Method for optimizing parameters of asymmetric fuel injection nozzle of high-pressure common-rail fuel injector

Abstract

The invention relates to a parameter optimization method for an asymmetric fuel injection nozzle of a high-pressure common-rail fuel injector. Comprising the following construction steps of 1, constructing a parameter optimization model of the asymmetric fuel injection nozzle of the high-pressure common-rail fuel injector, 2, designing an optimization process of a fire lizard algorithm, and 3, testing the fuel injection quantity and the fuel injection rate of the high-pressure common-rail fuel injector through a high-pressure fuel injection test system. By means of the fire lizard algorithm, parameters of the asymmetric fuel injection nozzle of the high-pressure common-rail fuel injector can be optimized, the fuel injection quantity and the fuel injection rate of the high-pressure common-rail fuel injector can be increased, the use performance of the high-pressure common-rail fuel injector can be improved, and meanwhile the energy efficiency of a diesel engine is improved.

Description

Technical Field

[0001] The present invention relates to a method for optimizing parameters of a high-pressure common rail fuel injector, and in particular to a method for optimizing parameters of a high-pressure common rail fuel injector nozzle based on a Salamander algorithm. Background Art

[0002] Faced with growing energy demands and environmental pollution, diesel engines, with their high energy efficiency and fuel economy, have gained a central position in the fields of special equipment and heavy-duty transportation. Currently, research on high-pressure diesel engines is insufficient, and related equipment is unable to meet the actual needs of the government, enterprises, and the military industry. Therefore, optimizing high-pressure injection systems is crucial, and injector optimization is crucial for achieving high performance. The injector is a core component of a diesel engine's high-pressure common rail system. The injector's structural parameters play a decisive role in fuel flow characteristics. Conventional symmetrical injectors are prone to flow separation under high-frequency injection conditions, resulting in uneven fuel distribution. As system pressure continues to increase, cavitation occurs, compromising diesel engine reliability. Furthermore, research on the optimization of asymmetric injectors is insufficient. Asymmetric injectors can effectively suppress cavitation. Therefore, optimizing the geometric parameters of asymmetric common rail injectors can facilitate uniform fuel jet distribution. To improve the optimization of the geometric parameters of asymmetric nozzles for high-pressure common rail injectors, a high-precision and high-efficiency intelligent optimization algorithm is required. Currently, commonly used intelligent algorithms include genetic algorithms, particle swarm optimization algorithms, and simulated annealing algorithms, which are prone to falling into local optimality during injection parameter optimization. Summary of the Invention

[0003] The present invention addresses the above-mentioned problems in the prior art and provides a method for optimizing the asymmetric nozzle parameters of high-pressure common rail injectors based on the Salamander algorithm. By using the designed Salamander algorithm, the asymmetric nozzle parameters of high-pressure common rail injectors are optimized, thereby improving the optimization effect of the asymmetric nozzle parameters of high-pressure common rail injectors, thereby increasing the service life of high-pressure common rail injectors and improving the energy efficiency of diesel engines.

[0004] The technical solution of the present invention comprises the following steps:

[0005] Step 1: Construct a parameter optimization model for asymmetric nozzles of high-pressure common rail injectors

[0006] Construct the optimization objective function, taking the injection amount Q p and injection rate v p At the same time, the optimization goal is to achieve the maximum, with the oil inlet diameter D in , oil outlet diameter D out , needle valve maximum lift L max and nozzle diameter D oAs constraints, the model is as follows:

[0007] Max h(x)=[h1(x),h2(x)]' (1)

[0008] Where, h1(x)=Q p (x), h2(x)=v p (x).

[0009] st0.2mm≤D in ≤0.4mm;

[0010] 0.2mm≤D out ≤0.3mm;

[0011] 0.25mm≤L max ≤0.35mm;

[0012] 0.10mm≤D o ≤0.20mm.

[0013] Step 2: Design the optimization process of the Salamander algorithm

[0014] In order to improve the parameter optimization effect of asymmetric nozzles of high-pressure common rail injectors, the Salamander algorithm is proposed. The algorithm flow is as follows:

[0015] Step 2-1: Initialization of parameters

[0016] Initialize the parameters of the Salamander algorithm, including the population size N, the maximum number of iterations I max Parameter range represents the minimum value of the asymmetric nozzle parameter, Indicates the maximum value of the asymmetric nozzle parameter.

