Dual-mode PID (Proportion Integration Differentiation) real-time correction method for injection quantity of ejector based on virtual prediction of pressure fluctuation

Abstract

The invention aims to provide an ejector injection quantity dual-mode PID real-time correction method based on pressure fluctuation virtual prediction, and belongs to the field of engines. Firstly, pressure sensors are installed at the 1 / 4 position and the 3 / 4 position of a high-pressure pipeline, the backflow flow of an ejector is measured, an ARIMA model is constructed to conduct virtual prediction of pressure fluctuation at the 3 / 4 position to achieve replacement of the sensors, and a support vector machine model is adopted to predict the backflow flow. And then a flow model of the high-pressure pipeline is constructed, and the total fuel quantity flowing out of the high-pressure pipeline is calculated by combining the pressure, measured in real time, of the 1 / 4 position and the pressure, virtually predicted, of the 3 / 4 position. The actual injection quantity of the injector is obtained by making a difference between the actual injection quantity and the predicted reflux quantity, and the injection pulse width is corrected in real time by proposing a dual-mode PID according to the deviation between the actual injection quantity and the target injection quantity, so that closed-loop control of the injection quantity is realized. According to the method, the complexity and cost of multiple sensors are avoided through virtual prediction, and the prediction precision under the complex working condition is improved through a double-model prediction mode.

Description

Technical Field

[0001] The present invention relates to an engine injection device, in particular to an injector control method. Background Art

[0002] Common rail fuel injection technology is a key technology in internal combustion engines and is widely used in dual-fuel engines and low-carbon fuel engines. With the development of engine technology, especially in high-power marine engines, the requirements for precise control of fuel injection are getting higher and higher. Traditional fuel injection system control methods mostly rely on open-loop control or use closed-loop control based on feedback variables such as engine speed.

[0003] In modern high-pressure fuel systems, due to the nonlinear pressure fluctuations involved in the injection process, fluid dynamics effects in the pipeline, and injector response delays, traditional control methods are difficult to provide sufficient accuracy and fast response under complex working conditions. Especially in ultra-high pressure injection environments, these dynamic factors make it more difficult to control the injection volume of the system, resulting in uneven injection in each cylinder, which in turn affects the overall performance and emissions of the engine. Summary of the invention

[0004] The purpose of the present invention is to provide a dual-mode PID real-time correction method for injector injection quantity based on virtual prediction of pressure fluctuations, which can solve the problems that the injection quantity cannot be accurately predicted online in real time and the injection quantity cannot be accurately closed-loop controlled.

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

[0006] The present invention provides a dual-mode PID real-time correction method for injector injection quantity based on virtual prediction of pressure fluctuation, which is characterized by comprising the following steps:

[0007] (1) The common rail pipe and the injector are connected through a high-pressure pipeline. Pressure sensors are installed at 1 / 4 and 3 / 4 of the high-pressure pipeline respectively. Under different injection conditions, the fuel pressure fluctuations at the above two locations are measured, and the return flow rate of the injector is measured by a flow meter at the same time. After obtaining the pressure fluctuation data set and the return flow data set, the pressure sensor at 3 / 4 of the high-pressure pipeline is removed;

[0008] (2) An automatic regression integrated moving average ARIMA pressure fluctuation prediction model is constructed with the pressure fluctuation at 1 / 4 as input and the pressure fluctuation at 3 / 4 as output. Parameter identification is performed based on the pressure fluctuation data set. The ARIMA pressure fluctuation prediction model performs a virtual prediction of the pressure fluctuation at 3 / 4 of the high-pressure pipeline when the high-pressure fuel system is actually working. A support vector machine reflux flow prediction model is constructed with the injection operating conditions as input and the reflux flow as output. The model is trained based on the reflux flow data set.

[0009] (3) A high-pressure pipeline flow model is established based on the pressure fluctuation transmission relationship between 1 / 4 and 3 / 4 of the high-pressure pipeline. The pressure signal at 1 / 4 of the high-pressure pipeline actually measured and the pressure at 3 / 4 predicted by the ARIMA pressure fluctuation prediction model are used as inputs of the high-pressure pipeline flow model. The flow of the high-pressure pipeline is calculated, and the amount of fuel flowing out of the high-pressure pipeline is obtained by integration, which is taken as the sum of the injector injection amount and the reflux amount.

