Fuel closed loop control method

Abstract

The application provides a fuel closed-loop control method, and belongs to the technical field of engine control. The fuel closed-loop control method comprises the following steps: calculating a short-term fuel correction value through proportional integral control based on an actual air-fuel ratio and a required air-fuel ratio in a current period; in response to meeting a self-learning activation condition, updating a stored long-term fuel correction component of each preset working condition interval based on a preset number of periods of short-term fuel correction values and working conditions, wherein the working conditions are a three-dimensional feature vector composed of an engine speed, an engine intake air volume and an engine coolant temperature; taking the working conditions in the current period as input, calculating a probability proportion of the working conditions in each preset working condition interval based on a Gaussian function; obtaining a long-term fuel correction value of the working conditions through weighted summation and a change rate limit by using the probability proportion of each preset working condition interval and the corresponding long-term fuel correction component; and obtaining a final fuel correction value by using the short-term fuel correction value and the long-term fuel correction value.

Description

Technical Field

[0001] This invention relates to the field of engine control technology, and in particular to a fuel closed-loop control method. Background Technology

[0002] Closed-loop air-fuel ratio control refers to a control strategy where the engine electronic control unit (ECU) dynamically adjusts the fuel injection quantity based on real-time feedback signals of oxygen content in the exhaust gas, precisely controlling the air-fuel ratio near its theoretical value (approximately 14.7:1). This includes Short-Term Fuel Adjustment (STFT) and Long-Term Fuel Adjustment (LTFT). STFT involves real-time, rapid adjustments to the fuel injection quantity (within ±10%) based on oxygen sensor signals. The adjustment value fluctuates rapidly between positive and negative percentages. LTFT involves the ECU observing the long-term trend of the short-term fuel adjustment. If the STFT continuously adjusts towards either enrichment or leaning, the ECU learns this deviation and accordingly modifies its base calculation value for the fuel injection quantity. LTFT is used to compensate for long-term, slow changes in the system (such as slight injector clogging or slight airflow meter drift). Traditional LTFT can only perform two-dimensional learning based on engine speed and load, and the learned value only participates in correction after the coolant temperature reaches a certain level. Summary of the Invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art, and proposes a fuel closed-loop control method based on neural network self-learning that introduces the water temperature dimension.

[0004] In a first aspect, embodiments of the present invention provide a fuel closed-loop control method, comprising: calculating a short-term fuel correction value based on the actual air-fuel ratio and the demand air-fuel ratio of the current period using proportional-integral control; in response to satisfying a self-learning activation condition, updating the stored long-term fuel correction component for each preset operating condition interval based on the short-term fuel correction value and operating condition for a preset number of periods, wherein the operating condition is a three-dimensional feature vector composed of engine speed, engine intake air volume, and engine coolant temperature; using the operating condition of the current period as input, calculating the probability proportion of the operating condition in each preset operating condition interval based on a Gaussian function; using the probability proportion of each preset operating condition interval and the corresponding long-term fuel correction component, weighted summing and then obtaining the long-term fuel correction value of the operating condition through a rate of change constraint; and using the short-term fuel correction value and the long-term fuel correction value to obtain the final fuel correction value.

[0005] According to an embodiment of the present invention, a short-term fuel correction value is calculated by proportional-integral control based on the actual air-fuel ratio and the demand air-fuel ratio of the current cycle. The calculation includes: calculating the equivalence ratio deviation between the demand equivalence ratio and the actual equivalence ratio based on the actual air-fuel ratio and the demand air-fuel ratio, wherein the ratio of the actual air-fuel ratio to the theoretical air-fuel ratio is defined as the actual equivalence ratio, and the ratio of the demand air-fuel ratio to the theoretical air-fuel ratio is defined as the demand equivalence ratio; multiplying the equivalence ratio deviation by a proportional coefficient to obtain the proportional gain; multiplying the equivalence ratio deviation by an integral coefficient and adding it to the short-term fuel correction value of the previous cycle to obtain the integral gain; and adding the proportional gain and the integral gain to a preset reference value to obtain the short-term fuel correction value.

[0006] According to an embodiment of the present invention, the self-learning activation conditions include: the engine speed is greater than the preset minimum operating speed; the engine is in normal fuel injection state; and the engine has no fuel injection fault.

[0007] According to an embodiment of the present invention, updating the long-term fuel correction component of each preset operating condition interval includes: recording the number of calculations of the short-term fuel correction value in response to satisfying the self-learning activation condition; calculating the average short-term fuel correction value based on the cumulative number of calculations in response to the number of calculations exceeding a preset minimum threshold; limiting the average short-term fuel correction value to an upper and lower limit based on a two-dimensional mapping table of engine speed and engine intake air volume to obtain a self-learning requirement value; calculating the average operating condition corresponding to the short-term fuel correction value to obtain a self-learning operating condition feature vector; and calculating the linear model output value of each preset operating condition interval as a long-term fuel correction component based on the self-learning requirement value and the corresponding self-learning operating condition feature vector.

[0008] According to an embodiment of the present invention, there are 15 preset operating condition intervals. The long-term fuel correction component of each preset operating condition interval is stored in a non-volatile memory, which remains valid when the controller is powered off and then powered on again. It is updated when the self-learning activation condition is met and participates in the weighted summation calculation after being read.

