Cross-domain multi-uncertainty cooperative scheduling method for port integrated energy system and ferry logistics system

Abstract

The invention discloses a port integrated energy system and ferry logistics system cooperative scheduling method under cross-domain multiple uncertainty, and relates to the field of port integrated energy and ferry logistics, and the method comprises the steps: optimizing the accuracy of a port energy production and logistics operation multiple uncertainty representation model; determining a scheduling optimization model in a multi-uncertainty scene; scheduling constraint optimization is carried out based on interval modeling; dynamic balance between energy consumption of the port ferry logistics system under multiple uncertainties of logistics operation and energy supply of the port comprehensive energy system under multiple uncertainties of energy production is realized; establishing a collaborative scheduling robust optimization model of the port integrated energy system and the ferry logistics system; and solving to obtain a port integrated energy system and ferry logistics system collaborative robust scheduling scheme through a constructed distributed collaborative optimization solving framework combining a column and constraint generation algorithm and a self-adaptive alternating direction multiplier method. According to the invention, the cooperation efficiency, economy and robustness of the port can be improved.

Description

Technical Field

[0001] The present invention relates to the field of port integrated energy and ferry logistics, and in particular to a collaborative scheduling method for a port integrated energy system and a ferry logistics system under cross-domain multiple uncertainties. Background Art

[0002] As key hubs for waterborne passenger transportation, ferry ports are gradually transitioning from traditional logistics management models to one that integrates integrated energy and ferry logistics through the electrification and hydrogenation of their logistics facilities and the coordinated promotion of integrated energy equipment. Port integrated energy systems integrate diverse energy resources, including renewable energy, cogeneration units, and energy storage equipment, to provide a stable energy supply for ferry logistics systems and infrastructure. In this context, research on the coordinated scheduling of port integrated energy systems and ferry logistics systems, and exploring the potential for efficient synergy to enhance port energy efficiency and operational effectiveness, is a key path to achieving green and low-carbon port development.

[0003] Currently, the port's integrated energy system and ferry logistics system are still primarily operated independently. This fragmented management model makes it difficult for the energy system to timely match the dynamic energy consumption needs of the logistics system, hindering the improvement of the port's overall energy efficiency. By optimizing the coordinated scheduling of the port's integrated energy system and ferry logistics system, the matching of logistics facility energy consumption with energy equipment production capacity can be effectively improved, avoiding overload problems caused by concentrated equipment operations during peak periods, thereby significantly improving the port's overall energy efficiency. For example, by optimizing logistics operation plans to match the operating hours of high-energy-consuming equipment with peak renewable energy output periods, reliance on traditional fossil fuels can be effectively reduced.

[0004] However, in actual operations, port energy production and logistics operations face multiple complex uncertainties, exacerbating the difficulty of coordinated scheduling. In port integrated energy systems, the output of renewable energy units such as photovoltaic and wind turbines fluctuates significantly. Deviations from forecasted output can lead to imbalances in energy supply and demand. Infrastructure load demands are based on day-ahead forecasts, and deviations from actual demand can lead to energy redundancy or shortages. In port ferry logistics systems, due to natural conditions such as weather and tides, actual ferry arrival times often differ from planned times, potentially triggering chain reactions such as berthing delays and delays in ferry operations. Furthermore, the high frequency and short turnaround times of ferry logistics operations exacerbate the risk of uncertainty propagation, potentially leading to a precipitous drop in logistics efficiency and the failure of energy supply and demand planning. Furthermore, the multi-energy coupling between the port integrated energy system and the ferry logistics system results in the dynamic coupling of multiple uncertainties in energy production and logistics operations, further exacerbating the difficulty of coordinated scheduling between the two systems.

[0005] As a core vehicle for short- and medium-distance water passenger transport, ferries, with their characteristically regular inter-port voyages and short-distance operations, have become a pioneer in the clean energy transition for ships. Zero-carbon emission vessels, exemplified by electric ferries, have established a large-scale application system. Their standardized operational process encompasses three key steps: shore power charging (i.e., rapid recharging via the port's energy system after berthing), passenger ferrying (i.e., efficient connection via electric / hydrogen shuttles), and port departure scheduling (i.e., energy disconnection and subsequent departure). Berth allocation strategies, shore power scheduling timing, and shuttle bus resource allocation constitute multi-dimensional decision variables that directly impact ferry dwell time in port, passenger service efficiency, and overall logistics energy intensity. For example, if shore power supply fails to coincide with peak renewable energy output periods, charging will rely on high-carbon backup power sources. Delays in shuttle bus scheduling can lead to passenger delays, further increasing energy consumption for auxiliary lighting, climate control, and other infrastructure. Therefore, the construction of a coordinated scheduling model for the port's integrated energy system and ferry logistics system requires breakthroughs in two key technical aspects: first, multi-stage energy consumption modeling, to quantify the dynamic energy consumption of the ferry from berthing and charging, passenger ferrying to departure and sailing, especially the coupling relationship between the time-varying characteristics of the charging load and the start-stop frequency of the ferry bus; second, coordinated optimization of facilities, to establish a resource linkage mechanism of berth-shore power-ferry bus, and to achieve dynamic coordination of energy equipment output, charging demand and logistics operation rhythm through time matching.

[0006] Existing research has mostly focused on optimizing uncertainty in a single area, either energy or logistics, making it difficult to address the complexity of the coordinated scheduling scenario of port integrated energy systems and ferry logistics systems. Although some studies have attempted to consider multi-domain uncertainty, their research subjects have mainly focused on port container logistics scenarios and have three significant limitations: First, the model is not adaptable enough. Existing methods focus on the static coupling optimization of container logistics and electricity, and are not adapted to the differentiated needs of ferry ports in the process of electrification and hydrogen energy (such as short-cycle high-frequency charging and dynamic coordinated scheduling of electric and hydrogen ferries). Second, they do not fully consider the multi-energy coupling characteristics, ignore the spatiotemporal interaction between multi-energy flows and logistics in ports, and fail to fully tap the multi-energy complementarity potential of port integrated energy systems. Third, there is a lack of a coordinated scheduling framework for port integrated energy systems and ferry logistics systems that is adaptable to high-uncertainty scenarios. In particular, there is a lack of systematic representation and modeling of the multiple uncertainties and dynamic relationships between energy production and logistics operations, as well as a lack of robustness and economic efficiency in system scheduling solutions under high-uncertainty scenarios. Summary of the Invention

[0007] The coordinated scheduling of port integrated energy systems and ferry logistics systems under cross-domain multiple uncertainties faces many challenges. It is urgent to construct a coordinated scheduling method for port integrated energy systems and ferry logistics systems that integrates cross-domain multiple uncertainty representation, dynamic coupling correlation analysis and distributed collaborative optimization technology to improve the coordinated efficiency, economy and robustness of port integrated energy systems and ferry logistics systems, provide theoretical support for the stable, efficient and sustainable operation of ports, and promote the economical, efficient and stable operation of ports under multiple uncertainties in energy production and logistics operations.

[0008] The purpose of the present invention is to be solved by the following technical solutions:

[0009] A collaborative scheduling method for a port integrated energy system and a ferry logistics system under cross-domain and multiple uncertainties, including:

[0010] Construct a multiple uncertainty characterization model for port energy production and logistics operations; optimize the accuracy of the multiple uncertainty characterization model for port energy production and logistics operations using the uncertainty fluctuation interval definition method based on convolutional neural networks and bidirectional long short-term memory networks; determine the scheduling optimization model under multiple uncertainty scenarios with the goal of economic optimization; optimize scheduling constraints based on interval modeling; achieve a dynamic balance between the energy consumption of the port ferry logistics system under multiple uncertainties in logistics operations and the energy supply of the port integrated energy system under multiple uncertainties in energy production based on the dynamic coupling association form between cross-domain multiple uncertainties; establish a robust optimization model for the coordinated scheduling of the port integrated energy system and the ferry logistics system; and obtain a robust scheduling solution for the coordinated scheduling of the port integrated energy system and the ferry logistics system through the constructed distributed collaborative optimization solution framework that combines the column and constraint generation algorithm and the adaptive alternating direction multiplier method.

[0011] The present invention has the following advantages over the prior art:

[0012] 1. The collaborative scheduling method of the present invention combines port integrated energy scheduling technology, ferry logistics scheduling technology, and robust optimization technology, systematically integrating cross-domain dynamic correlation analysis of multiple uncertainties such as renewable energy fluctuations and ferry arrival deviations. By tapping the complementary potential of multiple energy flows and the flexible adjustment capabilities of logistics, it achieves spatiotemporal collaborative optimization of integrated energy supply and ferry logistics operation needs, and generates a collaborative robust scheduling plan for the port integrated energy system and ferry logistics system that takes into account both economy and efficiency.

[0013] 2. The present invention solves two core problems in the prior art: the separate optimization of the port integrated energy system and the ferry logistics system, and the lack of cross-domain multiple uncertainty collaborative modeling. It provides a method for collaborative scheduling of the port integrated energy system and the ferry logistics system under cross-domain multiple uncertainties. This method breaks through the bottlenecks of delayed energy response and limited logistics efficiency under the traditional independent scheduling mode by integrating cross-domain multiple uncertainty dynamic characterization, multi-energy flow and logistics spatiotemporal coupling analysis, and distributed robust optimization technology. It realizes efficient collaborative scheduling of the port integrated energy system and the ferry logistics system under multiple uncertainties of energy production and logistics operations, improves the economic efficiency, robustness and overall energy efficiency of port operations, and provides technical support for the low-carbon transformation of ports. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Attachment Figure 1 A flow chart of the collaborative scheduling method for the port integrated energy system and the ferry logistics system under cross-domain multiple uncertainties provided by the present invention from the perspective of architecture and principle;

[0015] Attachment Figure 2 The uncertain fluctuation interval definition process based on convolutional neural network and bidirectional long short-term memory network of the present invention;

[0016] Attachment Figure 3 The principle of interval modeling method;

[0017] Attachment Figure 4 This is a schematic diagram of the collaborative scheduling solution process for the port integrated energy system and the ferry logistics system under cross-domain multiple uncertainties of the present invention;

[0018] Attachment Figure 5 A comparison chart of wind power output using the existing technology and the method of the present invention in energy scheduling solutions;

[0019] Attachment Figure 6 A comparison chart of photovoltaic output power using the existing technology and the method of the present invention in energy scheduling schemes;

[0020] Attachment Figure 7 A comparison chart of the power of infrastructure electric loads using the existing technology and the method of the present invention in an energy scheduling solution;

[0021] Attachment Figure 8 A comparison chart of cooling load power of infrastructure using the existing technology and the method of the present invention in an energy scheduling solution;

[0022] Attachment Figure 9 A comparison chart of the heat load power of infrastructure using the existing technology and the method of the present invention in the energy scheduling scheme;

[0023] Attachment Figure 10 A comparison chart of the electric power purchased from the external power grid using the existing technology and the method of the present invention in the energy scheduling scheme;

[0024] Attachment Figure 11 A comparison chart of electric power consumption of electric boilers using the existing technology and the method of the present invention in energy scheduling schemes;

[0025] Attachment Figure 12 A comparison chart of electric refrigeration power consumption using the existing technology and the method of the present invention in the energy scheduling scheme;

[0026] Attachment Figure 13 A comparison chart of combined cooling, heating and power generation power using the existing technology and the method of the present invention in an energy scheduling solution;

[0027] Attachment Figure 14 A comparison chart of battery charging and discharging power using the existing technology and the method of the present invention in an energy scheduling solution;

[0028] Attachment Figure 15 A comparison chart of the electric power consumption of electrolyzers using the existing technology and the method of the present invention in the energy scheduling scheme;

[0029] Attachment Figure 16 A comparison chart of hydrogen storage and release rates of hydrogen storage tanks using the prior art and the method of the present invention in an energy scheduling solution;

[0030] Attachment Figure 17 A comparison chart of heat storage and heat release power of heat storage tanks using the existing technology and the method of the present invention in an energy scheduling solution;

[0031] Attachment Figure 18 A comparison chart of hydrogen consumption rates of hydrogen-powered shuttle vehicles using the existing technology and the method of the present invention in a logistics scheduling solution;

[0032] Attachment Figure 19 A comparison chart of the electric power consumption of electric shuttle vehicles using the existing technology and the method of the present invention in a logistics scheduling solution;

[0033] Attachment Figure 20 A comparison chart of shore power consumption using the existing technology and the method of the present invention in a logistics scheduling solution;

[0034] Attachment Figure 21 This is a flow chart of a collaborative scheduling method for a port integrated energy system and a ferry logistics system under cross-domain multiple uncertainties according to the present invention. DETAILED DESCRIPTION

[0035] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that these embodiments are illustrative rather than restrictive, and the scope of protection of the present invention cannot be limited by the following embodiments.

[0036] The collaborative scheduling method of the present invention integrates cross-domain multiple uncertainty dynamic characterization, multi-energy flow and logistics spatiotemporal coupling analysis and distributed robust optimization technology to propose an economical, efficient and stable cross-domain multiple uncertainty collaborative scheduling method for the port integrated energy system and the ferry logistics system.

[0037] The present invention discloses a collaborative scheduling method for a port integrated energy system and a ferry logistics system under cross-domain multiple uncertainties, such as Figure 21 As shown, including:

[0038] S1. Construct a multiple uncertainty characterization model for port energy production and logistics operations; optimize the accuracy of the multiple uncertainty characterization model for port energy production and logistics operations using the uncertainty fluctuation interval definition method based on convolutional neural networks and bidirectional long short-term memory networks; S2. Determine the scheduling optimization model under multiple uncertainty scenarios with the goal of economic optimization; optimize scheduling constraints based on interval modeling; achieve a dynamic balance between the energy consumption of the port ferry logistics system under multiple uncertainties in logistics operations and the energy supply of the port integrated energy system under multiple uncertainties in energy production based on the dynamic coupling association form between cross-domain multiple uncertainties; S3. Establish a robust optimization model for the coordinated scheduling of the port integrated energy system and the ferry logistics system; obtain a coordinated robust scheduling scheme for the port integrated energy system and the ferry logistics system through the constructed distributed collaborative optimization solution framework combining the column and constraint generation algorithm and the adaptive alternating direction multiplier method.

[0039] In one embodiment, constructing a multiple uncertainty representation model for port energy production and logistics operations includes:

[0040] Based on the box uncertainty set and considering the volatility, the uncertain output characteristics of renewable energy units and the uncertain load characteristics of infrastructure, a multiple uncertainty representation model for port energy production is constructed. The formula of the multiple uncertainty representation model for port energy production is:

[0041]

[0042]

[0043] Where, and They represent the renewable energy equipment χ in period t respectively RES The actual output and predicted output; represents the renewable energy equipment χ in period t RES Output fluctuation range; Ω RES represents a collection of renewable energy devices; and They represent the actual and predicted values ​​of energy type load σ in period t respectively; represents the fluctuation amplitude of energy type load during period t; e, h, and c represent electricity, heating, and cooling, respectively;

[0044] According to the changes in ferry arrival time, disembarkation time, and power consumption when the ferry arrives at the port, the formula for the multiple uncertainty representation model of logistics operations is:

[0045]

[0046]

[0047]

[0048]

[0049]

[0050]

[0051] Where, and They represent the actual arrival time and the planned arrival time of the i-th ferry respectively; represents the maximum fluctuation of the actual arrival time of the i-th ferry; N FV represents the number of arriving ferries within the scheduling period; U represents the fluctuation range; Indicates the deviation between the actual arrival time and the planned arrival time; and They represent the actual disembarkation time and planned disembarkation time of passengers on the i-th ferry respectively; represents the maximum fluctuation of the time it takes for passengers to disembark from the i-th ferry; Indicates the deviation between the actual disembarkation time and the planned time; and represents the starting time of passenger ferrying for the i-th ferry; and They represent the actual state of charge and planned state of charge of the i-th ferry when it arrives at the port; represents the maximum fluctuation value of the state of charge when the i-th ferry arrives at the port; Indicates the deviation of the actual state of charge from the planned state of charge; and denote the charging start and end time of the i-th ferry respectively; represents the departure time of the i-th ferry.