[0010] (4) The reflux volume is calculated based on the injection operating conditions in each cycle, and the actual injection volume of the injector is obtained by subtracting it from the amount of fuel flowing out of the high-pressure pipeline. Based on the deviation between the actual injection volume and the target injection volume, the injection pulse width of the next cycle is corrected through the dual-mode PID to achieve closed-loop control of the injection volume.

[0011] The present invention may also include:

[0012] 1. In step (1), the operating conditions include rail pressure conditions and injection pulse width conditions. The rail pressure conditions include all rail pressure conditions with an interval of 10 MPa between the idle rail pressure and the rated rail pressure. The injection pulse width conditions include all injection pulse width conditions with an interval of 0.1 ms between the minimum start-up pulse width and the rated pulse width.

[0013] 2. The specific method for constructing the ARIMA pressure fluctuation prediction model in step (2) is to construct the autoregressive part, the difference part and the moving average part of the ARIMA model in sequence:

[0014] (21) Construct the autoregressive part:

[0015] φ(B)=1-φ1B-φ2B 2 -…-φ p B p

[0016] In the formula, φ i is the autoregressive coefficient in the autoregressive part; B is the lag operator, defined as p is the order of the autoregressive term;

[0017] (22) Construct the differential part and use the second-order difference to process the pressure fluctuation:

[0018]

[0019] Where P 1 / 4 (t) and P 3 / 4 (t) are the time series data of pressure fluctuation at 1 / 4 and 1 / 4, respectively;

[0020] (23) Construct the moving average part:

[0021] θ(B)=1+θ1B+θ2B2 +…+θ p B q

[0022] In the formula, θ i is the moving average coefficient in the autoregressive part, and q is the order of the moving average term;

[0023] (24) Couple the various models and construct the ARIMA model:

[0024]

[0025] In the formula, ε t is the white noise error term; β i is the influence of the pressure fluctuation time series at 1 / 4 on the pressure fluctuation at 3 / 4, specifically expressed as β i =(β1,β2,…β m ), m is the maximum order of the input sequence lag considered;

[0026] (25) Based on the maximum likelihood estimation method, the model coefficient φ i ,θ i and β i The model can predict the pressure fluctuation at 3 / 4 by identifying

[0027]

[0028] 3. The method for constructing the high-pressure pipeline flow model in step (3) is specifically as follows:

[0029] Establish the momentum balance equation in the high-pressure pipeline:

[0030]

[0031] Where ρ is the fuel density, u is the flow rate, P is the pressure, D is the pipe diameter, and γ is the wall shear stress;

[0032] Integrate the distance L between 1 / 4 and 3 / 4, and then multiply by the pipeline cross-sectional area A to obtain the mass flow rate out of the high-pressure pipeline:

[0033]

[0034] In the formula, ξ is a constant related to the speed of sound, and t is time;

[0035] The fuel mass of the high pressure pipeline is expressed as:

[0036] M pipe-out =∫Q pipe-out dt.

[0037] 4. The method for real-time correction of injection quantity by dual-mode PID in step (4) is specifically as follows:

[0038] According to the current deviation between the actual injection amount and the target injection amount, the dual-mode PID switches between the coarse adjustment mode and the fine adjustment mode to calculate the injection pulse width correction amount:

[0039]

[0040] Where Δu(k) is the injection pulse width correction value at the kth injection, e(k) is the deviation between the actual injection amount and the target injection amount at the kth injection, and e # is the error threshold, is the correction coefficient in coarse adjustment mode, is the correction coefficient in fine tuning mode;

[0041] The dual-mode PID outputs the control signal for the kth injection based on the calculated injection pulse width correction value:

[0042] u(k)=u(k-1)+Δu(k).

[0043] The advantages of the present invention are:

[0044] (1) The present invention only requires one pressure sensor to be installed in the high-pressure pipeline, and can virtually predict the pressure fluctuation at the other end based on the measured pressure fluctuation, thus avoiding the complexity and cost of adding redundant sensors in actual working conditions;

[0045] (2) The present invention estimates the injection amount in real time by the difference between the predicted total fuel amount and the reflux amount. This dual-model prediction method improves the prediction accuracy of the injection amount under complex working conditions.

[0046] (3) The dual-mode PID of the present invention can automatically adjust the control correction amount of the injection pulse width according to the injector error threshold, quickly respond to complex dynamic working condition changes, and improve the follow-up response speed of the system control. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a flow chart of the present invention;

[0048] Figure 2 A diagram of the experimental setup used to obtain the dataset;

[0049] Figure 3 This is the dual-mode PID control block diagram. DETAILED DESCRIPTION

[0050] The present invention is described in more detail below with reference to the accompanying drawings:

[0051] Combination Figure 1-3 The present invention provides a dual-mode PID real-time correction method for injector injection quantity based on virtual prediction of pressure fluctuation, comprising the following steps:

[0052] Step S1, experimental device layout as follows Figure 2 As shown, pressure sensors are installed at 1 / 4 and 3 / 4 of the high-pressure pipeline between the common rail pipe and the injector, respectively. Under different injection conditions, the fuel pressure fluctuations at the above two locations are measured, and the return flow rate of the injector is measured by a flow meter at the same time. After obtaining the pressure fluctuation data set and the return flow data set, the pressure sensor at 3 / 4 of the high-pressure pipeline is removed.

[0053] Step S2, construct an automatic regressive integrated moving average (ARIMA) pressure fluctuation prediction model with the pressure fluctuation at 1 / 4 as input and the pressure fluctuation at 3 / 4 as output, perform parameter identification based on the pressure fluctuation data set, and the ARIMA pressure fluctuation prediction model is used for virtual prediction of the pressure fluctuation at 3 / 4 of the high-pressure pipeline when the high-pressure fuel system is actually working; construct a support vector machine reflux flow prediction model with the injection working condition as input and the reflux flow as output, and perform training based on the reflux flow data set.

[0054] Step S3, according to the Euler partial differential equation of generalized mass and momentum, a high-pressure pipeline flow model based on the pressure fluctuation transmission relationship at 1 / 4 and 3 / 4 of the high-pressure pipeline is established, the pressure signal at 1 / 4 of the high-pressure pipeline actually measured and the pressure at 3 / 4 predicted in real time by the ARIMA pressure fluctuation prediction model are used as inputs of the high-pressure pipeline flow model, the flow of the high-pressure pipeline is calculated, and the amount of fuel flowing out of the high-pressure pipeline is obtained by integration, which is taken as the sum of the injector injection amount and the reflux amount.

[0055] Step S4, based on the support vector machine reflux prediction model, the reflux amount is calculated according to the injection working conditions in each cycle, and the actual injection amount of the injector is obtained by subtracting it from the amount of fuel flowing out of the high-pressure pipeline. Based on the deviation between the actual injection amount and the target injection amount, the injection pulse width of the next cycle is corrected through the dual-mode PID to achieve closed-loop control of the injection amount, such as Figure 3 shown.

[0056] The data set working conditions in step S1 are specifically:

[0057] The operating conditions include rail pressure and injection pulse width, wherein the rail pressure conditions include all rail pressure conditions between the idle rail pressure and the rated rail pressure with an interval of 10 MPa, and the injection pulse width conditions include all injection pulse width conditions between the minimum start-up pulse width and the rated pulse width with an interval of 0.1 ms.

[0058] The ARIMA pressure fluctuation prediction model constructed in step S2 is specifically as follows:

[0059] The three main parts of the ARIMA model, namely the autoregressive part, the difference part and the moving average part, are constructed in sequence:

[0060] Step (a), construct the autoregressive part, expressed as:

[0061] φ(B)=1-φ1B-φ2B 2 -…-φ p B p (1)

[0062] In the formula, φ i is the autoregressive coefficient in the autoregressive part, describing the relationship between the current pressure fluctuation value and the pressure fluctuation value at the previous i time points; B is the lag operator, defined as p is the order of the autoregressive term.