[0009] According to an embodiment of the present invention, based on the self-learning requirement value and the corresponding self-learning operating condition feature vector, the linear model output value of each preset operating condition interval is calculated as a long-term fuel correction component, including: normalizing the self-learning operating condition feature vector to obtain a normalized feature vector, wherein the engine speed component, engine intake air volume component, and engine coolant temperature component of the self-learning operating condition feature vector are divided by the corresponding preset speed reference value, preset intake air volume reference value, and preset coolant temperature reference value, respectively; multiplying the normalized feature vector by the self-learning requirement value and adding the preset feature reference value correction value of the corresponding preset operating condition interval to obtain the corrected normalized vector of each preset operating condition interval; performing an inner product operation on the corrected normalized vector of each preset operating condition interval and the normalized feature vector to obtain the linear basic value of each preset operating condition interval; and adding the linear basic value of each preset operating condition interval to the corresponding preset linear correction value to obtain the linear model output value of each preset operating condition interval.

[0010] According to an embodiment of the present invention, taking the current cycle operating condition as input, the probability proportion of the current cycle operating condition in each preset operating condition interval is calculated based on a Gaussian function, including: based on engine operating condition data, dividing each preset operating condition interval into a feature mean vector and a feature standard deviation vector, the feature mean vector including the mean engine speed, mean engine intake air volume, and mean engine coolant temperature in the preset operating condition interval, and the feature standard deviation vector including the standard deviation of engine speed, standard deviation of engine intake air volume, and standard deviation of engine coolant temperature in the preset operating condition interval; for each preset operating condition interval, calculating the difference between the current operating condition and the feature mean vector, dividing it by the feature standard deviation vector to obtain a standardized difference vector for calculating the Gaussian function value; and using the Gaussian function value of each preset operating condition interval to divide by the sum of Gaussian function values ​​to obtain the probability proportion of the current operating condition in each preset operating condition interval.

[0011] According to an embodiment of the present invention, the rate of change limitation includes: calculating the difference between the long-term fuel trim value of the current period and the final long-term fuel trim value of the previous period; if the difference is greater than a preset minimum change amount and less than a preset maximum change amount, then the long-term fuel trim value of the current period is used as the final long-term fuel trim value; if the difference is greater than a preset maximum change amount, then the final long-term fuel trim value of the previous period plus the preset maximum change amount is used as the final long-term fuel trim value; if the difference is less than a preset minimum change amount, then the final long-term fuel trim value of the previous period plus the preset minimum change amount is used as the final long-term fuel trim value.

[0012] According to an embodiment of the present invention, obtaining a final fuel correction value using a short-term fuel correction value and a long-term fuel correction value includes: multiplying the short-term fuel correction value and the long-term fuel correction value to obtain the final fuel correction value.

[0013] Secondly, the present invention provides a fuel closed-loop control system capable of implementing the aforementioned fuel closed-loop control method, comprising: a short-term correction module for calculating a short-term fuel correction value based on the actual air-fuel ratio and the demand air-fuel ratio of the current period using proportional-integral control; a probability calculation module for calculating the probability proportion of the operating condition in each preset operating condition interval based on a Gaussian function, using the current operating condition as input, in response to satisfying the self-learning activation condition, where the operating condition is a three-dimensional feature vector composed of engine speed, engine intake air volume, and engine coolant temperature; a long-term correction module for obtaining the long-term fuel correction value of the operating condition by weighted summation of the probability proportion of each preset operating condition interval and the corresponding long-term fuel correction component, and limiting the rate of change; and a combination correction module for obtaining the final fuel correction value using the short-term fuel correction value and the long-term fuel correction value.

[0014] A third aspect of the present invention provides an electronic device comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors perform the above-described fuel closed-loop control method.

[0015] A fourth aspect of the present invention also provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, cause the processor to perform the above-described fuel closed-loop control method.

[0016] The fuel closed-loop control method provided by this invention introduces a three-dimensional feature vector composed of engine coolant temperature, engine speed, and intake air volume for Gaussian probability weighted self-learning. Since it covers the operating condition identification of the coolant temperature dimension, it at least partially solves the problem that traditional two-dimensional LTFT cannot be effectively corrected in the low temperature and cold start stages, and achieves an improvement in fuel closed-loop control accuracy across the entire temperature range. Attached Figure Description

[0017] Figure 1 A schematic flowchart of a fuel closed-loop control method provided in an embodiment of the present invention;

[0018] Figure 2 This is a schematic diagram of the overall process provided for an embodiment of the present invention;

[0019] Figure 3 This is a flowchart illustrating the method for calculating the long-term fuel correction component in step S3 of this embodiment of the invention.

[0020] Figure 4 This is a structural block diagram of a fuel closed-loop control system provided in an embodiment of the present invention;

[0021] Figure 5 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0022] To enable those skilled in the art to better understand the technical solutions of the present invention, exemplary embodiments of the present invention are described below in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0023] Where there is no conflict, the various embodiments of the present invention and the features thereof may be combined with each other.

[0024] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.

[0025] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used herein, the singular forms “a” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms “comprising” and / or “made of” are used in this specification, the presence of the stated feature, integral, step, operation, element, and / or component is specified, but the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof is not excluded. Terms such as “connected” or “linked” are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect.

[0026] Unless otherwise specified, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having the meaning consistent with their meaning in the context of the relevant art and the invention, and will not be interpreted as having an idealized or overly formal meaning unless expressly so defined herein.

[0027] In the technical solution of this invention, the collection, storage, use, processing, transmission, provision, and disclosure of user personal information all comply with relevant laws and regulations and do not violate public order and good morals. The use of user data in this technical solution follows relevant national laws and regulations (e.g., the "Information Security Technology - Personal Information Security Specification"). For example: appropriate measures are taken for personal information access control; restrictions are imposed on the display of personal information; the purpose of using personal information does not exceed the scope of direct or reasonable association; and explicit identity targeting is eliminated when using personal information to avoid precisely locating a specific individual.