[0017] The population initialization is generated as follows:

[0018]

[0019] Where R represents the Latin hypercube sampling matrix and ⊙ represents element-by-element multiplication.

[0020] The initial fitness is calculated based on the digital twin model. The formula is as follows:

[0021]

[0022] Where g i represents the comprehensive fitness of the i-th individual, D S represents the Sauter mean diameter, μm; Q pi Indicates the actual injection amount corresponding to the i-th group of parameters, mm3 / ms;Q pt Indicates the actual injection amount, mm 3 / ms.

[0023] Step 2-2: Calculate the energy value of each individual using the following formula:

[0024]

[0025] Where g min Indicates the minimum fitness value of the current population, g max Indicates the maximum fitness value of the current population.

[0026] Step 2.3: Classify by energy:

[0027] When E i >μ E +0.5σ E , the i-th individual is the elite individual and is directly retained;

[0028] When μ E -0.5σ E ≤E i ≤μ E +0.5σ E , the i-th individual is a medium-energy individual and should undergo differential evolution;

[0029] When E i <μ E -0.5σ E , the i-th individual is a low-energy individual and should regenerate its broken limbs, thereby improving the global optimization ability of the algorithm.

[0030] Step 2.4: Perform differential evolution on medium energy individuals. For each medium energy individual x i , and evolve according to the following formula:

[0031] x i,new =x i +λ(x best -x d ) (5)

[0032] Where x i,new represents the updated individual; λ represents the scaling factor; x best Represents the elite individual with the largest energy value; x d represents a randomly selected reference individual in the population.

[0033] Step 2.5: Regeneration of limbs for low-energy individuals

[0034] First, for each constraint parameter m, calculate its sensitivity η m :

[0035]

[0036] Where Δ represents the perturbation step size, Δ=0.01(UB m -LB m ), Represents the unit basis vector (the mth component is 1 and the rest are 0).

[0037] Select the parameter with the highest sensitivity for mutation:

[0038] l=argmax(η m ) (7)

[0039] Next, calculate the decay probability:

[0040]

[0041] When t is small, P s (t) is a high probability mutation, mainly for exploration; when t is close to T, P s (t) is a low-probability mutation, mainly for development. When a mutation is triggered, noise is added to the parameter l:

[0042]

[0043] Where ρ represents the adjustment coefficient, σ l represents the standard deviation; Represents a Gaussian distributed random number with mean 0.

[0044] Finally, when the mutation exceeds the boundary [LB l ,UB l ], then regenerate the parameter:

[0045]

[0046] Where, Represents a uniformly distributed random number.

[0047] Step 2.6: When the fuel temperature change ΔT>10K and the rail pressure fluctuation ΔP>5MPa are detected, the optimization parameters will be adjusted in real time to maintain the performance of the Salamander algorithm. The parameter adjustment formula is as follows:

[0048] Δx=αΔT+βΔP (11)

[0049] Where α represents the temperature factor and β represents the pressure factor.

[0050] Step 2.7: In each iteration, the population is updated and evaluated to ensure the correctness of the optimization direction and convergence.

[0051] First, merge the new population and merge the three types of individuals into a new generation population:

[0052]

[0053] Where, represents the synthetic new generation population; Indicates elite population; represents the medium energy population; Represents a low-energy population.

[0054] Second, check the boundaries and truncate the parameters of each individual to ensure that they do not exceed the feasible range:

[0055]

[0056] Where x i,j represents the jth parameter of the i-th individual.

[0057] Finally, recalculate the fitness g i .

[0058] Step 2.8: When or t≥I max When , output the global optimal solution; otherwise, return to step 2.1.

[0059] Step 3: Test the injection quantity and injection rate of the high-pressure common rail injector through the high-pressure fuel injection test system to verify the effect of the parameter optimization of the asymmetric injection nozzle of the high-pressure common rail injector.

[0060] The advantages and effects of the present invention are as follows:

[0061] The Salamander algorithm constructed by the present invention optimizes the parameters of asymmetric nozzles for high-pressure common rail injectors. The designed Salamander algorithm achieves superior optimization results. Consequently, the injection volume and injection rate of the high-pressure common rail injector can be effectively improved after parameter optimization. DETAILED DESCRIPTION

[0062] Example

[0063] The technical solution of the present invention comprises the following steps:

[0064] Step 1: Construct a parameter optimization model for asymmetric nozzles of high-pressure common rail injectors

[0065] Construct the optimization objective function, taking the injection amount Q p and injection rate v p At the same time, the optimization goal is to achieve the maximum, with the oil inlet diameter D in , oil outlet diameter D out , needle valve maximum lift L maxand nozzle diameter D o As constraints, the model is as follows:

[0066] Max h(x)=[h1(x),h2(x)]' (1)

[0067] Where, h1(x)=Q p (x), h2(x)=v p (x).

[0068] st0.2mm≤D in ≤0.4mm;

[0069] 0.2mm≤D out ≤0.3mm;

[0070] 0.25mm≤L max ≤0.35mm;

[0071] 0.10mm≤D o ≤0.20mm.

[0072] Step 2: Design the optimization process of the Salamander algorithm

[0073] In order to improve the parameter optimization effect of asymmetric nozzles of high-pressure common rail injectors, the Salamander algorithm is proposed. The algorithm flow is as follows:

[0074] Step 2-1: Initialization of parameters

[0075] Initialize the parameters of the Salamander algorithm, including the population size N, the maximum number of iterations I max Parameter range represents the minimum value of the asymmetric nozzle parameter, Indicates the maximum value of the asymmetric nozzle parameter.

[0076] The population initialization is generated as follows:

[0077]

[0078] Where R represents the Latin hypercube sampling matrix and ⊙ represents element-by-element multiplication.

[0079] The initial fitness is calculated based on the digital twin model. The formula is as follows:

[0080]

[0081] Where g i represents the comprehensive fitness of the i-th individual, D S represents the Sauter mean diameter, μm; Q piIndicates the actual injection amount corresponding to the i-th group of parameters, mm 3 / ms;Q pt Indicates the actual injection amount, mm 3 / ms.

[0082] Step 2-2: Calculate the energy value of each individual using the following formula:

[0083]

[0084] Where g min Indicates the minimum fitness value of the current population, g max Indicates the maximum fitness value of the current population.

[0085] Step 2.3: Classify by energy:

[0086] When E i >μ E +0.5σ E , the i-th individual is the elite individual and is directly retained;

[0087] When μ E -0.5σ E ≤E i ≤μ E +0.5σ E , the i-th individual is a medium-energy individual and should undergo differential evolution;

[0088] When E i <μ E -0.5σ E , the i-th individual is a low-energy individual and should regenerate its broken limbs, thereby improving the global optimization ability of the algorithm.

[0089] Step 2.4: Perform differential evolution on medium energy individuals. For each medium energy individual x i , and evolve according to the following formula:

[0090] x i,new =x i +λ(x best -x d ) (5)

[0091] Where x i,new represents the updated individual; λ represents the scaling factor; x best Represents the elite individual with the largest energy value; x d represents a randomly selected reference individual in the population.

[0092] Step 2.5: Regeneration of limbs for low-energy individuals

[0093] First, for each constraint parameter m, calculate its sensitivity η m:

[0094]

[0095] Where Δ represents the perturbation step size, Δ=0.01(UB m -LB m ), Represents the unit basis vector (the mth component is 1 and the rest are 0).

[0096] Select the parameter with the highest sensitivity for mutation:

[0097] l=argmax(η m ) (7)

[0098] Next, calculate the decay probability:

[0099]

[0100] When t is small, P s (t) is a high probability mutation, mainly for exploration; when t is close to T, P s (t) is a low-probability mutation, mainly for development. When a mutation is triggered, noise is added to the parameter l:

[0101]

[0102] Where ρ represents the adjustment coefficient, σ l represents the standard deviation; Represents a Gaussian distributed random number with mean 0.

[0103] Finally, when the mutation exceeds the boundary [LB l ,UB l ], then regenerate the parameter:

[0104]

[0105] Where, Represents a uniformly distributed random number.

[0106] Step 2.6: When the fuel temperature change ΔT>10K and the rail pressure fluctuation ΔP>5MPa are detected, the optimization parameters will be adjusted in real time to maintain the performance of the Salamander algorithm. The parameter adjustment formula is as follows:

[0107] Δx=αΔT+βΔP (11)

[0108] Where α represents the temperature factor and β represents the pressure factor.

[0109] Step 2.7: In each iteration, the population is updated and evaluated to ensure the correctness of the optimization direction and convergence.

[0110] First, merge the new population and merge the three types of individuals into a new generation population:

[0111]

[0112] Where, represents the synthetic new generation population; Indicates elite population; represents the medium energy population; Represents a low-energy population.

[0113] Second, check the boundaries and truncate the parameters of each individual to ensure that they do not exceed the feasible range:

[0114]

[0115] Where x i,j represents the jth parameter of the i-th individual.

[0116] Finally, recalculate the fitness g i .

[0117] Step 2.8: When or t≥I max When , output the global optimal solution; otherwise, return to step 2.1.

[0118] Step 3: Test the injection quantity and injection rate of the high-pressure common rail injector through the high-pressure fuel injection test system to verify the effect of the parameter optimization of the asymmetric injection nozzle of the high-pressure common rail injector.

[0119] The specific embodiments are as follows:

[0120] In the embodiment, a high-pressure common rail injector is selected as a research object, and its parameters are optimized using the Salamander algorithm. The parameters of the Salamander algorithm are set as follows: N=100, λ=0.5, and an initial decay probability of 0.45.

[0121] The injection quantity and injection rate of the high-pressure common rail injector were tested using a high-pressure fuel injection test system to verify the injection performance of the asymmetric nozzle of the high-pressure common rail injector before and after optimization.

[0122] The structural parameters of the asymmetric common rail injector nozzle before and after optimization are shown in Table 1. The injection volume and injection rate of the common rail injector before and after optimization are shown in Table 2. As can be seen from Table 2, the optimization of the injector nozzle geometry significantly improves both the injection volume and injection rate, thereby enhancing the injection performance of the common rail injection and verifying the optimization performance of the Salamander algorithm.

[0123] Table 1 Structural parameters of high-pressure common rail injector nozzle before and after optimization

[0124] parameter Before optimization After optimization <![CDATA[Inlet diameter D in > 0.25 0.32 <![CDATA[Outlet diameter D out > 0.22 0.24 <![CDATA[Maximum lift L of the needle valve max > 0.27 0.30 <![CDATA[Jet hole diameter D o > 0.12 0.14

[0125] Table 2 Injection amount and injection rate of high common rail injector before and after optimization

[0126]

[0127] The pipeline network of a catalytic cracking unit in a petrochemical company was selected as the research object. A system of integrity evaluation indicators was constructed, as shown in Table 1.

[0128] Table 1: Pipeline network integrity evaluation index system for a petrochemical company's catalytic cracking unit

[0129]

[0130] Table 2 Weights of pipeline network integrity evaluation indicators

[0131]

[0132] By collecting data related to the integrity evaluation indicators of the petrochemical company's catalytic cracking unit from January 2020 to October 2021, a total of 22 sets of data were collected on a monthly basis. The first 16 sets of data were used as verification samples, and the next 6 sets of data were used as test samples. The trained evaluation model was used to test the last 6 sets of data. The test results are shown in Table 3.

[0133] Table 3 Evaluation results of test samples

[0134] Sample No. Output evaluation value Integrity Level Evaluation error Runtime 1 0.82 good 4.3％ 5.43s 2 0.74 middle 3.7％ 5.15s 3 0.79 middle 4.0％ 4.89s 4 0.80 good 3.7％ 5.03s 5 0.82 good 4.1％ 4.88s 6 0.77 good 3.9％ 5.05s

[0135] The results in Table 2 show that the error in pipeline network integrity assessment using the Gabor wavelet neural network optimized by the improved chicken swarm algorithm ranges from 3.7% to 4.1%, demonstrating that this evaluation method has high accuracy. Furthermore, the runtime for pipeline network integrity assessment using this evaluation method ranges from 4.89 seconds to 5.43 seconds, demonstrating that this evaluation method also has very high efficiency. These results demonstrate that the pipeline network integrity evaluation model constructed in this invention can efficiently and accurately assess the integrity level of a pipeline network, thereby providing a favorable theoretical basis for pipeline network maintenance and ensuring its safety and reliability.

Claims

1. A method for optimizing the parameters of asymmetric nozzles of a high-pressure common rail injector, characterized in that The steps include: Step 1: Construct a parameter optimization model for asymmetric nozzles of high-pressure common rail injectors Construct the optimization objective function, taking the injection amount Q p and injection rate v p At the same time, the optimization goal is to achieve the maximum, with the oil inlet diameter D in , oil outlet diameter D out , needle valve maximum lift L max and nozzle diameter D o As constraints, the model is as follows: Max h(x)=[h1(x),h2(x)]' (1) Where, h1(x)=Q p (x), h2(x)=v p (x). s.t.0.2mm≤D in ≤0.4mm; 0.2mm≤D out ≤0.3mm; 0.25mm≤L max ≤0.35mm; 0.10mm≤D o ≤0.20mm。 Step 2: Design the optimization process of the Salamander algorithm Step 3: Test the injection quantity and injection rate of the high-pressure common rail injector through the high-pressure fuel injection test system to verify the effect of the parameter optimization of the asymmetric injection nozzle of the high-pressure common rail injector.

2. The method for optimizing the parameters of asymmetric nozzles of a high-pressure common rail injector according to claim 1, characterized in that The optimization process steps for designing the Salamander algorithm are as follows: Step 2-1: Initialization of parameters Initialize the parameters of the Salamander algorithm, including the population size N, the maximum number of iterations I max Parameter range represents the minimum value of the asymmetric nozzle parameter, Indicates the maximum value of the asymmetric nozzle parameter. The population initialization is generated as follows: Where R represents the Latin hypercube sampling matrix and ⊙ represents element-by-element multiplication. The initial fitness is calculated based on the digital twin model. The formula is as follows: Where g i represents the comprehensive fitness of the i-th individual, D S represents the Sauter mean diameter, μm; Q pi Indicates the actual injection amount corresponding to the i-th group of parameters, mm 3 / ms;Q pt Indicates the actual injection amount, mm 3 / ms. Step 2-2: Calculate the energy value of each individual using the following formula: Where g min Indicates the minimum fitness value of the current population, g max Indicates the maximum fitness value of the current population. Step 2.3: Classify by energy: When E i >μ E +0.5σ E , the i-th individual is the elite individual and is directly retained; When μ E -0.5σ E ≤E i ≤μ E +0.5σ E , the i-th individual is a medium-energy individual and should undergo differential evolution; When E i <μ E -0.5σ E , the i-th individual is a low-energy individual and should regenerate its broken limbs, thereby improving the global optimization ability of the algorithm. Step 2.4: Perform differential evolution on medium energy individuals. For each medium energy individual x i , and evolve according to the following formula: x i,new =x i +λ(x best -x d ) (5) Where x i,new represents the updated individual; λ represents the scaling factor; x best Represents the elite individual with the largest energy value; x d represents a randomly selected reference individual in the population. Step 2.5: Regeneration of limbs for low-energy individuals First, for each constraint parameter m, calculate its sensitivity η m : Where Δ represents the perturbation step size, Δ=0.01(UB m -LB m ), Represents the unit basis vector (the mth component is 1 and the rest are 0). Select the parameter with the highest sensitivity for mutation: l=argmax(η m ) (7) Next, calculate the decay probability: When t is small, P s (t) is a high probability mutation, mainly for exploration; when t is close to T, P s (t) is a low-probability mutation, mainly for development. When a mutation is triggered, noise is added to the parameter l: Where ρ represents the adjustment coefficient, σ l represents the standard deviation; Represents a Gaussian distributed random number with mean 0. Finally, when the mutation exceeds the boundary [LB l ,UB l ], then regenerate the parameter: Where, Represents a uniformly distributed random number. Step 2.6: When the fuel temperature change ΔT>10K and the rail pressure fluctuation ΔP>5MPa are detected, the optimization parameters will be adjusted in real time to maintain the performance of the Salamander algorithm. The parameter adjustment formula is as follows: Δx=αΔT+βΔP (11) Where α represents the temperature factor and β represents the pressure factor. Step 2.7: In each iteration, the population is updated and evaluated to ensure the correctness of the optimization direction and convergence. First, merge the new population and merge the three types of individuals into a new generation population: Where, represents the synthetic new generation population; Indicates elite population; represents the medium energy population; Represents a low-energy population. Second, check the boundaries and truncate the parameters of each individual to ensure that they do not exceed the feasible range: Where x i,j represents the jth parameter of the i-th individual. Finally, recalculate the fitness g i . Step 2.8: When or t≥I max When , output the global optimal solution; otherwise, return to step 2.1.

Images