CN112258014A - Clustering and grouping-based risk discrimination analysis method for heat exchangers - Google Patents

Clustering and grouping-based risk discrimination analysis method for heat exchangers Download PDF

Info

Publication number
CN112258014A
CN112258014A CN202011113817.1A CN202011113817A CN112258014A CN 112258014 A CN112258014 A CN 112258014A CN 202011113817 A CN202011113817 A CN 202011113817A CN 112258014 A CN112258014 A CN 112258014A
Authority
CN
China
Prior art keywords
heat exchanger
group
medium
corresponding value
characteristic
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
Application number
CN202011113817.1A
Other languages
Chinese (zh)
Inventor
李春树
李涛
孙全胜
李洪涛
郭拂娟
张中洋
王艳丽
李梦瑶
高丽岩
张雅贤
曹德成
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.)
China Petroleum and Chemical Corp
Original Assignee
China Petroleum and Chemical Corp
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 China Petroleum and Chemical Corp filed Critical China Petroleum and Chemical Corp
Priority to CN202011113817.1A priority Critical patent/CN112258014A/en
Publication of CN112258014A publication Critical patent/CN112258014A/en
Pending legal-status Critical Current

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
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/20—Ensemble learning

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Human Resources & Organizations (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Software Systems (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Medical Informatics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Heat-Exchange Devices With Radiators And Conduit Assemblies (AREA)

Abstract

A cluster grouping and group-based risk discrimination analysis method for heat exchangers comprises the following steps: selecting a sample heat exchanger group, and determining a clustering analysis variable and a discrimination analysis variable of a heat exchanger; establishing a data mapping rule under each characteristic variable, and carrying out numerical processing on the sample heat exchanger according to the data mapping rule to be used as a training data set; performing clustering analysis on the training data set by adopting a K-means algorithm, and dividing a heat exchanger group; establishing a heat exchanger discrimination function based on the divided heat exchanger groups and the known characteristic variables, and determining a discrimination criterion; and judging a group where the newly designed heat exchanger is located according to a heat exchanger group judgment criterion, searching the newly designed heat exchanger and the heat exchanger with the closest characteristic distance in the group, and performing risk management and control on the newly designed heat exchanger according to the running risk state of the heat exchanger with the closest characteristic distance. The method has the advantages of rich and comprehensive training data, concise and reasonable group division program based on the characteristic variables and convenient use.

Description

Clustering and grouping-based risk discrimination analysis method for heat exchangers
Technical Field
The invention relates to a risk discrimination analysis method of a heat exchanger. In particular to a clustering and grouping-based risk discrimination analysis method for heat exchangers.
Background
The petrochemical industry is a high-energy-consumption industry, and little improvement in any production process can bring huge economic benefits. The heat exchanger is one of the most common devices in petrochemical production, not only serves as a widely-used device for ensuring normal operation of a specific process flow, but also is an important device for developing and utilizing industrial secondary energy to realize waste heat recovery.
The reliability guarantee and control of equipment can be realized in the design stage of the heat exchanger, the possible failures of the heat exchanger in the use process can be comprehensively analyzed, methods and measures for avoiding the failures are provided, and the intrinsic safety of the heat exchanger is ensured; according to the theory of risk engineering, the risk level is evaluated, and measures are taken to control the risk level.
Obviously, the operating condition of the heat exchanger can be simulated by the process condition of possible operation of the 'designed heat exchanger' and the characteristics of the selected materials, the possible risks of the heat exchanger can be analyzed, and corresponding risk precautionary measures can be taken, so that the failure possibility of the heat exchanger is reduced in the design stage, and further the risk of the heat exchanger is reduced.
However, this method has too much theoretical analysis and simulation, and the actual application of the equipment in the industrial field has a lot of uncertain factors, i.e. the theory can not replace the field, so the field factor needs to be fully considered through the integrity management applied in the design stage of the heat exchanger, and the simulation and the field actual are combined.
Although the heat exchanger in the design stage has no actual operation condition, the heat exchanger with characteristics similar to the characteristics of the heat exchanger in the aspects of structure, material, operation condition, medium characteristics and the like can be searched, and whether the heat exchanger is designed reasonably can be known according to the risk state of the 'operation heat exchanger' closest to the 'design heat exchanger', and the risk can be reduced by what measure. And by knowing the position of the heat exchanger closest to the heat exchanger in an equipment management system and the management measures applied to the heat exchanger, the management measures in the aspects of manufacturing, running and overhauling can be well established in advance for the heat exchanger so as to realize the high-reliability running of the heat exchanger.
Disclosure of Invention
The invention aims to solve the technical problem of providing a clustering and grouping method of a heat exchanger and a risk discrimination and analysis method based on the clustering and grouping method of the heat exchanger, which can determine a group of the heat exchanger more intuitively and simply and carry out risk discrimination and analysis on a newly designed heat exchanger based on a group which is already classified.
The technical scheme adopted by the invention is as follows: a cluster grouping and group-based risk discrimination analysis method for heat exchangers comprises the following steps:
1) selecting a sample heat exchanger group, and determining a clustering analysis variable and a discrimination analysis variable of a heat exchanger;
2) establishing a data mapping rule under each characteristic variable, and carrying out numerical processing on the sample heat exchanger according to the data mapping rule to be used as a training data set;
3) performing clustering analysis on the training data set by adopting a K-means algorithm, and dividing a heat exchanger group;
4) establishing a heat exchanger discrimination function based on the divided heat exchanger groups and the known characteristic variables, and determining a discrimination criterion;
5) and judging a group where the newly designed heat exchanger is located according to a heat exchanger group judgment criterion, searching the newly designed heat exchanger and the heat exchanger with the closest characteristic distance in the group, and performing risk management and control on the newly designed heat exchanger according to the running risk state of the heat exchanger with the closest characteristic distance.
According to the cluster grouping and group-based risk discrimination analysis method of the heat exchanger, a sample heat exchanger group is selected, and a cluster analysis variable and a discrimination analysis variable of the heat exchanger are determined according to the sample heat exchanger group; establishing a data mapping rule under each characteristic variable, and carrying out numerical processing on the sample heat exchanger according to the data mapping rule to be used as a training data set; performing clustering analysis on the training data set by adopting a K-means algorithm, and dividing a heat exchanger group; establishing a heat exchanger discrimination function based on the divided heat exchanger groups and known discrimination variables, and determining a discrimination criterion; and calculating the heat exchanger closest to the newly designed heat exchanger and the group, and performing design risk management and control according to the running risk state of the heat exchanger closest to the newly designed heat exchanger. The invention has the following advantages:
1. the selected sample heat exchanger group covers heat exchangers with various structural forms in the current petrochemical device, and training data are rich and comprehensive; the selected 7 characteristic variables respectively represent the structural characteristics of the heat exchanger, the operating characteristics of the heat exchanger, the corrosion mechanism of the heat exchanger and the production importance of the heat exchanger, and the group division program based on the characteristic variables is simple and reasonable and is convenient to use.
2. The heat exchanger discrimination analysis method is used for the risk discrimination of the heat exchanger in the design stage, the heat exchanger with similar characteristics in the aspects of structure, material, operation condition, medium characteristic and the like of the newly designed heat exchanger can be searched, and then whether the heat exchanger is designed reasonably or not is judged according to the risk state of the running heat exchanger closest to the designed heat exchanger, and risk reduction measures are given. By knowing the position of the most similar heat exchanger to the designed heat exchanger in an equipment management system and the management measures applied to the most similar heat exchanger, the management and control measures in the aspects of manufacturing, running and overhauling can be well established in advance for designing the heat exchanger so as to realize the high-reliability running of the heat exchanger.
Drawings
FIG. 1 is a flow chart of a cluster grouping and group-based risk discrimination analysis method of a heat exchanger according to the present invention.
Detailed Description
The cluster clustering and group-based risk discrimination analysis method of the heat exchanger of the present invention is described in detail below with reference to the embodiments and the accompanying drawings.
As shown in fig. 1, the cluster clustering and group-based risk discrimination analysis method for heat exchangers according to the present invention includes the following steps:
1) selecting a sample heat exchanger group, and determining a clustering analysis variable and a discrimination analysis variable of a heat exchanger; the method comprises the following steps:
(1) all types of heat exchangers (such as 842 heat exchangers) in an refining enterprise are taken as a sample heat exchanger group, and the enterprise covers the production device of the current refining process;
(2) selecting 7 characteristic variables of the number of heat exchange tubes, the material of tube passes of the heat exchanger, the outlet temperature of a tube pass medium, the inlet temperature of a shell pass medium, the operating pressure, the corrosion level of the medium and the key degree of the heat exchanger to the device, wherein the 7 characteristic variables respectively represent the characteristics of the heat exchanger, the corrosion mechanism of the heat exchanger and the production importance of the heat exchanger; and taking the 7 characteristic variables as the variables of the heat exchanger clustering analysis and the discriminant analysis.
2) Establishing a data mapping rule under each characteristic variable, and carrying out numerical processing on the sample heat exchanger according to the data mapping rule to be used as a training data set;
because the corresponding values under each index are not all numerical values, a data mapping rule needs to be determined, that is, each selected characteristic variable is subjected to de-dimensioning, and the mapping rule is 1 to 5 in sequence from high to low according to the degree, so that a mapping relation of data is formed. The establishing of the data mapping criterion under each characteristic variable comprises the following steps:
(1) the characteristic variable of the number of the heat exchange tubes is used for carrying out numerical processing on the number of the heat exchange tubes in the heat exchanger and converting the number into a numerical value between 1 and 5; namely:
the number of the heat exchange tubes is less than 100, and the corresponding value is 1; the number of the heat exchange tubes is more than or equal to 100 and less than 400, and the corresponding value is 2; the number of the heat exchange tubes is more than or equal to 400 and less than 1000, and the corresponding value is 3; the number of the heat exchange tubes is more than or equal to 1000 and less than 2000, and the corresponding value is 4; the number of the heat exchange tubes is more than or equal to 2000, and the corresponding value is 5.
(2) The tube side medium outlet temperature characteristic variable is used for carrying out numerical processing on the tube side medium outlet temperature in the heat exchanger, and correspondingly converting the tube side medium outlet temperature into a numerical value between 1 and 5 according to the temperature from low to high; namely:
the outlet temperature of the tube pass is less than 50 ℃ when the medium is circulating water, the outlet temperature of the tube pass is less than 100 ℃ when the medium is non-circulating water, and the corresponding value is 1; when the medium is circulating water, the outlet temperature of the tube pass is more than or equal to 50 ℃ and less than 70 ℃, and when the medium is non-circulating water, the outlet temperature of the tube pass is more than or equal to 100 ℃ and less than 250 ℃, and the corresponding value is 2; when the medium is circulating water, the outlet temperature of the tube pass is more than or equal to 70 ℃ and less than 100 ℃, when the medium is non-circulating water, the outlet temperature of the tube pass is more than or equal to 250 ℃ and less than 300 ℃, and the corresponding value is 4; when the medium is circulating water, the outlet temperature of the tube pass is more than or equal to 100 ℃ and less than 150 ℃, when the medium is non-circulating water, the outlet temperature of the tube pass is more than or equal to 300 ℃ and less than 400 ℃, and the corresponding value is 4; the outlet temperature of the tube pass is more than 150 ℃ when the medium is circulating water, the outlet temperature of the tube pass is more than or equal to 400 ℃ when the medium is non-circulating water, and the corresponding value is 5.
(3) Performing numerical treatment on the inlet temperature of the shell pass medium in the heat exchanger according to the characteristic variable of the inlet temperature of the shell pass medium, and correspondingly converting the inlet temperature of the shell pass medium into a numerical value between 1 and 5 according to the temperature from low to high; namely:
the inlet temperature of the shell side medium is less than 100 ℃, and the corresponding value is 1; the operation temperature is more than or equal to 100 ℃ and less than 200 ℃, and the corresponding value is 2; the operating temperature is more than or equal to 200 ℃ and less than 300 ℃, and the corresponding value is 3; the operating temperature is more than or equal to 300 ℃ and less than 400 ℃, and the corresponding value is 4; the operating temperature is equal to or greater than 400 ℃ with a corresponding value of 5.
(4) Comparing the shell pass operating pressure characteristic variable with the heat exchanger tube pass operating pressure, taking the maximum operating pressure between the shell pass operating pressure characteristic variable and the heat exchanger tube pass operating pressure as a heat exchanger operating pressure characteristic vector, carrying out numerical processing, and correspondingly converting the operating pressure from low to high into a numerical value between 1 and 5; namely:
the corresponding value for an operating pressure of less than 1.5MPa is 1; the operating pressure is more than or equal to 1.5MPa and less than 3.4MPa, and the corresponding value is 2; the operating pressure is more than or equal to 3.4MPa and less than 5MPa, and the corresponding value is 3; the operating pressure is more than or equal to 5MPa and less than 10MPa, and the corresponding value is 4; the operating pressure is equal to or greater than 10MPa, corresponding to a value of 5.
(5) Performing numerical treatment according to the tube bundle material corrosion resistance of the heat exchanger tube pass and the shell pass medium under the operation temperature and pressure, and correspondingly converting the tube bundle material corrosion resistance from weak to strong into a numerical value between 1 and 5; namely:
the material grade of the heat exchange steel pipe used in the pipe bundle is No. 10 steel, and the corresponding value is 1; the grade of the material of the heat exchange steel pipe used in the pipe bundle is 20 steel or NS1 steel for resisting sulfuric acid dew point corrosion, and the corresponding value is 2; the heat exchange steel pipe used in the pipe bundle is made of No. 10 steel, the heat exchange pipe is coated with an anticorrosive coating, the heat exchange steel pipe used in the pipe bundle is made of No. 10 steel pipe subjected to cold drawing, the heat exchange steel pipe is made of No. 10 aluminized steel, and the corresponding value is 3; the heat exchange steel pipe is made of austenitic stainless steel with the grade of 321, the heat exchange steel pipe is made of ferritic stainless steel, the heat exchange steel pipe is made of steel with the grade of 20, and the heat exchange pipe is coated with an anticorrosive coating, and the corresponding value is 4; the heat exchange steel pipe is made of nickel and alloy steel, the heat exchange steel pipe is made of Incoloy 825 steel, the heat exchange steel pipe is made of titanium alloy steel, and the corresponding value is 5.
(6) Comparing the corrosivity of the heat exchanger tube side medium to the heat exchanger tube bundle metal at the operating temperature with the corrosivity of the shell side medium to the heat exchanger tube bundle metal at the operating temperature, taking the medium with the highest corrosivity between the two as a medium corrosivity grade characteristic variable, and correspondingly converting the medium corrosivity into a numerical value between 1 and 5 from weak to strong according to the medium corrosivity; namely:
the medium is deaerated water, softened water, boiler water, steam, a factory product and does not have corrosive medium, and the corresponding value is 1; the medium is crude oil subjected to electric desalting treatment, and the corresponding value is 2; the medium is stable tower top oil gas, separation tower top oil gas, fuel gas, normal line oil in an atmospheric tower, reduced line oil in a vacuum tower and newly-prepared hydrogen, and the corresponding value is 3; the medium is liquefied gas, atmospheric tower bottom oil, vacuum tower bottom oil, lean amine liquid, rich amine liquid and sulfolane, and the corresponding value is 4; the medium is hydrogenation reaction product, reforming reaction product, fractionating tower top oil gas, absorbing tower top oil gas, analyzing tower top oil gas, normal pressure tower top circulating oil gas, residual oil discharged from catalytic cracking unit, acidic water containing ammonium hydrogen sulfide concentration greater than 15000ppm, circulating water, coking unit purified water, circulating hydrogen and medium conversion gas, and the corresponding value is 5.
(7) The key degree characteristic variable of the heat exchanger to the device is correspondingly converted into a numerical value between 1 and 5 according to the equipment failure intensity grade of the heat exchanger, namely the influence degree of the heat exchanger on a production device and a related production device in which the heat exchanger is positioned when the heat exchanger is in failure shutdown is from low to high; namely:
the product quality, the process operation and other equipment are not influenced after the heat exchanger leaks, and the corresponding value is 1; after the heat exchanger leaks, the product quality and the process operation are not influenced, but media are caused to flow in series and pollute the media on the other side, so that the long-term operation risk of the equipment is increased, and the corresponding value is 2; only the normal production and process operation of the production device is affected after the heat exchanger leaks, the product quality is unqualified, and the corresponding value is 3; after the heat exchanger leaks, the local stop of the production device or the sudden stop of a large unit of the device is caused, and the corresponding value is 4; after the heat exchanger leaks, the production device stops running or more than two sets of related production devices fluctuate abnormally, and the corresponding value is 5.
Carrying out numerical processing between 1 and 5 on each heat exchanger in the sample heat exchanger group according to a data mapping criterion, and forming a training data set after the numerical processing of the sample heat exchangers: r ═ R1,R2,…,Ri,…,RnN is the number of sample heat exchangers, and Ri＝{ri1,ri2,…,ri7And assigning values to 7 characteristic variables of the ith heat exchanger in the sample heat exchanger to form a characteristic vector.
If 842 sample heat exchanger groups are processed numerically according to a data mapping criterion to form a training data set: r ═ R1,R2,…,Ri,…,R842}。
3) Performing clustering analysis on the training data set by adopting a K-means algorithm, and dividing a heat exchanger group; the method comprises the following steps:
(1) set the sample heat exchanger groups as k groups, note
Figure BDA0002729554330000042
GjFor the jth heat exchanger group, randomly selecting k heat exchangers from the training data set as a clustering center, and recording as { U1,U2,…,Uj,…,UkThe feature vectors of k cluster centers are represented as: u shapej＝{uj1,uj2,…,uj7},j∈[1,k]，UjA feature vector of a jth clustering center;
(2) respectively calculating each heat exchanger R in sample heat exchanger groupiAnd the distance between the feature vectors Uj of the k cluster centers:
Figure BDA0002729554330000041
d(Ri,Uj) Is the distance from the ith heat exchanger to the jth cluster center, ri,tIs the t component, u, in the characteristic vector of the ith heat exchangerj,tThe t component of the feature vector of the jth cluster center;
each heat exchanger RiEigenvectors U to k clustering centersjThe distances between the heat exchangers are respectively sequenced from small to large, and the heat exchanger with the minimum distance to the jth clustering center is recorded as
Figure BDA00027295543300000513
And the heat exchanger group is combined into the heat exchanger group where the clustering center is positioned
Figure BDA00027295543300000510
And correcting the cluster center value of each heat exchanger group:
Figure BDA0002729554330000051
wherein N isjIs the jth heat exchanger group GjThe number of the middle heat exchangers;
(3) judging the convergence of clustering calculation, wherein the convergence judgment formula is as follows:
Figure BDA0002729554330000052
if Y is converged, namely the clustering center is not changed any more, the clustering analysis is finished, the heat exchanger group division is finished, otherwise, the step (2) is returned until Y is converged, and the final group division is carried out
Figure BDA00027295543300000511
If 842 sample heat exchanger groups are processed numerically according to a data mapping criterion to form a training data set, after clustering analysis, 7 heat exchanger groups are divided
Figure BDA00027295543300000512
4) Establishing a heat exchanger discrimination function based on the divided heat exchanger groups and the known characteristic variables, and determining a discrimination criterion; the method comprises the following steps:
(1) taking the characteristic variable in the step 1) as a discrimination variable, taking data in k heat exchanger groups in the clustering result in the step 3) as k groups of sample data sets, wherein the sample sets are 7-dimensional and comprise N samples of k classes, and N is1，N2，…,Nj,…,Nk，N＝N1+N2+…+Nk，
Figure BDA0002729554330000053
And the data vector of the alpha heat exchanger of the j heat exchanger group.
(2) Constructing a heat exchanger group discrimination function:
y(x)＝c1x1+…c7x7＝c′x
wherein c ═ c1，…，c7) ' is the coefficient of discriminant function, x ═ x1，…，x7) ' is the discriminating variable of the heat exchanger;
calculating y (x) in each heat exchanger group
Figure BDA0002729554330000058
Sample mean vector of
Figure BDA0002729554330000054
Calculating y (x) in each heat exchanger group
Figure BDA0002729554330000059
Sample covariance matrix of
σj 2＝c′s(j)c
Wherein the content of the first and second substances,
Figure BDA0002729554330000055
sample mean vectors of all heat exchangers in the jth heat exchanger group are obtained; s(j)A sample covariance array of all heat exchangers in the jth heat exchanger group is obtained;
Figure BDA0002729554330000056
is y (x) sample mean vector at jth heat exchanger group; sigmaj 2Is the sample covariance matrix at the jth heat exchanger group.
(3) Determining coefficient vector of discriminant function, and calculating each heat exchanger group GjIntra-group dispersion matrix E:
Figure BDA0002729554330000057
calculate each heat exchanger group GjCovariance matrix a between:
Figure BDA0002729554330000061
wherein the content of the first and second substances,
Figure BDA0002729554330000062
the mean vector of the total heat exchanger sample is obtained; q. q.siPositive weighting coefficient, qi＝Nj-1；
Calculating the generalized characteristic root lambda of the covariance matrix A relative to the dispersion matrix E and the characteristic vector p corresponding to the generalized characteristic root lambda so as to ensure that
Figure BDA0002729554330000063
When the maximum value is reached, the feature vector p is the coefficient vector c of the constructed discriminant function y (x).
Constructing k groups G of heat exchangersjThe discriminant function of (c):
yj(x)＝c(j)′x
wherein: lambda is a generalized characteristic root of the covariance matrix A relative to the dispersion matrix E; p is a characteristic vector corresponding to lambda; y isj(x) Is the j discriminant equation; c. C(j)Is the coefficient of the j-th discriminant equation.
(4) Determining the criterion of heat exchanger group, and determining the heat exchanger to be judged
Figure BDA0002729554330000064
The k values are calculated by being brought into a discriminant function:
yj(x*)＝c(j)′x*，j＝1，…，k
sample mean vector of all heat exchangers in the ith heat exchanger group
Figure BDA0002729554330000065
Carry into the discriminant function to calculate:
Figure BDA0002729554330000066
and (3) calculating:
Figure BDA0002729554330000067
if it is
Figure BDA0002729554330000068
Then x is judged*∈Gγ；
Wherein x*The characteristic vector of the heat exchanger to be distinguished is obtained; y isj(x*) Is x*Substituting the calculated value into the j discrimination function;
Figure BDA0002729554330000069
the sample mean vector of the heat exchanger in the ith heat exchanger group in the k heat exchanger groups is obtained;
Figure BDA00027295543300000610
sample mean vector for ith heat exchanger group
Figure BDA00027295543300000611
Substituting the calculated value into the j discrimination function; lambda [ alpha ]jIs the jth characteristic value;
Figure BDA00027295543300000612
the weighted distances from the discrimination calculation value of the ith heat exchanger cluster center to the k discrimination function values of the heat exchanger to be discriminated are respectively calculated, and the weight is lambdaj；
Figure BDA00027295543300000613
Is k in number
Figure BDA00027295543300000614
The smallest value of; gγTo a minimum value of
Figure BDA00027295543300000615
The corresponding heat exchanger group.
According to the steps of establishing the heat exchanger discrimination method, establishing a discrimination analysis function based on 7 heat exchanger groups of 842 heat exchangers:
Figure BDA00027295543300000616
5) and judging a group where the newly designed heat exchanger is located according to a heat exchanger group judgment criterion, searching the newly designed heat exchanger and the heat exchanger with the closest characteristic distance in the group, and performing risk management and control on the newly designed heat exchanger according to the running risk state of the heat exchanger with the closest characteristic distance. The principle of designing a new heat exchanger is as follows:
(1) assigning values to each characteristic variable of the 7 characteristic variables of the newly designed heat exchanger according to the data mapping rule in the step 2) to form a characteristic vector of the newly designed heat exchanger
Figure BDA0002729554330000071
Figure BDA0002729554330000072
Judging the group to which the heat exchanger belongs according to the heat exchanger judgment criterion in the step 4);
for example, in a new installation in a petrochemical enterprise, 16 heat exchangers of the design are randomly extracted, and a grading table is formed according to data mapping criteria, as shown in table 1:
TABLE 1
Figure BDA0002729554330000073
For example, scoring results of randomly extracting 16 designed heat exchangers from a newly built device in a petrochemical enterprise are respectively substituted into 7 discrimination equations for calculation, the group to which the heat exchangers belong is judged according to the discrimination criteria, and the results are shown in table 2:
TABLE 2
Random extraction heat exchanger Group to which they belong Random extraction heat exchanger Group to which they belong
Reaction product and circulating hydrogen heat exchanger 4 Heat exchanger for bottom oil and hydrogenated residual oil of hydrogen sulfide stripping tower 2
Reaction product and mixed hydrogen oil heat exchanger 4 Raw oil and hydrogenation residual oil heat exchanger 2
Heat exchanger for hot high-pressure gas and mixed hydrogen oil 4 Naphtha water cooler 5
High heatGas-distributing and circulating hydrogen heat exchanger 4 Heat exchanger for middle section reflux and raw oil 4
Heat high-pressure gas and circulating hydrogen heat exchanger 4 Diesel oil side-stream tower bottom reboiler 5
Hot low-temperature gas/cold low-temperature oil heat exchanger 7 Diesel oil and cold low oil separating heat exchanger 2
Reaction lean amine liquid cooler 5 Diesel oil and hot water heat exchanger 6
Water cooler on top of stripping tower for removing hydrogen sulfide 1 Low-gas-separation desulfurization amine liquid cooler 5
(2) Searching the heat exchanger with the newly designed heat exchanger and the heat exchanger with the closest characteristic distance in the group, and calculating the formula:
Figure BDA0002729554330000081
wherein the content of the first and second substances,
Figure BDA0002729554330000082
for the eigenvectors of the newly designed heat exchanger determined to belong to the group of class j heat exchangers,
Figure BDA0002729554330000083
is a component of the feature vector;
Figure BDA0002729554330000084
the heat exchanger is the ith heat exchanger in the jth type heat exchanger group;
Figure BDA0002729554330000085
the distance from the newly designed heat exchanger to the ith heat exchanger in the jth heat exchanger group is calculated;
will be provided with
Figure BDA0002729554330000086
Arranged from small to large, smallest
Figure BDA0002729554330000087
The corresponding heat exchanger is the heat exchanger which has the closest characteristic distance with the newly designed heat exchanger, namely the most similar heat exchanger;
for example, after search calculation, 16 designed heat exchangers randomly extracted from a new installation of a petrochemical enterprise obtain a heat exchanger with the same group as each newly designed heat exchanger and the closest characteristic distance, i.e., the most similar heat exchanger, and give the risk level of the most similar heat exchanger to each newly designed heat exchanger, the result is shown in table 3:
TABLE 3
Random oil taking design heat exchanger Group to which they belong Heat exchanger with nearest characteristic distance Risk rating
Heat exchanger 1 1 3# atmospheric and vacuum pressure E-104B Middle risk
Heat exchanger 2 7 2# coking E-103 Middle risk
Heat exchanger 3 4 Air separation E-101 Middle risk
Heat exchanger 4 5 3# atmospheric and vacuum pressure E-202E Low risk
Heat exchanger 5 1 2# atmospheric and vacuum E-119 High and high risk
Heat exchanger 6 4 Catalytic converter 2017 Middle risk
Heat exchanger 7 5 1# hydrocracking E201 Middle risk
Heat exchanger 8 1 3# diesel hydrogenation E-202 Middle risk
Heat exchanger 9 4 1# hydrocracking E401 Low risk
Heat exchanger 10 4 1# coking E-104 Middle risk
Heat exchanger 11 5 Desulfurization E3403 Low risk
Heat exchanger 12 4 Wax oil hydrogenation E-203 Middle risk
Heat exchangeDevice 13 5 2# hydrocracking E-205 Middle risk
Heat exchanger 14 2 1# coking E-123/2 Middle risk
Heat exchanger 15 6 Wax oil hydrogenation E-211 Low risk
Heat exchanger 16 5 2# diesel hydrogenation E-104 Middle risk
Heat exchanger 17 4 Reforming E-305 Low risk
Heat exchanger 18 5 2# coking E-107 Middle risk
Heat exchanger 19 4 Reforming E-511 Low risk
(3) And (3) performing risk management and control on designing a new heat exchanger according to the running risk state of the heat exchanger closest to the heat exchanger, wherein the risk grade of the heat exchanger closest to the heat exchanger is a management and control measure of the newly designed heat exchanger with medium-high risk or high risk:
considering the upgrading of the tube side material of the heat exchanger, or carrying out an anti-corrosion coating on the tube bundle, or carrying out design correction by taking the heat exchanger with the lowest failure possibility in the group where the newly designed heat exchanger is positioned as a reference; selecting a tube bundle type which is not easy to scale or taking a heat exchanger tube bundle type with the best scale control in a group where a newly designed heat exchanger is located as design correction; the heat exchange tube and the tube mouth of the heat exchanger are connected with the tube plate by adopting high-reliability expansion welding combination; the heat exchange process is provided with a cross line and an inlet and outlet valve, and the material of the cross line is consistent with that of an inlet and outlet pipeline.
As can be seen from table 3, in 16 design heat exchangers randomly extracted from a new installation of a petrochemical enterprise, 6 heat exchangers closest to each other have a risk level of medium high risk or high risk, and therefore risk management and control are required in the design process.

Claims (6)

1. A cluster grouping and group-based risk discrimination analysis method for heat exchangers is characterized by comprising the following steps:
1) selecting a sample heat exchanger group, and determining a clustering analysis variable and a discrimination analysis variable of a heat exchanger;
2) establishing a data mapping rule under each characteristic variable, and carrying out numerical processing on the sample heat exchanger according to the data mapping rule to be used as a training data set;
3) performing clustering analysis on the training data set by adopting a K-means algorithm, and dividing a heat exchanger group;
4) establishing a heat exchanger discrimination function based on the divided heat exchanger groups and the known characteristic variables, and determining a discrimination criterion;
5) and judging a group where the newly designed heat exchanger is located according to a heat exchanger group judgment criterion, searching the newly designed heat exchanger and the heat exchanger with the closest characteristic distance in the group, and performing risk management and control on the newly designed heat exchanger according to the running risk state of the heat exchanger with the closest characteristic distance.
2. The cluster grouping and group-based risk discrimination analysis method of the heat exchanger according to claim 1, wherein the step 1) comprises:
(1) all types of heat exchangers in an refining enterprise are taken as sample heat exchanger groups, and the enterprise covers a production device of the current refining process;
(2) selecting 7 characteristic variables of the number of heat exchange tubes, the material of tube passes of the heat exchanger, the outlet temperature of a tube pass medium, the inlet temperature of a shell pass medium, the operating pressure, the corrosion level of the medium and the key degree of the heat exchanger to the device, wherein the 7 characteristic variables respectively represent the characteristics of the heat exchanger, the corrosion mechanism of the heat exchanger and the production importance of the heat exchanger; and taking the 7 characteristic variables as the variables of the heat exchanger clustering analysis and the discriminant analysis.
3. The cluster grouping and group-based risk discrimination analysis method of heat exchangers according to claim 1, wherein the step 2) of establishing the data mapping criterion under each characteristic variable comprises:
(1) the characteristic variable of the number of the heat exchange tubes is used for carrying out numerical processing on the number of the heat exchange tubes in the heat exchanger and converting the number into a numerical value between 1 and 5; namely:
the number of the heat exchange tubes is less than 100, and the corresponding value is 1; the number of the heat exchange tubes is more than or equal to 100 and less than 400, and the corresponding value is 2; the number of the heat exchange tubes is more than or equal to 400 and less than 1000, and the corresponding value is 3; the number of the heat exchange tubes is more than or equal to 1000 and less than 2000, and the corresponding value is 4; the number of the heat exchange tubes is more than or equal to 2000, and the corresponding value is 5;
(2) the tube side medium outlet temperature characteristic variable is used for carrying out numerical processing on the tube side medium outlet temperature in the heat exchanger, and correspondingly converting the tube side medium outlet temperature into a numerical value between 1 and 5 according to the temperature from low to high; namely:
the outlet temperature of the tube pass is less than 50 ℃ when the medium is circulating water, the outlet temperature of the tube pass is less than 100 ℃ when the medium is non-circulating water, and the corresponding value is 1; when the medium is circulating water, the outlet temperature of the tube pass is more than or equal to 50 ℃ and less than 70 ℃, and when the medium is non-circulating water, the outlet temperature of the tube pass is more than or equal to 100 ℃ and less than 250 ℃, and the corresponding value is 2; when the medium is circulating water, the outlet temperature of the tube pass is more than or equal to 70 ℃ and less than 100 ℃, when the medium is non-circulating water, the outlet temperature of the tube pass is more than or equal to 250 ℃ and less than 300 ℃, and the corresponding value is 4; when the medium is circulating water, the outlet temperature of the tube pass is more than or equal to 100 ℃ and less than 150 ℃, when the medium is non-circulating water, the outlet temperature of the tube pass is more than or equal to 300 ℃ and less than 400 ℃, and the corresponding value is 4; when the medium is circulating water, the outlet temperature of the tube side is more than 150 ℃, when the medium is non-circulating water, the outlet temperature of the tube side is more than or equal to 400 ℃, and the corresponding value is 5;
(3) performing numerical treatment on the inlet temperature of the shell pass medium in the heat exchanger according to the characteristic variable of the inlet temperature of the shell pass medium, and correspondingly converting the inlet temperature of the shell pass medium into a numerical value between 1 and 5 according to the temperature from low to high; namely:
the inlet temperature of the shell side medium is less than 100 ℃, and the corresponding value is 1; the operation temperature is more than or equal to 100 ℃ and less than 200 ℃, and the corresponding value is 2; the operating temperature is more than or equal to 200 ℃ and less than 300 ℃, and the corresponding value is 3; the operating temperature is more than or equal to 300 ℃ and less than 400 ℃, and the corresponding value is 4; the operating temperature is greater than or equal to 400 ℃ and the corresponding value is 5;
(4) comparing the shell pass operating pressure characteristic variable with the heat exchanger tube pass operating pressure, taking the maximum operating pressure between the shell pass operating pressure characteristic variable and the heat exchanger tube pass operating pressure as a heat exchanger operating pressure characteristic vector, carrying out numerical processing, and correspondingly converting the operating pressure from low to high into a numerical value between 1 and 5; namely:
the corresponding value for an operating pressure of less than 1.5MPa is 1; the operating pressure is more than or equal to 1.5MPa and less than 3.4MPa, and the corresponding value is 2; the operating pressure is more than or equal to 3.4MPa and less than 5MPa, and the corresponding value is 3; the operating pressure is more than or equal to 5MPa and less than 10MPa, and the corresponding value is 4; the corresponding value of the operating pressure of 10MPa or more is 5;
(5) performing numerical treatment according to the tube bundle material corrosion resistance of the heat exchanger tube pass and the shell pass medium under the operation temperature and pressure, and correspondingly converting the tube bundle material corrosion resistance from weak to strong into a numerical value between 1 and 5; namely:
the material grade of the heat exchange steel pipe used in the pipe bundle is No. 10 steel, and the corresponding value is 1; the grade of the material of the heat exchange steel pipe used in the pipe bundle is 20 steel or NS1 steel for resisting sulfuric acid dew point corrosion, and the corresponding value is 2; the heat exchange steel pipe used in the pipe bundle is made of No. 10 steel, the heat exchange pipe is coated with an anticorrosive coating, the heat exchange steel pipe used in the pipe bundle is made of No. 10 steel pipe subjected to cold drawing, the heat exchange steel pipe is made of No. 10 aluminized steel, and the corresponding value is 3; the heat exchange steel pipe is made of austenitic stainless steel with the grade of 321, the heat exchange steel pipe is made of ferritic stainless steel, the heat exchange steel pipe is made of steel with the grade of 20, and the heat exchange pipe is coated with an anticorrosive coating, and the corresponding value is 4; the heat exchange steel pipe is made of nickel and alloy steel, the heat exchange steel pipe is made of Incoloy 825 steel, the heat exchange steel pipe is made of titanium alloy steel, and the corresponding value is 5;
(6) comparing the corrosivity of the heat exchanger tube side medium to the heat exchanger tube bundle metal at the operating temperature with the corrosivity of the shell side medium to the heat exchanger tube bundle metal at the operating temperature, taking the medium with the highest corrosivity between the two as a medium corrosivity grade characteristic variable, and correspondingly converting the medium corrosivity into a numerical value between 1 and 5 from weak to strong according to the medium corrosivity; namely:
the medium is deaerated water, softened water, boiler water, steam, a factory product and does not have corrosive medium, and the corresponding value is 1; the medium is crude oil subjected to electric desalting treatment, and the corresponding value is 2; the medium is stable tower top oil gas, separation tower top oil gas, fuel gas, normal line oil in an atmospheric tower, reduced line oil in a vacuum tower and newly-prepared hydrogen, and the corresponding value is 3; the medium is liquefied gas, atmospheric tower bottom oil, vacuum tower bottom oil, lean amine liquid, rich amine liquid and sulfolane, and the corresponding value is 4; the medium is hydrogenation reaction product, reforming reaction product, fractionating tower top oil gas, absorbing tower top oil gas, analyzing tower top oil gas, normal pressure tower top circulating oil gas, residual oil discharged from catalytic cracking unit, acid water containing ammonium hydrogen sulfide with concentration greater than 15000ppm, circulating water, coking unit purified water, circulating hydrogen and medium conversion gas, and the corresponding value is 5;
(7) the key degree characteristic variable of the heat exchanger to the device is correspondingly converted into a numerical value between 1 and 5 according to the equipment failure intensity grade of the heat exchanger, namely the influence degree of the heat exchanger on a production device and a related production device in which the heat exchanger is positioned when the heat exchanger is in failure shutdown is from low to high; namely:
the product quality, the process operation and other equipment are not influenced after the heat exchanger leaks, and the corresponding value is 1; after the heat exchanger leaks, the product quality and the process operation are not influenced, but media are caused to flow in series and pollute the media on the other side, so that the long-term operation risk of the equipment is increased, and the corresponding value is 2; only the normal production and process operation of the production device is affected after the heat exchanger leaks, the product quality is unqualified, and the corresponding value is 3; after the heat exchanger leaks, the local stop of the production device or the sudden stop of a large unit of the device is caused, and the corresponding value is 4; after the heat exchanger leaks, the production device stops running, or more than two sets of related production devices fluctuate abnormally, and the corresponding value is 5;
carrying out numerical processing between 1 and 5 on each heat exchanger in the sample heat exchanger group according to a data mapping criterion, and forming a training data set after the numerical processing of the sample heat exchangers: r ═ R1,R2,…,Ri,…,RnN is the number of sample heat exchangers, and Ri＝{ri1,ri2,…,ri7And assigning values to 7 characteristic variables of the ith heat exchanger in the sample heat exchanger to form a characteristic vector.
4. The cluster grouping and group-based risk discrimination analysis method of heat exchangers according to claim 1, wherein the step 3) comprises:
(1) set the sample heat exchanger groups as k groups, note
Figure FDA0002729554320000035
UjFor the jth heat exchanger group, randomly selecting k heat exchangers from the training data set as a clustering center, and recording as { U1,U2,…,Uj,…,UkThe feature vectors of k cluster centers are represented as: u shapej＝{uj1,uj2,…,uj7},j∈[1,k]，UjAs the jth cluster centerA feature vector;
(2) respectively calculating each heat exchanger R in sample heat exchanger groupiAnd k feature vectors U of cluster centersjThe distance between:
Figure FDA0002729554320000031
d(Ri,Uj) Is the distance from the ith heat exchanger to the jth cluster center, ri,tIs the t component, u, in the characteristic vector of the ith heat exchangerj,tThe t component of the feature vector of the jth cluster center;
each heat exchanger RiThe distances between the characteristic vectors Uj of the k clustering centers are respectively sequenced from small to large, and the heat exchanger with the minimum distance to the jth clustering center is marked as
Figure FDA0002729554320000032
Is merged into the heat exchanger group where the clustering center is positioned
Figure FDA0002729554320000036
And correcting the cluster center value of each heat exchanger group:
Figure FDA0002729554320000033
wherein N isjIs the jth heat exchanger group GjThe number of the middle heat exchangers;
(3) judging the convergence of clustering calculation, wherein the convergence judgment formula is as follows:
Figure FDA0002729554320000034
if Y is converged, namely the clustering center is not changed any more, the clustering analysis is finished, the heat exchanger group division is finished, otherwise, the step (2) is returned until Y is converged, and finallyGroup partitioning
Figure FDA0002729554320000049
5. The cluster grouping and group-based risk discrimination analysis method of heat exchangers according to claim 1, wherein the step 4) comprises:
(1) taking the characteristic variable in the step 1) as a discrimination variable, taking data in k heat exchanger groups in the clustering result in the step 3) as k groups of sample data sets, wherein the sample sets are 7-dimensional and comprise N samples of k classes, and N is1，N2，…,Nj,…,Nk，N＝N1+N2+…+Nk，
Figure FDA0002729554320000041
A data vector of an alpha heat exchanger of a jth heat exchanger group;
(2) constructing a heat exchanger group discrimination function:
y(x)＝c1x1+…c7x7＝c′x
wherein c ═ c1，…，c7) ' is the coefficient of discriminant function, x ═ x1，…，x7) ' is the discriminating variable of the heat exchanger;
calculating y (x) in each heat exchanger group
Figure FDA00027295543200000410
Sample mean vector of
Figure FDA0002729554320000042
Calculating y (x) in each heat exchanger group
Figure FDA00027295543200000411
Sample covariance matrix of
σj 2＝c′s(j)c
Wherein the content of the first and second substances,
Figure FDA0002729554320000043
sample mean vectors of all heat exchangers in the jth heat exchanger group are obtained; s(j)A sample covariance array of all heat exchangers in the jth heat exchanger group is obtained;
Figure FDA0002729554320000044
is y (x) sample mean vector at jth heat exchanger group; sigmaj 2Is y (x) sample covariance matrix at jth heat exchanger group;
(3) determining coefficient vector of discriminant function, and calculating each heat exchanger group GjIntra-group dispersion matrix E:
Figure FDA0002729554320000045
calculate each heat exchanger group GjCovariance matrix a between:
Figure FDA0002729554320000046
wherein the content of the first and second substances,
Figure FDA0002729554320000047
the mean vector of the total heat exchanger sample is obtained; q. q.siPositive weighting coefficient, qi＝Nj-1；
Calculating the generalized characteristic root lambda of the covariance matrix A relative to the dispersion matrix E and the characteristic vector p corresponding to the generalized characteristic root lambda so as to ensure that
Figure FDA0002729554320000048
Reaching the maximum value, the characteristic vector p is the coefficient vector c of the constructed discriminant function y (x);
constructing k groups G of heat exchangersjThe discriminant function of (c):
yj(x)＝c(j)′x
wherein: lambda is a generalized characteristic root of the covariance matrix A relative to the dispersion matrix E; p is a characteristic vector corresponding to lambda; y isj(x) Is the j discriminant equation; c. C(j)Is the coefficient of the j-th discriminant equation;
(4) determining the criterion of heat exchanger group, and determining the heat exchanger to be judged
Figure FDA0002729554320000051
The k values are calculated by being brought into a discriminant function:
yj(x*)＝c(j)′x*，j＝1，…，k
sample mean vector of all heat exchangers in the ith heat exchanger group
Figure FDA0002729554320000052
Carry into the discriminant function to calculate:
Figure FDA0002729554320000053
and (3) calculating:
Figure FDA0002729554320000054
if it is
Figure FDA0002729554320000055
Then x is judged*∈Gγ；
Wherein x*The characteristic vector of the heat exchanger to be distinguished is obtained; y isj(x*) Is x*Substituting the calculated value into the j discrimination function;
Figure FDA0002729554320000056
the sample mean vector of the heat exchanger in the ith heat exchanger group in the k heat exchanger groups is obtained;
Figure FDA0002729554320000057
sample mean vector for ith heat exchanger group
Figure FDA0002729554320000058
Substituting the calculated value into the j discrimination function; lambda [ alpha ]jIs the jth characteristic value;
Figure FDA0002729554320000059
the weighted distances from the discrimination calculation value of the ith heat exchanger cluster center to the k discrimination function values of the heat exchanger to be discriminated are respectively calculated, and the weight is lambdaj；
Figure FDA00027295543200000510
Is k in number
Figure FDA00027295543200000511
The smallest value of; gγTo a minimum value of
Figure FDA00027295543200000512
The corresponding heat exchanger group.
6. The cluster grouping and group-based risk discrimination analysis method of heat exchangers according to claim 1, wherein the principle of designing a new heat exchanger in step 5) comprises:
(1) assigning values to each characteristic variable of the 7 characteristic variables of the newly designed heat exchanger according to the data mapping rule in the step 2) to form a characteristic vector of the newly designed heat exchanger
Figure FDA00027295543200000513
Figure FDA00027295543200000514
Judging the group to which the heat exchanger belongs according to the heat exchanger judgment criterion in the step 4);
(2) searching the newly designed heat exchanger and the heat exchanger with the closest characteristic distance in the group:
Figure FDA00027295543200000515
wherein the content of the first and second substances,
Figure FDA00027295543200000516
for the eigenvectors of the newly designed heat exchanger determined to belong to the group of class j heat exchangers,
Figure FDA00027295543200000517
is a component of the feature vector;
Figure FDA00027295543200000518
the heat exchanger is the ith heat exchanger in the jth type heat exchanger group;
Figure FDA00027295543200000519
the distance from the newly designed heat exchanger to the ith heat exchanger in the jth heat exchanger group is calculated;
will be provided with
Figure FDA00027295543200000520
Arranged from small to large, smallest
Figure FDA00027295543200000521
The corresponding heat exchanger is the heat exchanger closest to the characteristic distance of the newly designed heat exchanger;
(3) designing risk management and control of the new heat exchanger according to the running risk state of the heat exchanger closest to the heat exchanger, wherein management and control measures are as follows:
considering the upgrading of the tube side material of the heat exchanger, or carrying out an anti-corrosion coating on the tube bundle, or carrying out design correction by taking the heat exchanger with the lowest failure possibility in the group where the newly designed heat exchanger is positioned as a reference; selecting a tube bundle type which is not easy to scale or taking a heat exchanger tube bundle type with the best scale control in a group where a newly designed heat exchanger is located as design correction; the heat exchange tube and the tube mouth of the heat exchanger are connected with the tube plate by adopting high-reliability expansion welding combination; the heat exchange process is provided with a cross line and an inlet and outlet valve, and the material of the cross line is consistent with that of an inlet and outlet pipeline.
CN202011113817.1A 2020-10-17 2020-10-17 Clustering and grouping-based risk discrimination analysis method for heat exchangers Pending CN112258014A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011113817.1A CN112258014A (en) 2020-10-17 2020-10-17 Clustering and grouping-based risk discrimination analysis method for heat exchangers

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011113817.1A CN112258014A (en) 2020-10-17 2020-10-17 Clustering and grouping-based risk discrimination analysis method for heat exchangers

Publications (1)

Publication Number Publication Date
CN112258014A true CN112258014A (en) 2021-01-22

Family

ID=74244620

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011113817.1A Pending CN112258014A (en) 2020-10-17 2020-10-17 Clustering and grouping-based risk discrimination analysis method for heat exchangers

Country Status (1)

Country Link
CN (1) CN112258014A (en)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003028855A (en) * 2001-06-15 2003-01-29 Hitachi Software Eng Co Ltd Method for evaluation and display of clustered result
CN103399852A (en) * 2013-06-27 2013-11-20 江南大学 Multi-channel spectrum clustering method based on local density estimation and neighbor relation spreading
CN105403777A (en) * 2015-09-09 2016-03-16 华北电力大学 Classification and discrimination method of aging state of composite insulator
US20170091937A1 (en) * 2014-06-10 2017-03-30 Ventana Medical Systems, Inc. Methods and systems for assessing risk of breast cancer recurrence
CN110544373A (en) * 2019-08-21 2019-12-06 北京交通大学 truck early warning information extraction and risk identification method based on Beidou Internet of vehicles
CN110766200A (en) * 2019-09-23 2020-02-07 广东工业大学 Method for predicting generating power of wind turbine generator based on K-means mean clustering
US20200104630A1 (en) * 2018-09-30 2020-04-02 Sas Institute Inc. Distributable classification system
CN111175046A (en) * 2020-03-18 2020-05-19 北京工业大学 Rolling bearing fault diagnosis method based on manifold learning and s-k-means clustering

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003028855A (en) * 2001-06-15 2003-01-29 Hitachi Software Eng Co Ltd Method for evaluation and display of clustered result
CN103399852A (en) * 2013-06-27 2013-11-20 江南大学 Multi-channel spectrum clustering method based on local density estimation and neighbor relation spreading
US20170091937A1 (en) * 2014-06-10 2017-03-30 Ventana Medical Systems, Inc. Methods and systems for assessing risk of breast cancer recurrence
CN105403777A (en) * 2015-09-09 2016-03-16 华北电力大学 Classification and discrimination method of aging state of composite insulator
US20200104630A1 (en) * 2018-09-30 2020-04-02 Sas Institute Inc. Distributable classification system
CN110544373A (en) * 2019-08-21 2019-12-06 北京交通大学 truck early warning information extraction and risk identification method based on Beidou Internet of vehicles
CN110766200A (en) * 2019-09-23 2020-02-07 广东工业大学 Method for predicting generating power of wind turbine generator based on K-means mean clustering
CN111175046A (en) * 2020-03-18 2020-05-19 北京工业大学 Rolling bearing fault diagnosis method based on manifold learning and s-k-means clustering

Similar Documents

Publication Publication Date Title
Cao et al. Hierarchical hybrid distributed PCA for plant-wide monitoring of chemical processes
CN106845796B (en) One kind is hydrocracked flow product quality on-line prediction method
WO2006031750A2 (en) Application of abnormal event detection technology to hydrocracking units
CN112989635B (en) Integrated learning soft measurement modeling method based on self-encoder diversity generation mechanism
CN112749849A (en) Integrated learning online prediction method for key parameters of continuous catalytic reforming process
CN106709654A (en) Global operating condition evaluating and quality tracing method for hydrocracking process
CN113592359A (en) Health degree evaluation method and device for power transformer
CN114372693B (en) Transformer fault diagnosis method based on cloud model and improved DS evidence theory
CN112465010A (en) DKPCA and KFDA-based transformer fault detection and classification method
CN108491503B (en) Method and system for determining fault type of transformer based on data analysis
CN115881238A (en) Model training method, transformer fault diagnosis method and related device
Meng et al. Two-level comprehensive energy-efficiency quantitative diagnosis scheme for ethylene-cracking furnace with multi-working-condition of fault and exception operation
CN112258014A (en) Clustering and grouping-based risk discrimination analysis method for heat exchangers
Kushwaha et al. Performance evaluation of bagasse-based cogeneration power generation plant utilizing IFLT, IF-FMEA and IF-TOPSIS approaches
CN101727609A (en) Pyrolyzate yield forecasting method based on support vector machine
Paiva et al. Improvement of the monochlorobenzene separation process through heat integration: A sustainability-based assessment
CN114118292B (en) Fault classification method based on linear discriminant neighborhood preserving embedding
Hu et al. Techno-economic analysis on an electrochemical non-oxidative deprotonation process for ethylene production from Ethane
CN112560924B (en) Propylene rectifying tower state monitoring method based on dynamic internal slow feature analysis
CN115327081A (en) Transformer fault diagnosis method for improved goblet sea squirt group optimization support vector machine
CN111598305B (en) Light hydrocarbon separation device running state optimization and prediction method
CN110458408B (en) Method for analyzing influence consequence of typical fault on mobile equipment and device
CN107871054A (en) Oil refining atmospheric and vacuum distillation unit nertralizer Selection Method and nertralizer composition based on AHP Field Using Fuzzy Comprehensive Assessments
Sadighi et al. Modeling and optimizing a vacuum gas oil hydrocracking plant using an artificial neural network
CN112881827B (en) Oil-immersed transformer fault diagnosis method based on improved grey correlation analysis

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