 1 The present invention provides a method for online real-time prediction of fuel injector injection law and evaluation of nozzle aging state, comprising the following steps:

[0042] Step S1, installing a new injector on a fuel injection system test bench, collecting excitation current, injector inlet pressure, injection rate and injection amount of single injection, pre-main injection and pre-main-post injection under different working conditions, and establishing an injection characteristic database before aging.

[0043] Step S2, constructing a mathematical model of the fuel injection system, calibrating the model based on the injection characteristic database, in the model, changing the injector nozzle diameter to 90%, 80%, 70%, 60%, 50% of the original nozzle diameter, so as to simulate the injector under different nozzle aging conditions, using the modified model, calculating and generating the excitation current, injector inlet pressure, injection rate and injection amount of a single injection, pre-main injection, pre-main-post injection under different working conditions, and establishing an injection characteristic database under nozzle aging conditions.

[0044] In step S3, an RNN-based injection rate prediction model is constructed with the excitation current and the injector inlet pressure as input and the injection rate as output. The prediction model is trained simultaneously using the injection characteristic database before aging and the injection characteristic database under nozzle aging conditions to achieve injection characteristic prediction before and after aging of the injector.

[0045] Step S4, based on the excitation current signal measured in real time by the fuel injection system and the pressure signal at the injector inlet, the injection rate prediction model established in step S3 is used to perform online real-time prediction of the injection rate, the real-time injection amount is obtained by integration, and the nozzle aging status is evaluated.

[0046] The specific working conditions are as follows:

[0047] In the single injection data set, the operating conditions include pressure and injection pulse width, wherein the pressure conditions include all pressure conditions with an interval of 10 MPa between the idle pressure and the rated pressure, and the injection pulse width conditions include all injection pulse width conditions with an interval of 0.2 ms between the minimum injection start pulse width and the rated pulse width;

[0048] In the multiple injection data set, the pre-main injection working conditions include pressure, pre-injection pulse width, pre-main injection interval and main injection pulse width, and the pre-main-post injection working conditions include pressure, pre-injection pulse width, pre-main injection interval, main injection pulse width, main-post injection interval and post injection pulse width, wherein the pressure condition includes all pressure conditions between the idle pressure and the rated pressure with an interval of 20MPa, the pre-injection pulse width condition includes all pre-injection pulse width conditions between the minimum start-injection pulse width and the maximum pre-injection pulse width with an interval of 0.05ms, and the main The spray pulse width conditions include all main spray pulse width conditions with an interval of 0.2ms between the minimum start-injection pulse width and the rated pulse width. The post-injection pulse width conditions include all post-injection pulse width conditions with an interval of 0.05ms between the minimum start-injection pulse width and the maximum post-injection pulse width. The pre-main spray interval conditions include all interval conditions with an interval of 1ms between the minimum pre-main spray interval and the maximum pre-main spray interval. The main-post-injection interval conditions include all interval conditions with an interval of 1ms between the minimum main-post-injection interval and the maximum main-post-injection interval.

[0049] The construction of the injection rate prediction model in step S3 specifically includes:

[0050] Step (a), taking the excitation current and injector inlet pressure data in the database as the input data set, taking the injection rate as the output data set, and performing batch normalization:

[0051]

[0052] In the formula, x * is the processed data; represents the mean of unnormalized data; represents the standard deviation of the data; ε is the margin used to prevent division by zero.

[0053] Step (b), constructing the framework of the injection rate prediction model.

[0054] The input layer is established, and the input X(t) is the combination of the excitation current and the injector inlet pressure:

[0055]

[0056] In the formula, I drive (t) is the excitation current of the injector at time t, P in (t) is the value of the injector inlet pressure at time t.

[0057] The hidden state of each time step of the model is established through a loop structure:

[0058] h t =σ(W h ·h t-1 +W x ·X(t)+bh )

[0059] In the formula, h t is the hidden state at time t, W h is the weight matrix between different hidden states, W x is the weight matrix input to the hidden state, b h is the bias vector of the hidden state, and σ is the tanh activation function.

[0060] Establish the output layer and define the output Q inj (t) Determined by the current implicit state:

[0061] Q inj (t) = W y ·h t +b y

[0062] Where W y is the weight matrix from hidden state to output layer, b y is the bias vector of the output layer.

[0063] Step (c) establishes a loss function, which is defined as the mean square error between the injection rate predicted by the model and the true value:

[0064]

[0065] In the formula, N is the number of samples in the data set, T is the time series length of each sample, For time t i The predicted injection rate, For time t i The actual injection rate.

[0066] The evaluation of the nozzle aging state in step S4 specifically includes:

[0067] The maximum injection rate will be predicted and total injection volume The maximum injection rate before aging and total injection volume As a ratio, we use the mean of the two quantities η total Comprehensively evaluate the aging degree of the injector nozzle, the formula is:

[0068]

[0069] The aging state of the nozzle is determined by η total Determine that when 90% < η total When ≤100%, the aging state is excellent; when 80%<η total When ≤90%, the aging state is good; when 70%<η totalWhen ≤80%, the aging state is medium; when 60%<η total When ≤70%, the aging state is poor; when η total When ≤60%, the aging state is extremely poor.

