Diesel engine DOC outlet temperature control method
By simplifying the DOC model and design of the diesel engine, combined with the internal point method optimization, the problem of insufficient accuracy of the DOC outlet temperature control of the diesel engine is solved, and the stable temperature control and the improvement of DPF thermal regeneration efficiency are achieved.
Patent Information
- Application Number
- CN202410995119.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-24
- Publication Date
- 2024-11-15
AI Technical Summary
The prior art lacks accuracy in the DOC outlet temperature control of diesel engines, resulting in limited DPF thermal regeneration efficiency and safety, and insufficient calibration of model parameters leads to deviations in simulation and experimental data.
The gas phase temperature equation in the DOC model of the diesel engine was simplified by the feature line method, and the parameter calibration was performed using the backward difference formula and the Levinberg-Marquard method. The Longbug observer and the Liyapunov function were designed to ensure the stability of the observer. , and multi-objective optimization is carried out through the in-point method to build a nonlinear model prediction control system.
The precise control of the DOC outlet temperature of the diesel engine is achieved, and it is stable between 585℃ and 615℃, which improves the DPF thermal regeneration efficiency and safety, and reduces the deviation between the model and the experimental data.
Smart Images

Figure CN118959130A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a method in the technical field of diesel engine exhaust aftertreatment, and in particular to a method for controlling the DOC outlet temperature of a diesel engine. Background Art
[0002] Diesel particulate filter (DPF) is the core aftertreatment device of modern diesel engines to meet stringent emission standards. Its working principle is to trap the exhaust particulate matter in the wall-flow DPF and remove it through regular thermal regeneration when the engine is running to maintain low exhaust back pressure and ensure engine performance. To achieve thermal regeneration of DPF, fuel post-injection can be performed in the exhaust upstream of the diesel oxidation catalyst (DOC) or in the cylinder. These additional injected fuels are mainly converted into hydrocarbon reducing agents (HC), which increase the diesel DOC outlet temperature to about 600°C through oxidation exothermic reaction in the diesel DOC, thereby achieving DPF thermal regeneration. Accurate control of the set point operating temperature of the diesel DOC outlet is crucial to maintaining efficient DPF regeneration. However, due to the complex thermal response dynamic characteristics of the diesel DOC and the variability of driving and environmental conditions, improper control of HC injection may cause the internal temperature of the DPF to rise and damage the filter; or insufficient control may extend the thermal regeneration time and affect fuel economy.
[0003] Many scholars have conducted in-depth research on the optimization and closed-loop control strategies of diesel engine DOC. Eck et al. applied a control algorithm to adjust the diesel engine DOC outlet temperature of a small commercial vehicle diesel engine. The algorithm combined model-based control and gain scheduling techniques, and evaluated its performance on an engine test bench and chassis dynamometer. However, the system requires complex calculations and may not be applicable to all practical situations. Ning et al. proposed a composite controller combined with improved active disturbance suppression with high efficiency. The controller adopts model feedforward and time delay compensation strategies, and verifies its control performance of diesel engine DOC output temperature under steady-state and new European dynamic cycle conditions in the high-fidelity simulation of the Gamma Technologies-Power (GT-Power) model. It performs well and has effectiveness and robustness. However, when Ning et al. used GT-Power software for modeling, the model parameters used the software default values, resulting in deviations between the simulation results and the actual test data. Tiexiong et al. constructed a second-order diesel engine DOC model to achieve precise control of the outlet temperature. They proposed a control strategy that combines internal model control design and extended state observer, effectively integrating proportional integral differential (PID) feedback mechanism with active disturbance suppression. The strategy has been verified by test cycles and performed well. The partial differential equation based on the second-order diesel engine DOC model used in the research is a hyperbolic partial differential equation. Although the equations are not simplified, which may lead to a slower solution speed, the research results ensure the accuracy of the control effect. Bencherif et al. used a model predictive control method for DPF thermal control regeneration. Simulation tests were carried out and compared with traditional PID control. The results showed that the use of model predictive control method can obtain better control performance, with a maximum temperature deviation of about 33°C.
[0004] Diesel engine DOC models can be divided into three categories: first-order lumped parameter models, second-order lumped parameter models, and high-order lumped parameter models. The first-order diesel engine DOC model is faster in calculation speed but lower in accuracy, while the second-order diesel engine DOC model introduces gas-solid phase temperature difference and complex partial differential equations to improve accuracy, but the difficulty and time of solving the model increase. Although the high-order model pursues higher accuracy, it is complex and time-consuming to solve, and is not applicable in fast HC injection control. Research on diesel engine DOC control strategies has made some progress, but there are also several limitations. Some algorithms perform well in laboratory environments, but may be limited by complexity and versatility in practical applications. Insufficient parameter calibration may cause deviations between model simulation and experimental data. It has not been fully tested in more dynamic working condition verification. Simplified models may not accurately capture complex system responses, so their accuracy and robustness need to be improved before practical application. Therefore, accurate management and control of diesel engine DOC outlet gas temperature is crucial to ensure the efficiency and safety of regeneration. Summary of the invention
[0005] The technical problem to be solved by the present invention is to overcome the problem of inaccurate control of the DOC outlet temperature of a diesel engine in the prior art, and to provide a method for controlling the DOC outlet temperature of a diesel engine.
