| The environmental conditions during Floating Production Storage and Offloading Units(FPSO)offloading operation are more severe than those in the port.Moreover,the movement of multiple floating bodies makes the working conditions more complicated,and it is prone to accidents such as collisions and leakage.In order to eliminate the hidden dangers in the bud,the risk assessment of FPSO offloading operation is carried out.The accident evolution law and development process in the offloading operation process are analyzed in depth,and high-incidence accidents are early warning.The goal is to provide security for FPSO offloading operation.In view of the shortcomings in the current research status of FPSO offloading operation risk,this paper carries out the following three aspects:(1)Existing FPSO offloading operation risk researchs only consider the accident as an orderly occurrence of single event or superposition of potential factors.To solve this problem,a FPSO offloading operation accident evolution law research method based on Functional Resonance Accident Model(FRAM)is proposed.The model fully reveals the dynamics and coupling relationships between offloading systemic risk factors.First,the human factors,technical factors,and organizational factors having the potential to cause the accident are identified based on common performance conditions of FRAM.Then,functional performance changes are evaluated to determine the key operations and accident evolution process.Finally,prevention and control barriers against accidents are established.Taking the "FPSO+Cargo Transfer Vessel(CTV)+Ordinary Tanker" offloading operation as an example,the accident evolution law was studied.The results show that the berthing and transfer stage is the key operation links for high accidents.And the comprehensive quality of the operators,the berthing technology of CTV,and the communication effect between CTV and FPSO are the key factors inducing accidents.(2)FPSO offloading operation has problems such as difficulty in determining failure probability,missing fault data and related information,etc.Aiming at the above problems,a risk assessment method based on intuitionistic fuzzy set and Vlsekriterijumska Optimizacija I Kompromisno Resenje(VIKOR)is proposed to realize the hierarchical supervision of risk factors of FPSO offloading operation.Firstly,the scoring expert and the evaluation criteria are objectively weighted by the trust function and the entropy method,and the evaluation matrix is obtained.Secondly,the optimal and worst items of each risk factor under the evaluation criteria are determined.Finally,the risk ranking is based on the distance between the evaluation value and the optimal item.The evaluation method divides the risk factors of "FPSO+CTV+Ordinary Tanker" offloading operation into three levels.There are 7 primary risks,11 secondary risks,and 7 tertiary risks.(3)Existing researchs can't realize real-time early warning of FPSO offloading operation accidents.And the early warning model mostly uses a single data set for training,resulting in poor generalization ability.Regarding the issue above,a collision warning model for FPSO offloading operation based on cross-validation Support Vector Machine(SVM)is proposed.Firstly,the sample data is divided into multiple subsets,and any subset is retained as a test sample for repeated training each time.Then,the parameter value in the parameter space of SVM with the smallest error is found,and the collision accident warning model of FPSO offloading operation is obtained.Finally,the real-time monitoring data of the on-site wind and wave current is input into the collision accident warning model to predict the movement trend of tanker and to warn the dangerous situation.The mean square error of the predicted ordinary tanker surge amplitude is 0.53 and the predicted ordinary tanker sway amplitude is 0.41,which are superior to the conventional SVM model.Therefore,the cross-validation SVM model has better generalization ability and accuracy. |
PDF Full Text Request