CN111027733A - Petrochemical device product yield optimization method based on big data technology - Google Patents
Petrochemical device product yield optimization method based on big data technology Download PDFInfo
- Publication number
- CN111027733A CN111027733A CN201811175992.6A CN201811175992A CN111027733A CN 111027733 A CN111027733 A CN 111027733A CN 201811175992 A CN201811175992 A CN 201811175992A CN 111027733 A CN111027733 A CN 111027733A
- Authority
- CN
- China
- Prior art keywords
- product yield
- neural network
- data
- optimal
- optimization method
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Mining
Abstract
The invention discloses a petrochemical device product yield optimization method based on big data technology, which comprises the following steps: collecting historical production data of the device, and cleaning and setting the collected data to obtain a data sample for optimizing the product yield; carrying out correlation analysis on the data sample, and screening out process parameters related to the product yield; establishing a neural network model for describing the relation between the product yield and relevant process parameters by using a neural network algorithm, taking the data sample as a training sample, and training the coefficient of the neural network model by using a genetic algorithm to generate a product yield prediction model; under the constraint condition of the device, aiming at maximizing economic benefit, the product yield prediction model is utilized to determine the optimal product yield and the value of the adjustable operation variable in production under the optimal product yield.
Description
Technical Field
The invention belongs to the field of application of big data technology, and particularly relates to a petrochemical device product yield optimization method based on big data technology.
Background
The petrochemical industry is a basic industry, provides matched articles and services for agriculture, energy, traffic, machinery, electronics, textiles, light industry, buildings, building materials and other industrial and agricultural industries and daily life of people, plays a significant role in national economy, and is a supporting industry of China. Petrochemical production is a series of complex processes combining physical and chemical reactions, and some processes lack an effective and strict description of mechanism. In order to overcome the defect that the petrochemical industry is difficult to analyze problems by adopting the traditional technical means, the introduction of the big data technology is a new idea and method for solving the defect.
In 2001, a study by Gartner reported the first appearance of a "big data" concept. In 2012, the term big data is increasingly mentioned, which is used to describe and define the massive data generated in the information explosion era and to name the technical development and innovation related to the data. More and more government, enterprise, etc. organizations are becoming aware that data is becoming their most important assets and that data analysis capabilities are becoming their core competence. In 2013, 3, 22 months, the Aubama government announces that 2 hundred million dollars are invested to pull the development of big data related industries, and the big data strategy is upgraded to the national strategy. The Oubama government defines data as "future new petroleum" and indicates that the ability of a country to possess the scale, activity and explain applications of data will become an important component of the integrated national force; in the future, the possession and control of data will become even another national core asset beyond the terrestrial, marine, and air rights. This is again a "violent leap" at the national level into the information field following the initiative of the U.S. government in 1993 on the "information highway" program at 9 months.
The big data technology also shows a rapid development trend in China, and the big data technology applied to the industries of the financial industry, the internet, the communication, the electronic commerce and the like in China has good effect.
In recent years, the web retail third party transaction platform and the e-commerce website such as the kyoto, the naobao and the like are developed vigorously, and a large number of operators, consumers and related goods and services are gathered, and thus a large amount of data is generated. By utilizing a big data technology, deep mining and deep analysis are carried out on online shopping, online consumption, online payment data and the like, a large amount of valuable information and statistical rules can be found, and positive effects are generated on layout, health and ordered development of future internet economy is promoted, electronic commerce activities of operators and consumers are further standardized, and the macroscopic regulation and control and supervision of the country to the field are enhanced.
In the petrochemical field, experimental exploration is carried out abroad, for example, great amount of pipeline sensing data are analyzed by British oil (BP) company, and the correlation between the pipeline pressure data and the pipeline corrosion degree is found and can be used as the representation of the pipeline corrosion degree. Based on the historical data of crude oil transportation, the crude oil is classified according to the corrosion capability of the pipeline, so that the crude oil transportation is better arranged, and the risk of pipeline corrosion is reduced.
For the petroleum industry, domestic petroleum enterprises will apply more new technologies to various fields such as strategic decision, scientific and technological research and development, production and operation, safety and environmental protection, and the like, and the purpose is to mine more wealth and value from large data resources. The big data application is the necessity of deep informatization and deep fusion of IT and business in the petroleum industry, and has wider and wider application prospect in the petroleum and petrochemical industry in China. With the gradual reduction of petroleum reserves, the exploration and development difficulty in the industrial chain of the petroleum and petrochemical industry is increasing day by day, and the maturity of informatization becomes a primary factor influencing the growth range of the industry.
The petrochemical industry has the characteristics of complicated raw material properties and complex production process, which is also the reason that the big data technology is not generally applied in the petrochemical industry. At present, the research of applying big data technology in the petrochemical engineering device yield optimization in China is in a blank stage, completely depends on the experience of experts, and can not simply and effectively optimize the yield and determine the specific operation condition when the optimal yield is not determined.
Disclosure of Invention
In order to solve the above technical problems, the present invention provides a method for optimizing the yield of a petrochemical plant based on big data technology, so as to solve the problems in the prior art that the yield cannot be optimized simply and effectively and the specific operating conditions for the optimal yield cannot be determined because the yield is optimized completely depending on the expertise.
The purpose of the invention is realized by the following technical scheme:
a petrochemical device product yield optimization method based on big data technology comprises the following steps:
collecting historical production data of the device, and cleaning and setting the collected data to obtain a data sample for optimizing the product yield;
carrying out correlation analysis on the data sample, and screening out process parameters related to the product yield, wherein the process parameters comprise raw oil property parameters influencing the product yield and adjustable operation variables in production;
establishing a neural network model for describing the relation between the product yield and relevant process parameters by using a neural network algorithm, taking the data sample as a training sample, and training the coefficient of the neural network model by using a genetic algorithm to generate a product yield prediction model;
under the constraint condition of the device, aiming at maximizing economic benefit, the product yield prediction model is utilized to determine the optimal product yield and the value of the adjustable operation variable in production under the optimal product yield.
Preferably, the historical production data includes production process data and analytical assay data.
Preferably, cleansing and tuning the collected data includes deleting raw data in the data having an error exceeding a given threshold.
Preferably, a rough set algorithm is adopted to perform correlation analysis on the data samples, and process parameters related to the product yield are screened out according to the magnitude of the correlation coefficient.
Preferably, the neural network algorithm is a BP neural network algorithm.
The method comprises the following steps of establishing a neural network model for describing the relation between the product yield and relevant process parameters by using a neural network algorithm, training the coefficient of the neural network model by using a genetic algorithm by using the data sample as a training sample, and generating a product yield prediction model, wherein the method specifically comprises the following steps:
selecting different hidden layer neuron numbers to establish different neural network models, introducing the data samples into the neural network models for training, determining the optimal hidden layer neuron number through detecting model precision, determining the structure of the optimal BP neural network according to the optimal hidden layer neuron number, and establishing the optimal BP neural network model;
and training the coefficients of the optimal BP neural network model by using a genetic algorithm to generate a product yield prediction model.
Preferably, the number of hidden layer neurons is from 6 to 25.
In addition, according to another embodiment of the present invention, preferably, the product yield prediction model is tested, and only when the precision of the product yield prediction model meets a preset precision condition, the product yield prediction model is used to determine the optimal product yield and the value of the controllable operation variable in production at the optimal product yield.
Preferably, the device constraints are within the allowable operating range of the device.
Preferably, the operating scheme to obtain the optimal product yield is shown graphically.
Compared with the prior art, the invention has the following advantages or beneficial effects:
1) the method comprises the steps of establishing a neural network model for describing the relation between the product yield and relevant process parameters by using a neural network algorithm, training coefficients of the neural network model by using a genetic algorithm by using a data sample as a training sample, and generating a product yield prediction model; under the constraint condition of the device, the optimal product yield and the operation condition for obtaining the optimal product yield are determined by utilizing the product yield prediction model with the aim of maximizing economic benefits, so that difficult mechanism analysis is avoided, the method is visual and effective, and a convenient method is provided for product distribution optimization of a petrochemical device.
2) The invention preferably adopts a rough set method to screen the properties and the operation variables of the raw oil which affect the product yield, and can carry out correlation analysis on the data samples without any prior knowledge except the data samples. And fitting a prediction model of the product yield by adopting a BP neural network, and optimizing neural network parameters and adjustable operating conditions by adopting a genetic algorithm so as to obtain an operating scheme of the optimal product yield.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention.
FIG. 1 is a diagram of a neural network topology employed in the present invention;
FIG. 2 is a flow chart of a genetic algorithm employed in the present invention;
FIG. 3 is a flow chart of a method according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating the relationship between the number of hidden layer neurons and the mean square error for the embodiment shown in FIG. 3;
FIG. 5 is a flow chart of a method according to another embodiment of the present invention.
Detailed Description
The following detailed description of the embodiments of the present invention will be provided with reference to the drawings and examples, so that how to apply the technical means to solve the technical problems and achieve the technical effects can be fully understood and implemented. It should be noted that, as long as there is no conflict, the embodiments and the features of the embodiments of the present invention may be combined with each other, and the technical solutions formed are within the scope of the present invention.
In the following description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced without some of these specific details or with other methods described herein.
The technical scheme of the invention is further explained by combining the attached drawings.
Fig. 1 is a diagram showing a topology structure of a neural network used in the present invention. The method abstracts the human brain neuron network from the information processing angle, establishes a certain simple model, and forms different networks according to different connection modes. A neural network is an operational model, which is formed by a large number of nodes (neurons) connected to each other. Each node represents a particular output function, called the stimulus function. Every connection between two nodes represents a weighted value, called weight, for the signal passing through the connection, which is equivalent to the memory of the artificial neural network. The output of the network is different according to the connection mode of the network, the weight value and the excitation function. The network itself is usually an approximation to some algorithm or function in nature, and may also be an expression of a logic strategy.
The neural network algorithm calculates new input and output results according to part of input-output data rules, has strong self-learning and self-training capabilities, is a new information processing model simulating the biological nervous system, and has a unique structure, so people expect that the neural network algorithm can solve the problems which are difficult to solve by using the traditional method.
In practical application, the neural network can be divided into a BP network, a radial basis network, a recursive network and the like, and the BP neural network is adopted as an industrial production process with complicated internal mechanism according to the need. The BP neural network is able to learn and store a large number of input-output pattern mappings without prior disclosure of mathematical equations describing such mappings. The BP neural network model topological structure comprises an input layer, a hidden layer and an output layer.
The BP algorithm consists of two processes, forward computation of the data stream (forward propagation) and back propagation of the error signal. In forward propagation, the propagation direction is input layer → hidden layer → output layer, and the state of each layer of neurons only affects the next layer of neurons. If the desired output is not available at the output layer, the back propagation flow of the error signal is reversed. By alternately carrying out the two processes, an error function gradient descending strategy is executed in the weight vector space, and a group of weight vectors are dynamically and iteratively searched, so that the network error function reaches the minimum value, and the information extraction and memory processes are finished.
Fig. 2 is a flow chart of a genetic algorithm, function optimization is a classical application field of the genetic algorithm and is also a common means for evaluating the performance of the genetic algorithm, and various complex test functions are constructed at present: continuous and discrete functions, convex and concave functions, low and high dimensional functions, unimodal and multimodal functions, and the like. For some nonlinear, multi-model and multi-target function optimization problems, other optimization methods are difficult to solve, and a genetic algorithm can conveniently obtain a better result. The parameter optimization based on the neural network is carried out by using a genetic algorithm through verification, and the result accords with the actual operation result. The specific algorithm flow is common knowledge of those skilled in the art, and will not be described herein (please refer to fig. 2).
There are several key elements in the flow of genetic algorithms:
1) the fitness value evaluates a function. This function is the key of the algorithm and is used to calculate the fitness of each individual in the population. That is, whether the evaluation score of the bred offspring is good, general or poor is quantified by the function.
2) Selecting an operation rule: also known as a selection operator. The purpose of selection is to inherit the optimized individuals (or solutions) directly to the next generation or to generate new individuals by pairwise crossing and then to inherit to the next generation. The selection operation is based on fitness evaluation of individuals in the population. The most common is the roulette algorithm, which is also the selection algorithm used in this embodiment, and this algorithm is simple and effective. There are currently more than 10 selection algorithms, and various services can be selected as required.
Fig. 3 is a flowchart of a method according to an embodiment of the invention. The invention mainly utilizes the neural network and the genetic algorithm, and determines the optimal yield of the product and the operation condition when the optimal yield is determined by the BP neural network model optimized by the genetic algorithm. The method mainly comprises the following steps:
s11: and collecting historical production data of the device, cleaning and setting the collected data, and obtaining a big data sample for product yield analysis and optimization.
The historical production data mainly comprises production process data and analysis and test data, data with significant errors in the original data are removed when the collected data are cleaned and set so as to reduce the influence of random errors on measured values, and when variables which are not measured exist, the variables are estimated and supplemented.
S12: and carrying out correlation analysis on the data sample, and screening out process parameters having great influence on the product yield according to the magnitude of the correlation coefficient. The process parameters include parameters describing the properties of the crude oil and manipulated variables which can be controlled during production. Since the rough set algorithm does not require any a priori knowledge other than the data samples to perform the correlation analysis on the data samples, the rough set algorithm is preferably used for the data correlation analysis in the present embodiment.
S13: and establishing a neural network model for describing the relation between the product yield and relevant process parameters by using a BP neural network algorithm, taking the data sample as a training sample, and training the coefficient of the neural network model by using a genetic algorithm to generate a product yield prediction model.
Generally, the BP neural network model employs a three-layer network structure, i.e., an input layer, a hidden layer, and an output layer. The neurons in the input layer are the process parameters screened in the previous step, and the number of neurons in the output layer is usually 1 (i.e., gasoline yield in this example). The neural network model sets the maximum iteration times and the learning rate during learning, and trains the BP neural network by adopting an L-M algorithm. The L-M (Levenberg-Marquardt) algorithm is a nonlinear least square algorithm and is based on the improvement of Gauss-Newton algorithm, and the improved L-M algorithm has stronger adaptability and faster convergence rate on the least square problem with high nonlinearity degree, so the L-M algorithm is commonly used for parameter fitting, parameter estimation and function approximation in engineering technology and scientific experiments.
The most important task of the BP neural network is to determine the number of hidden layer neurons, because the number of hidden layer neurons has a great influence on the prediction accuracy of the BP neural network: the number of the neurons is too many, the training time is increased, and the network is easy to over-fit; the number of nodes is too small, the network cannot learn well, and the training times need to be increased. The hidden layer neuron number calculation formula is as follows:
in the formula: h is the number of hidden layer neurons; m is the number of neurons in the input layer; n is the number of neurons in the output layer; l is a constant between 1 and 10.
Due to the massive parallel distribution structure and the nonlinear dynamic characteristic of the BP neural network, the ideal effect is difficult to obtain in actual calculation through the above formula. In order to find out the optimal network structure, in this embodiment, the number of hidden layer neurons is gradually increased from 6 to 25 (as shown in fig. 4), different numbers of hidden layer neurons are selected to establish different BP neural network models, the selected training data is introduced into the BP neural network model for training, and then the model accuracy is checked, for example, the mean square error of each verification sample is compared, so as to determine the optimal number of hidden layer neurons, thereby determining the structure of the optimal BP neural network and establishing the optimal BP neural network model. FIG. 4 is a diagram illustrating the relation between the number of hidden layer neurons and the mean square error according to the present embodiment.
In this embodiment, in the gasoline yield optimization process, a genetic algorithm is adopted to determine the number of initial populations and set the number of iterations. In the process of calculating the extreme value, the trained neural network prediction result is used as an individual fitness value, and the calculation formula is as follows:
Fi＝ygasoline yield i
In the formula: fiFitness value for individual i; y isGasoline yield iAnd (4) obtaining the gasoline yield of the individual i by the neural network prediction value.
Selecting individuals by adopting a roulette method, wherein the individuals with larger fitness are selected to enter the next generation, and the calculation formula is as follows:
fi＝k/Fi
the smaller the fitness value, the better, so before individual selection, the reciprocal of the fitness value is calculated to obtain the processed individual fitness. In the formula: k is a coefficient, fiIs the processed fitness value, f, of the individual ijIs the fitness value of the individual j after treatment, N is the number of population individuals, piFor each individual quiltProbability of inheritance into the next generation population.
The cross operation is a main mode for generating new individuals, but the cross operation of randomly selecting individuals can cause deletion of effective genes, and the cross probability is generally between 0.6 and 0.9; the mutation operation can provide new genes for the population, and the mutation probability is usually small and generally ranges from 0 to 0.1.
S14: under the constraint condition of the device, aiming at maximizing economic benefit, the product yield prediction model is utilized to determine the optimal product yield and the value of the adjustable operation variable in production under the optimal product yield.
In the present embodiment, the device constraint condition refers to an operation range allowed by the device. Specifically, according to the production safety requirement, setting a regulation range for a variable related to the product yield prediction model (wherein an uncontrollable variable is set as a fixed value), such as: the reaction temperature was controlled within the range of + -5 ℃. And when the product yield prediction model is operated, the regulating and controlling ranges of all variables are used as constraint conditions, the optimal product yield is searched under the constraint conditions, and the values of all the controllable operation variables under the optimal product yield are obtained.
S15: the operating regime to achieve the optimum product yield is shown for the plant operator.
In this embodiment, the predicted value of the product yield, the calculation result of the operation optimization, and the operation scheme (including the operation conditions) for obtaining the optimal product yield are graphically displayed, so as to provide a reference for the operator to optimize the operation of the apparatus.
Fig. 5 is a flowchart of a method according to another embodiment of the present invention. The embodiment is different from the first embodiment, and in the embodiment, the method further includes the step of taking the data sample as a test sample, testing the product yield prediction model, and then judging whether the precision of the product yield prediction model meets a preset precision condition. And only when the precision of the product yield prediction model meets a preset precision condition, determining the optimal product yield and the value of the adjustable operation variable in production under the optimal product yield by using the product yield prediction model so as to ensure the accuracy of the result. Otherwise, returning to the previous step to execute the previous step again.
In this embodiment, the method mainly includes the following steps:
s21: and collecting historical production data of the device, cleaning and setting the collected data, and obtaining a big data sample for product yield analysis and optimization.
The historical production data mainly comprises production process data and analysis and test data, data with significant errors in the original data are removed when the collected data are cleaned and set so as to reduce the influence of random errors on measured values, and when variables which are not measured exist, the variables are estimated and supplemented.
S22: and carrying out correlation analysis on the data sample, and screening out process parameters having great influence on the product yield according to the magnitude of the correlation coefficient. The process parameters include parameters describing the properties of the crude oil and manipulated variables which can be controlled during production. Since the rough set algorithm does not require any a priori knowledge other than the data samples to perform the correlation analysis on the data samples, the rough set algorithm is preferably used for the data correlation analysis in the present embodiment.
S23: and establishing a neural network model for describing the relation between the product yield and relevant process parameters by using a BP neural network algorithm, taking the data sample as a training sample, and training the coefficient of the neural network model by using a genetic algorithm to generate a product yield prediction model.
S24: and testing the product yield prediction model by taking the data sample as a test sample. And judging whether the precision of the product yield prediction model meets a preset precision condition. If the precision of the product yield prediction model meets the preset precision condition, continuing to execute the step S25; otherwise, the process returns to step S23 to retrain the product yield prediction model.
S25: under the constraint condition of the device, aiming at maximizing economic benefit, the product yield prediction model is utilized to determine the optimal product yield and the value of the adjustable operation variable in production under the optimal product yield.
In the present embodiment, the device constraint condition refers to an operation range allowed by the device.
S26: the operating regime to achieve the optimum product yield is shown for the plant operator.
In this embodiment, the predicted value of the product yield, the calculation result of the operation optimization, and the operation scheme (including the operation conditions) for obtaining the optimal product yield are graphically displayed, so as to provide a reference for the operator to optimize the operation of the apparatus.
The invention solves the problems that the yield optimization in the prior art completely depends on the expert experience, the yield can not be simply and effectively optimized, and the specific operation condition for the optimal yield can not be determined. Establishing a neural network model for describing the relation between the product yield and relevant process parameters by using a neural network algorithm, taking the data sample as a training sample, and training the coefficient of the neural network model by using a genetic algorithm to generate a product yield prediction model; under the constraint condition of the device, the optimal product yield and the operation condition for obtaining the optimal product yield are determined by utilizing the product yield prediction model with the aim of maximizing economic benefits, so that difficult mechanism analysis is avoided, the method is visual and effective, and a convenient method is provided for product distribution optimization of a petrochemical device.
It should be noted that, although the embodiments of the present invention are described above, the descriptions are only for the convenience of understanding the present invention and are not intended to limit the present invention. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (10)
1. A petrochemical device product yield optimization method based on big data technology comprises the following steps:
collecting historical production data of the device, and cleaning and setting the collected data to obtain a data sample for optimizing the product yield;
carrying out correlation analysis on the data sample, and screening out process parameters related to the product yield, wherein the process parameters comprise raw oil property parameters influencing the product yield and adjustable operation variables in production;
establishing a neural network model for describing the relation between the product yield and relevant process parameters by using a neural network algorithm, taking the data sample as a training sample, and training the coefficient of the neural network model by using a genetic algorithm to generate a product yield prediction model;
under the constraint condition of the device, aiming at maximizing economic benefit, the product yield prediction model is utilized to determine the optimal product yield and the value of the adjustable operation variable in production under the optimal product yield.
2. The optimization method according to claim 1,
the historical production data includes production process data and analytical assay data.
3. The optimization method according to claim 1,
cleansing and tuning the collected data includes deleting raw data in the collected data that has an error exceeding a given threshold.
4. The optimization method according to claim 1,
and performing correlation analysis on the data sample by adopting a rough set algorithm, and screening out process parameters related to the product yield according to the magnitude of the correlation coefficient.
5. The optimization method according to claim 1,
the neural network algorithm is a BP neural network algorithm.
6. The optimization method according to claim 5,
establishing a neural network model for describing the relation between the product yield and relevant process parameters by using a neural network algorithm, training the coefficients of the neural network model by using the data sample as a training sample and using a genetic algorithm to generate a product yield prediction model, and the method comprises the following steps of:
selecting different hidden layer neuron numbers to establish different neural network models, introducing the data samples into the neural network models for training, determining the optimal hidden layer neuron number through detecting model precision, determining the structure of the optimal BP neural network according to the optimal hidden layer neuron number, and establishing the optimal BP neural network model;
and training the coefficients of the optimal BP neural network model by using a genetic algorithm to generate a product yield prediction model.
7. The optimization method of claim 6, wherein different neural network models are created by selecting different hidden layer neuron numbers, wherein the hidden layer neuron numbers are from 6 to 25.
8. The optimization method of claim 1, further comprising:
and testing the product yield prediction model, and determining the optimal product yield and the value of the adjustable operation variable in production at the optimal product yield by using the product yield prediction model only when the precision of the product yield prediction model meets the preset precision condition.
9. The optimization method according to claim 1,
the device constraints are within the allowable operating range of the device.
10. The optimization method according to claim 1,
the operating scheme to achieve the optimal product yield is shown graphically.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811175992.6A CN111027733A (en) | 2018-10-10 | 2018-10-10 | Petrochemical device product yield optimization method based on big data technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811175992.6A CN111027733A (en) | 2018-10-10 | 2018-10-10 | Petrochemical device product yield optimization method based on big data technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111027733A true CN111027733A (en) | 2020-04-17 |
Family
ID=70191628
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811175992.6A Pending CN111027733A (en) | 2018-10-10 | 2018-10-10 | Petrochemical device product yield optimization method based on big data technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111027733A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111598306A (en) * | 2020-04-22 | 2020-08-28 | 汉谷云智(武汉)科技有限公司 | Oil refinery production plan optimization method and device |
| CN112464554A (en) * | 2020-11-03 | 2021-03-09 | 桂林理工大学 | Operating parameter optimization method of gasoline refining equipment |
| CN113420799A (en) * | 2021-06-10 | 2021-09-21 | 北京宜能高科科技有限公司 | Sample enhancement method, model training method and system |
| CN115496426A (en) * | 2022-11-16 | 2022-12-20 | 承德石油高等专科学校 | Industrial big data intelligent analysis decision-making method and system |
| CN115879638A (en) * | 2022-12-30 | 2023-03-31 | 东北石油大学 | Carbon emission prediction method for oil field transfer station system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090182693A1 (en) * | 2008-01-14 | 2009-07-16 | Halliburton Energy Services, Inc. | Determining stimulation design parameters using artificial neural networks optimized with a genetic algorithm |
| CN103077267A (en) * | 2012-12-28 | 2013-05-01 | 电子科技大学 | Parameter sound source modeling method based on improved BP (Back Propagation) neural network |
| CN104778338A (en) * | 2015-05-03 | 2015-07-15 | 长春工业大学 | Optimization method for set value of low-energy-consumption molecular distillation process |
-
2018
- 2018-10-10 CN CN201811175992.6A patent/CN111027733A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090182693A1 (en) * | 2008-01-14 | 2009-07-16 | Halliburton Energy Services, Inc. | Determining stimulation design parameters using artificial neural networks optimized with a genetic algorithm |
| CN103077267A (en) * | 2012-12-28 | 2013-05-01 | 电子科技大学 | Parameter sound source modeling method based on improved BP (Back Propagation) neural network |
| CN104778338A (en) * | 2015-05-03 | 2015-07-15 | 长春工业大学 | Optimization method for set value of low-energy-consumption molecular distillation process |
Non-Patent Citations (2)
| Title |
|---|
| 王江荣: "基于主成分分析和遗传算法的模糊神经网络在矿井突水判断中的应用", 《计量技术》 * |
| 赵媛媛: "数据挖掘技术在MIP工艺汽油收率优化中的应用", 《中国优秀博硕士学位论文全文数据库（硕士）》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111598306A (en) * | 2020-04-22 | 2020-08-28 | 汉谷云智(武汉)科技有限公司 | Oil refinery production plan optimization method and device |
| CN111598306B (en) * | 2020-04-22 | 2023-07-18 | 汉谷云智(武汉)科技有限公司 | Method and device for optimizing production plan of oil refinery |
| CN112464554A (en) * | 2020-11-03 | 2021-03-09 | 桂林理工大学 | Operating parameter optimization method of gasoline refining equipment |
| CN113420799A (en) * | 2021-06-10 | 2021-09-21 | 北京宜能高科科技有限公司 | Sample enhancement method, model training method and system |
| CN115496426A (en) * | 2022-11-16 | 2022-12-20 | 承德石油高等专科学校 | Industrial big data intelligent analysis decision-making method and system |
| CN115879638A (en) * | 2022-12-30 | 2023-03-31 | 东北石油大学 | Carbon emission prediction method for oil field transfer station system |
| CN115879638B (en) * | 2022-12-30 | 2023-06-09 | 东北石油大学 | Carbon emission prediction method for oilfield transfer station system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111027733A (en) | Petrochemical device product yield optimization method based on big data technology | |
| Naderi et al. | Novel statistical forecasting models for crude oil price, gas price, and interest rate based on meta-heuristic bat algorithm | |
| Stepchenko et al. | Nonlinear, non-stationary and seasonal time series forecasting using different methods coupled with data preprocessing | |
| CN109242558A (en) | A kind of management data prediction technique, readable storage medium storing program for executing and forecasting system neural network based | |
| Buragohain | Adaptive network based fuzzy inference system (ANFIS) as a tool for system identification with special emphasis on training data minimization | |
| Jafarzadeh-Ghoushchi et al. | Performance study of artificial neural network modelling to predict carried weight in the transportation system | |
| Alhroob et al. | Adaptive fuzzy map approach for accruing velocity of big data relies on fireflies algorithm for decentralized decision making | |
| Júnior et al. | An approach for evolving neuro-fuzzy forecasting of time series based on parallel recursive singular spectrum analysis | |
| Haixiang et al. | Optimizing reservoir features in oil exploration management based on fusion of soft computing | |
| Divasón et al. | PSO-PARSIMONY: A method for finding parsimonious and accurate machine learning models with particle swarm optimization. Application for predicting force–displacement curves in T-stub steel connections | |
| Wang et al. | Mutual information-weighted principle components identified from the depth features of stacked autoencoders and original variables for oil dry point soft sensor | |
| Horák et al. | Comparison of neural networks and regression time series when predicting the export development from the USA to PRC | |
| Gu et al. | Fuzzy time series forecasting based on information granule and neural network | |
| Sugumar et al. | A technique to stock market prediction using fuzzy clustering and artificial neural networks | |
| Jadli et al. | A Novel LSTM-GRU-Based Hybrid Approach for Electrical Products Demand Forecasting. | |
| Wang et al. | Color trend prediction method based on genetic algorithm and extreme learning machine | |
| Chen et al. | The dynamic extensions of fuzzy grey cognitive maps | |
| Luo | Application of BP neural network in economic management of coastal area | |
| Fu et al. | Prediction of financial economic time series based on group intelligence algorithm based on machine learning | |
| Ahmad | Analyzing different machine learning techniques for stock market prediction | |
| Miskony et al. | A randomized algorithm for prediction interval using RVFL networks ensemble | |
| Pandey et al. | Framework for Stock Market Prediction using Deep Learning Technique | |
| Jadhav et al. | Prediction of earnings per share for industry | |
| Tao et al. | Applying Machine Learning to Chemical Industry: A Self-Adaptive GA-BP Neural Network-Based Predictor of Gasoline Octane Number | |
| Скакаліна | Investigation of intelligent technologies for formation forecasting models= Дослідження інтелектуальних технологій для моделей прогнозування пласта |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200417 |
|
| RJ01 | Rejection of invention patent application after publication |