[0006] In order to solve the above technical problems, the present invention is implemented by adopting the following technical solutions: the method for controlling the DOC outlet temperature of a diesel engine comprises the following steps:
[0007] 1) According to the characteristics of the diesel engine DOC model, the model is simplified. The gas phase temperature equation in the model is simplified using the characteristic line method, and only the heat generated by HC injected into the diesel engine DOC is considered to simplify the solid phase temperature equation in the model;
[0008] 2) Based on the simplified diesel engine DOC model, the backward difference formula is used to solve the model, and the Levenberg-Marquardt method is used to calibrate the chemical reaction parameters at different temperatures;
[0009] 3) According to the characteristics of the diesel engine DOC model, the Lumberg observer is designed, and the Lyapunov function is designed to ensure the stability of the observer; based on the Lyapunov stability theory, the appropriate value range of the observer gain is calculated;
[0010] 4) Since the diesel engine DOC outlet temperature control is a multi-objective optimization problem, the cost function is designed according to the characteristics of the diesel engine DOC model, and the interior point method is used to solve the multi-objective optimization problem;
[0011] 5) Prove the stability of the diesel engine DOC outlet temperature nonlinear model predictive control system.
[0012] According to the characteristics of the diesel engine DOC model described in the technical solution, the model is simplified by using the characteristic line method to simplify the gas phase temperature equation in the model, and only the heat generated by HC injected into the diesel engine DOC is considered to simplify the solid phase temperature equation in the model, specifically:
[0013] Gas phase temperature equation: Considering the accumulation, convection and gas-solid heat transfer of flowing gas, the differential equation is constructed in combination with enthalpy balance as follows:
[0014]
[0015] Where: T is the gas phase temperature, T s is the solid phase temperature, k h is the heat transfer coefficient, ρ is the gas density, c p is the specific heat capacity of the gas, ε g is the pore area ratio of the catalyst coating, and u is the air velocity;
[0016] The characteristic line method is used to introduce new variables to simplify the equation. The characteristic line is a curve defined by the following equation:
[0017]
[0018] According to the above formula, we can get
[0019]
[0020] Where f(T) is an arbitrary function about T, so we introduce the variable ξ
[0021]
[0022] From the chain rule we can conclude that
[0023]
[0024] Finally, the simplified formula can be obtained in the defined characteristic line ξ=z-ut direction as follows:
[0025]
[0026] Solid phase temperature equation: Considering the enthalpy balance of the solid phase, including accumulation, axial heat conduction, gas-solid heat transfer, heat exchange with the surrounding environment (usually negligible for full-scale monomers, more important for small laboratory reactors) and heat sources from catalytic reactions, the differential equation is constructed in combination with the enthalpy balance as follows:
[0027]
[0028] Where: is the effective specific heat capacity of the solid phase, ρ s is the solid surface density, ΔH is the standard reaction enthalpy;
[0029] Because the DOC model in the present invention does not consider the heat released by the oxidation of HC, NO, and CO, and only considers the heat generated by HC injected into the DOC of the diesel engine, Ignoring this in the formula, the formula can be simplified to:
[0030]
[0031] Where: m HC is the HC injection amount, LHV is the lower heating value of the diesel engine, η HC is the conversion efficiency of HC. The conversion efficiency of HC is about 95% or more when the exhaust temperature is 300°C. Assuming η HC 95%;
[0032] In summary, the simplified diesel engine DOC control model can be obtained in the direction of the defined characteristic line ξ=z-ut.
[0033]
[0034] The backward difference formula described in the technical solution is used to solve the model, specifically:
[0035] The backward difference formula (BDF) is a numerical differentiation method used to find the numerical solution of ordinary differential equations. The calculation process is as follows:
[0036] First, choose appropriate initial values and determine the order of the differential equation;
[0037] During the model construction process, z = 0 is the gas inlet position, z = L is the gas outlet position, and the temperature boundary condition is
[0038] T z=0 =T in (t)
[0039]
[0040] Then, according to the selected BDF order, a suitable difference formula is selected. In each step, the value of the next time step is calculated using the BDF formula according to the function value and derivative value at the current moment. This includes calculating the weight coefficient and constructing the interpolation polynomial based on the data of the previous time steps.
[0041] Finally, by iterating this process until the desired time point or time period is reached, the numerical solution of the differential equation is obtained; it is worth noting that the stability and accuracy of the backward difference formula are affected by the choice of time step and the characteristics of the problem being solved, so the parameters need to be carefully adjusted to obtain accurate results.
[0042] The Levenberg-Marquardt method described in the technical solution is used to calibrate the chemical reaction parameters at different temperatures, specifically:
[0043] 1) Based on the given initial parameter values, calculate the current value of the objective function and its gradient;
[0044] 2) By introducing an adjustment parameter, combining the steepest descent method and the Gauss-Newton method, a new parameter vector is generated and the corresponding objective function value is calculated;
[0045] 3) Compare the new and old objective function values. If the new value is smaller, accept the new parameter vector and reduce the adjustment parameter.
[0046] 4) Otherwise, the old parameter vector is retained and the adjustment parameter is increased, and this process is repeated until the set convergence condition or number of iterations is reached.
