CN113669142A - Method and device for estimating mass flow of soot in original exhaust of engine - Google Patents

Method and device for estimating mass flow of soot in original exhaust of engine Download PDF

Info

Publication number
CN113669142A
CN113669142A CN202111023733.3A CN202111023733A CN113669142A CN 113669142 A CN113669142 A CN 113669142A CN 202111023733 A CN202111023733 A CN 202111023733A CN 113669142 A CN113669142 A CN 113669142A
Authority
CN
China
Prior art keywords
engine
mass flow
neural network
soot
parameter set
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202111023733.3A
Other languages
Chinese (zh)
Other versions
CN113669142B (en
Inventor
刘杰
王梅俊
夏消消
程欢
李林
白桃李
郑攀
周坤诚
李芳�
陈玉俊
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Dongfeng Commercial Vehicle Co Ltd
Original Assignee
Dongfeng Commercial Vehicle Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Dongfeng Commercial Vehicle Co Ltd filed Critical Dongfeng Commercial Vehicle Co Ltd
Priority to CN202111023733.3A priority Critical patent/CN113669142B/en
Publication of CN113669142A publication Critical patent/CN113669142A/en
Application granted granted Critical
Publication of CN113669142B publication Critical patent/CN113669142B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F01—MACHINES OR ENGINES IN GENERAL; ENGINE PLANTS IN GENERAL; STEAM ENGINES
    • F01N—GAS-FLOW SILENCERS OR EXHAUST APPARATUS FOR MACHINES OR ENGINES IN GENERAL; GAS-FLOW SILENCERS OR EXHAUST APPARATUS FOR INTERNAL COMBUSTION ENGINES
    • F01N11/00—Monitoring or diagnostic devices for exhaust-gas treatment apparatus, e.g. for catalytic activity
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F01—MACHINES OR ENGINES IN GENERAL; ENGINE PLANTS IN GENERAL; STEAM ENGINES
    • F01N—GAS-FLOW SILENCERS OR EXHAUST APPARATUS FOR MACHINES OR ENGINES IN GENERAL; GAS-FLOW SILENCERS OR EXHAUST APPARATUS FOR INTERNAL COMBUSTION ENGINES
    • F01N9/00—Electrical control of exhaust gas treating apparatus
    • F01N9/002—Electrical control of exhaust gas treating apparatus of filter regeneration, e.g. detection of clogging
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F01—MACHINES OR ENGINES IN GENERAL; ENGINE PLANTS IN GENERAL; STEAM ENGINES
    • F01N—GAS-FLOW SILENCERS OR EXHAUST APPARATUS FOR MACHINES OR ENGINES IN GENERAL; GAS-FLOW SILENCERS OR EXHAUST APPARATUS FOR INTERNAL COMBUSTION ENGINES
    • F01N9/00—Electrical control of exhaust gas treating apparatus
    • F01N9/005—Electrical control of exhaust gas treating apparatus using models instead of sensors to determine operating characteristics of exhaust systems, e.g. calculating catalyst temperature instead of measuring it directly
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F01—MACHINES OR ENGINES IN GENERAL; ENGINE PLANTS IN GENERAL; STEAM ENGINES
    • F01N—GAS-FLOW SILENCERS OR EXHAUST APPARATUS FOR MACHINES OR ENGINES IN GENERAL; GAS-FLOW SILENCERS OR EXHAUST APPARATUS FOR INTERNAL COMBUSTION ENGINES
    • F01N2900/00—Details of electrical control or of the monitoring of the exhaust gas treating apparatus
    • F01N2900/06—Parameters used for exhaust control or diagnosing
    • F01N2900/16—Parameters used for exhaust control or diagnosing said parameters being related to the exhaust apparatus, e.g. particulate filter or catalyst
    • F01N2900/1606—Particle filter loading or soot amount
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Chemical & Material Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Mechanical Engineering (AREA)
  • Combustion & Propulsion (AREA)
  • Biophysics (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Computational Linguistics (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Chemical Kinetics & Catalysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Combined Controls Of Internal Combustion Engines (AREA)

Abstract

The application relates to a method and a device for estimating the mass flow of soot in the original exhaust of an engine, which relate to the technical field of engine monitoring, and the method comprises the following steps: monitoring the working state of an engine and various soot mass flow influence parameters to obtain a corresponding first parameter set; constructing a first neural network model according to the plurality of first parameter sets and the mass flow of the soot in the primary exhaust of the engine obtained through monitoring; monitoring current soot mass flow influence parameters through the current working state of the engine to obtain a first parameter set corresponding to the current working state of the engine; and obtaining the mass flow of the soot in the original exhaust of the engine according to a first parameter set corresponding to the current working state of the engine and the first neural network model. The method and the device build a first neural network model according to the monitored engine working state and various soot mass flow influence parameters, estimate and obtain the mass flow of the soot in the original row of the engine by means of the first neural network model, and improve the accuracy of estimation to a certain extent.