Claims

1. A method for online real-time prediction of fuel injector injection law and evaluation of nozzle aging state, characterized by: The following steps are involved: (1) Collect the excitation current, injector inlet pressure, injection rate and injection amount of the injector for single injection, pre-main injection and pre-main-post injection under different engine operating conditions to establish an injection characteristic database before aging; (2) Construct a mathematical model of the fuel injection system and calibrate the model based on the injection characteristic database. In the model, the injector nozzle diameter is changed to 90%, 80%, 70%, 60%, and 50% of the original nozzle diameter to simulate the injector under different nozzle aging conditions. The modified model is used to calculate the excitation current, injector inlet pressure, injection rate, and injection amount of single injection, pre-main injection, and pre-main-post injection under different working conditions, and establish an injection characteristic database under nozzle aging conditions; (3) With the excitation current and the injector inlet pressure as input and the injection rate as output, an RNN-based injection rate prediction model is constructed. The prediction model is trained simultaneously using the injection characteristic database before aging and the injection characteristic database after nozzle aging to achieve the injection characteristic prediction of the injector before and after aging. (4) Based on the excitation current signal and the pressure signal at the injector inlet measured in real time by the fuel injection system, the injection rate prediction model established in step (3) is used to perform online real-time prediction of the injection rate, and the real-time injection amount is obtained by integration to evaluate the nozzle aging status.

2. The method for online real-time prediction of fuel injector injection law and evaluation of nozzle hole aging state according to claim 1 is characterized by: The different working conditions in step (1) are specifically: In the single injection data set, the operating conditions include pressure conditions and injection pulse width conditions. The pressure conditions include all pressure conditions with an interval of 10 MPa from the idle pressure to the rated pressure, and the injection pulse width conditions include all injection pulse width conditions with an interval of 0.2 ms from the minimum start-up pulse width to the rated pulse width. The pre-main injection working conditions include pressure conditions, pre-injection pulse width conditions, pre-main injection interval conditions and main injection pulse width conditions. The pre-main-post injection working conditions include pressure conditions, pre-injection pulse width conditions, pre-main injection interval conditions, main injection pulse width conditions, main-post injection interval conditions and post injection pulse width conditions. The pressure conditions include all pressure conditions between the idle pressure and the rated pressure with an interval of 20MPa. The pre-injection pulse width conditions include all pre-injection pulse widths between the minimum start-injection pulse width and the maximum pre-injection pulse width with an interval of 0.05ms. Conditions, main spray pulse width conditions include all main spray pulse width conditions with an interval of 0.2ms between the minimum start-injection pulse width and the rated pulse width, post-injection pulse width conditions include all post-injection pulse width conditions with an interval of 0.05ms between the minimum start-injection pulse width and the maximum post-injection pulse width, pre-main spray interval conditions include all interval conditions with an interval of 1ms between the minimum pre-main spray interval and the maximum pre-main spray interval, main-post-injection interval conditions include all interval conditions with an interval of 1ms between the minimum main-post-injection interval and the maximum main-post-injection interval.

3. The method for online real-time prediction of fuel injector injection law and evaluation of nozzle hole aging state according to claim 1 is characterized by: The method for constructing the injection rate prediction model in step (3) specifically includes: (31) The excitation current and injector inlet pressure data in the database are used as input data sets, and the injection rate is used as output data sets, and batch normalization is performed: In the formula, x * is the processed data; represents the mean of unnormalized data; represents the standard deviation of the data; ε is the margin used to prevent division by zero; (32) Constructing the framework of injection rate prediction model: The input layer is established, and the input X(t) is the combination of the excitation current and the injector inlet pressure: In the formula, I drive (t) is the excitation current of the injector at time t, P in (t) is the value of the injector inlet pressure at time t; The hidden state of each time step of the model is established through a loop structure: h t =σ(W h ·h t-1 +W x ·X(t)+b h ) In the formula, h t is the hidden state at time t, W h is the weight matrix between different hidden states, W x is the weight matrix input to the hidden state, b h is the bias vector of the hidden state, σ is the tanh activation function; Establish the output layer and define the output Q inj (t) Determined by the current implicit state: Q inj (t)=W y ·h t +b y Where W y is the weight matrix from hidden state to output layer, b y is the bias vector of the output layer; (33) A loss function is established, which is defined as the mean square error between the injection rate predicted by the model and the true value: In the formula, N is the number of samples in the data set, T is the time series length of each sample, For time t i The predicted injection rate, For time t i The actual injection rate.

4. The method for online real-time prediction of fuel injector injection law and evaluation of nozzle hole aging state according to claim 1 is characterized by: The method for evaluating the nozzle aging state in step (4) specifically includes: The maximum injection rate will be predicted and total injection volume The maximum injection rate before aging and total injection volume As a ratio, we use the mean of the two quantities η total Comprehensive evaluation of the aging degree of the injector nozzle: When 90%<η total When ≤100%, the aging state is excellent; when 80%<η total When ≤90%, the aging state is good; when 70%<η total When ≤80%, the aging state is medium; when 60%<η total When ≤70%, the aging state is poor; when η total When ≤60%, the aging state is extremely poor.

Images