CN-121980398-A - Fault identification method and device of ship tail gas aftertreatment system and ship tail gas aftertreatment system

CN121980398ACN 121980398 ACN121980398 ACN 121980398ACN-121980398-A

Abstract

The application discloses a fault identification method and device of a ship tail gas aftertreatment system and the ship tail gas aftertreatment system, and aims to solve the problems that the adaptability to the composite fault of multivariable coupling is insufficient and accurate identification is impossible. The fault identification method of the ship tail gas aftertreatment system comprises the steps of sampling a plurality of operation parameters of the ship tail gas aftertreatment system within a preset time period to obtain a plurality of operation parameter time sequences, acquiring fault sensitivity of each operation parameter according to the plurality of operation parameter time sequences, representing the correlation between the operation parameters and faults, and constructing a fault identification model according to operation data corresponding to the fault sensitivity parameters to obtain fault information corresponding to abnormal working conditions. The fault identification method of the ship tail gas aftertreatment system accurately locates key sensitive parameter groups causing faults, and therefore accurate analysis of multivariable coupling faults is achieved.

Inventors

LI XINGYI
KONG WENJIE
LIU TIANYANG
CHU XUEJIAO
Tie Zhengze
Zuo Jinchao
WANG CHUNYUAN
Xiong bibo
KUANG TIANYANG
HU SHENGKAI
WU YUXIANG

Assignees

中国船舶集团有限公司第七一一研究所

Dates

Publication Date: 20260505
Application Date: 20251218

Claims (16)

1. The fault identification method of the ship tail gas aftertreatment system is characterized by comprising the following steps of: Sampling a plurality of operating parameters X i of the ship tail gas aftertreatment system within a preset time period to obtain a plurality of operating parameter time sequences Wherein i is more than or equal to 1 and less than or equal to m, m is the number of operating parameters, and n is the time point number of the time sequence; Acquiring fault sensitivity of each operation parameter according to the plurality of operation parameter time sequences, wherein the fault sensitivity represents the correlation between the operation parameter and a fault; And constructing a fault identification model according to the operation data corresponding to the fault sensitivity parameters to obtain fault information corresponding to the abnormal working conditions, wherein the fault sensitivity parameters are operation parameters with the fault sensitivity greater than or equal to a first preset threshold.
2. The fault identification method as claimed in claim 1, wherein said obtaining fault sensitivity of each of said operating parameters from said plurality of operating parameter time series comprises: Performing cluster analysis on the plurality of operation parameter time sequences respectively to obtain a plurality of operation parameter characteristic time sequences; classifying the characteristic time sequences of the operating parameters to obtain a normal working condition sequence X normal and an abnormal working condition sequence X fault , wherein X normal and X fault are m multiplied by n matrixes; Respectively normalizing X normal and X fault to obtain corresponding normalized matrix And ; According to respectively And Calculating to obtain the association degree between every two operation parameters at each time point of the normal working condition and the abnormal working condition, and obtaining a normal working condition association degree matrix and an abnormal working condition association degree matrix; And calculating according to the normal working condition association degree matrix and the abnormal working condition association degree matrix to obtain the fault sensitivity of each operation parameter.
3. The fault identification method as claimed in claim 2, wherein said classifying the plurality of operation parameter feature time sequences to obtain a normal operating condition sequence X normal and an abnormal operating condition sequence X fault comprises: respectively counting the maintenance requirement probabilities corresponding to the characteristic time sequences of the plurality of operating parameters according to the historical maintenance records of the ship tail gas aftertreatment system; If the maintenance requirement probability is greater than or equal to a second preset threshold, determining that the corresponding operation parameter characteristic time sequence belongs to a normal working condition sequence X normal , and if the maintenance requirement probability is less than the second preset threshold, determining that the corresponding operation parameter characteristic time sequence belongs to an abnormal working condition sequence X fault .
4. The fault identification method as claimed in claim 2, wherein the normalizing the X normal and the X fault to obtain the corresponding normalized matrices X ́ normal and X ́ fault respectively comprises: Calculation of , Wherein, the Is the value of the ith parameter in X normal or X fault at the kth time point, i is more than or equal to 1 and less than or equal to m, and k is more than or equal to 1 and less than or equal to n; As the mean of the ith parameter in X normal or X fault at n time points, The value of the ith parameter at the kth time point in X ́ normal or X ́ fault .
5. The method for identifying a fault as claimed in claim 2, wherein the calculating according to X ́ normal and X ́ fault respectively obtains the correlation between every two operation parameters at each time point of the normal condition and the abnormal condition, and obtains a normal condition correlation matrix and an abnormal condition correlation matrix, includes: Calculation of ; Wherein, the For parameters at the kth time point in X ́ normal or X ́ fault And parameters And j is equal to or greater than 1 and is equal to or less than m, Δmax is the global maximum difference of X ́ normal or X ́ fault , Δmin is the global minimum difference of X ́ normal or X ́ fault , , Ρ is a preset resolution factor; the degree of association between the parameters in X ́ normal forms a normal condition degree of association matrix The degree of association between the parameters in X ́ fault forms an abnormal condition degree of association matrix 。
6. The method for identifying faults as claimed in claim 5, wherein said calculating fault sensitivities of the respective operating parameters based on the normal condition association degree matrix and the abnormal condition association degree matrix comprises: Calculating a relevance change matrix between the normal working condition relevance matrix and the abnormal working condition relevance matrix ; The sum of Δr row i vectors is calculated as the fault sensitivity S i of the operating parameter X i .
7. The fault identification method of any one of claims 1-6, wherein the fault identification model is constructed based on a machine learning algorithm from operational data corresponding to the fault-sensitive parameters.
8. The fault identification method of claim 7, wherein the fault identification model determines a final fault class based on voting results of all decision trees, wherein the voting results , For the j-th decision tree, N () is the indicator function.
9. The fault identification method of claim 1, further comprising, prior to sampling the plurality of operating parameters X i of the marine exhaust aftertreatment system for a predetermined period of time: and establishing an operation condition virtual library, wherein the operation condition virtual library is used for storing each condition of the ship tail gas aftertreatment system and operation parameters corresponding to each condition.
10. The fault identification method as claimed in claim 1 or 9, characterized in that, If the ship tail gas aftertreatment system is a selective catalytic reduction system, the operating parameters include at least two of a feed pump operating state, a bypass valve state, a reactor inlet temperature, a reactor outlet temperature, a reactor pressure difference, a reducing agent flow, a reactor inlet NOx concentration and a reactor outlet NOx concentration; If the marine tail gas post-treatment system is the waste gas cleaning system ‌, the operation parameters comprise at least two of a flue gas valve state, a seawater pump operation state, a tail gas SO 2 content, a tail gas CO 2 content, a reactor pressure difference, a system 0.1% mode exhaust sulfur-carbon ratio, a system 0.5% mode exhaust sulfur-carbon ratio, a water quality analyzer turbidity value, a water quality analyzer PH value and a water pump outlet pressure.
11. A fault identification device for a marine exhaust aftertreatment system, comprising: The sampling module is used for sampling a plurality of operating parameters X i of the ship tail gas aftertreatment system within a preset time period to obtain a plurality of operating parameter time sequences Wherein i is more than or equal to 1 and less than or equal to m, m is the number of operating parameters, and n is the time point number of the time sequence; The parameter fault sensitivity analysis module is used for acquiring fault sensitivity of each operation parameter according to the plurality of operation parameter time sequences, and the fault sensitivity represents the correlation between the operation parameter and the fault; The fault identification module is used for constructing a fault identification model according to the operation data corresponding to the fault sensitivity parameters to obtain fault information corresponding to the abnormal working conditions, wherein the fault sensitivity parameters are operation parameters with the fault sensitivity being greater than or equal to a first preset threshold value.
12. The fault identification device of claim 11, wherein the parameter fault sensitivity analysis module comprises: The clustering sub-module is used for carrying out clustering analysis on the plurality of operation parameter time sequences respectively to obtain a plurality of operation parameter characteristic time sequences; The classification sub-module is used for respectively counting maintenance requirement probabilities corresponding to the plurality of operation parameter feature time sequences according to the historical maintenance records of the ship tail gas aftertreatment system, determining that the corresponding operation parameter feature time sequences belong to a normal working condition sequence X normal if the maintenance requirement probability is greater than or equal to a first preset threshold value, and determining that the corresponding operation parameter feature time sequences belong to an abnormal working condition sequence X fault if the maintenance requirement probability is less than the first preset threshold value, wherein X normal and X fault are m multiplied by n matrixes; The standardized submodules are used for respectively standardizing X normal and X fault to obtain corresponding standardized matrixes X ́ normal and X ́ fault ; The parameter association degree analysis submodule is used for calculating and obtaining association degrees between every two operation parameters at each time point of a normal working condition and an abnormal working condition according to X ́ normal and X ́ fault respectively, and obtaining a normal working condition association degree matrix and an abnormal working condition association degree matrix; And the fault sensitivity calculation sub-module is used for calculating the fault sensitivity of each operation parameter according to the normal working condition association degree matrix and the abnormal working condition association degree matrix.
13. The fault identification device according to claim 11 or 12, wherein the fault identification module is specifically configured to construct the fault identification model based on a machine learning algorithm according to the operation data corresponding to the fault sensitive parameter.
14. A marine exhaust aftertreatment system, comprising: at least one processor, and A memory communicatively coupled to at least one of the processors, wherein, The memory stores instructions executable by the processor for execution by the processor to implement the fault identification method of the marine exhaust aftertreatment system of any one of claims 1-10.
15. A computer readable storage medium storing computer instructions for execution by the computer to implement the fault identification method of the marine exhaust gas aftertreatment system of any one of claims 1-10.
16. A computer program product comprising instructions which, when executed by a computer device, cause the computer device to perform the method of fault identification of a marine exhaust aftertreatment system according to any one of claims 1-10.

Description

Fault identification method and device of ship tail gas aftertreatment system and ship tail gas aftertreatment system Technical Field The application relates to the technical field of ship fault diagnosis, in particular to a fault identification method and device of a ship tail gas aftertreatment system, a computer readable storage medium, a computer program product and the ship tail gas aftertreatment system. Background Along with the enforcement of emission standards in the global sea area by the International Maritime Organization (IMO) and the annex VI TIER III of the antifouling convention, the control of emission of nitrogen oxides (NOx) and sulfur oxides (SOx) in ship exhaust gas has become a key task for the shipping industry to meet environmental requirements, and the ship exhaust gas post-treatment equipment has become a core technical means for achieving the aim. In current marine Exhaust gas aftertreatment technology systems, a selective catalytic Reduction system (SCR, selective Catalytic Reduction) and an Exhaust gas cleaning system (EGC, exhaust GAS CLEANING) are mainstream solutions. However, the complex operating conditions and harsh environments encountered during vessel sailing present a significant challenge to the stable operation of SCR/EGC systems. On one hand, when the ship sails, the load of the main engine changes along with the sailing state (such as sailing speed adjustment, cargo load change, sea condition fluctuation and the like) to cause continuous fluctuation of key parameters such as tail gas temperature, flow, pollutant concentration and the like to form dynamic working condition fluctuation, and on the other hand, the high salt spray characteristic in the marine environment is easy to corrode equipment parts, and the maintenance resources are limited during the sailing of the ship, so that frequent and comprehensive equipment overhaul is difficult to realize. Under the superposition influence of the multiple severe conditions, faults such as urea crystallization, catalyst poisoning, unbalanced washing liquid and the like frequently occur in the SCR/EGC system. Industry data show that the annual average failure rate of the SCR/EGC system exceeds 12%, wherein about 30% of failures belong to unknown failures which cannot be identified by the traditional diagnosis method, and the failures not only can cause equipment failure and influence the emission of tail gas of a ship to reach standards, but also can seriously threaten the environment-friendly compliance and navigation safety of the ship, and bring economic loss and compliance risks to shipping enterprises. More importantly, the fault discrimination technology of the ship tail gas aftertreatment system at the present stage is difficult to effectively cope with the fault challenges, and the market application prospect of the ship tail gas aftertreatment system is severely restricted. The defects of the existing fault judging technology are mainly characterized in that firstly, the traditional fault judging is based on a fixed threshold alarming mechanism, namely, by setting a fixed normal range of each parameter, alarming is triggered when the parameter exceeds the range, but due to continuous change of related parameters such as tail gas temperature and the like caused by fluctuation of a ship host load, the fixed threshold cannot adapt to the dynamic change rule of the parameter, so that the false alarm rate of the alarming mechanism is up to 40%, and misjudgment or missed judgment of an actual fault is caused, secondly, the faults of a ship tail gas aftertreatment system are not caused by single parameter abnormality, most faults belong to composite faults, the change rule and the coupling relation of a plurality of parameters are needed to be comprehensively analyzed to accurately judge, and the judging rule can only be established aiming at the known single faults or simple associated faults, so that the adaptability of the composite faults to multivariable coupling is insufficient, and the diagnosis requirement of the complex faults is difficult to be met. Under the background, providing a ship tail gas aftertreatment system intelligent diagnosis technology with composite fault and unknown fault recognition capability has become an urgent need in the field of shipping industry and ship environmental protection equipment. Disclosure of Invention In the summary, a series of concepts in a simplified form are introduced, which will be further described in detail in the detailed description. The summary of the application is not intended to define the key features and essential features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter. In view of the problems existing at present, the application provides a fault identification method of a ship tail gas aftertreatment system, which comprises the following steps: Sampling a plurality of operating