Description

Method and device for estimating mass flow of soot in original exhaust of engine
Technical Field
The application relates to the technical field of engine monitoring, in particular to a method and a device for estimating the mass flow of soot in the original exhaust of an engine.
Background
With the increasing attention on environmental protection and the increasing requirements of regulations, the problem of diesel engine emission is increasingly prominent, and the requirements on engine aftertreatment are increasingly strict. DOC + DPF is currently the primary means of reducing particulate emissions from diesel engines. As the DPF becomes more soot trapped, it may cause increased engine air system drag, combustion degradation, and increased emissions. Therefore, regeneration of the DPF carrier is required when soot is accumulated to a certain degree. The manner of DPF regeneration is largely divided into two broad categories, passive regeneration and active regeneration.
Passive regeneration reduces soot particulates trapped by the DPF by reacting the soot trapped by the DPF with NOx and oxygen in the exhaust gases. The active regeneration is to inject fuel oil through an injection system arranged at the front end of the DOC, and oxidize the fuel oil through the DOC to release a large amount of heat, so that the exhaust temperature is improved to oxidize soot particles, and the soot particles trapped by the DPF are effectively reduced.
Active regeneration is the mode that soot particles are caught to the reduction DPF that mainly adopts in the diesel engine aftertreatment, adopts active regeneration, will control exhaust temperature through the carbon loading condition of control fuel injection quantity and DPF entrapment, if the carbon loading is more, the temperature is higher will cause DPF carrier to burn out, and the carbon loading is less then can sacrifice the fuel economy of engine. Therefore, the regeneration timing is determined by the soot mass (carbon loading) in the DPF, and safety and economy are ensured.
At the present stage, the estimation result of the DPF carbon loading capacity estimation method is susceptible to external environment, and the estimation result has a large error. Therefore, the method is based on a new engine raw-exhaust soot mass flow estimation technology to meet the current estimation requirement.
Disclosure of Invention
The application provides an engine original-exhaust soot mass flow estimation method and device, which are characterized in that an engine working state and various soot mass flow influence parameters are obtained according to historical monitoring, a first neural network model is built, and then the engine original-exhaust soot mass flow is estimated and obtained according to the engine working state and the various soot mass flow influence parameters obtained through real-time monitoring by means of the first neural network model, so that the estimation accuracy is improved to a certain extent, and a guarantee is provided for later-stage processing work.
In a first aspect, the present application provides a method for estimating soot mass flow in an engine raw exhaust, the method comprising the steps of:
monitoring the working state of an engine and various soot mass flow influence parameters to obtain a corresponding first parameter set;
constructing a first neural network model according to a plurality of different first parameter sets and the corresponding mass flow of the soot in the original exhaust of the engine obtained by monitoring;
monitoring current soot mass flow influence parameters through the current working state of the engine to obtain the first parameter set corresponding to the current working state of the engine;
obtaining the corresponding mass flow of the soot in the original exhaust of the engine according to the first parameter set corresponding to the current working state of the engine and the first neural network model; wherein the content of the first and second substances,
the first neural network model is used for simulating the corresponding relation between the first parameter set and the mass flow of soot in the primary exhaust of the engine.
Specifically, the first parameter set comprises engine speed, engine in-cylinder IMEP, engine EGR opening, engine cooling water temperature, engine in-cylinder injection rail pressure, EGR pipe rear end cylinder inlet air temperature, air throttle rear end cylinder inlet air temperature, ambient humidity and air throttle rear end intake oxygen concentration.
Specifically, the constructing of the first neural network model according to the plurality of different first parameter sets and the corresponding engine raw exhaust soot mass flow obtained through monitoring includes the following steps:
discretizing the data in the first parameter set;
performing normalization processing on the first parameter set subjected to discretization processing;
and constructing a first neural network model according to a plurality of different discretized and normalized first parameter sets and corresponding monitored engine raw-row soot mass flow, and by matching with weight values distributed to different types of parameters in the first parameter set.
Specifically, the discretization processing of the data in the first parameter set includes the following steps:
comparing the data in the first parameter set with a plurality of corresponding data value intervals according to the data type;
and taking the intermediate value of the data value interval as the discretization data value of the data of the corresponding data type in the first parameter set.
Specifically, the normalization processing of the first parameter set after the discretization processing includes the following steps:
calculating to obtain a normalized data value of the data in the first parameter set according to the discretization data value of the data in the first parameter set, the maximum value and the minimum value of the corresponding data type; wherein the content of the first and second substances,
the normalized data value ranges from 0 to 1.
Specifically, the obtaining of the corresponding mass flow of soot in the original exhaust of the engine according to the first parameter set corresponding to the current working state of the engine and the first neural network model includes the following steps:
calculating to obtain a corresponding first neural network model output value according to the first parameter set corresponding to the current working state of the engine and the first neural network model;