[0047] According to the characteristics of the diesel engine DOC model, the technical solution designs a Lumburger observer and designs a Lyapunov function to ensure the stability of the observer. Based on the Lyapunov stability theory, the appropriate value range of the Lumburger observer gain is calculated, which is:
[0048] By properly selecting the observer gain, the rapid convergence of the estimation error can be ensured, thereby improving the accuracy and reliability of the diesel engine DOC temperature estimation. The Lumburg observer is designed as follows:
[0049]
[0050] Where: and It's T and T s The observed value, L1 and L2 are the gains of the observer;
[0051] Define the error as and Then the dynamic equation of the error is:
[0052]
[0053] Where b1, b2, b3> 0, and
[0054]
[0055] Choose the Lyapunov function as:
[0056]
[0057] Its derivative is
[0058]
[0059] So -(b1+L1)·e T 2 and are all negative values. The impact is not To be positive, you also need to ensure that:
[0060]
[0061] This can be achieved by choosing L1 and L2 large enough to cover the The positive impact caused
[0062] If the above equation holds, the selected Lyapunov function and its derivatives are finite and non-negative; this ensures that the estimation error will asymptotically approach zero, ensuring that the estimated state will eventually converge to the true value through the observer.
[0063]
[0064] The technical solution uses the interior point method to solve the multi-objective optimization problem, specifically:
[0065] The present invention proposes a nonlinear model predictive control method for large-scale system control problems. This method comprehensively uses the characteristic line method to deal with a complex system composed of hyperbolic partial differential equations (PDEs). The characteristic line method limits the solution direction of the hyperbolic partial differential equation to ξ=z-ut on the Z axis according to the actual flow of gas in the diesel engine DOC. The application of this method reduces the calculation amount of the entire partial differential equation system.
[0066] Combining the characteristics of the diesel engine DOC model and the constraints within the limited prediction range, the continuous-time optimal control problem can be expressed as
[0067]
[0068] Where: cost(T,T s ) is the cost function, u(t) is the control input time series, u(t) min is the minimum value of the control input, u(t) max is the maximum value of the control input, t r To control the start time, t p is the control end time, T is the total control time;
[0069]
[0070] Where: error1 = T out -873, error2=T s out -873,T out is the outlet gas temperature, T s out is the outlet solid phase temperature;
[0071] The optimization problem of the present invention is solved by the interior point method, which is an efficient optimization algorithm for solving linear, quadratic and nonlinear optimization problems;
[0072] 1) The algorithm first selects an initial point inside the feasible region;
[0073] 2) By introducing a barrier function, which tends to infinity on the boundary of the feasible region, the search path is restricted to the inside of the feasible region; the barrier function problem is solved by nonlinear optimization methods, such as Newton's method;
[0074] As the iteration proceeds, the parameter of the barrier function is reduced, so that the search path gradually approaches the boundary of the feasible region until the convergence condition is met. The interior point method has significant computational advantages.
[0075] The interior point method is usually more efficient than other algorithms (such as the simplex method), especially for large-scale problems; secondly, the method can usually converge to the optimal solution globally, and it can converge stably regardless of the type of problem and the initial point; in addition, the interior point method is applicable to a variety of optimization problems and can handle different types of constraints. Due to the introduction of the barrier function, the search path of the algorithm is smooth inside the feasible domain, which is conducive to the calculation of the gradient and Hessian matrix.
[0076] The technical solution described in the technical solution proves the stability of the diesel engine DOC outlet temperature nonlinear model prediction control system, specifically:
[0077] according to It can be seen from the function that the cost function is designed according to the control target at the outlet of the diesel engine DOC. After the terminal constraint is introduced, further proof is needed to ensure the stability of the closed-loop system;
[0078] According to the cost function, it can be further split into
[0079]
[0080] Where: u(t)=[u1,...u m ,u m+1 ,...u p ], m and p are the model prediction control domains, i=1,...,p;
[0081] Terminal constraints ensure that the state or output of the system converges to a specific value at the end of the prediction horizon; the following terminal constraints can be used to ensure the stability of closed-loop control systems
[0082]
[0083] Where: t k The control sequence calculated at each moment is u(t)=[u1,...u m ,u m+1 ,...u p+1 ];
[0084] If the equation satisfies both the cost function and the above equation, it means that the entire diesel engine DOC partial differential control system is stable. The following is the proof process;
[0085] Assume that at time t k The calculated optimal control sequence u(t)=[u1,...u m ,u m+1 ,...up ], at time point t k+1 The optimal control sequence calculated at is u(t)=[u2,...u m ,u m+1 ,...u p+1 ], then there is no need to calculate t k+1 The optimal control sequence at The calculation formula is
[0086]
[0087] According to the above formula, we can infer that Further deduction shows that
[0088] if converges to a steady-state positive value, then It can be determined that under the optimal control sequence, the calculated value of the output will converge to a steady-state value that is not equal to the calculated value of the cost function. At the same time, because the calculated value of the cost function can be 0, there will also be Converges to 0, converges to the calculated value of the output will converge to a steady-state value equal to the steady-state value corresponding to the cost function; then the optimal control sequence will definitely satisfy the following inequality
[0089]
[0090] therefore is a non-increasing positive function, and the output converges to the set point, so is a Lyapunov function; the above process proves the stability of the diesel engine DOC outlet temperature nonlinear model predictive control system.