[0052] In one embodiment, the accuracy of the multiple uncertainty characterization model for port energy production and logistics operations is optimized using an uncertainty fluctuation interval definition method based on a convolutional neural network and a bidirectional long short-term memory network, including:

[0053] Key features are extracted from historical data through convolutional neural networks to provide input for the prediction of bidirectional long short-term memory networks. Collaborative management information based on historical data is collected to obtain the uncertainty fluctuation range of the multiple uncertainty characterization model of port energy production and logistics operations, and the box boundaries are dynamically adjusted to form a generalized box uncertainty set.

[0054] In one embodiment, with economic optimization as the goal, determining a scheduling optimization model under multiple uncertainty scenarios includes:

[0055] Based on the external energy procurement cost, equipment operation and maintenance cost, energy storage degradation cost, pollution treatment cost, in-port energy sales income and demand response subsidy, and with the principle of minimizing the port comprehensive energy scheduling cost under multiple uncertainties in energy production, the objective function of the port comprehensive energy system scheduling optimization model under multiple uncertainties in energy production is constructed:

[0056]

[0057]

[0058]

[0059]

[0060]

[0061]

[0062] Where C buy 、C om 、C deg 、C ep 、C se and C dr They represent the external energy procurement cost, equipment operation and maintenance cost, energy storage degradation cost, pollution treatment cost, in-port energy sales income, and demand response subsidy during the dispatch period; and They represent the prices of electricity purchased from and sold to the grid during period t; Indicates the unit price of natural gas purchased from the gas grid; and They represent the power purchased and sold from the power grid during period t respectively; They represent the gas purchase rate from the gas network during period t, m 3 / h; Δt E Indicates the unit duration of energy dispatch; c om,χ represents the unit operation and maintenance price of energy equipment χ; Indicates that the output of renewable energy units is uncertain; represents the output power of energy device χ; Ω represents the set of energy devices; represents the energy storage device χ ESS Unit degradation cost of and Represent the energy storage device χ ESS Charging and discharging power; Ω ESS Represents a collection of energy storage devices; represents the emission tax rate or emission right unit price for pollutant τ; ε e,τ and ε g,τ They represent the grid emission factor and the gas pollutant τ emission factor respectively; Q τ Conversion factor indicating the pollution equivalent value of τ; CO2, SO2 and NO x represent carbon dioxide, sulfur dioxide, and nitrogen oxide gases respectively; Indicates the unit price of energy sold within the port's integrated energy system; represents the energy consumption of the port ferry logistics system under multiple uncertainties; represents the energy consumption of port infrastructure under multiple uncertainties; T p and T v They represent the peak shaving and valley filling demand time periods published by the power grid; and They represent the unit price of demand response subsidies corresponding to peak shaving and valley filling periods respectively; The power purchased from external sources in the initial dispatch plan of the port integrated energy system under a deterministic scenario without considering demand response; Indicates the output power of energy equipment other than renewable energy units.

[0063] Based on the operating costs of all shore power, electric shuttles, and hydrogen shuttles within the scheduling period, and taking minimizing the port ferry logistics scheduling cost under multiple uncertainties of logistics operations as the principle, the objective function of the port ferry logistics system scheduling optimization model under multiple uncertainties of logistics operations is constructed:

[0064] C fo =C CI +C ESB +C HSB ;

[0065]

[0066]

[0067]

[0068]

[0069]

[0070] Where C fo represents the joint operation cost of logistics facilities during the scheduling period; C CI 、C ESB and C HSB are the operating costs of all shore power, electric shuttles, and hydrogen shuttles during the dispatch period; c CI,on 、c ESB,on and c HSB,on are the basic operating costs of a single shore power, electric shuttle, and hydrogen shuttle respectively; c ESB,w and c HSB,w represent the unit cost of ferrying operations of electric shuttle and hydrogen shuttle respectively; Indicates that the j-th shore power station serves the i-th ferry in period t. 0-1 state variable; and They represent the state variables of the mth electric shuttle and the qth hydrogen shuttle serving the i-th ferry during period t; Δt L Indicates the unit duration of logistics scheduling; C ec and C dp They represent the logistics energy cost and delay compensation cost of the scheduling period respectively; c da 、c ds and c dd denote the unit penalty costs of ferry delayed arrival, delayed departure and delayed departure respectively; and They represent the actual departure time and the latest scheduled departure time of the i-th ferry respectively; represents the energy consumption of the port ferry logistics system under multiple uncertainties; Indicates the deviation between the actual arrival time and the planned arrival time; Indicates the deviation between the actual disembarkation time and the planned time; It indicates the unit price of energy sold within the port's integrated energy system.

[0071] In one embodiment, the constraints of the port integrated energy system scheduling optimization model under multiple uncertainties in energy production include: multi-energy supply and demand balance, external energy network supply constraints and energy equipment operation constraints; among them, the power generation, hydrogen production, heat production and cooling production of the port integrated energy system in each time period meet the power, hydrogen energy, heat energy and cooling energy needs of the port ferry logistics system and infrastructure in each time period; the external energy network supply constraints constrain the energy purchase and sales interaction between the port integrated energy system and the external energy network, including the power supply capacity of the external power grid, the power returned to the power grid is limited by the maximum transmission power of the power cable, and the gas supply capacity of the external natural gas network is limited by the maximum gas supply rate; ensure that the energy storage equipment operates within a safe range, and the energy level of the energy storage equipment at the end of the scheduling cycle is consistent with that at the beginning of the scheduling.

[0072] The constraints of the port ferry logistics system scheduling optimization model under the multiple uncertainties of logistics operations include: ferry berthing constraints, ferry logistics facility allocation constraints and ferry logistics facility operation constraints; among them, the port applies the continuous berth allocation method to meet the first-come, first-served principle; the ferry berthing time is later than the arrival time to ensure the feasibility of the scheduling plan; the berthing position and berthing time of any two ferries in the same time period are formulated based on the principle of avoidance; the ferry logistics facility allocation constraints include independent allocation of a single facility and joint allocation constraints between facilities; the ferry logistics facility operation constraints include work efficiency, operating power and battery capacity constraints.

[0073] In one embodiment, performing scheduling constraint optimization based on interval modeling includes:

[0074] The interval modeling method is used to reduce the dimensionality of the three-dimensional integer variables in the port ferry logistics system scheduling optimization model to optimize the scheduling optimization model. The two-dimensional binary time variable is introduced to define the start time, processing time, and end time of each ferry berthing, charging, and ferrying. The constraints of the port ferry logistics system scheduling optimization model under multiple uncertainties of logistics operations are optimized, and the constraint condition formula of the optimized port ferry logistics system scheduling optimization model is obtained:

[0075]

[0076]

[0077]

[0078]

[0079]

[0080]

[0081]

[0082]

[0083] Where, Indicates whether the i-th ferry starts the facility ω service phase during time t; Indicates whether the i-th ferry has completed the facility ω service phase during time period t; represents the set of positive integers; and presence(ω i ) indicate whether the i-th ferry accepts the logistics facility ω service during period t; and Indicates whether the i-th ferry accepts the logistics facility ω service during the period t1+1, t1, t3, t2-1 and t2 respectively; and Indicates whether the i-th ferry berths during period t2, period t2-1 and period t respectively; Indicates whether the sth ferry is berthing during time period t; Indicates whether the i-th ferry accepts the logistics facility ω service during time period t'; represents the 0-1 state variable that the zth facility ω serves the i-th ferry at time t'; and They represent whether the i-th ferry is in the charging, berthing, electric ferrying and hydrogen ferrying stages during time period t; M represents an infinite constant, l i represents the starting berth position of ferry i, ι i represents the length of ferry i, l max Indicates the length of the dock shoreline, and They represent the maximum and minimum number of electric shuttle buses and hydrogen shuttle buses that can be allocated to the i-th ferry, and denote the time when the logistics facility ω starts and ends serving the i-th ferry, T represents the time that the logistics facility ω serves the i-th ferry, L Represents a logistics scheduling cycle, and Indicates whether the i-th ferry is berthed at berth b during period t2 and period t2-1 respectively; represents a 0-1 state variable indicating that the zth facility ω serves the i-th ferry at time t; The 0-1 state variable represents the j-th shore power station serving the i-th ferry in period t; and are the 0-1 state variables representing the mth electric shuttle and the qth hydrogen shuttle serving the i-th ferry in period t.

[0084] In one embodiment, based on the dynamic coupling association between multiple uncertainties across domains, achieving a dynamic balance between the energy consumption of the port ferry logistics system under multiple uncertainties in logistics operations and the energy supply of the port integrated energy system under multiple uncertainties in energy production includes:

[0085] Energy production, conversion, and storage equipment are dynamically dispatched through the energy production of the port integrated energy system to meet the actual energy consumption of the port ferry logistics system under multiple uncertainties in logistics operations. Based on the multiple uncertainties in energy production, the dynamic coupling correlation form between multiple uncertainties across multiple fields is determined, and the expression of the dynamic coupling correlation form between multiple uncertainties across multiple fields is:

[0086]

[0087] Where Λ and Υ represent the energy conversion and storage matrices, respectively; and Respectively represent the charging and discharging power of the battery; and Respectively represent the hydrogen storage and release rates of the hydrogen storage tank; and Respectively represent the storage and release power of the heat storage tank; and They represent the power purchased and sold from the power grid during period t, represents the gas purchase rate from the gas network during period t, and represents the actual electricity and hydrogen demand of the port ferry logistics system during period t, and represent the actual electricity, heating and cooling loads of the port infrastructure during period t, represents the transmission power loss from the port integrated energy system to the ferry logistics system during period t, represents the hydrogen loss from the port integrated energy system to the ferry logistics system during period t, and They represent the losses in the process of transmitting electricity, heat and cold from the port integrated energy system to the logistics system during period t.

[0088] In one embodiment, establishing a robust optimization model for coordinated scheduling of the port integrated energy system and the ferry logistics system includes:

[0089] A dual incentive mechanism including load adjustment incentives and grid demand response incentives is introduced to make the port integrated energy system pay load adjustment fees to the ferry logistics system. The objective function of the coordinated scheduling of the port integrated energy system and the ferry logistics system is constructed. Based on the parameter matrix of the port integrated energy system or the ferry logistics system scheduling, a robust optimization model for the coordinated scheduling of the port integrated energy system and the ferry logistics system under cross-domain multiple uncertainties is proposed. The formula of the robust optimization model for the coordinated scheduling of the port integrated energy system and the ferry logistics system is:

[0090]

[0091] Where c τ 、b τ d τ 、r τ 、h τ 、D τ 、R τ 、F τ 、H τ 、L τ and Yτ Represents the parameter matrix of different port integrated energy systems or ferry logistics system scheduling; U τ Indicates uncertain scheduling scenario; x τ 、y τ and u τ They represent the scheduling variable matrix composed of binary variables, continuous variables and uncertain variables respectively; Ω represents the scheduling variable matrix composed of x τ and u τ The uncertain scenarios constituted by PE and PL represent the port integrated energy system and port ferry logistics system respectively.

[0092] In one embodiment, the scheduling variable matrix and the expression of the uncertain scenario are:

[0093]

[0094]

[0095]

[0096] Where x PE 、y PE and u PE They represent the decision variable matrices composed of integer variables, continuous variables, and uncertain variables in the robust optimization problem of port integrated energy system scheduling; x PL 、y PL and u PL They represent the decision variable matrices composed of integer variables, continuous variables and uncertain variables in the robust optimization problem of port ferry logistics system scheduling; and Respectively represent each period and each time period The matrix of integer variables formed; and Respectively represent each period Each time period Each time period and each time period The integer variable matrix composed of n ESB,m 、n HSB,q 、x i 、f i and l represent the periods Each time period Each ferry x is , each ferryφ is and ferries i The matrix of integer variables formed; and Respectively represent each period Each time period Each time period Each time period and each time period The continuous variable matrix formed by t ber,beg , t ω,beg , t ω,end , t dep 、 ζ ESB,m ,ζ HSB,q 、 and Respectively, each ferry Ferries Ferries Ferries Each time period Each time period Each time period Each time period Each time period Each time period Each time period and each time period The continuous variable matrix formed by and Respectively represent each period and each time period The uncertain variable matrix formed; and Respectively, each ferry Ferries and ferries The uncertain variable matrix formed;

[0097] in, represents the energy storage device χ ESS The charging status of , if charged, it is 1, otherwise it is 0; Indicates the electricity purchase status, 1 if electricity is purchased, otherwise 0; The state variable representing the berthing work of the i-th ferry at berth b during period t; The 0-1 state variable represents the j-th shore power station serving the i-th ferry in period t; and They represent the state variables of the ferry operation of the mth electric shuttle and the qth hydrogen shuttle serving the i-th ferry during period t; represents the actual passenger capacity of the mth electric shuttle bus during period t; represents the passenger capacity of the qth hydrogen shuttle bus in period t; x is The state variables of the berthing positions of any ferry i and ferry s; φ is The state variable of the berthing time of any ferry i and s; l i represents the starting berth position of ferry i; represents the output power of energy equipment χ; and Represent the energy storage device χ ESS The charging and discharging power; Indicates the power or rate of energy purchased from the energy network; represents the time when the i-th ferry starts berthing; represents the starting time of the phase when the i-th ferry is served by facility ω; represents the end time of the period in which the i-th ferry is served by facility ω; represents the actual departure time of the i-th ferry; represents the state of charge of the battery of the i-th ferry during period t; and They represent the charge state of the mth electric shuttle and the hydrogen charge state of the qth hydrogen shuttle at the end of period t respectively; and They represent the actual energy consumption of the jth shore power station, the mth electric shuttle bus, and the qth hydrogen shuttle bus during period t respectively; and represents the actual electricity and hydrogen demand of the port ferry logistics system during period t; Indicates that the output of renewable energy units is uncertain; It represents the energy consumption of port infrastructure under multiple uncertainties; represents the actual arrival time of the i-th ferry; represents the actual disembarkation time for passengers of the i-th ferry; represents the actual state of charge of the i-th ferry when it arrives at the port.

[0098] In one embodiment, a distributed collaborative optimization solution framework combining a column and constraint generation algorithm and an adaptive alternating direction multiplier method is constructed to solve a coordinated robust scheduling scheme for the port integrated energy system and the ferry logistics system, including:

[0099] Based on the column and constraint generation algorithm, the robust optimization model of the coordinated scheduling of the port integrated energy system and the ferry logistics system is reconstructed into a main problem and sub-problems. In the first stage, the deterministic matrix is ​​used to solve the main problem, and the scheduling objectives and integer variable matrices of the port integrated energy system and the ferry logistics system are obtained. In the second stage, the KKT condition is used to convert the continuous variable correlation function in the objective function into a max-minc τ T y τ Transformed into a maximum value problem and related to the uncertain variable function maxb τ T u τCombined, and then constitute a sub-problem; based on the integer variable matrix obtained in the first stage, calculate the uncertain variable matrix and the continuous variable matrix; using the uncertain variable matrix as a parameter, the adaptive alternating direction multiplier method is used to perform distributed collaborative scheduling optimization based on the dynamic balance consensus of energy supply and demand between the port integrated energy system and the ferry logistics system; in each iteration, the port integrated energy system optimizes its own scheduling based on the energy consumption of the ferry logistics system obtained from the previous iteration, and sends the updated energy supply plan to the port ferry logistics system; the port ferry logistics system optimizes its own scheduling based on the latest energy supply plan, and sends the updated logistics energy demand back to the port integrated energy system; based on the original residual and the dual residual, the penalty operator matrix and the Lagrange multiplier matrix are updated:

[0100]

[0101] Where, and Respectively represent The residuals and dual residuals of energy supply and demand variables at the iteration; and denote the smoothing parameters of the original residual norm and the dual residual norm respectively; and Respectively represent The sub-penalty operator matrix and the Lagrange multiplier matrix, and Respectively represent The sub-penalty operator matrix and the Lagrange multiplier matrix, u τ represents the scheduling variable matrix composed of uncertain variables; and Respectively represent The smoothed values ​​of residual norm and dual residual norm of energy supply and demand variables of the iteration; Indicates the The smoothed value of the residual norm of the energy supply and demand variables of the iteration, Indicates the The energy variables of the port integrated energy system supplying the ferry logistics system are obtained by the iteration. and Respectively represent Second and The energy demand variables of the ferry logistics system obtained by the iteration are: ε1 and ε2 represent the parameters of the penalty operator matrix adjustment, ρ and They represent the lower and upper bound matrices of the penalty operator respectively; the iteration stops when the convergence condition is met, and the final coordinated robust scheduling scheme of the port integrated energy system and the ferry logistics system is obtained.

[0102] In order to facilitate understanding of the above technical solutions of the present invention, the above technical solutions of the present invention are further explained from the perspectives of architecture and principle as follows.

[0103] Step A: Clarify the composition of the port integrated energy system equipment and ferry logistics system facilities. Step B: Construct a multiple uncertainty characterization model for port energy production. Step C: Construct a multiple uncertainty characterization model for port logistics operations. Step D: Propose an uncertain fluctuation interval definition method based on convolutional neural networks and bidirectional long short-term memory networks. Step E: Establish a port integrated energy system scheduling optimization model under multiple uncertainties in energy production. Step F: Establish a port ferry logistics system scheduling optimization model under multiple uncertainties in logistics operations. Step G: Propose a scheduling constraint optimization method based on interval modeling. Step H: Construct a dynamic coupling association model between cross-domain multiple uncertainties. Step I: Establish a robust optimization model for the coordinated scheduling of the port integrated energy system and the ferry logistics system. Step J: Propose a distributed collaborative optimization solution framework that combines column and constraint generation algorithms and adaptive alternating direction multiplier method. Step K: Solve the models in steps B, C, E, F, H, and I to obtain a robust scheduling solution for the coordinated scheduling of the port integrated energy system and the ferry logistics system. The specific flow chart is as follows. Figure 1 shown.

[0104] In step A, the port's integrated energy system equipment includes energy production, conversion and storage equipment, and various types of equipment work together to produce electricity, heat, cold and hydrogen energy; the port's ferry logistics system facilities include shore power and ferry vehicles, which are jointly dispatched to complete ferry berthing and charging and passenger ferrying operations based on the ferry's arrival time, departure time, battery power and number of passengers.

[0105] In step B, the multiple uncertainties of port energy production are mainly manifested as the difference between the actual output of renewable energy units and infrastructure load values ​​and the predicted output of renewable energy units and infrastructure load values. The output of port renewable energy units, represented by wind power and photovoltaic power, is affected by intermittent and unstable wind speed and radiation intensity, and has significant volatility. At the same time, even in the absence of logistics operations, the port still has multiple energy infrastructure loads such as electric lighting, hot water, and air conditioning, which have obvious volatility. Considering volatility, the uncertain output characteristics of renewable energy units and the uncertain load characteristics of infrastructure can be represented based on the box uncertainty set as Equations (1) and (2):

[0106]

[0107]

[0108] In formulas (1)–(2): and They represent the renewable energy equipment χ in period t respectivelyRES Actual and predicted output, MW; represents the renewable energy equipment χ in period t RES Output fluctuation range, MW; Ω RES represents a collection of renewable energy devices; and They represent the actual and predicted values ​​of energy type load in period t, MW; represents the fluctuation amplitude of energy type load during period t, in MW; e, h, and c represent electricity, heating, and cooling, respectively.

[0109] In step C, the multiple uncertainties in port logistics operations are mainly manifested as the deviation between the actual ferry arrival time, passenger disembarkation time, and power consumption at arrival and the planned ferry arrival time, passenger disembarkation time, and power consumption at arrival provided by the shipping company. Ferry logistics operations mainly include the ferry arriving at the terminal, berthing, connecting to shore power and charging, serving passengers disembarking and ferrying, and ferry departure. These operations are completed by the ferry logistics system. Usually, the scheduling plan is formulated in advance based on the ferry arrival plan provided by the shipping company. However, due to the many uncertainties in actual operations, it will not only have a direct impact on the logistics operations of a single ferry, but may also further affect the operation scheduling of subsequent arriving ferries through cumulative effects. The arrival of the ferry is the starting point of the ferry logistics operation. If the ferry arrives later than the planned time, the logistics operation will be delayed, which will in turn affect the scheduling of subsequent ferries; while arriving early may increase the scheduling pressure of berth resources, which will in turn affect shore power connection and subsequent operations. The arrival time variation can be defined by the difference between the actual arrival time and the planned time, as shown in formula (3). In addition, after the ferry docks, passengers disembark. Changes in disembarkation time will affect the departure time of the ferry vehicles and affect the subsequent vehicle allocation and operations. The change in disembarkation time can be defined by the difference between the actual disembarkation time and the planned time, as shown in formula (4). In addition, after the ferry docks, it is connected to the shore power and starts charging. The departure time depends on the charging time, which is related to the battery power when the ferry arrives at the port. However, if the actual battery power is lower than the planned value provided by the shipping company when the ferry arrives at the port, the charging time in the port will be extended, which may cause the ferry to be delayed in leaving the port, and the power demand will increase. The change in power when the ferry arrives at the port can be defined by the difference between the actual state of charge at arrival and the planned state of charge, as shown in formula (5). The formula is as follows:

[0110]

[0111]

[0112]

[0113]

[0114] In formulas (3)–(5): and They represent the actual arrival time and the planned arrival time of the i-th ferry respectively; represents the maximum fluctuation of the actual arrival time of the i-th ferry; N FV represents the number of arriving ferries in a scheduling cycle; U represents the fluctuation range; represents the deviation between the actual arrival time and the planned arrival time, h; and They represent the actual and planned disembarkation time of passengers on the i-th ferry, h; represents the maximum fluctuation of the time it takes for passengers to disembark from the i-th ferry, h; It represents the deviation between the actual disembarkation time and the planned time, h. and represents the starting time of passenger ferrying of the i-th ferry, h; and They represent the actual state of charge and planned state of charge of the i-th ferry when it arrives at the port; represents the maximum fluctuation value of the state of charge when the i-th ferry arrives at the port; Indicates the deviation of the actual state of charge from the planned state of charge; and denote the charging start and end time of the i-th ferry, h; represents the departure time of the i-th ferry, h.

[0115] In step D, the uncertainty fluctuation interval definition method based on convolutional neural networks and bidirectional long short-term memory networks is used to determine the boundary of the box uncertainty set by predicting the uncertainty fluctuation interval, so as to improve the accuracy of uncertainty characterization modeling and overcome the limitation of traditional boundary setting based on fixed multipliers, which leads to overly conservative optimization results.

[0116] Long-term short-term memory (LSTM) networks are widely used in prediction due to their strong time series data processing capabilities. The LSTM network structure consists of four parts: a forget gate, an input gate, an output gate, and an update module, which collaboratively manage the input, storage, update, and output of information, as expressed in Equation (6). The collaboratively managed information is not only the original input historical data, but also the key information after the historical data is processed by the network. In order to capture the long-term dependencies in the time series data, the internal state representation after extraction and processing mainly includes: 1) Long-term dependencies: These include abstract features in the historical data that have an impact on the current prediction task over a long time span. For example, trends, cyclical changes, or abnormal fluctuations in time series data may appear as weak signals at a certain moment but have long-term cumulative effects. 2) Abstract representation of effective features: The network maps the original input into more abstract and discriminative representations through multiple nonlinear transformations. These representations condense the multi-level features of the original data, such as patterns, rhythms, and rates of change, while compressing information and eliminating some redundant or noisy data. 3) Information filtered by the "forget gate" and "input gate": This part of the information represents the historical data features and background knowledge that are most useful for future predictions. The information here is no longer a raw value, but a signal that has been weighted, accumulated, and smoothed to help the model capture long-term dependencies. 4) Contextual information associated with the current prediction task: This part of the information can be seen as a "summary" of past data. It reflects the content in the historical process that is instructive for current situational decision-making, such as the implicit laws of the data, dynamic change trends, etc. In general, the "information" here includes not only historical data, but also representations of higher dimensions and predictive value for the task that have been formed through multiple layers of abstraction and refinement. This information is collaboratively managed, dynamically saved, and updated to support accurate predictions of future states.

[0117] The prediction process of the long short-term memory network can be simplified as formula (7). However, the long short-term memory network is based on a unidirectional time series for prediction, and it is difficult to fully analyze the complex relationship between input and output data and weather and environmental parameters with time variability. It has obvious limitations in capturing the dynamic changes of the front-to-back relationship in the time series, especially when used to predict the output of renewable energy units and infrastructure loads with significant uncertainty. The ability to characterize the fluctuation range is relatively limited. In comparison, the bidirectional long short-term memory network combines the forward and reverse long short-term memory networks, uses the past and future time series inputs, and enhances the research ability of the dynamic change characteristics of the forward and reverse sequence moments reflecting the fluctuation range by learning the front-to-back coupling characteristics of the time series. The prediction process of the bidirectional long short-term memory network can be expressed as formula (7). At the same time, in order to further enhance the prediction effect, a data preprocessing link based on the convolutional neural network is added, and the convolutional neural network is introduced to extract key features from historical data to provide high-quality input for the prediction. Solve the uncertain fluctuation interval process, such as Figure 2 shown.

[0118]

[0119]

[0120]

[0121] In formulas (6)–(8): and Respectively represent the inputs of the forget gate, input gate, and output gate during period t; Indicates updating the input gate during period t; and Represent the storage status of period t and period t-1 respectively; Represents historical data for period t; and represents the hidden state of the long short-term memory network unit in period t and period t-1 respectively; σ is the Sigmoid function, which is used to normalize the value to between 0 and 1; tanh represents the hyperbolic function, which converts the value to the range of -1 to 1; π fi ,π ii ,π oi ,π iu represents the weight; q fi ,q ii and q oi Represents different bias terms; and Represent the hidden states of the forward and reverse long short-term memory network units in time period t respectively; and are the hidden states of the forward and reverse long short-term memory network units in period t-1, respectively.

[0122] A method for defining uncertainty fluctuation intervals based on convolutional neural networks and bidirectional long short-term memory networks (LSTMs) was developed to obtain the uncertainty fluctuation interval range. The box boundaries were then dynamically adjusted to form a generalized box uncertainty set. This method effectively improved the accuracy of the uncertainty model and reduced the conservatism of scheduling optimization, providing support for robust optimization of the coordinated scheduling of port integrated energy systems and ferry logistics systems.

[0123] In step E, the port integrated energy system scheduling optimization model under multiple uncertainties in energy production includes the port integrated energy system scheduling objective function and scheduling constraints. The port integrated energy system scheduling objective function is to minimize the port integrated energy scheduling cost under multiple uncertainties in energy production. The port integrated energy scheduling cost mainly includes external energy procurement costs, equipment operation and maintenance costs, energy storage degradation costs, pollution treatment costs, port energy sales revenue, and demand response subsidies. Among them, the energy procurement cost is composed of the purchase costs of external electricity and natural gas. Taking into account the two-way interaction with the power grid, the port integrated energy system can also obtain revenue by selling surplus electricity to the external network, as expressed in formula (9). The equipment operation and maintenance cost is closely related to the equipment working status and output level, as expressed in formula (10). Energy storage equipment degrades faster with frequent charging and discharging, and the energy storage degradation cost increases with the increase of charging and discharging cycles, as expressed in formula (11). The port integrated energy system needs to pay pollution treatment fees for pollutants emitted during operation, as expressed in formula (12). The production capacity of the port integrated energy system is mainly sold to the port ferry logistics system and infrastructure operators to obtain energy sales revenue, as expressed in formula (13). In addition, since the port integrated energy system is connected to the power grid, as a demand-side adjustable resource, it can respond to the grid demand response incentive to adjust the electricity purchase demand during the specified peak-shaving and valley-filling periods to obtain demand response subsidies, which is expressed as formula (14). The formula is as follows:

[0124]

[0125]

[0126]

[0127]

[0128]

[0129]

[0130] In formulas (9)–(14): C buy 、C om 、C deg 、C ep 、C se and C drThey represent the external energy procurement cost, equipment operation and maintenance cost, energy storage degradation cost, pollution treatment cost, in-port energy sales income and demand response subsidy of a dispatch cycle, in yuan. and The prices of electricity purchased from and sold to the grid during period t, respectively, are RMB / MWh; Indicates the unit price of natural gas purchased from the gas grid, yuan / m 3 ; and They represent the power purchased and sold from the grid during period t, in MW; They represent the gas purchase rate from the gas network during period t, m 3 / h;c om,χ Indicates the unit operation and maintenance price of energy equipment χ, RMB / MWh or RMB / m 3 ; represents the uncertain output of renewable energy units, MW; Indicates the output power of energy equipment χ, MW or m 3 / h; Ω represents the energy equipment collection; c deg,χESS represents the energy storage device χ ESS Unit degradation cost, RMB / MWh or RMB / m 3 ; and Represents ESS equipment χ ESS Charging and discharging power, MW or m 3 / h;Ω ESS Represents a collection of energy storage devices; represents the emission tax rate or emission right price for pollutant τ, RMB / tCO2; ε g,τ and ε g,τ Indicates the grid emission factor and gas pollutant τ emission factor, t / MWh, t / m 3 ;Q τ The conversion factor representing the pollution equivalent value of τ is based on the potential harmfulness, environmental impact and ecotoxicity of the pollutant τ, which is quantified into CO2 equivalent; CO2, SO2 and NO x represent carbon dioxide, sulfur dioxide, and nitrogen oxide gases respectively; Indicates the unit price of energy sold within the port's integrated energy system, RMB / MWh or RMB / m 3 ; represents the energy consumption of the port ferry logistics system under multiple uncertainties, MW or m 3 / h; represents the energy consumption of port infrastructure under multiple uncertainties, MW; T p and T v They represent the peak shaving and valley filling demand time periods published by the power grid, h; and They represent the unit price of demand response subsidies corresponding to peak shaving and valley filling periods, in yuan / MWh; They represent the power purchased from external sources in the initial dispatch plan of the port integrated energy system in a deterministic scenario without considering demand response, MW; Δt E Indicates the unit duration of energy dispatch, h.

[0131] The scheduling constraints of the port integrated energy system include the balance of multi-energy supply and demand, the supply constraints of the external energy network, and the operation constraints of energy equipment. The power generation, hydrogen production, heat production, and cooling production of the port integrated energy system in each period must meet the power, hydrogen energy, heat energy, and cooling energy demand of the port ferry logistics system and infrastructure in each period, as shown in Equations (15)–(18). The supply constraints of the external energy network constitute an important constraint on the energy purchase and sales interaction between the port integrated energy system and the external energy network, including the power supply capacity of the external power grid and the power returned to the power grid are limited by the maximum transmission power of the power cable, and the gas supply capacity of the external natural gas network is limited by the maximum gas supply rate, as shown in Equations (19)–(20). The scheduling of the port integrated energy system needs to strictly limit the output, ramp rate, and charging and discharging capacity and storage capacity of energy storage equipment to ensure that the equipment operates within a safe range, as shown in Equations (21)–(23). At the same time, it is stipulated that the energy level of the energy storage equipment at the end of the scheduling cycle is consistent with that at the beginning of the scheduling to ensure the continuity and reliability of subsequent scheduling, as shown in Equation (24). The formula is as follows:

[0132]

[0133]

[0134]

[0135]

[0136]

[0137]

[0138]

[0139]

[0140]

[0141]

[0142] In formulas (15)–(24): and They represent the electricity, heat and cooling power generated by the port integrated energy system during period t, in MW; represents the hydrogen production rate of the port integrated energy system during period t, m 3 / h; represents the transmission power loss from the port integrated energy system to the ferry logistics system during period t, MW; represents the hydrogen loss from the port integrated energy system to the ferry logistics system during period t, m 3 / h; and They represent the losses in the process of transmitting heat and cold from the port integrated energy system to the logistics system during period t, MW; and They represent the actual electricity, heating and cooling loads of the port infrastructure during period t, in MW; and represents the actual electricity and hydrogen demand of the port ferry logistics system during period t, in MW or m 3 / h; and Respectively represent the maximum power supply power of the power grid and the maximum acceptable power return power, MW; Indicates the electricity purchase status, 1 if electricity is purchased, otherwise 0; Indicates the maximum gas supply rate of the natural gas network, m 3 / h; and Respectively represent the minimum and maximum output of energy equipment χ, MW; and Represents gas equipment χ g Minimum and maximum ramp rates, MW / h; Ω g Represents a collection of gas equipment; and Represent the energy storage device χ ESS Maximum charge and discharge power, MW or m 3 / h; represents the energy storage device χ ESS The charging status of , if charged, it is 1, otherwise it is 0; and Represent the energy storage device χ ESS Upper and lower limits of storable energy, MWh or m 3 ; and They represent the energy storage equipment χ at the end and beginning of scheduling respectively. ESS Energy stored, MWh or m 3 ; and Represents ESS equipment χ ESS The charging and discharging power; and They represent the power purchased and sold from the power grid during period t respectively; They represent the gas purchase rate from the gas network during period t; Indicates the maximum gas supply rate of the gas network; represents the output power of energy equipment χ during period t; and They represent the gas equipment χ in period t and period t-1 respectively. g Output power; represents the energy storage device χ during period t ESS Store energy.

[0143] In step F, the port ferry logistics system scheduling optimization model under multiple uncertainties of logistics operations includes the port ferry logistics system scheduling objective function and scheduling constraints.

[0144] The objective function of the port ferry logistics system scheduling is to minimize the port ferry logistics scheduling cost under multiple uncertainties of logistics operations. The port ferry logistics scheduling cost includes the joint operation cost of logistics facilities, logistics energy cost and delay compensation cost. Among them, the joint operation cost of logistics facilities covers the occupation, startup and operation costs of various logistics facilities in the port ferry logistics system, including shore power and shuttle buses, as shown in Equations (25)-(28). The logistics energy cost is the cost of the port ferry logistics system purchasing energy from the port integrated energy system, as shown in Equation (29). The delay compensation cost is used to quantify the impact of the uncertainty of arrival time and passenger disembarkation time on logistics scheduling, mainly including: first, the waiting cost caused by shore power waiting for the delayed ferry; second, the waiting operation cost caused by the delay of shuttle bus departure due to the failure of passengers to disembark as planned; third, the delay departure compensation required to be paid due to the delayed departure of the ferry due to improper logistics scheduling. However, if the actual departure time of the ferry is earlier than the planned time, no compensation is required. In addition, the uncertainty of the battery state of charge when the ferry is berthing is reflected in the changes in shore power consumption and the changes in the ferry's departure time. Therefore, the delay compensation cost is expressed as formula (30), which is as follows:

[0145] C fo =C CI +C ESB +C HSB ; (25)

[0146]

[0147]

[0148]

[0149]

[0150]

[0151] In formulas (25)–(30): C fo represents the joint operation cost of logistics facilities in a scheduling cycle, RMB; C CI 、C ESB and C HSB They represent the operating costs of all shore power, electric shuttles, and hydrogen shuttles in a scheduling cycle, in yuan; c CI,on 、c ESB,on and c HSB,on represents the basic operating cost of a single shore power, electric shuttle bus and hydrogen shuttle bus, RMB / unit; c ESB,w and c HSB,w Represent the unit cost of ferrying operation of electric shuttle and hydrogen shuttle respectively, RMB / h; The 0-1 state variable represents that the j-th shore power station serves the i-th ferry in period t, which is 1 if the ferry is in service and 0 otherwise; and They represent the 0-1 state variables of the mth electric shuttle and the qth hydrogen shuttle serving the i-th ferry during period t, where ferrying is 1 and outage is 0; Δt L Indicates the unit duration of logistics scheduling, h; C ec and C dp They represent the logistics energy cost and delay compensation cost of a scheduling cycle, RMB; c da 、c ds and c dd represents the unit penalty cost of ferry delayed arrival, delayed departure and delayed departure, respectively, in yuan / h; and They represent the actual departure time and the latest scheduled departure time of the i-th ferry, h.

[0152] The constraints for port ferry logistics scheduling include ferry berthing constraints, ferry logistics facility allocation constraints, and facility operation constraints. The continuous berth allocation method, which is widely used in ports, is adopted to meet the first-come, first-served principle. Ferry berthing is subject to the order of arrival time of each ferry and the ferry length constraints. That is, the berthing position of each ferry must be within the total length of the shoreline, and the berthing positions of any two ferries cannot overlap, see formula (31). The ferry berthing time is later than its arrival time to ensure the feasibility of the scheduling plan, see formula (32). The berthing position and berthing time of any two ferries in the same time period must not conflict, see formula (33). Ferry logistics facility allocation constraints include independent allocation of a single facility and joint allocation constraints between facilities. Specifically, each logistics facility can only serve one ferry in each time period, see formula (34); the number of dispatched facilities in each time period cannot exceed the total number of facilities, see formula (35); to ensure the rationality and orderliness of shore power allocation, the number of shore power supplies serving each ferry in each time period shall not exceed one, and the shore power allocation shall follow the order of ferry arrival, and the allocation process shall not overlap, see formula (36); the number of shuttle buses allocated to each ferry is limited, see formula (37); the number of passengers carried by electric shuttle buses and hydrogen shuttle buses shall be allocated according to the number of ferry passengers they serve, see formula (38). The operation of ferry logistics facilities must meet constraints such as work efficiency, operating power, and battery capacity, including: the passenger capacity of each electric ferry and hydrogen ferry must meet upper and lower limit constraints, see formula (39); the charging power and charge state of each electric ferry in each period are subject to upper and lower limit constraints, and to facilitate the scheduling of the next cycle, the charge state value at the end of the scheduling is required to be the same as that at the beginning of the scheduling, see formula (40); the hydrogenation rate and charge state of each hydrogen ferry are subject to upper and lower limit constraints, and the charge state at the end of the scheduling is also required to be the same as that at the beginning of the scheduling, see formula (41); the charging power of each ferry (i.e., the power discharged from the shore power) is subject to the upper and lower limit constraints of the shore power supply power, and it is also necessary to ensure that the ferry is fully charged before leaving the port, see formula (42), as follows:

[0153]

[0154]

[0155]

[0156]

[0157]

[0158]

[0159]

[0160]

[0161]

[0162]

[0163]

[0164]

[0165] In formulas (31)–(42): l i and l s denote the starting berth positions of ferry i and ferry s, m; ι i represents the length of ferry i (including the safe berthing distance between ferries), m; l max Indicates the length of the dock shoreline, m; represents the berthing time of ferry s, h; x is represents the berthing position of any ferry i and ferry s, which is a 0-1 state variable. If ferry i is berthed at the left shoreline of ferry s, it is 1, otherwise it is 0; φ is represents the berthing time of any ferry i and s. It is a 0-1 state variable. If ferry i berths before ferry s, the value is 1, otherwise it is 0. M represents an infinite constant. The 0-1 state variable represents whether the zth facility ω serves the i-th ferry in time period t, which is 1 if it serves, and 0 otherwise; N is a 0-1 state variable representing the berthing work of the i-th ferry at berth b during period t. If the work is in progress, it is 1, otherwise it is 0; CI 、N ESB and N HSB They represent the total number of shore power, electric shuttle buses and hydrogen shuttle buses in the port ferry logistics system, respectively; and They represent the maximum and minimum number of electric shuttle buses and hydrogen shuttle buses that can be allocated to the i-th ferry, respectively; and They represent the actual passenger capacity and rated passenger capacity of the mth electric shuttle bus in time period t, respectively; and They represent the passenger capacity and rated passenger capacity of the qth hydrogen shuttle bus in period t, respectively; and They represent the power consumption of the mth electric shuttle and the jth shore power supply in period t, in MW; represents the hydrogen refueling rate of the qth hydrogen shuttle vehicle in period t, m 3 / h; and Respectively represent the upper and lower limits of the charging power of the mth electric shuttle in each time period, MW; and They represent the upper and lower limits of the hydrogen refueling rate of the qth hydrogen shuttle bus in each period, m 3 / h; and They represent the charge state of the mth electric shuttle and the hydrogen charge state of the qth hydrogen shuttle at the end of period t respectively; represents the time when the i-th ferry ends charging, h; represents the state of charge of the i-th ferry at the end of charging; and They represent the upper and lower limits of the state of charge of the mth electric shuttle bus and the state of charge at the beginning and end of the dispatch; and They represent the upper and lower limits of the hydrogen charge state of the qth hydrogen shuttle vehicle, as well as the hydrogen charge state values ​​at the beginning and end of scheduling; and They represent the upper and lower limits of the power consumption of the j-th shore power station in each time period, in MW; and They represent the upper and lower limits of the state of charge of the i-th ferry when charging is completed; represents the time when the i-th ferry begins to berth, h; Ψ PL Represents a collection of equipment types for the port ferry logistics system; Indicates that the shore power supply of the j'th station is serving the sth ferry in period t. It is a 0-1 state variable, which is 1 if the ferry is serving and 0 otherwise. represents the number of passengers carried by ferry i.

[0166] In step G, the scheduling constraint optimization method based on interval modeling is used to reduce the dimensionality of a large number of three-dimensional integer variables involved in the port ferry logistics system scheduling optimization model, so as to optimize the expression of the scheduling constraint model and thereby improve the solution efficiency of the port ferry logistics system scheduling optimization model. The principle of interval modeling is as follows Figure 3 As shown. Assume that α and β are different processing stages, start(α), end(α) and size(α) represent the start time, end time and duration of job α respectively. For any α, start(α)+size(α)=end(α); presence(α) and presence(β) represent whether α and β are performed in time interval t, and take 1 if they are performed, otherwise take 0. Based on this, the time variable is introduced and To define the start time, processing time, and end time of each ferry's berthing, charging, and ferrying, see Equation (43). The various logistics stages are interrelated, but the logistics operations for different ferries are independent of each other, as shown in Equation (44), as follows:

[0167]

[0168] In formulas (43) and (44): and start(ω i ) denotes the starting time of the service phase of the i-th ferry by facility ω, h; and size(ω i ) denotes the duration of the period in which the i-th ferry is served by facility ω, h; and end(ω i ) denotes the end time of the period in which the i-th ferry is served by facility ω; and denote the time when the i-th ferry starts electric ferrying and hydrogen ferrying, h; represents the time when the i-th ferry ends its berthing, h.

[0169] Furthermore, the time variable can be represented by a two-dimensional binary variable and Definition, which respectively indicate whether the stage in which the i-th ferry is served by the facility ω starts, is in progress, and ends the operation.

[0170] Therefore, the original constraint conditions (34)–(38) containing three-dimensional binary variables can be optimized into (45)–(49). Equation (45) indicates that each logistics stage of the i-th ferry only starts and ends once, and the start and end times correspond to and The time when a non-zero value appears. Then the operation time of the logistics phase of the i-th ferry is The sum is expressed as formula (46). Formula (47) ensures that the operation of the i-th ferry is not interrupted, so, The starting state of the phase can be and end state Determine. Formula (48) constrains that the berths allocated to the ferry cannot exceed the length of the shoreline, and the order of ferry berthing depends on the order of arrival. Formula (49) constrains that the charging status of the i-th ferry in period t is equal to the sum of the shore power operation status serving it, and the berthed ferry can be in charging or non-charging state. In addition, the number of electric ferries and hydrogen ferries serving the ferry is respectively subject to the time limit constraint of the ferry phase and the number of electric ferries and hydrogen ferries that can be allocated. This constraint also reflects the difference between the introduced variables and the original three-dimensional binary variables. The relationship between them. As the model is simplified, the optimal solution can be obtained through the commercial solver Gurobi. The sub-steps to obtain the optimal solution include: model preprocessing (establishing a mathematical model and simplifying it, removing redundant constraints, bounding variables, scaling coefficients, and merging some variables to reduce model complexity), relaxation solution and linearization (relaxing integer constraints, and using mathematical transformations in linearization to simplify calculations), heuristic solution (using heuristic algorithms such as greedy methods, local search, and simulated annealing to quickly obtain an initial feasible solution and provide an upper bound on the optimization target value for pruning and reducing the search space), branch and bound (traversing the search tree through node splitting and pruning, pruning nodes that cannot improve the global optimal solution, reducing the size of the search tree, and improving the efficiency of global search), branch and cut (dynamically adding cutting planes to enhance model compactness, thereby improving pruning and solution speed), iteration and convergence judgment (continuously updating the bounds to narrow the gap between the upper and lower bounds. When the difference between the upper and lower bounds is within the preset error range or the maximum number of iterations is reached, the current optimal solution is judged to meet the convergence conditions), output and post-processing (outputting the optimal solution and verifying and performing sensitivity analysis on it). The formula is as follows:

[0171]

[0172]

[0173]

[0174]

[0175]

[0176] In formulas (45)–(49): Indicates whether the i-th ferry starts the facility ω service phase in time period t, 1 if it starts, 0 otherwise; Indicates whether the i-th ferry has completed the facility ω service phase in period t, if completed, it is 1, otherwise 0; represents the set of positive integers; and presence(ω i ) indicate whether the i-th ferry accepts the logistics facility ω service in period t, 1 if it accepts, 0 otherwise; and Indicates whether the i-th ferry accepts the logistics facility ω service during period t1+1, period t1, period t3, period t2-1 and period t2 respectively, if accepted, it is 1, otherwise 0; and Indicates whether the i-th ferry is berthing during period t2, period t2-1 and period t, respectively. If berthing, it is 1, otherwise it is 0; Indicates whether the sth ferry is berthing during time period t, if yes, it is 1, otherwise it is 0; Indicates whether the i-th ferry accepts the logistics facility ω service at time t', if yes, it is 1, otherwise 0; The 0-1 state variable represents whether the zth facility ω serves the i-th ferry at time t', which is 1 if it serves, and 0 otherwise; and They respectively indicate whether the i-th ferry is in the charging, berthing, electric ferrying and hydrogen ferrying stages during period t, which is 1 if yes and 0 otherwise.

[0177] In step H, the dynamic coupling relationship between multiple uncertainties across domains is reflected in the dynamic balance between the energy consumption of the port ferry logistics system under multiple uncertainties in logistics operations and the energy supply of the port comprehensive energy system under multiple uncertainties in energy production.

[0178] The energy demand of the port ferry logistics system is mainly composed of the ferry charging power and the energy consumption of the ferry bus. In order to cope with the uncertainty of logistics operations, the ferry logistics system scheduling plan usually needs to be dynamically adjusted. This adjustment will cause a deviation between the energy demand of the actual scheduling plan and the energy demand of the original scheduling plan, as shown in Equation (50). After the ferry docks, it is connected to the shore power for charging. The charging power can be expressed by the shore power consumption and the ferry charge state, as shown in Equation (51). The energy consumption of the electric ferry is the charging power when it is in the off-service charging state, and the charging process of the electric ferry can be described by the charge state at the end of the t-1 period and the change in the charge state during the t period. The change reflects the change in the charge of the electric ferry during the off-service charging or ferry operation, which is closely related to key factors such as the ferry duration, passenger capacity, and driving distance, as shown in Equation (52). