Collaborative Optimization Of Maintenance And Scheduling For Ship-loading Operation System In Dry Bulk Port

With the development of the dry bulk shipping market,a dry bulk port,which is an important node of the global dry bulk supply chain,its throughput maintains at a high level,making the handling operation quite busy.Since the handling equipment of the dry bulk port is large and complex,its failure will stop the operation of the dry bulk port and impact the whole dry bulk supply chain.To reduce the influences of handling equipment failure,this study mined the equipment fault characteristics of ship-loading operation system in dry bulk port,proposed an “equipment-level to flow-level to system-level” operation and maintenance optimization method based on the data mining results,and built a collaborative optimization model for maintenance and scheduling of ship-loading operation system,fulfilling optimization of the system scheduling and equipment maintenance at the same time.Firstly,aiming at the issues that a number of “equipment downtime reasons” data missed in the data chassis of ship-loading operation system in dry bulk port,a deep learning model based on high-level feature fusion was established.Specifically,a 1D-CNN sub-model,a Vanilla LSTM sub-model,and a VGG16 sub-model were used to process the structured nontime series data,structured time series data,and unstructured image data respectively.An LRCN sub-model was developed to extract features from unstructured video data.After that,the extracted features were fused,and “equipment downtime reasons” were recognized,so as to complete the data chassis cleaning.Also,the handling equipment fault feature was extracted by data mining.Secondly,an “equipment-level to flow-level to system-level” operation and maintenance optimization model of ship-loading operation system in dry bulk port was established.To solve the optimization model,a Variable Lead-Lag Time Window-based solving method is proposed.Concretely,based on the handling equipment fault feature extracted by data mining,an evolution model of a single equipment failure rate was constructed to achieve operation and maintenance optimization at equipment-level.Then,according to the Variable Lead-Lag Time Window method,the predictive maintenance interval length time threshold of the equipment was determined to further optimize the equipment availability.Based on it,the Theory of Constraint was used to determine the priority of equipment in series-parallel flow,and an opportunistic maintenance strategy was proposed thus achieving operation and maintenance optimization at flow-level.Considering the interplay between operation flows in the shiploading operation system,the flow-level operation and maintenance scheme was revised thus achieving the operation and maintenance optimization at system-level.Thirdly,a collaborative optimization model of maintenance and scheduling for shiploading operation system in dry bulk port was established to minimize the total cost of maintenance and scheduling.A two-layer algorithm based on the opportunistic maintenance strategy was proposed to solve the optimization model,so as to complete the collaborative optimization of maintenance and scheduling.Finally,the ship-loading operation system of the phase II project for a dry bulk port in northern China was taken as an example to verify the effectiveness of the proposed method.The results show that the deep learning model based on high-level feature fusion can effectively clean the data chassis of the ship-loading operation system.The proposed “equipment-level to flow-level to system-level” operation and maintenance optimization method,as well as the collaborative optimization model of maintenance and scheduling,can effectively reduce the impact of equipment failure on the port operation and improve its stability.The deep learning model based on high-level feature fusion constructed in this study can provide a theoretical foundation for the processing of multi-source heterogeneous port data.The proposed “equipment-level to flow-level to system-level” operation and maintenance optimization method can provide a theoretical basis for the operation and maintenance of the ship-loading operation system.Besides,the proposed collaborative optimization model of maintenance and scheduling can provide decision support for scheduling of ship-loading operation system.