[0091] Compared with the prior art, the present invention has the following beneficial effects:
[0092] 1. The diesel engine DOC outlet temperature control method described in the present invention is based on the partial differential equations of the gas flow and heat transfer process of the diesel engine DOC, and the characteristic line method is used to simplify the change of the gas phase temperature. In contrast, the present invention ignores the influence of HC, CO and NO in the exhaust gas on the system temperature, making the model more simplified;
[0093] 2. In order to achieve efficient nonlinear model predictive control, a diesel engine DOC outlet temperature control method described in the present invention constructs a Lumberg observer for the diesel engine DOC and designs a Lyapunov function to ensure the stability of the observer. The appropriate value range of the observer gain is determined through the Lyapunov stability theory, thereby minimizing the cost function within a limited prediction range and complying with the established constraints, and obtaining the optimal HC injection sequence;
[0094] 3. In order to meet the requirements of active DPF regeneration, a diesel engine DOC outlet temperature control method described in the present invention verifies the accuracy of the diesel engine DOC model and calculates the injection concentration of the inlet HC using a nonlinear model predictive control. Through this optimization control, the diesel engine DOC outlet temperature is successfully stably controlled between 585°C and 615°C. BRIEF DESCRIPTION OF THE DRAWINGS
[0095] The present invention will be further described below in conjunction with the accompanying drawings:
[0096] Figure 1 This is the working principle diagram of diesel engine DOC;
[0097] Figure 2 This is the chemical reaction diagram of diesel engine DOC;
[0098] Figure 3 It is a flow chart of a method for controlling the DOC outlet temperature of a diesel engine according to the present invention;
[0099] Figure 4 It is a flow chart of the diesel engine DOC nonlinear model prediction control in a diesel engine DOC outlet temperature control method according to the present invention;
[0100] Figure 5 A diesel engine torque speed percentage curve diagram under WHTC conditions in a diesel engine DOC outlet temperature control method according to the present invention;
[0101] Figure 6 A curve diagram of HC injection amount and exhaust mass flow rate in a diesel engine DOC outlet temperature control method according to the present invention;
[0102] Figure 7 It is an air velocity curve diagram of a diesel engine DOC in a diesel engine DOC outlet temperature control method according to the present invention;
[0103] Figure 8 A diesel engine DOC inlet and outlet temperature curve diagram in a diesel engine DOC outlet temperature control method according to the present invention;
[0104] Fig. 9 A comparison curve diagram of the experimental value model calculation value and the observed value in a diesel engine DOC outlet temperature control method according to the present invention;
[0105] Fig.10 It is a curve diagram of HC injection amount after optimization in a method for controlling the outlet temperature of a diesel engine DOC according to the present invention;
[0106] Fig.11It is a graph showing the optimized diesel engine DOC inlet temperature, outlet temperature and observed temperature in a diesel engine DOC outlet temperature control method according to the present invention;
[0107] Fig.12 A frequency distribution diagram of model calculated temperature and observed temperature in a diesel engine DOC outlet temperature control method according to the present invention;
[0108] Fig.13 This is a gas phase temperature distribution diagram of a diesel engine DOC after optimization in a diesel engine DOC outlet temperature control method described in the present invention. DETAILED DESCRIPTION
[0109] The present invention is described in detail below in conjunction with the accompanying drawings:
[0110] See also Figure 1 , the diesel oxidation catalyst (diesel DOC) plays a key role in the diesel engine emission reduction system, converting hydrocarbons and carbon monoxide in the exhaust. To ensure high efficiency, temperature control is essential. Real-time monitoring of the temperature sensor and closed-loop control strategy ensure the stable operation of the system.
[0111] See also Figure 2 , the exhaust gas first enters the diesel engine DOC, during which the HC, CO and particulate matter in the exhaust gas are effectively transferred to the catalyst surface. The diesel engine DOC is coated with catalysts composed of precious metals such as platinum or palladium. These catalysts can catalyze oxidation reactions and convert pollutants into more harmless substances. CO is oxidized into CO2 under the catalytic action, HC is oxidized into CO2 and H2O, some particulate matter is also oxidized, and some NO is oxidized into NO2. This series of reactions significantly reduces the content of harmful substances in the exhaust, converting them into carbon dioxide and water vapor, and converting some particulate matter into smaller particles or gaseous substances.
[0112] See also Figure 3 , a method for controlling the DOC outlet temperature of a diesel engine comprises the following steps:
[0113] 1. According to the characteristics of the diesel engine DOC model, the model is simplified. The gas phase temperature equation in the model is simplified using the characteristic line method, and only the heat generated by HC injected into the DOC is considered to simplify the solid phase temperature equation in the model;
[0114] According to the characteristics of the diesel engine DOC model, the model is simplified by using the characteristic line method to simplify the gas phase temperature equation in the model, and only considering the heat generated by HC injected into the DOC, the solid phase temperature equation in the model is simplified, specifically:
[0115] Gas phase temperature equation: Considering the accumulation, convection and gas-solid heat transfer of flowing gas, the differential equation is constructed in combination with enthalpy balance as follows:
[0116]
[0117] Where: T is the gas phase temperature, T s is the solid phase temperature, k h is the heat transfer coefficient, ρ is the gas density, c p is the specific heat capacity of the gas, ε g is the pore area ratio of the catalyst coating, and u is the air velocity;
[0118] The characteristic line method is used to introduce new variables to simplify the equation. The characteristic line is a curve defined by the following equation:
[0119]
[0120] According to the above formula, we can get
[0121]
[0122] Where f(T) is an arbitrary function about T, so we introduce the variable ξ
[0123]
[0124] From the chain rule we can conclude that
[0125]
[0126] Finally, the simplified formula can be obtained in the defined characteristic line ξ=z-ut direction as follows:
[0127]
[0128] Solid phase temperature equation: Considering the enthalpy balance of the solid phase, including accumulation, axial heat conduction, gas-solid heat transfer, heat exchange with the surrounding environment (usually negligible for full-scale monomers, more important for small laboratory reactors) and heat sources from catalytic reactions, the differential equation is constructed in combination with the enthalpy balance as follows:
[0129]
[0130] Where: is the effective specific heat capacity of the solid phase, ρ s is the solid surface density, ΔH is the standard reaction enthalpy;
[0131] Because the diesel engine DOC model in the present invention does not consider the heat released by the oxidation of HC, NO, and CO, and only considers the heat generated by HC injected into the diesel engine DOC, Ignoring this in the formula, the formula can be simplified to:
[0132]
[0133] Where: m HC is the HC injection amount, LHV is the lower heating value of the diesel engine, η HC is the conversion efficiency of HC. The conversion efficiency of HC is about 95% or more when the exhaust temperature is 300°C. Assuming η HC is 95%.
[0134] In summary, the simplified diesel engine DOC control model can be obtained in the direction of the defined characteristic line ξ=z-ut.
[0135]
[0136] The parameters of the diesel engine DOC control model are shown in Table 1.
[0137] Table 1 DOC control model parameters
[0138]
[0139] 2. Based on the simplified diesel engine DOC model, the backward difference formula is used to solve the model, and the Levenberg-Marquardt method is used to calibrate the chemical reaction parameters at different temperatures;
[0140] 1) The backward difference formula is used to solve the model, specifically:
[0141] The backward difference formula (BDF) is a numerical differentiation method used to find the numerical solution of ordinary differential equations. The calculation process is as follows:
[0142] First, choose appropriate initial values and determine the order of the differential equation;
[0143] During the model construction process, z = 0 is the gas inlet position, z = L is the gas outlet position, and the temperature boundary condition is
[0144] T z=0 =T in (t)
[0145]
[0146] Then, according to the selected BDF order, a suitable difference formula is selected. In each step, the value of the next time step is calculated using the BDF formula according to the function value and derivative value at the current moment. This includes calculating the weight coefficient and constructing the interpolation polynomial based on the data of the previous time steps.
[0147] Finally, by iterating this process until the desired time point or time period is reached, the numerical solution of the differential equation is obtained; it is worth noting that the stability and accuracy of the backward difference formula are affected by the choice of time step and the characteristics of the problem being solved, so the parameters need to be carefully adjusted to obtain accurate results;
[0148] 2) The Levenberg-Marquardt method is used to calibrate the chemical reaction parameters at different temperatures, specifically:
[0149] (1) Based on the given initial parameter values, calculate the current value of the objective function and its gradient;
[0150] (2) By introducing an adjustment parameter, combining the steepest descent method and the Gauss-Newton method, a new parameter vector is generated and the corresponding objective function value is calculated;
[0151] (3) Compare the new and old objective function values. If the new value is smaller, accept the new parameter vector and reduce the adjustment parameter.
[0152] (4) Otherwise, the old parameter vector is retained and the adjustment parameter is increased, and this process is repeated until the set convergence condition or number of iterations is reached;
[0153] The key to the Levenberg-Marquardt method is to balance the advantages and disadvantages of the steepest descent method and the Gauss-Newton method by continuously adjusting the adjustment parameters, thereby achieving efficient parameter identification;
[0154] 3. According to the characteristics of diesel engine DOC model, the Lumberg observer is designed, and the Lyapunov function is designed to ensure the stability of the observer; based on the Lyapunov stability theory, the appropriate value range of the Lumberg observer gain is calculated;
[0155] The Lumberg observer is designed, and the Lyapunov function is designed to ensure the stability of the observer; based on the Lyapunov stability theory, the appropriate value range of the observer gain is calculated, which is specifically:
[0156] By selecting the observer gain, the rapid convergence of the estimation error can be ensured, thereby improving the accuracy and reliability of the diesel engine DOC temperature estimation. The Lumburg observer is designed as follows
[0157]
[0158] Where: and It's T and T s The observed value, L1 and L2 are the gains of the observer;
[0159] Define the error as and Then the dynamic equation of the error is:
[0160]
[0161] Where b1, b2, b3> 0, and
[0162]
[0163] Choose the Lyapunov function as:
[0164]
[0165] Its derivative is
[0166]
[0167] So -(b1+L1)·e T 2 and are all negative values. The impact is not To be positive, you also need to ensure that:
[0168]
[0169] This can be achieved by choosing L1 and L2 large enough to cover the The positive impact caused;
[0170] If the above equation holds, the selected Lyapunov function and its derivatives are finite and non-negative; this ensures that the estimation error will asymptotically approach zero, ensuring that the estimated state will eventually converge to the true value through the observer.
[0171]
[0172] The appropriate value range of the Lumberg observer gain is calculated.
[0173] 4. Since the diesel engine DOC outlet temperature control is a multi-objective optimization problem, the cost function is designed according to the characteristics of the diesel engine DOC model, and the interior point method is used to solve the multi-objective optimization problem;
[0174] The cost function is designed according to the characteristics of the diesel engine DOC model, and the interior point method is used to solve the multi-objective optimization problem, specifically:
[0175] The present invention proposes a nonlinear model predictive control method for large-scale system control problems. This method comprehensively uses the characteristic line method to deal with a complex system composed of hyperbolic partial differential equations (PDEs). The characteristic line method limits the solution direction of the hyperbolic partial differential equation to ξ=z-ut on the Z axis according to the actual flow of gas in the diesel engine DOC. The application of this method reduces the calculation amount of the entire partial differential equation system.
[0176] Combining the characteristics of the diesel engine DOC model and the constraints within the limited prediction range, the continuous-time optimal control problem can be expressed as
[0177]
[0178] Where: cost(T,T s ) is the cost function, u(t) is the control input time series, u(t) min is the minimum value of the control input, u(t) max is the maximum value of the control input, t r To control the start time, t p is the control end time, T is the total control time;
[0179]
[0180] Where: error1 = T out -873, error2=T s out -873,T out is the outlet gas temperature, T s out is the outlet solid phase temperature;
[0181] The optimization problem of the present invention is solved by the interior point method, which is an efficient optimization algorithm used to solve linear, quadratic and nonlinear optimization problems.
[0182] 1) The algorithm first selects an initial point inside the feasible region;
[0183] 2) By introducing a barrier function, which tends to infinity on the boundary of the feasible region, the search path is restricted to the inside of the feasible region; the barrier function problem is solved by nonlinear optimization methods, such as Newton's method;
[0184] As the iteration proceeds, the parameters of the barrier function are reduced so that the search path gradually approaches the boundary of the feasible region until the convergence condition is met. The interior point method has significant computational advantages.
[0185] The interior point method is usually more efficient than other algorithms (such as the simplex method), especially for large-scale problems. Secondly, the method can usually converge to the optimal solution globally. In addition, it can converge stably regardless of the type of problem and the initial point. In addition, the interior point method is applicable to a variety of optimization problems and can handle different types of constraints. Due to the introduction of the barrier function, the search path of the algorithm is smooth inside the feasible domain, which is conducive to the calculation of the gradient and Hessian matrix.
[0186] Diesel engine DOC outlet temperature control flow chart as follows Figure 4 As shown;
[0187] 5. Prove the stability of the diesel engine DOC outlet temperature nonlinear model predictive control system;
[0188] The above-mentioned proof of the stability of the diesel engine DOC outlet temperature nonlinear model predictive control system is specifically as follows:
[0189] according to It can be seen from the function that the cost function is designed according to the control target at the outlet of the diesel engine DOC. After the terminal constraint is introduced, further proof is needed to ensure the stability of the closed-loop system;
[0190] According to the cost function, it can be further split into
[0191]
[0192] Where: u(t)=[u1,...u m ,u m+1 ,...u p ], m and p are the model prediction control domains, i=1,...,p;
[0193] Terminal constraints ensure that the state or output of the system converges to a specific value at the end of the prediction horizon; the following terminal constraints can be used to ensure the stability of closed-loop control systems
[0194]
[0195] Where: t k The control sequence calculated at each moment is u(t)=[u1,...u m ,u m+1 ,…u p+1 ];
[0196] If the equation satisfies both the cost function and the above equation, it means that the entire diesel engine DOC partial differential control system is stable. The following is the proof process;
[0197] Assume that at time t k The calculated optimal control sequence u(t)=[u1,…um ,u m+1 ,...u p ], at time point t k+1 The optimal control sequence calculated at is u(t)=[u2,...u m ,u m+1 ,...u p+1 ], then there is no need to calculate t k+1 The optimal control sequence at The calculation formula is
[0198]
[0199] According to the above formula, we can infer that Further deduction shows that
[0200] if converges to a steady-state positive value, then It can be determined that under the optimal control sequence, the calculated value of the output will converge to a steady-state value that is not equal to the calculated value of the cost function. At the same time, because the calculated value of the cost function can be 0, there will also be Converges to 0, converges to the calculated value of the output will converge to a steady-state value equal to the steady-state value corresponding to the cost function; then the optimal control sequence will definitely satisfy the following inequality
[0201]
[0202] therefore is a non-increasing positive function, and the output converges to the set point, so is a Lyapunov function; the above process proves that the diesel engine DOC outlet temperature nonlinear model predictive control system in this paper is stable.
[0203] Example:
[0204] To verify the present invention, a diesel engine meeting the National VI emission standard was used to conduct a bench test under the World Harmonized Light Vehicle Test Cycle (WHTC) condition. The diesel engine torque speed percentage under the WHTC condition is shown in Figure 5 , HC injection amount and scheduled mass flow rate see Figure 6 , air speed see Figure 7 , diesel engine DOC inlet and outlet temperatures see Figure 8 ;
[0205] According to the steps of the present invention and the characteristics of experimental data, a MATLAB program was independently written to solve the diesel engine DOC model, calibrate the chemical reaction related parameters, and apply the Lumburg observer algorithm and nonlinear model predictive control algorithm;
[0206] The engine test data ( Figure 6 , Figure 7 and Figure 8 ) Input the written MATLAB program for calculation, and the comparison between the experimental value model calculation value and the observed value is shown in Fig. 9 , the optimized HC injection amount is shown in Fig.10 , the optimized diesel engine DOC inlet temperature, outlet temperature and observed temperature are shown in Fig.11 , the frequency distribution diagram of model calculated temperature and observed temperature is shown in Fig.12 ;
[0207] The gas phase temperature distribution inside the DOC catalyst of the diesel engine after optimization is shown in Fig.13 , the model calculation value shows that 96.03% is within the target temperature range, indicating that the model can accurately control the temperature in most cases, and the distribution is relatively concentrated, indicating that its control stability and reliability are high; however, the mean deviates from the target temperature midpoint by -3.49℃, and the standard deviation is 56.75℃, reflecting that most data are close to the target range; among the observed values, 79.49% of the data fall within the target temperature range, but the distribution is relatively dispersed and the volatility is large. The mean deviates from the target temperature midpoint by -2.49℃, and the standard deviation is 56.81℃, indicating that the accuracy and stability of the observed data are lower than the model calculation value;
[0208] In general, the model calculation values performed better than the observed values within the target temperature range, with higher accuracy and stability, providing a reliable basis for temperature control;
[0209] In order to meet the requirements of active DPF regeneration, the present invention calculates the injection concentration of inlet HC based on nonlinear model predictive control and uses it for model calculation; through this optimization control, the outlet temperature of the diesel engine DOC can be stably controlled between 585°C and 615°C;
[0210] Although the embodiments of the present invention have been disclosed in the specification and the embodiments, they are not limited to the contents described. They can be widely applied to various related fields and can be modified as needed. Therefore, the present invention is not limited by specific details or illustrations without exceeding the general concept of the claims and their equivalents.
Claims
1. A method for controlling the DOC outlet temperature of a diesel engine, characterized in that: The method for controlling the DOC outlet temperature of a diesel engine comprises the following steps: 1) According to the characteristics of the diesel engine DOC model, the model is simplified. The gas phase temperature equation in the model is simplified using the characteristic line method, and only the heat generated by HC injected into the diesel engine DOC is considered to simplify the solid phase temperature equation in the model; 2) Based on the simplified diesel engine DOC model, the backward difference formula is used to solve the model, and the Levenberg-Marquardt method is used to calibrate the chemical reaction parameters at different temperatures; 3) According to the characteristics of diesel engine DOC model, the Lumberg observer is designed, and the Lyapunov function is designed to ensure the stability of the observer; Based on Lyapunov stability theory, the appropriate range of observer gain is calculated. 4) Since the diesel engine DOC outlet temperature control is a multi-objective optimization problem, the cost function is designed according to the characteristics of the diesel engine DOC model, and the interior point method is used to solve the multi-objective optimization problem; 5) Prove the stability of the diesel engine DOC outlet temperature nonlinear model predictive control system.
2. A method for controlling the DOC outlet temperature of a diesel engine according to claim 1, characterized in that: According to the characteristics of the diesel engine DOC model, the model is simplified by using the characteristic line method to simplify the gas phase temperature equation in the model, and only the heat generated by HC injected into the diesel engine DOC is considered to simplify the solid phase temperature equation in the model, specifically: Gas phase temperature equation: Considering the accumulation, convection and gas-solid heat transfer of flowing gas, the differential equation is constructed in combination with enthalpy balance as follows: Where: T is the gas phase temperature, T s is the solid phase temperature, k h is the heat transfer coefficient, ρ is the gas density, c p is the specific heat capacity of the gas, ε g is the pore area ratio of the catalyst coating, and u is the air velocity; The characteristic line method is used to introduce new variables to simplify the equation. The characteristic line is a curve defined by the following equation: According to the above formula, we can get Where f(T) is an arbitrary function about T, so we introduce the variable ξ From the chain rule we can conclude that Finally, the simplified formula can be obtained in the defined characteristic line ξ=z-ut direction as follows: Solid phase temperature equation: Considering the enthalpy balance of the solid phase, including accumulation, axial heat conduction, gas-solid heat transfer, heat exchange with the surrounding environment (usually negligible for full-scale monomers, more important for small laboratory reactors) and heat sources from catalytic reactions, the differential equation is constructed in combination with the enthalpy balance as follows: Where: is the effective specific heat capacity of the solid phase, ρ s is the solid surface density, ΔH is the standard reaction enthalpy; Because the DOC model in the present invention does not consider the heat released by the oxidation of HC, NO, and CO, and only considers the heat generated by HC injected into the DOC of the diesel engine, Ignoring this in the formula, the formula can be simplified to: Where: m HC is the HC injection amount, LHV is the lower heating value of the diesel engine, η HC is the conversion efficiency of HC. The conversion efficiency of HC is about 95% or more when the exhaust temperature is 300°C. Assuming η HC 95%; In summary, the simplified diesel engine DOC control model can be obtained in the direction of the defined characteristic line ξ=z-ut.
3. A method for controlling the DOC outlet temperature of a diesel engine according to claim 1, characterized in that: The backward difference formula is used to solve the model, specifically: The backward difference formula (BDF) is a numerical differentiation method used to find the numerical solution of ordinary differential equations. The calculation process is as follows: First, choose appropriate initial values and determine the order of the differential equation; During the model construction process, z = 0 is the gas inlet position, z = L is the gas outlet position, and the temperature boundary condition is T z=0 =T in (t) Then, according to the selected BDF order, a suitable difference formula is selected. In each step, the value of the next time step is calculated using the BDF formula according to the function value and derivative value at the current moment. This includes calculating the weight coefficient and constructing the interpolation polynomial based on the data of the previous time steps. Finally, by iterating this process until the desired time point or time period is reached, the numerical solution of the differential equation is obtained; it is worth noting that the stability and accuracy of the backward difference formula are affected by the choice of time step and the characteristics of the problem being solved, so the parameters need to be carefully adjusted to obtain accurate results.
4. A method for controlling the DOC outlet temperature of a diesel engine according to claim 1, characterized in that: The Levenberg-Marquardt method is used to calibrate the chemical reaction parameters at different temperatures, specifically: 1) Based on the given initial parameter values, calculate the current value of the objective function and its gradient; 2) By introducing an adjustment parameter, combining the steepest descent method and the Gauss-Newton method, a new parameter vector is generated and the corresponding objective function value is calculated; 3) Compare the new and old objective function values. If the new value is smaller, accept the new parameter vector and reduce the adjustment parameter. 4) Otherwise, the old parameter vector is retained and the adjustment parameter is increased, and this process is repeated until the set convergence condition or number of iterations is reached.
5. A method for controlling the DOC outlet temperature of a diesel engine according to claim 1, characterized in that: According to the characteristics of the diesel engine DOC model, the Lumberg observer is designed, and the Lyapunov function is designed to ensure the stability of the observer; based on the Lyapunov stability theory, the appropriate value range of the Lumberg observer gain is calculated, which is specifically: By properly selecting the observer gain, the rapid convergence of the estimation error can be ensured, thereby improving the accuracy and reliability of the diesel engine DOC temperature estimation. The Lumburg observer is designed as follows: Where: and It's T and T s The observed value, L1 and L2 are the gains of the observer; Define the error as and Then the dynamic equation of the error is: Where b1, b2, b3> 0, and Choose the Lyapunov function as: Its derivative is So -(b1+L1)·e T 2 and are all negative values. The impact is not To be positive, you also need to ensure that: This can be achieved by choosing L1 and L2 large enough to cover the The positive impact caused If the above equation holds, the selected Lyapunov function and its derivatives are finite and non-negative; this ensures that the estimation error will asymptotically approach zero, ensuring that the estimated state will eventually converge to the true value through the observer.
6. A method for controlling the DOC outlet temperature of a diesel engine according to claim 1, characterized in that: The interior point method is used to solve the multi-objective optimization problem, specifically: The present invention proposes a nonlinear model predictive control method for large-scale system control problems. This method comprehensively uses the characteristic line method to deal with a complex system composed of hyperbolic partial differential equations (PDEs). The characteristic line method limits the solution direction of the hyperbolic partial differential equation to ξ=z-ut on the Z axis according to the actual flow of gas in the diesel engine DOC. The application of this method reduces the calculation amount of the entire partial differential equation system. Combining the characteristics of the diesel engine DOC model and the constraints within the limited prediction range, the continuous-time optimal control problem can be expressed as Where: cost(T,T s ) is the cost function, u(t) is the control input time series, u(t) min is the minimum value of the control input, u(t) max is the maximum value of the control input, t r To control the start time, t p is the control end time, T is the total control time; Where: error1 = T out -873, error2=T s out -873,T out is the outlet gas temperature, T s out is the outlet solid phase temperature; The optimization problem of the present invention is solved by the interior point method, which is an efficient optimization algorithm for solving linear, quadratic and nonlinear optimization problems; 1) The algorithm first selects an initial point inside the feasible region; 2) By introducing a barrier function, which tends to infinity on the boundary of the feasible domain, the search path is restricted to the inside of the feasible domain; The barrier function problem is solved by nonlinear optimization methods, such as Newton's method; As the iteration proceeds, the parameter of the barrier function is reduced, so that the search path gradually approaches the boundary of the feasible region until the convergence condition is met. The interior point method has significant computational advantages. Interior point methods are often more efficient than other algorithms (such as the simplex method), especially for large problems; Second, the method can usually converge to the optimal solution globally. In addition, it can converge stably regardless of the problem type and the initial point. In addition, the interior point method is applicable to a variety of optimization problems and can handle different types of constraints. Due to the introduction of the barrier function, the search path of the algorithm is smooth inside the feasible domain, which is conducive to the calculation of gradients and Hessian matrices.
7. A method for controlling the DOC outlet temperature of a diesel engine according to claim 1, characterized in that: The above-mentioned proof of the stability of the diesel engine DOC outlet temperature nonlinear model predictive control system is specifically as follows: according to It can be seen from the function that the cost function is designed according to the control target at the outlet of the diesel engine DOC. After the terminal constraint is introduced, further proof is needed to ensure the stability of the closed-loop system; According to the cost function, it can be further split into Where: u(t)=[u1,...u m ,u m+1 ,...u p ], m and p are the model prediction control domains, i=1,...,p; Terminal constraints ensure that the state or output of the system converges to a specific value at the end of the prediction horizon; the following terminal constraints can be used to ensure the stability of closed-loop control systems Where: The control sequence u(t)=[u1,...u m ,u m+1 ,...u p+1 ]; If the equation satisfies both the cost function and the above equation, it means that the entire diesel engine DOC partial differential control system is stable. The following is the proof process; Assume that at time t k The calculated optimal control sequence u(t)=[u1,...u m ,u m+1 ,...u p ], at time point t k+1 The optimal control sequence calculated at is u(t)=[u2,...u m ,u m+1 ,...u p+1 ], then there is no need to calculate t k+1 The optimal control sequence at The calculation formula is According to the above formula, we can infer that Further deduction shows that if converges to a steady-state positive value, then It can be determined that under the optimal control sequence, the calculated value of the output will converge to a steady-state value that is not equal to the calculated value of the cost function. At the same time, because the calculated value of the cost function can be 0, there will also be Converges to 0, converges to the calculated value of the output will converge to a steady-state value equal to the steady-state value corresponding to the cost function; then the optimal control sequence will definitely satisfy the following inequality therefore is a non-increasing positive function, and the output converges to the set point, so is a Lyapunov function; the above process proves the stability of the diesel engine DOC outlet temperature nonlinear model predictive control system.
Citation Information
Cited By
Diesel engine SCR high-order model solving method
CN118780078A