[0063] Step (b) constructs the difference part to eliminate the non-stationarity in the time series. The pressure fluctuation is processed using the second-order difference:

[0064]

[0065] Where P 1 / 4 (t) and P 3 / 4 (t) are the time series data of pressure fluctuation at 1 / 4 and 1 / 4, respectively.

[0066] Step (c), construct the moving average part, expressed as:

[0067] θ(B)=1+θ1B+θ2B 2 +…+θ p B q (3)

[0068] In the formula, θ i is the moving average coefficient in the autoregressive part, describing the relationship between the current pressure fluctuation value and the pressure fluctuation error at the previous i time points, and q is the order of the moving average term;

[0069] Step (d), couple the various models and build an ARIMA model:

[0070]

[0071] In the formula, ε t is the white noise error term; β i is the influence of the pressure fluctuation time series at 1 / 4 on the pressure fluctuation at 3 / 4, specifically expressed as β i =(β1,β2,…β m ), m is the maximum order of lag of the input sequence considered.

[0072] Step (e), based on the maximum likelihood estimation method, the model coefficient φ i ,θ i and β iThe model can predict the pressure fluctuation at 3 / 4 by identifying

[0073]

[0074] The high-pressure pipeline flow model is constructed in step S3 as follows:

[0075] Establish the momentum balance equation in the high-pressure pipeline:

[0076]

[0077] Where ρ is the fuel density, u is the flow rate, P is the pressure, D is the pipe diameter, and γ is the wall shear stress.

[0078] Integrate the distance L between 1 / 4 and 3 / 4, and then multiply by the pipeline cross-sectional area A to obtain the mass flow rate out of the high-pressure pipeline:

[0079]

[0080] Where ξ is a constant related to the speed of sound, and t is time.

[0081] The fuel mass of the high pressure pipeline is expressed as:

[0082] M pipe-out =∫Q pipe-out dt (8)

[0083] The real-time correction of the injection amount by the dual-mode PID in step S4 is specifically as follows:

[0084] According to the deviation between the actual injection amount and the target injection amount, the dual-mode PID switches between the coarse adjustment mode and the fine adjustment mode to calculate the injection pulse width correction amount. The specific formula is as follows:

[0085]

[0086] Where Δu(k) is the injection pulse width correction value at the kth injection, e(k) is the deviation between the actual injection amount and the target injection amount at the kth injection, and e # is the error threshold, is the correction coefficient in coarse adjustment mode, is the correction factor in fine tuning mode.

[0087] The dual-mode PID outputs the control signal for the kth injection based on the calculated injection pulse width correction value:

[0088] u(k)=u(k-1)+Δu(k) (10).

Claims

1. A dual-mode PID real-time correction method for injector injection quantity based on virtual prediction of pressure fluctuation, characterized by: The following steps are involved: (1) The common rail pipe and the injector are connected through a high-pressure pipeline. Pressure sensors are installed at 1 / 4 and 3 / 4 of the high-pressure pipeline respectively. Under different injection conditions, the fuel pressure fluctuations at the above two locations are measured, and the return flow rate of the injector is measured by a flow meter at the same time. After obtaining the pressure fluctuation data set and the return flow data set, the pressure sensor at 3 / 4 of the high-pressure pipeline is removed; (2) An automatic regression integrated moving average (ARIMA) pressure fluctuation prediction model is constructed with the pressure fluctuation at 1 / 4 as input and the pressure fluctuation at 3 / 4 as output. Parameter identification is performed based on the pressure fluctuation data set. The ARIMA pressure fluctuation prediction model performs a virtual prediction of the pressure fluctuation at 3 / 4 of the high-pressure pipeline when the high-pressure fuel system is actually working. A support vector machine reflux flow prediction model with injection conditions as input and reflux flow as output was constructed, and trained based on the reflux flow data set. (3) A high-pressure pipeline flow model is established based on the pressure fluctuation transmission relationship between 1 / 4 and 3 / 4 of the high-pressure pipeline. The pressure signal at 1 / 4 of the high-pressure pipeline actually measured and the pressure at 3 / 4 predicted by the ARIMA pressure fluctuation prediction model are used as inputs of the high-pressure pipeline flow model. The flow of the high-pressure pipeline is calculated, and the amount of fuel flowing out of the high-pressure pipeline is obtained by integration, which is taken as the sum of the injector injection amount and the reflux amount. (4) The reflux volume is calculated based on the injection operating conditions in each cycle, and the actual injection volume of the injector is obtained by subtracting it from the amount of fuel flowing out of the high-pressure pipeline. Based on the deviation between the actual injection volume and the target injection volume, the injection pulse width of the next cycle is corrected through the dual-mode PID to achieve closed-loop control of the injection volume.

2. The method for real-time correction of injector injection quantity dual-mode PID based on virtual prediction of pressure fluctuation according to claim 1 is characterized by: The operating conditions in step (1) include rail pressure conditions and injection pulse width conditions. The rail pressure conditions include all rail pressure conditions with an interval of 10 MPa between the idle rail pressure and the rated rail pressure. The injection pulse width conditions include all injection pulse width conditions with an interval of 0.1 ms between the minimum start-up pulse width and the rated pulse width.

3. The method for real-time correction of injector injection quantity dual-mode PID based on virtual prediction of pressure fluctuation according to claim 1 is characterized by: The specific method for constructing the ARIMA pressure fluctuation prediction model in step (2) is to construct the autoregressive part, the difference part and the moving average part of the ARIMA model in sequence: (21) Construct the autoregressive part: φ(B)=1-φ1B-φ2B 2 -…-f p B p In the formula, φ i is the autoregressive coefficient in the autoregressive part; B is the lag operator, defined as p is the order of the autoregressive term; (22) Construct the differential part and use the second-order difference to process the pressure fluctuation: Where P 1 / 4 (t) and P 3 / 4 (t) are the time series data of pressure fluctuation at 1 / 4 and 1 / 4, respectively; (23) Construct the moving average part: θ(B)=1+θ1B+θ2B 2 +…+θ p B q In the formula, θ i is the moving average coefficient in the autoregressive part, and q is the order of the moving average term; (24) Couple the various models and construct the ARIMA model: In the formula, ε t is the white noise error term; β i is the influence of the pressure fluctuation time series at 1 / 4 on the pressure fluctuation at 3 / 4, specifically expressed as β i =(β1,β2,…β m ), m is the maximum order of the input sequence lag considered; (25) Based on the maximum likelihood estimation method, the model coefficient φ i ,θ i and β i The model can predict the pressure fluctuation at 3 / 4 by identifying 4. The method for real-time correction of injector injection quantity dual-mode PID based on virtual prediction of pressure fluctuation according to claim 1 is characterized by: The method for constructing the high-pressure pipeline flow model in step (3) is specifically as follows: Establish the momentum balance equation in the high-pressure pipeline: Where ρ is the fuel density, u is the flow rate, P is the pressure, D is the pipe diameter, and γ is the wall shear stress; Integrate the distance L between 1 / 4 and 3 / 4, and then multiply by the pipeline cross-sectional area A to obtain the mass flow rate out of the high-pressure pipeline: In the formula, ξ is a constant related to the speed of sound, and t is time; The fuel mass of the high pressure pipeline is expressed as: M pipe-out =∫Q pipe-out dt。 5. The method for real-time correction of injector injection quantity dual-mode PID based on virtual prediction of pressure fluctuation according to claim 1 is characterized by: The method for real-time correction of injection quantity by dual-mode PID in step (4) is specifically as follows: According to the current deviation between the actual injection amount and the target injection amount, the dual-mode PID switches between the coarse adjustment mode and the fine adjustment mode to calculate the injection pulse width correction amount: Where Δu(k) is the injection pulse width correction value at the kth injection, e(k) is the deviation between the actual injection amount and the target injection amount at the kth injection, and e # is the error threshold, is the correction coefficient in coarse adjustment mode, is the correction coefficient in fine tuning mode; The dual-mode PID outputs the control signal for the kth injection based on the calculated injection pulse width correction value: u(k)=u(k-1)+Δu(k).

Images