[0028] CN111412074A provides a self-learning method for long-term fuel trim of a gasoline engine. It determines whether the self-learning operating condition is met; if so, it enters the self-learning operating condition stabilization stage. It then determines whether the self-learning stabilization stage has ended; if so, it enters the self-learning storage activation stage. Based on the average engine speed, average load, and average short-term fuel trim over a given time period, it calculates the engine speed and load, and then updates the long-term fuel trim value. However, this patent only considers the influence of engine speed and load on the learning value, neglecting coolant temperature and thus having no effect at low temperatures.

[0029] To address at least one of the technical problems existing in the aforementioned related technologies, the present invention provides a self-learning control method for fuel closed-loop control based on neural networks. Figure 1 This is a schematic flowchart of a fuel closed-loop control method provided in an embodiment of the present invention, as shown below. Figure 1 As shown, this embodiment of the invention provides a fuel closed-loop control method, comprising: S1, calculating a short-term fuel correction value based on the actual air-fuel ratio and the demand air-fuel ratio of the current period using proportional-integral control; S2, in response to meeting the self-learning activation condition, updating the stored long-term fuel correction component for each preset operating condition interval based on the short-term fuel correction value and operating condition for a preset number of periods, wherein the operating condition is a three-dimensional feature vector composed of engine speed, engine intake air volume, and engine coolant temperature; S3, using the operating condition of the current period as input, calculating the probability proportion of the operating condition in each preset operating condition interval based on a Gaussian function; S4, using the probability proportion of each preset operating condition interval and the corresponding long-term fuel correction component, weighted summing and then obtaining the long-term fuel correction value of the operating condition through a rate of change constraint; S5, using the short-term fuel correction value and the long-term fuel correction value to obtain the final fuel correction value.

[0030] Figure 2 This is a schematic diagram of the overall process provided for an embodiment of the present invention, such as... Figure 2 As shown, firstly, through fuel closed-loop control, short-term fuel correction values ​​are generated in real time using PID regulation, while simultaneously collecting operating condition signals such as engine speed, intake air volume, and coolant temperature. Next, a series of self-learning activation conditions are used to determine whether to allow entry into self-learning mode. Once activated, the probability weights of the current operating condition and 15 preset regions are calculated based on a Gaussian function, and the long-term fuel correction increment for each region is obtained by combining a linear model. Finally, the self-learning results are processed to complete the closed loop from real-time control to intelligent learning and then to application, enabling the engine to continuously self-calibrate.

[0031] It should be noted that the input signals required by this invention are obtained through a general acquisition method: the engine coolant temperature T_EngineCoolant is obtained through a sensor; the engine speed n_EngSpd is obtained through a sensor; the engine intake air volume rho_CylAir is obtained through other modules of the controller; the actual air-fuel ratio r_FrontO2lambda is obtained through an oxygen sensor; and the controller chip's task execution cycle T_10ms = 10ms. This method includes an intake air volume control module, which receives signals from the above modules.

[0032] Through the embodiments of this invention, three-dimensional feature vectors are used to simultaneously capture engine speed, load, and thermal state, providing a more accurate description of operating conditions compared to a two-dimensional approach based on engine speed and load. Gaussian probability calculation followed by weighting avoids excessive fluctuations in correction values ​​due to interval switching, ensuring continuous and smooth correction values ​​as operating conditions change. This achieves the technical effect of incorporating a water temperature dimension into three-dimensional learning based on a neural network, allowing self-learned values ​​to still participate in calculations during low-temperature operation and cold start phases, thus improving the accuracy of fuel closed-loop control.

[0033] Based on the above embodiments, a short-term fuel correction value is calculated using proportional-integral control based on the actual air-fuel ratio and the demand air-fuel ratio for the current cycle. This includes: calculating the equivalence ratio deviation between the demand equivalence ratio and the actual equivalence ratio based on the actual air-fuel ratio and the demand air-fuel ratio, wherein the ratio of the actual air-fuel ratio to the theoretical air-fuel ratio is defined as the actual equivalence ratio, and the ratio of the demand air-fuel ratio to the theoretical air-fuel ratio is defined as the demand equivalence ratio; multiplying the equivalence ratio deviation by a proportional coefficient to obtain the proportional gain; multiplying the equivalence ratio deviation by an integral coefficient and adding it to the short-term fuel correction value for the previous cycle to obtain the integral gain; and adding the proportional gain and the integral gain to a preset benchmark value to obtain the short-term fuel correction value.

[0034] In this embodiment, the ratio of the actual air-fuel ratio r_FrontO2lambdaAct to the theoretical air-fuel ratio (a fixed value, the theoretical value of the ratio of air to fuel mass when gasoline is completely burned) is set as the actual equivalence ratio r_FrontO2Act, and the calculation method is as follows: r_FrontO2req=r_FrontO2lambdaAct / 14.7.

[0035] Set the demand air-fuel ratio r_FrontO2lambdaReq (the air-fuel ratio expected to be achieved under the current operating conditions), which is the ratio to the theoretical air-fuel ratio, defined as the demand equivalence ratio r_FrontO2Req, and calculated as follows: r_FrontO2act=r_FrontO2lambdaReq / 14.7.

[0036] The deviation between the demand value and the actual equivalent value is calculated as follows:

[0037] r_FrontO2error=r_FrontO2req-r_FrontO2act.

[0038] Subsequently, based on the demand and actual values, PI closed-loop control is implemented to calculate the equivalence ratio closed-loop value r_FrontO2Dsrd for short-term fuel correction:

[0039] (1) Calculation of P term: Set the coefficient of P term to K_Prop, then the gain of P term r_FrontO2Prop is calculated as follows: r_FrontO2Prop=K_Prop×r_FrontO2error.

[0040] (2) Calculation of I term: Set the coefficient of I term as K_Intg, and the closed-loop equivalent value of the closed-loop demand at the previous time step as r_rontO2Dsrd(z), then the calculation formula of I gain r_FrontO2Intg is as follows:

[0041] r_FrontO2Intg=K_Intg×r_FrontO2error+r_rontO2Dsrd(z).

[0042] (3) The final equivalent ratio closed-loop value r_FrontO2PI is calculated as follows:

[0043] r_FrontO2PI=r_FrontO2Prop+r_FrontO2Intg+1.

[0044] The above Short-Term Fuel Adjustment (STFT) is used as the input for the self-learning value.

[0045] In this embodiment, the P-term and I-term coefficients are typically tuned through engine bench calibration tests. The PI controller uses the equivalence ratio deviation as an error signal to adjust the fuel injection quantity in real time. Its control objective is to make the actual equivalence ratio approach the required equivalence ratio, i.e., the deviation tends to zero. The P-term obtains the proportional gain by multiplying the deviation by the proportional coefficient, achieving a rapid response to the deviation. However, using the P-term alone will result in a steady-state error, meaning that the deviation cannot automatically return to zero after the system stabilizes. The I-term accumulates historical errors by multiplying the deviation by the integral coefficient and adding the PI output value of the previous cycle. As time progresses, the correction amount continuously increases until the deviation is completely eliminated, thereby compensating for the steady-state error defect of the P-term. The final PI output is the sum of the P-term gain, the I-term gain, and the reference value of 1. Since this output participates in the final fuel correction as a multiplicative correction coefficient, using 1 as the reference ensures that the correction coefficient remains neutral when the deviation is zero, i.e., no correction is performed. When the deviation is positive, the output is greater than 1 to enrich the fuel, and when the deviation is negative, the output is less than 1 to lean the fuel. In addition, the controller does not include the D term. The oxygen sensor signal itself contains high-frequency noise. Introducing the D term would amplify the noise component, which would lead to unstable control. Therefore, a PI control structure is adopted to balance response speed and control accuracy.

[0046] Through the embodiments of the present invention, based on the equivalence ratio deviation between the actual air-fuel ratio and the demand air-fuel ratio, a rapid response to the deviation is achieved through a proportional term (P term), and a steady-state error is eliminated through an integral term (I term, combined with the closed-loop value at the previous moment). The two are added together and then a reference value of 1 is added to obtain the equivalence ratio closed-loop value of the short-term fuel correction. This value is the short-term fuel correction, which is used to fine-tune the fuel injection quantity in real time and serves as the input signal for subsequent neural network self-learning.

[0047] Based on the above embodiments, the self-learning activation conditions include: the engine speed is greater than the preset minimum operating speed; the engine is in normal fuel injection state; and the engine has no fuel injection fault.

[0048] In this embodiment of the invention, the minimum engine operating speed is set to n_EngSpdRunning(rpm) = 50. The engine is considered to be running when the engine speed meets the following condition: n_EngSpd > n_EngSpdRunning. The self-learning activation state STFT_Enable is considered to be 1 when the following conditions are met: the current engine is in normal fuel injection mode; the current engine is running; and no fuel injection-related faults are activated. Self-learning is only allowed when the engine is running stably and normally, solving the problem of erroneous learning under abnormal operating conditions and ensuring the reliability of the self-learning process.

[0049] Figure 3 This is a flowchart illustrating the calculation method for the long-term fuel correction component in step S3 of this embodiment of the invention. Figure 3As shown, updating the long-term fuel correction component for each preset operating condition interval includes: S31, recording the number of calculations of the short-term fuel correction value in response to meeting the self-learning activation condition; S32, calculating the average short-term fuel correction value based on the cumulative number of calculations when the number of calculations exceeds a preset minimum threshold; S33, applying upper and lower limits to the average short-term fuel correction value based on a two-dimensional mapping table of engine speed and engine intake air volume to obtain the self-learning requirement value; S34, calculating the average operating condition corresponding to the short-term fuel correction value to obtain the self-learning operating condition feature vector; S35, calculating the linear model output value of each preset operating condition interval as the long-term fuel correction component based on the self-learning requirement value and the corresponding self-learning operating condition feature vector.

[0050] In this embodiment, the calculation process for the self-learning value input includes:

[0051] (1) Average the input values:

[0052] When STFT_Enable is 1, the controller runs once, accumulating the equivalent ratio closed-loop value. The accumulated value is defined as r_FrontO2PISum, and the accumulated number of times is Cntsum. The initial average value of the equivalent ratio closed-loop is defined as r_FrontO2PIMeanRaw, and the calculation method is as follows:

[0053] ;

[0054] ;

[0055] .

[0056] Define the equivalent ratio closed-loop average value r_FrontO2PIMean, and set the minimum number of closed-loop runs Cnt_minComb=100. When Cnt_Sum>Cnt_minComb is satisfied, output r_FrontO2PIMean, which is calculated as follows: r_FrontO2PIMean=r_FrontO2PIMeanRaw.

[0057] (2) Input value restrictions:

[0058] Based on engine speed n_EngSpd and engine intake air volume rho_CylAir as the X-axis and Y-axis respectively, the maximum equivalence ratio limit value r_FrontO2DsrdMax is set, and the speed-intake air volume-maximum equivalence ratio mapping table is shown in Table 1.

[0059] Table 1:

[0060]

[0061] The speed-intake volume-minimum equivalence ratio mapping table is shown in Table 2.

[0062] Table 2:

[0063]

[0064] The constrained value is defined as the self-learning requirement value r_FrontO2Dsrd.

[0065] As mentioned above, the two-dimensional mapping table includes a maximum equivalence ratio limit table and a minimum equivalence ratio limit table. Both tables use engine speed as the first dimension and engine intake air volume as the second dimension. Optionally, the upper and lower limits of each speed-intake air volume are pre-calibrated through bench tests. When the short-term fuel correction average exceeds the maximum limit corresponding to the current operating point, the self-learning demand value is clamped to the maximum limit. When it is lower than the minimum limit, the self-learning demand value is clamped to the minimum limit. If it is between the two, the original value remains unchanged.