and calculating to obtain the engine original-row soot mass flow corresponding to the current working state of the engine according to the first neural network model output value, the maximum value of the engine original-row soot mass flow and the minimum value of the engine original-row soot mass flow.
In a second aspect, the present application provides an engine raw soot mass flow estimation device, comprising:
the parameter detection module is used for monitoring the working state of the engine and various soot mass flow influence parameters to obtain a corresponding first parameter set;
the parameter detection module is also used for monitoring current soot mass flow influence parameters through the current working state of the engine to obtain the first parameter set corresponding to the current working state of the engine;
the neural network construction module is used for constructing a first neural network model according to a plurality of different first parameter sets and the corresponding engine raw-row soot mass flow obtained through monitoring;
the first estimation module is used for obtaining the corresponding mass flow of the soot discharged by the primary exhaust of the engine according to the first parameter set corresponding to the current working state of the engine and the first neural network model; wherein the content of the first and second substances,
the first neural network model is used for simulating the corresponding relation between the first parameter set and the mass flow of soot in the primary exhaust of the engine.
Specifically, the first parameter set comprises engine speed, engine in-cylinder IMEP, engine EGR opening, engine cooling water temperature, engine in-cylinder injection rail pressure, EGR pipe rear end cylinder inlet air temperature, air throttle rear end cylinder inlet air temperature, ambient humidity and air throttle rear end intake oxygen concentration.
Specifically, the neural network building module includes:
a discretization sub-module for discretizing the data in the first parameter set;
the normalization processing submodule is used for performing normalization processing on the first parameter set subjected to discretization processing;
and the neural network construction submodule is used for constructing a first neural network model according to a plurality of different discretized and normalized first parameter sets and corresponding monitored engine raw exhaust carbon smoke mass flow, and by matching with weight values distributed to different types of parameters in the first parameter sets.
Further, the first estimation module is further configured to calculate and obtain a corresponding first neural network model output value according to the first parameter set corresponding to the current working state of the engine and the first neural network model;
the first estimation module is further used for calculating and obtaining the engine original-row soot mass flow corresponding to the current working state of the engine according to the first neural network model output value, the maximum value of the engine original-row soot mass flow and the minimum value of the engine original-row soot mass flow.
The beneficial effect that technical scheme that this application provided brought includes:
this application is according to engine operating condition and the multiple soot mass flow influence parameter that historical monitoring obtained to establish a neural network model, and then with the help of a neural network model, according to the engine operating condition and the multiple soot mass flow influence parameter that real-time supervision obtained, the estimation obtains the former row soot mass flow of engine, improves the degree of accuracy of estimation to a certain extent, provides the guarantee for the processing work in later stage.
Drawings
Interpretation of terms:
DOC: a Diesel Oxidation Catalyst;
DPF: diesel Particulate Filter, Diesel Particulate Filter;
IMEP: an indicative Mean Effective Pressure;
EGR: exhaust Gas Re-circulation System.
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a flow chart of steps of a method for estimating soot mass flow in an engine block provided in an embodiment of the present application;
FIG. 2 is a schematic flow chart of a method for estimating soot mass flow in an engine block provided in an embodiment of the present application;
FIG. 3 is a schematic structural diagram of a neural network model in the method for estimating soot mass flow in the engine block provided in the embodiment of the present application;
FIG. 4 is a block diagram of an engine raw soot mass flow estimation device provided in an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Embodiments of the present application will be described in further detail below with reference to the accompanying drawings.
The embodiment of the application provides an engine original-row soot mass flow estimation method and device, the engine working state and various soot mass flow influence parameters obtained through historical monitoring are obtained, a first neural network model is built, then the engine original-row soot mass flow of the engine is estimated and obtained through the first neural network model according to the engine working state and the various soot mass flow influence parameters obtained through real-time monitoring, the estimation accuracy is improved to a certain extent, and a guarantee is provided for later-stage processing work.
In order to achieve the technical effects, the general idea of the application is as follows:
a method for estimating the soot mass flow of an engine original exhaust comprises the following steps:
s1, monitoring the working state of the engine and various soot mass flow influence parameters to obtain a corresponding first parameter set;
s2, constructing a first neural network model according to the plurality of different first parameter sets and the corresponding engine raw-row soot mass flow obtained through monitoring;
s3, monitoring current soot mass flow influence parameters through the current working state of the engine, and obtaining the first parameter set corresponding to the current working state of the engine;
s4, obtaining the corresponding mass flow of the soot in the original row of the engine according to the first parameter set corresponding to the current working state of the engine and the first neural network model; wherein the content of the first and second substances,
the first neural network model is used for simulating the corresponding relation between the first parameter set and the mass flow of soot in the primary exhaust of the engine.
Embodiments of the present application will be described in further detail below with reference to the accompanying drawings.
In a first aspect, referring to fig. 1 to 3, an embodiment of the present application provides a method for estimating a soot mass flow in a raw exhaust of an engine, where the method includes the following steps:
s1, monitoring the working state of the engine and various soot mass flow influence parameters to obtain a corresponding first parameter set;
s2, constructing a first neural network model according to the plurality of different first parameter sets and the corresponding engine raw-row soot mass flow obtained through monitoring;
s3, monitoring current soot mass flow influence parameters through the current working state of the engine, and obtaining the first parameter set corresponding to the current working state of the engine;
s4, obtaining the corresponding mass flow of the soot in the original row of the engine according to the first parameter set corresponding to the current working state of the engine and the first neural network model; wherein the content of the first and second substances,
the first neural network model is used for simulating the corresponding relation between the first parameter set and the mass flow of soot in the primary exhaust of the engine.
The technical scheme of the embodiment of the application estimates the mass flow of the soot in the original exhaust of the engine based on the neural network:
firstly, selecting soot mass flow influence parameters which can describe the working state of an engine and influence soot mass flow externally as characteristic parameters of a neural network model to be trained, carrying out discretization processing on parameter data obtained by direct measurement or indirect calculation of a sensor according to the important degree of the parameters on the soot mass flow influence, and carrying out normalization processing on the discretized data according to reasonable minimum and maximum values to be used as a training sample of the neural network model to be trained;
secondly, the existing equipment capable of directly measuring or indirectly calculating the mass flow of the soot in the exhaust gas is used for collecting the mass flow of the soot in the original exhaust of the engine, discretizing the collected data according to the preset mass flow precision of the soot, and normalizing the discretized data according to the reasonable minimum value and the reasonable maximum value to be used as a training sample of the neural network model to be trained.
Establishing a neural network model (comprising an input layer, a hidden layer, an output layer, an activation function and a learning rate), updating the acquired and processed engine parameters under various working conditions and the actually measured engine raw soot mass flow samples by applying a BP (back propagation) algorithm to the weight and the offset value of each layer of neurons of the established neural network, and introducing a random gradient algorithm in the training of the neural network model in order to prevent the updating result from being in a local optimal state in the updating process, wherein the final purpose is to enable the error between the soot mass flow output by the neural network model and the actually measured soot mass flow to be in a given target range and be in an optimal state;
and finally, guiding the trained neural network model into an engine control unit, and estimating the mass flow of the soot in the original exhaust of the engine.
In the embodiment of the application, according to the engine operating condition and the multiple soot mass flow influence parameters that historical monitoring obtained to establish a first neural network model, and then with the help of a first neural network model, according to the engine operating condition and the multiple soot mass flow influence parameters that real-time supervision obtained, the estimation obtains the former row soot mass flow of engine, improves the degree of accuracy of estimation to a certain extent, provides the guarantee for the processing work in later stage.
It should be noted that, with the aid of the technical solution of the embodiment of the present application, the problems of large resource occupation, long time consumption, high error probability and poor engine transient description caused by calibrating MAP in the soot mass flow estimation of the original exhaust of the engine can be solved.
Specifically, the first parameter set comprises engine speed, engine in-cylinder IMEP, engine EGR opening, engine cooling water temperature, engine in-cylinder injection rail pressure, EGR pipe rear end cylinder inlet air temperature, air throttle rear end cylinder inlet air temperature, ambient humidity and air throttle rear end intake oxygen concentration.
Specifically, the constructing of the first neural network model according to the plurality of different first parameter sets and the corresponding engine raw exhaust soot mass flow obtained through monitoring includes the following steps:
discretizing the data in the first parameter set;
performing normalization processing on the first parameter set subjected to discretization processing;
and constructing a first neural network model according to a plurality of different discretized and normalized first parameter sets and corresponding monitored engine raw-row soot mass flow, and by matching with weight values distributed to different types of parameters in the first parameter set.
Specifically, the discretization processing of the data in the first parameter set includes the following steps:
comparing the data in the first parameter set with a plurality of corresponding data value intervals according to the data type;
and taking the intermediate value of the data value interval as the discretization data value of the data of the corresponding data type in the first parameter set.
When the discretization processing is performed, the specific operations are as follows:
taking an engine rotating speed parameter as an example, when discretizing the engine rotating speed parameter, considering that the calculation accuracy of an engine rotating speed sensor and the influence on the soot mass flow are very insignificant when the rotating speed fluctuates in a rotating speed range of 50-100rpm, discretizing the collected continuous engine rotating speed at a width of 50rpm for the engine rotating speed, wherein the discretizing method comprises the following steps:
Figure BDA0003240364060000101
wherein, engine _ speedoutThe output value of the discrete engine speed;
engine_speedsigfor the collected engine speed signals, the minimum and maximum speeds of the engine during normal operation are 650rpm and 3000rpm respectively;
625rpm to 675rpm, 675rpm to 725rpm, and 2975rpm to 3025rpm are examples of the intervals for discretization that are set as required.
For the external environment temperature parameters, because the change is slow in the actual operation process of the engine, the change amplitude is relatively small, and the influence of a small change range on the soot mass flow is not obvious;
in addition, the external environment temperature may be discretized in a width of 10 degrees celsius, and the discretization method refers to an engine rotational speed signal discretization method.
Specifically, the normalization processing of the first parameter set after the discretization processing includes the following steps:
calculating to obtain a normalized data value of the data in the first parameter set according to the discretization data value of the data in the first parameter set, the maximum value and the minimum value of the corresponding data type; wherein the content of the first and second substances,
the normalized data value ranges from 0 to 1, i.e., a value between 0 and 1.
Still taking the engine speed parameter as an example, when the engine speed parameter is normalized, the maximum-minimum normalization method is adopted to map the discrete data into the [0,1] interval. The engine speed discrete data is normalized as follows:
Figure BDA0003240364060000111
wherein, engine _ speedtraFor the engine speed data to be trained, the minimum and maximum engine speeds for normal engine operation were 650rpm and 3000rpm, respectively.
Specifically, the obtaining of the corresponding mass flow of soot in the original exhaust of the engine according to the first parameter set corresponding to the current working state of the engine and the first neural network model includes the following steps:
calculating to obtain a corresponding first neural network model output value according to the first parameter set corresponding to the current working state of the engine and the first neural network model;
and calculating to obtain the engine original-row soot mass flow corresponding to the current working state of the engine according to the first neural network model output value, the maximum value of the engine original-row soot mass flow and the minimum value of the engine original-row soot mass flow.
Based on the technical scheme of the embodiment of the application, during specific operation, the established neural network model is shown in fig. 3, and the model has three layers: input layer, hidden layer, output layer.
The input layer is the processing amount of ten characteristic parameters such as the engine speed, the IMEP value, the EGR opening degree, the cooling water temperature, the rail pressure, the air inlet temperature, the air inlet flow, the external environment temperature and humidity, the air inlet pipe oxygen concentration and the like.
The number of nodes of the hidden layer can be adjusted, but is more than or equal to the number of the characteristic parameters, and the hidden layer is a traditional technology in a neural network algorithm.
The output layer is a node, the output is a processing value for processing the soot mass flow of the original exhaust of the engine, and then the true value of the soot mass flow is obtained through inverse normalization processing. And transferring the layers through an activation function, setting an initial learning rate and training iteration times by adopting a sigmoid function as the activation function in the neural network model.
The operation of training the established neural network model is as follows:
firstly, initializing and randomly assigning initial weights and offset of each node of each layer, wherein the assignment range is [0,1], and the numerical values from 0 to 1.
Then updating the weighted values and the offset values of each layer through a BP algorithm according to the output obtained by current calculation and the error size of the actual measured engine original row mass flow normalization, in the process, in order to prevent the situation that the finally obtained weighted values and the offset values are in local optimization and cause that a model built by a training result cannot provide good original row soot mass flow accuracy, a random gradient algorithm is adopted while updating the weighted values and the offset values by utilizing the BP algorithm, and the gradient in the BP algorithm is randomly processed, so that the finally calculated weighted values and the offset values are closer to global optimization, and the model has good accuracy;
it should be noted that the global optimum refers to a set of values of the training result that can make the model calculate the result with the highest accuracy among different training results, i.e., the set of values is the best among all the training results.
Finally, the trained model is led into an engine control unit, the engine control unit also comprises a discretization processing and normalization processing part for the collected characteristic parameters of the engine speed, the IMEP value, the EGR opening degree, the cooling water temperature, the rail pressure, the inlet air temperature, the inlet air flow, the external environment temperature, the humidity and the like, and an inverse normalization processing part for the mass flow of the original exhaust carbon smoke of the model output result, wherein the inverse normalization processing formula is as follows:
x＝y*(xmax-xmin)+xmin
wherein x is an original-row soot mass flow inverse normalization value, namely a numerical value of the original-row soot mass flow estimated or obtained according to a neural network model, and xmaxIs the maximum value of the soot mass flow, x, of the original rowmin is the minimum value of the mass flow of the soot in the original row, and y is the output value of the neural network model.
In a second aspect, referring to fig. 4, an embodiment of the present application provides an engine raw soot mass flow estimation device based on the engine raw soot mass flow estimation method mentioned in the first aspect, the device includes:
the parameter detection module is used for monitoring the working state of the engine and various soot mass flow influence parameters to obtain a corresponding first parameter set;
the parameter detection module is also used for monitoring current soot mass flow influence parameters through the current working state of the engine to obtain the first parameter set corresponding to the current working state of the engine;
the neural network construction module is used for constructing a first neural network model according to a plurality of different first parameter sets and the corresponding engine raw-row soot mass flow obtained through monitoring;
the first estimation module is used for obtaining the corresponding mass flow of the soot discharged by the primary exhaust of the engine according to the first parameter set corresponding to the current working state of the engine and the first neural network model; wherein the content of the first and second substances,
the first neural network model is used for simulating the corresponding relation between the first parameter set and the mass flow of soot in the primary exhaust of the engine.
The technical scheme of the embodiment of the application estimates the mass flow of the soot in the original exhaust of the engine based on the neural network:
firstly, selecting soot mass flow influence parameters which can describe the working state of an engine and influence soot mass flow externally as characteristic parameters of a neural network model to be trained, carrying out discretization processing on parameter data obtained by direct measurement or indirect calculation of a sensor according to the important degree of the parameters on the soot mass flow influence, and carrying out normalization processing on the discretized data according to reasonable minimum and maximum values to be used as a training sample of the neural network model to be trained;
secondly, the existing equipment capable of directly measuring or indirectly calculating the mass flow of the soot in the exhaust gas is used for collecting the mass flow of the soot in the original exhaust of the engine, discretizing the collected data according to the preset mass flow precision of the soot, and normalizing the discretized data according to the reasonable minimum value and the reasonable maximum value to be used as a training sample of the neural network model to be trained.
Establishing a neural network model (comprising an input layer, a hidden layer, an output layer, an activation function and a learning rate), updating the acquired and processed engine parameters under various working conditions and the actually measured engine raw soot mass flow samples by applying a BP (back propagation) algorithm to the weight and the offset value of each layer of neurons of the established neural network, and introducing a random gradient algorithm in the training of the neural network model in order to prevent the updating result from being in a local optimal state in the updating process, wherein the final purpose is to enable the error between the soot mass flow output by the neural network model and the actually measured soot mass flow to be in a given target range and be in an optimal state;
and finally, guiding the trained neural network model into an engine control unit, and estimating the mass flow of the soot in the original exhaust of the engine.
In the embodiment of the application, according to the engine operating condition and the multiple soot mass flow influence parameters that historical monitoring obtained to establish a first neural network model, and then with the help of a first neural network model, according to the engine operating condition and the multiple soot mass flow influence parameters that real-time supervision obtained, the estimation obtains the former row soot mass flow of engine, improves the degree of accuracy of estimation to a certain extent, provides the guarantee for the processing work in later stage.
It should be noted that, with the aid of the technical solution of the embodiment of the present application, the problems of large resource occupation, long time consumption, high error probability and poor engine transient description caused by calibrating MAP in the soot mass flow estimation of the original exhaust of the engine can be solved.
Specifically, the first parameter set comprises engine speed, engine in-cylinder IMEP, engine EGR opening, engine cooling water temperature, engine in-cylinder injection rail pressure, EGR pipe rear end cylinder inlet air temperature, air throttle rear end cylinder inlet air temperature, ambient humidity and air throttle rear end intake oxygen concentration.
Specifically, the neural network building module includes:
a discretization sub-module for discretizing the data in the first parameter set;
the normalization processing submodule is used for performing normalization processing on the first parameter set subjected to discretization processing;
and the neural network construction submodule is used for constructing a first neural network model according to a plurality of different discretized and normalized first parameter sets and corresponding monitored engine raw exhaust carbon smoke mass flow, and by matching with weight values distributed to different types of parameters in the first parameter sets.
It should be noted that, a specific discretization processing mode is based on the discretization processing mode in the engine raw-exhaust soot mass flow estimation method mentioned in the first aspect;
also, the specific normalization processing manner is based on the normalization processing manner in the engine-on-exhaust soot mass flow estimation method mentioned in the first aspect.
Further, the first estimation module is further configured to calculate and obtain a corresponding first neural network model output value according to the first parameter set corresponding to the current working state of the engine and the first neural network model;
the first estimation module is further used for calculating and obtaining the engine original-row soot mass flow corresponding to the current working state of the engine according to the first neural network model output value, the maximum value of the engine original-row soot mass flow and the minimum value of the engine original-row soot mass flow.
It is noted that, in the present application, relational terms such as "first" and "second", and the like, are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The foregoing are merely exemplary embodiments of the present application and are presented to enable those skilled in the art to understand and practice the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (10)

1. An estimation method for soot mass flow of an engine original exhaust is characterized by comprising the following steps:
monitoring the working state of an engine and various soot mass flow influence parameters to obtain a corresponding first parameter set;
constructing a first neural network model according to a plurality of different first parameter sets and the corresponding mass flow of the soot in the original exhaust of the engine obtained by monitoring;
monitoring current soot mass flow influence parameters through the current working state of the engine to obtain the first parameter set corresponding to the current working state of the engine;
obtaining the corresponding mass flow of the soot in the original exhaust of the engine according to the first parameter set corresponding to the current working state of the engine and the first neural network model; wherein the content of the first and second substances,
the first neural network model is used for simulating the corresponding relation between the first parameter set and the mass flow of soot in the primary exhaust of the engine.
2. The engine on-line soot mass flow estimation method of claim 1, characterized in that:
the first parameter set comprises engine rotating speed, engine in-cylinder IMEP, engine EGR opening, engine cooling water temperature, engine in-cylinder injection rail pressure, EGR pipe rear end cylinder inlet air temperature, air throttle rear end cylinder inlet air temperature, environment humidity and air throttle rear end intake oxygen concentration.
3. The method for estimating soot mass flow in an engine block according to claim 1, wherein the step of constructing a first neural network model according to a plurality of different first parameter sets and corresponding soot mass flow in the engine block obtained by monitoring comprises the steps of:
discretizing the data in the first parameter set;
performing normalization processing on the first parameter set subjected to discretization processing;
and constructing a first neural network model according to a plurality of different discretized and normalized first parameter sets and corresponding monitored engine raw-row soot mass flow, and by matching with weight values distributed to different types of parameters in the first parameter set.
4. The engine on-line soot mass flow estimation method as claimed in claim 3, wherein the discretizing the data in the first set of parameters comprises the steps of:
comparing the data in the first parameter set with a plurality of corresponding data value intervals according to the data type;
and taking the intermediate value of the data value interval as the discretization data value of the data of the corresponding data type in the first parameter set.
5. The method for estimating soot mass flow in an engine block as claimed in claim 3, wherein the step of normalizing the discretized first set of parameters comprises the steps of:
calculating to obtain a normalized data value of the data in the first parameter set according to the discretization data value of the data in the first parameter set, the maximum value and the minimum value of the corresponding data type; wherein the content of the first and second substances,
the normalized data value ranges from 0 to 1.
6. The method for estimating soot mass flow of an engine original row according to claim 1, wherein the step of obtaining the corresponding soot mass flow of the engine original row according to the first parameter set corresponding to the current working state of the engine and the first neural network model comprises the steps of:
calculating to obtain a corresponding first neural network model output value according to the first parameter set corresponding to the current working state of the engine and the first neural network model;
and calculating to obtain the engine original-row soot mass flow corresponding to the current working state of the engine according to the first neural network model output value, the maximum value of the engine original-row soot mass flow and the minimum value of the engine original-row soot mass flow.
7. An engine in-line soot mass flow estimation device, comprising:
the parameter detection module is used for monitoring the working state of the engine and various soot mass flow influence parameters to obtain a corresponding first parameter set;
the parameter detection module is also used for monitoring current soot mass flow influence parameters through the current working state of the engine to obtain the first parameter set corresponding to the current working state of the engine;
the neural network construction module is used for constructing a first neural network model according to a plurality of different first parameter sets and the corresponding engine raw-row soot mass flow obtained through monitoring;
the first estimation module is used for obtaining the corresponding mass flow of the soot discharged by the primary exhaust of the engine according to the first parameter set corresponding to the current working state of the engine and the first neural network model; wherein the content of the first and second substances,
the first neural network model is used for simulating the corresponding relation between the first parameter set and the mass flow of soot in the primary exhaust of the engine.
8. The engine in-line soot mass flow estimation device of claim 7, characterized in that:
the first parameter set comprises engine rotating speed, engine in-cylinder IMEP, engine EGR opening, engine cooling water temperature, engine in-cylinder injection rail pressure, EGR pipe rear end cylinder inlet air temperature, air throttle rear end cylinder inlet air temperature, environment humidity and air throttle rear end intake oxygen concentration.
9. The engine on-line soot mass flow estimation device of claim 7, wherein the neural network construction module comprises:
a discretization sub-module for discretizing the data in the first parameter set;
the normalization processing submodule is used for performing normalization processing on the first parameter set subjected to discretization processing;
and the neural network construction submodule is used for constructing a first neural network model according to a plurality of different discretized and normalized first parameter sets and corresponding monitored engine raw exhaust carbon smoke mass flow, and by matching with weight values distributed to different types of parameters in the first parameter sets.
10. The engine in-line soot mass flow estimation device of claim 7, characterized in that:
the first estimation module is further used for calculating and obtaining a corresponding first neural network model output value according to the first parameter set corresponding to the current working state of the engine and the first neural network model;
the first estimation module is further used for calculating and obtaining the engine original-row soot mass flow corresponding to the current working state of the engine according to the first neural network model output value, the maximum value of the engine original-row soot mass flow and the minimum value of the engine original-row soot mass flow.
CN202111023733.3A 2021-08-31 2021-08-31 Method and device for estimating mass flow of soot in original exhaust of engine Active CN113669142B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111023733.3A CN113669142B (en) 2021-08-31 2021-08-31 Method and device for estimating mass flow of soot in original exhaust of engine

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111023733.3A CN113669142B (en) 2021-08-31 2021-08-31 Method and device for estimating mass flow of soot in original exhaust of engine

Publications (2)

Publication Number Publication Date
CN113669142A true CN113669142A (en) 2021-11-19
CN113669142B CN113669142B (en) 2022-12-06

Family

ID=78548097

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111023733.3A Active CN113669142B (en) 2021-08-31 2021-08-31 Method and device for estimating mass flow of soot in original exhaust of engine

Country Status (1)

Country Link
CN (1) CN113669142B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114033538A (en) * 2022-01-12 2022-02-11 潍柴动力股份有限公司 Double DPF regeneration control method and device and engine
CN115059533A (en) * 2022-08-16 2022-09-16 潍柴动力股份有限公司 Method, device, equipment and medium for determining original soot emission of engine

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102008014509A1 (en) * 2008-03-15 2009-02-19 Daimler Ag Method for determining mass of soot arranged in particle filter in exhaust gas system of internal combustion engine, involves determining parameter of particulate filter laden with soot using particulate specific input size
US20120023903A1 (en) * 2010-07-28 2012-02-02 Gm Global Technology Opoerations, Inc. Apparatus and method for monitoring regeneration frequency of a vehicle particulate filter
FR2994709A1 (en) * 2012-08-22 2014-02-28 Peugeot Citroen Automobiles Sa METHOD FOR CORRECTING A SOIL MASS ESTIMATION IN A PARTICLE FILTER
CN109035488A (en) * 2018-08-07 2018-12-18 哈尔滨工业大学（威海） Aero-engine time series method for detecting abnormality based on CNN feature extraction
CN110941917A (en) * 2019-12-17 2020-03-31 凯龙高科技股份有限公司 Diesel engine DPF carbon loading capacity calculation method based on pressure drop
CN112883653A (en) * 2021-03-19 2021-06-01 广西玉柴机器股份有限公司 Artificial intelligence-based modeling method for real-time engine model

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102008014509A1 (en) * 2008-03-15 2009-02-19 Daimler Ag Method for determining mass of soot arranged in particle filter in exhaust gas system of internal combustion engine, involves determining parameter of particulate filter laden with soot using particulate specific input size
US20120023903A1 (en) * 2010-07-28 2012-02-02 Gm Global Technology Opoerations, Inc. Apparatus and method for monitoring regeneration frequency of a vehicle particulate filter
FR2994709A1 (en) * 2012-08-22 2014-02-28 Peugeot Citroen Automobiles Sa METHOD FOR CORRECTING A SOIL MASS ESTIMATION IN A PARTICLE FILTER
CN109035488A (en) * 2018-08-07 2018-12-18 哈尔滨工业大学（威海） Aero-engine time series method for detecting abnormality based on CNN feature extraction
CN110941917A (en) * 2019-12-17 2020-03-31 凯龙高科技股份有限公司 Diesel engine DPF carbon loading capacity calculation method based on pressure drop
CN112883653A (en) * 2021-03-19 2021-06-01 广西玉柴机器股份有限公司 Artificial intelligence-based modeling method for real-time engine model

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114033538A (en) * 2022-01-12 2022-02-11 潍柴动力股份有限公司 Double DPF regeneration control method and device and engine
CN114033538B (en) * 2022-01-12 2022-03-11 潍柴动力股份有限公司 Double DPF regeneration control method and device and engine
CN115059533A (en) * 2022-08-16 2022-09-16 潍柴动力股份有限公司 Method, device, equipment and medium for determining original soot emission of engine

Also Published As

Publication number Publication date
CN113669142B (en) 2022-12-06

Similar Documents

Publication Publication Date Title
US10635976B2 (en) Machine learning system for estimating a temperature of an exhaust purification catalyst
CN113669142B (en) Method and device for estimating mass flow of soot in original exhaust of engine
CN102191979B (en) Engine-out NOx virtual sensor for an internal combustion engine
CN102691559B (en) The evaluation method of selective catalytic reduction system operating drop amount
CN111639683B (en) Method for evaluating nitrogen oxide emission grade of diesel vehicle based on OBD technology
Shen et al. Gaussian mixture model clustering-based knock threshold learning in automotive engines
CN106837496B (en) Engine particulate purifying regeneration control system
CN107577910A (en) A kind of Vehicular exhaust retrieving concentration method based on profound neutral net
CN104898641A (en) Vehicle fault detection method and apparatus
CN113379145B (en) Vehicle emission remote online monitoring platform, system and vehicle-mounted terminal
CN112983608A (en) Particle trap carbon load calculation method and system, controller and storage medium
CN113657651A (en) Diesel vehicle emission prediction method, medium and equipment based on deep migration learning
CN107023367B (en) A kind of SCR system of diesel engine ammonia input pickup fault diagnosis and fault tolerant control method
CN113530656A (en) DPF fault monitoring method and device
JP2020063699A (en) Control device of internal combustion engine and its control method, and learning model for controlling internal combustion engine and its learning method
CN116558834A (en) PEMS emission analysis method, PEMS emission analysis device, PEMS emission analysis computer equipment and PEMS emission analysis storage medium
Bellone et al. Comparison of CNN and LSTM for Modeling Virtual Sensors in an Engine
JP2008516141A (en) Improved exhaust gas temperature estimation method and apparatus, and internal combustion engine equipped with exhaust gas temperature estimation apparatus
JP2011132915A (en) Device for estimating physical quantity
US11561212B1 (en) Method for determining NOx sensor data falsification based on remote emission monitoring
CN110131017B (en) Method for predicting ash deposition amount of automobile particulate filter and regeneration system
Pla et al. Ammonia injection failure diagnostic and correction in engine after-treatment system by NOx and NH3 emissions observation
CN114704361A (en) DPF partial regeneration control method, device, electronic apparatus and storage medium
CN114282680A (en) Vehicle exhaust emission prediction method and system based on machine learning algorithm
CN103711548A (en) Method and system for preventing diesel engine particle trap from melting

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant