| Dangerous goods are essential raw materials for petrochemicals.new materials,fine chemicals and high-end manufacturing industries.In recent years,their types and circulation have been increasing,and the transportation and storage of dangerous goods in terminals has increased year by year.Dangerous good container(DGC),as the most important carrier for the transportation and storage of dangerous goods in terminals,have gradually increased their proportion in the terminal yard.As a result,the imbalance between the fixed storage capacity of the DGC yard and the increase in storage demand has become increasingly prominent.Once there is no safe and scientific storage space allocation strategy for DGC in the yard,it will not only cause potential safety accidents,but also cause more serious chain reactions and secondary disasters after the accident,which will eventually cause casualties,property losses,and environmental pollution.Due to the complex and abstract safety elements of the storage space allocation of DGC,and the lack of quantitative representation of the safety index,the current manual allocation of storage space for DGC in terminals generally has practical problems such as low storage space utilization and insufficient safety index.General safety accidents in the DGC yard have caused major secondary disasters due to unreasonable storage.For example,the Tianjin Port "812" explosion accident is caused by the high temperature and spontaneous combustion of nitrocellulose in the dangerous goods containers in the storage yard,and a chain reaction of disasters caused by a series of explosions.Therefore,under the premise that the design storage capacity of the terminals’s DGC yard remains unchanged,how to obtain the optimal safe storage space allocation decision for reducing the storage risk,improving the emergency response capacity of the accident and increasing the storage space utilization,has become one of the urgent problems to be solved in the current storage operations of DGC in terminals.In view of the above problems,the basic theoretical issues of disaster chain,safety elements,and storage strategies in the process of storage space allocation of DGC yard are firstly analyzed.By studying the risk factors and safe storage specifications of DGC,the evaluation system and evaluation index for the safe storage space allocation are constructed for quantifying the safety index of the storage space allocation of DGC.With the help of the quantitative characterization of the safety index by the safety evaluation system,a multiobjective optimization model and Markov model for the safe storage space allocation of DGC yard is established.In view of the dynamic uncertainty and high-dimensionality of the discrete state space and the action space in the model,an improved deep reinforcement learning algorithm is proposed,which can quickly solve the optimal safe storage space allocation decision.Finally,an application system for safe storage space allocation of DGC yard is constructed,and a comprehensive study is carried out on the allocation strategy of typical DGC and the balance between safety index and safe storage space by taking Shanghai Port as an application example.The main research contents of this dissertation are as follows:(1)From the perspective of safety,the basic theoretical issues of safe storage space allocation in DGC yard are deeply studied.DGC yard is a complex and huge system,and all links and elements are closely related to safety.Therefore,after summarizing and analyzing the definition,characteristics and complexity of safe storage space allocation of DGC yard,the accident forms and the evolution model of the disaster chain in the storage process of DGC are systematically studied,and on this basis,the six major safety factors such as isolation relationship,container attribute state,and container block attribute state in the storage process of DGC are further analyzed.According to the strategy and principles of safe storage for DGC,all relevant factors involved in the safe storage space allocation are determined as the theoretical basis for the subsequent chapters.(2)A study on the quantitative evaluation of the safe storage space allocation of DGC yard is carried out by using the safety evaluation theory.Based on historical experience analysis,the factors with the highest probability of occurrence of risk accidents are identified as the risk factors of the research.Combined with the safe storage specification,the evaluation index and measurement method of the safe storage space allocation of DGC yard are proposed.In order to quantify the safety index in the storage space allocation strategy of DGC,according to the characteristics of fuzzy interaction between risk factors and risk assessment in the index system,a quantitative evaluation method for the safe storage space allocation of DGC is designed.The case shows that the quantitative evaluation system constructed in this dissertation can objectively reflect the safety index of safe storage space allocation in DGC yard,and has important application value.(3)Considering the storage utilization and safety,the optimal decision-making problem of safe storage space allocation of DGC yard is studied deeply.A multi-objective optimization model is constructed to describe the safe storage space allocation process under the uncertainty of the arrival time of DGC,taking the cumulative reward of safe storage space and safety index of DGC yard as the optimization goal.Aiming at the large-scale discrete state space and action space of the DGC yard,the traditional optimization algorithm has limitations in solution speed and convergence.Therefore,the multi-objective optimization model is converted into a Markov decision model,and the state,action and reward function of the decision model are designed.Meanwhile,a fast solution method for safe storage space allocation of DGC yard is proposed based on the improved A2C algorithm.By introducing an action mask into the policy neural network of the traditional A2C algorithm,the action space is pruned according to the constraint conditions,and the infeasible actions in the safe space allocation process are shielded and the algorithm convergence is accelerated.Calculation examples show that the performance of the allocation strategy obtained by this method is better than that of manual allocation and other strategy-based DRL algorithms,and it can converge quickly.(4)Based on the above theoretical research results,an application system for safe storage space allocation of DGC yard is built,and the safe storage space allocation method studied in this dissertation is transplanted to the production business of Shanghai Port for example verification.In order to obtain the storage status information required to quantify the storage safety index,five types of situational awareness subsystems such as temperature perception of temperature-controlled container,dangerous gas perception,and meteorological perception are designed.By analyzing the storage space allocation of typical DGC in Shanghai Port,and the equilibrium relationship between safety index and safe storage space,the optimal weight ratio of safety index and safe storage space is determined to provide a basis for port staff to choose a reasonable storage strategyThe main research of this article has been deployed in December 2019 and applied to the space allocation management of dangerous goods container in the Shanghai Port.After two years of practical application in Shanghai Port,the performance and stability of the safe storage space allocation strategy for dangerous goods container have been fully verified.The average storage space utilization rate of DGC has increased by about 13%,and the average safety index has increased by about 12%,the amount of storage container increases by nearly 130,000 TEU every year.The research in this dissertation guarantees the demand for the increase in the volume of DGC in Shanghai Port,and has reference and demonstration significance for the improvement of the storage space utilization and safety index of the DGC yard in terminals. |
PDF Full Text Request