[0066] In this embodiment, the feature input and its dimension length are specified. Specifically, engine speed n_EngSpd, engine intake air volume rho_CylAir, and engine coolant temperature T_EngineCoolant are used as the three dimensions of the self-learning feature input. Therefore, the feature input vector is:

[0067] .

[0068] Each feature input dimension is set to be partitioned, and the number of partitions is defined as Num_Region=15.

[0069] Through the embodiments of the present invention, by accumulating short-term correction values ​​over multiple periods and taking the average value, and then using the result as the self-learning requirement value after limiting the amplitude according to the operating conditions, the influence of transient fluctuations is eliminated by utilizing steady-state statistics and boundary constraints, thus ensuring the stability and rationality of the self-learning input signal.

[0070] Based on the above embodiments, the number of preset operating condition intervals is 15. The long-term fuel correction component of each preset operating condition interval is stored in a non-volatile memory, which remains valid when the controller is powered off and then powered on again. It is updated when the self-learning activation condition is met and participates in the weighted summation calculation after being read.

[0071] Through the embodiments of the present invention, since the data is not lost when power is off and is updated after each self-learning, continuous learning and memory across driving cycles are achieved, avoiding the need to accumulate data from scratch with each startup, and improving the efficiency and consistency of long-term adaptation.

[0072] Based on the above embodiments, the linear model output value for each preset operating condition interval is calculated as a long-term fuel correction component based on the self-learning requirement value and the corresponding self-learning operating condition feature vector. This includes: normalizing the self-learning operating condition feature vector to obtain a normalized feature vector, wherein the engine speed component, engine intake air volume component, and engine coolant temperature component of the self-learning operating condition feature vector are divided by the corresponding preset speed reference value, preset intake air volume reference value, and preset coolant temperature reference value, respectively; multiplying the normalized feature vector by the self-learning requirement value and adding the preset feature reference value correction value for the corresponding preset operating condition interval to obtain the corrected normalized vector for each preset operating condition interval; performing an inner product operation on the corrected normalized vector and the normalized feature vector for each preset operating condition interval to obtain the linear basic value for each preset operating condition interval; and adding the linear basic value for each preset operating condition interval to the corresponding preset linear correction value to obtain the linear model output value for each preset operating condition interval.

[0073] In this embodiment, the feature input is normalized and a linear model is established, specifically including:

[0074] (1) Engine speed n_EngSpd, engine intake air volume rho_CylAir, and engine coolant temperature T_EngineCoolant are used as self-learning feature inputs, and the vector of the feature input baseline is defined as Var_NormLinear. The values ​​are defined as follows:

[0075] .

[0076] Define the feature baseline value correction value for each interval as r_LinearNormOffset_i, and look up the table based on the current region i_Region. The table is shown in Table 3.

[0077] Table 3:

[0078]

[0079] The normalized value of the feature baseline for each interval is defined as r_LinearIn_i, and the calculation formula is as follows:

[0080] r_LinearIn_i=Var_Indepedence / Var_NormLinear.

[0081] (2) The linear base value based on the characteristic baseline is r_LinearOutputBase, and the calculation formula is as follows:

[0082] .

[0083] (3) Define the linear correction value for each interval as r_LinearOffset_i. Based on the current region i_Region, look up the table. The table is shown in Table 4.

[0084] Table 4:

[0085]

[0086] (4) Define the linear output value for each interval as r_LinearOutput_i, and calculate it as follows:

[0087] r_LinearOutput_i=r_LinearOutputBase_i+r_LinearOffset_i.

[0088] It should be noted that Tables 3 and 4 define the feature baseline correction values ​​and linear correction values ​​for 15 preset operating condition intervals, respectively. Both are pre-calibrated and stored in non-volatile memory as adjustable parameters. The feature baseline correction values ​​are used to correct the baseline offset of intermediate results for each interval after multiplying the normalized feature vector by the self-learning requirement value. Their initial values ​​can be determined through offline clustering analysis or uniform grid partitioning. The linear correction values ​​are superimposed after the linear basic values ​​for each interval are calculated, and are used to compensate for systematic deviations within that interval that were not fully captured by previous steps. Together, these two tables constitute the learnable parameters of the piecewise linear model, allowing each interval to independently adjust its linear mapping relationship, thereby achieving modeling of all operating conditions.

[0089] In this invention, the working space is divided into multiple intervals, and a linear model is used to approximate the correction characteristics of each interval. First, the self-learning working condition feature vector is normalized to eliminate the influence of different dimensions and magnitudes. Second, the normalized feature vector is multiplied by the self-learning requirement value, and a feature benchmark correction value unique to that interval is added to couple the working condition location with the correction requirement. Then, the corrected vector is multiplied by the original normalized feature vector to complete the dimensionality reduction mapping from three-dimensional features to a single-valued correction. Finally, a preset linear correction value for that interval is added to compensate for systematic deviations not captured by previous steps. Through a series of linear transformations, refined compensation is achieved.

[0090] Based on the above embodiments, taking the current cycle operating condition as input, the probability proportion of the current cycle operating condition in each preset operating condition interval is calculated based on the Gaussian function. This includes: based on engine operating condition data, dividing the data into feature mean vectors and feature standard deviation vectors for each preset operating condition interval. The feature mean vector includes the mean engine speed, mean engine intake air volume, and mean engine coolant temperature in the preset operating condition interval. The feature standard deviation vector includes the standard deviation of engine speed, standard deviation of engine intake air volume, and standard deviation of engine coolant temperature in the preset operating condition interval. For each preset operating condition interval, the difference between the current operating condition and the feature mean vector is calculated, and divided by the feature standard deviation vector to obtain a standardized difference vector for calculating the Gaussian function value. The probability proportion of the current operating condition in each preset operating condition interval is obtained by dividing the Gaussian function value of each preset operating condition interval by the sum of the Gaussian function values.

[0091] In this embodiment, the probability calculation based on the Gaussian function includes:

[0092] (1) Define the mean of the three factors of the feature input as r_GSCenter, use the number of partitions as the X-axis and Y-axis, define the current region as i_Region, divide the region based on the engine operating condition data, look up the table based on the current region, and obtain the mean of the current corresponding feature input. The table is shown in Table 5.

[0093] Table 5:

[0094]

[0095] Define the independent variables for the current interval: mean engine speed i_n_EngSpdCenter, mean intake air volume i_rho_CylAirCenter, and coolant temperature i_T_EngineCoolantCenter. The vector value is expressed as follows:

[0096] .

[0097] (2) Define the standard deviation of the three factors of the feature input as r_GSSigma, use the number of partitions as the X-axis and Y-axis, define the current region as i_Region, and look up the table based on the current region to obtain the standard deviation of the corresponding feature input. The table is shown in Table 6.

[0098] Table 6:

[0099]

[0100] Define the independent variables for the current interval: the standard deviation of engine speed i_n_EngSpdSigma, the standard deviation of intake air volume i_rho_CylAirSigma, and the standard deviation of coolant temperature i_T_EngineCoolantSigma.

[0101] but The vector value is expressed as follows:

[0102] .

[0103] (3) Based on the equation of the Gaussian function, perform probability calculation. Calculate 15 times according to the number of partitions Num_Region=15. Define the inputs i_x1 and i_x2 as follows:

[0104] ;

[0105] ;

[0106] Define the probability of each interval as The calculation is as follows:

[0107] .

[0108] The total probability output for all intervals is defined as Var_Gauss, and the calculation formula is as follows:

[0109] .

[0110] The proportion of each of the 15 intervals is defined as a vector. The calculation formula is as follows:

[0111] .

[0112] If i is 1-15, then the proportion of all intervals is a vector. The expression is as follows:

[0113] r_GaussOut=(r_GaussOut_1, r_GaussOut_2, r_GaussOut_3,…, r_GaussOut_15).

[0114] After the probability calculation is completed, based on the proportion of each of the 15 intervals as the vector r_GaussOut_i and the linear output value r_LinearOutput_i of each interval, the long-term fuel trim value for each interval is defined as r_GSLinear_i, and the calculation formula is as follows:

[0115] r_GSLinear_i=r_GaussOut_i×r_LinearOutput_i.

[0116] Define the long-term fuel trim value for all intervals as a vector. The expression is as follows:

[0117] r_GSLinear=(r_GSLinear_1, r_GSLinear_2, r_GSLinear_3,…, r_GSLinear_15).

[0118] The vector is stored in the controller as non-volatile storage, so that the learned value can be saved even after the controller is powered off and then powered on again.

[0119] The long-term fuel trim value for the current operating condition is defined as r_GaussOutSum, and the calculation formula is as follows:

[0120] .

[0121] In this embodiment of the invention, a Gaussian function is used as the membership function, allowing a single operating point to simultaneously belong to multiple adjacent intervals. Furthermore, the membership decreases with increasing distance, thus avoiding control value jumps caused by hard switching. First, the feature mean vector and feature standard deviation vector of each preset operating interval are calibrated offline based on engine operating condition data, ensuring that each interval covers an operating condition subspace with similar correction characteristics. Second, the standardized distance from the current operating point to the center of each interval is calculated to eliminate the influence of differences in the dimensions and variances of different features. Then, the sum of the squares of the standardized distances is substituted into the Gaussian kernel function to obtain the Gaussian function value for each interval. Finally, the Gaussian function values ​​of all intervals are summed, and the sum is normalized by dividing the Gaussian value of each interval to obtain the probability proportion. Each operating point expresses its membership relationship to each interval in the form of a probability vector, providing weighting coefficients for subsequent weighted summation.

[0122] Based on the above embodiments, the rate of change limitation includes: calculating the difference between the long-term fuel trim value of the current period and the final long-term fuel trim value of the previous period; if the difference is greater than a preset minimum change amount and less than a preset maximum change amount, then the long-term fuel trim value of the current period is used as the final long-term fuel trim value; if the difference is greater than a preset maximum change amount, then the final long-term fuel trim value of the previous period plus the preset maximum change amount is used as the final long-term fuel trim value; if the difference is less than a preset minimum change amount, then the final long-term fuel trim value of the previous period plus the preset minimum change amount is used as the final long-term fuel trim value.

[0123] In this embodiment, define the current long-term fuel trim final output value as r_GaussOutSumFinal, and the long-term fuel trim final output value at the previous moment as r_GaussOutSumFinal(z). Set the maximum and minimum values of the long-term fuel trim value change at different moments to be r_RampMax and r_RampMin respectively. The output value deviation is r_GaussOutSumErr, and the calculation formula is as follows.

[0124] r_RampMax = 0.01; r_RampMin = -0.01;

[0125] r_GaussOutSumErr = r_GaussOutSumFinal - r_GaussOutSumFinal(z).

[0126] If the following conditions are met: r_RampMax ≥ r_GaussOutSumErr ≥ r_RampMin, then r_GaussOutSumFinal = r_GaussOutSum.

[0127] If the following conditions are met: r_RampMax < r_GaussOutSumErr, then r_GaussOutSumFinal = r_GaussOutSum(z) + r_RampMax.

[0128] If the following conditions are met: r_RampMin > r_GaussOutSumErr, then r_GaussOutSumFinal = r_GaussOutSum(z) + r_RampMin.

[0129] Through the embodiments of the present invention, by comparing the difference between the long-term correction value in the current cycle and the final value in the previous cycle, and limiting the amplitude according to the preset upper and lower limits of the change rate, the change amplitude of each correction amount is restricted, preventing the air-fuel ratio oscillation caused by the sudden change of the learning value, and improving the stability of the control.

[0130] Based on the above embodiments, using the short-term fuel trim value and the long-term fuel trim value to obtain the final fuel trim value, including: multiplying the short-term fuel trim value and the long-term fuel trim value to obtain the final fuel trim value.

[0131] In this embodiment, define the final fuel trim as r_FuelTrim, and the calculation formula is as follows:

[0132] r_FuelTrim = r_FrontO2PI × r_GaussOutSumFinal.

[0133] Through the embodiments of the present invention, the rapid dynamic response capability of short-term correction is retained, while the steady-state compensation of long-term correction is superimposed, achieving the final correction that complements the advantages of both.

[0134] Figure 4 A structural block diagram of a fuel closed-loop control system provided in an embodiment of the present invention is shown below. Figure 4 As shown, the present invention also provides a fuel closed-loop control system capable of implementing the above-mentioned fuel closed-loop control method, comprising: a short-term correction module 401, used to calculate a short-term fuel correction value based on the actual air-fuel ratio and the demand air-fuel ratio of the current period through proportional-integral control; a self-learning module 402, used to update the stored long-term fuel correction component of each preset operating condition interval based on the short-term fuel correction value and operating condition of a preset number of periods in response to the satisfaction of the self-learning activation condition, wherein the operating condition is a three-dimensional feature vector composed of engine speed, engine intake air volume and engine coolant temperature; a probability calculation module 403, used to calculate the probability proportion of the operating condition in each preset operating condition interval based on a Gaussian function, using the operating condition of the current period as input; a long-term correction module 404, used to obtain the long-term fuel correction value of the operating condition by weighted summation of the probability proportion of each preset operating condition interval and the corresponding long-term fuel correction component, and by limiting the rate of change; and a combination correction module 405, used to obtain the final fuel correction value by using the short-term fuel correction value and the long-term fuel correction value.

[0135] Based on the same inventive concept, embodiments of the present invention also provide an electronic device. Figure 5 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Figure 5 As shown, an embodiment of the present invention provides an electronic device including: one or more processors 101, a memory 102, and one or more I / O interfaces 103. The memory 102 stores one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement any of the fuel closed-loop control methods described in the above embodiments; the one or more I / O interfaces 103 are connected between the processors and the memory, configured to enable information interaction between the processors and the memory.

[0136] The processor 101 is a device with data processing capabilities, including but not limited to a central processing unit (CPU); the memory 102 is a device with data storage capabilities, including but not limited to random access memory (RAM, more specifically SDRAM, DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), and flash memory (FLASH); the I / O interface (read / write interface) 103 is connected between the processor 101 and the memory 102, and can realize information interaction between the processor 101 and the memory 102, including but not limited to a data bus (Bus).

[0137] In some embodiments, the processor 101, memory 102, and I / O interface 103 are interconnected via bus 104, and thus connected to other components of the computing device.

[0138] In some embodiments, the one or more processors 101 include a field-programmable gate array.

[0139] This invention also provides a computer-readable medium. The computer-readable medium stores a computer program, which, when executed by a processor, implements the steps of any of the fuel closed-loop control methods described in the above embodiments. The computer-readable storage medium may be volatile or non-volatile.

[0140] This invention also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code. When the computer-readable code is run in the processor of an electronic device, the processor in the electronic device executes the above-described fuel closed-loop control method.

[0141] Those skilled in the art will understand that all or some of the steps, systems, and apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software can be distributed on a computer-readable storage medium, which may include computer storage media (or non-transitory media) and communication media (or transient media).

[0142] As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable program instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technologies, portable compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable program instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0143] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0144] The computer program instructions used to perform the operations of this invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing state information from the computer-readable program instructions. This electronic circuitry can execute the computer-readable program instructions to implement various aspects of the invention.

[0145] The computer program product described herein can be implemented specifically through hardware, software, or a combination thereof. In one alternative embodiment, the computer program product is specifically embodied in a computer storage medium; in another alternative embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0146] Various aspects of the present invention are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0147] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0148] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0149] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction, which contains one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0150] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for illustrative purposes only and should be construed as such, and is not intended to be limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in conjunction with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in conjunction with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of the invention as set forth in the appended claims.

Claims

1. A fuel closed-loop control method, characterized in that, include: Based on the actual air-fuel ratio and demand air-fuel ratio of the current cycle, the short-term fuel correction value is calculated through proportional-integral control. In response to the fulfillment of the self-learning activation condition, the long-term fuel correction component of each preset operating condition interval is updated based on the short-term fuel correction value and operating condition of a preset number of cycles, wherein the operating condition is a three-dimensional feature vector composed of engine speed, engine intake air volume and engine coolant temperature. Using the current cycle's operating condition as input, the probability percentage of the operating condition in each preset operating condition interval is calculated based on a Gaussian function; By using the probability proportion of each preset operating condition interval and the corresponding long-term fuel correction component, the long-term fuel correction value of the operating condition is obtained by weighted summation and then subject to a rate of change limit. The final fuel correction value is obtained by using the short-term fuel correction value and the long-term fuel correction value.

2. The method according to claim 1, wherein, The calculation of short-term fuel correction values ​​based on the actual air-fuel ratio and demand air-fuel ratio of the current cycle, using proportional-integral control, includes: Based on the actual air-fuel ratio and the demand air-fuel ratio, the equivalence ratio deviation between the demand equivalence ratio and the actual equivalence ratio is calculated, wherein the ratio of the actual air-fuel ratio to the theoretical air-fuel ratio is defined as the actual equivalence ratio, and the ratio of the demand air-fuel ratio to the theoretical air-fuel ratio is defined as the demand equivalence ratio. Multiplying the equivalent ratio deviation by the proportionality coefficient yields the proportional term gain; Multiplying the equivalent ratio deviation by the integral coefficient and then adding it to the short-term fuel correction value of the previous cycle yields the integral term gain: The short-term fuel correction value is obtained by adding the proportional term gain, the integral term gain, and the preset reference value.

3. The method according to claim 1, wherein, The self-learning activation conditions include: the engine speed is greater than the preset minimum operating speed; the engine is in normal fuel injection mode; and the engine has no fuel injection fault.

4. The method according to claim 1, wherein, The updated long-term fuel correction components for each preset operating condition range include: In response to the satisfaction of the self-learning activation condition, the number of times the short-term fuel trim value is calculated is recorded; If the number of calculations exceeds a preset minimum threshold, a short-term fuel correction average is calculated based on the cumulative number of calculations. Based on a two-dimensional mapping table of engine speed and engine intake air volume, the upper and lower limits of the short-term fuel correction mean are imposed to obtain the self-learning demand value. Calculate the mean value of the operating conditions corresponding to the short-term fuel correction value to obtain the self-learning operating condition feature vector; Based on the self-learning requirement value and the corresponding self-learning operating condition feature vector, the linear model output value for each preset operating condition interval is calculated as the long-term fuel correction component.

5. The method according to claim 4, wherein, The number of preset operating condition intervals is 15. The long-term fuel correction component of each preset operating condition interval is stored in a non-volatile memory. It remains valid when the controller is powered off and then powered on again. It is updated when the self-learning activation condition is met and participates in the weighted summation calculation after being read.

6. The method according to claim 4, wherein, The step of calculating the linear model output value for each preset operating condition interval as the long-term fuel correction component based on the self-learning demand value and the corresponding self-learning operating condition feature vector includes: The self-learning operating condition feature vector is normalized to obtain a normalized feature vector, wherein the engine speed component, engine intake air volume component, and engine coolant temperature component of the self-learning operating condition feature vector are divided by the corresponding preset speed reference value, preset intake air volume reference value, and preset coolant temperature reference value, respectively. After multiplying the normalized feature vector by the self-learning requirement value, add the preset feature benchmark value correction value corresponding to the preset working condition interval to obtain the corrected normalized vector for each preset working condition interval. The inner product operation is performed between the corrected normalized vector and the normalized feature vector of each preset working condition interval to obtain the linear basic value of each preset working condition interval. The linear basic value for each preset working condition interval is added to the corresponding preset linear correction value to obtain the linear model output value for each preset working condition interval.

7. The method according to claim 1, wherein, The step of using the current cycle operating condition as input and calculating the probability proportion of the current cycle operating condition in each preset operating condition interval based on a Gaussian function includes: Based on engine operating condition data, a feature mean vector and a feature standard deviation vector are obtained for each preset operating condition interval. The feature mean vector includes the mean engine speed, mean engine intake air volume, and mean engine coolant temperature in the preset operating condition interval. The feature standard deviation vector includes the standard deviation of engine speed, standard deviation of engine intake air volume, and standard deviation of engine coolant temperature in the preset operating condition interval. For each preset working condition interval, calculate the difference between the current working condition and the feature mean vector, divide it by the feature standard deviation vector to obtain a standardized difference vector for calculating the Gaussian function value; By dividing the Gaussian function value of each preset working condition interval by the sum of the Gaussian function values, the probability percentage of the current working condition in each preset working condition interval can be obtained.

8. The method according to claim 1, wherein, The rate of change limit includes: Calculate the difference between the current long-term fuel trim value and the final long-term fuel trim value of the previous period; If the difference is greater than the preset minimum change and less than the preset maximum change, then the long-term fuel correction value of the current period shall be used as the final long-term fuel correction value. If the difference is greater than the preset maximum change, then the final long-term fuel correction value of the previous period plus the preset maximum change is used as the final long-term fuel correction value. If the difference is less than the preset minimum change, then the final long-term fuel correction value of the previous period plus the preset minimum change is used as the final long-term fuel correction value.

9. The method according to claim 1, wherein, The step of obtaining the final fuel trim value using the short-term fuel trim value and the long-term fuel trim value includes: The final fuel correction value is obtained by multiplying the short-term fuel correction value by the long-term fuel correction value.

10. A fuel closed-loop control system, characterized in that, Capable of implementing the fuel closed-loop control method as described in claim 1, comprising: The short-term correction module is used to calculate the short-term fuel correction value based on the actual air-fuel ratio and the demand air-fuel ratio of the current cycle through proportional-integral control. The self-learning module is used to update the long-term fuel correction component of each preset operating condition interval based on the short-term fuel correction value and operating condition of a preset number of cycles in response to the satisfaction of the self-learning activation condition. The operating condition is a three-dimensional feature vector composed of engine speed, engine intake air volume and engine coolant temperature. The probability calculation module is used to calculate the probability percentage of the current working condition in each preset working condition interval based on the Gaussian function, taking the current working condition as input. The long-term correction module is used to obtain the long-term fuel correction value of the operating condition by weighted summation of the probability proportion and the corresponding long-term fuel correction component of each preset operating condition interval and limiting the rate of change. The correction module is used to obtain the final fuel correction value by combining the short-term fuel correction value and the long-term fuel correction value.

11. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 9.

12. A computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 9.

Images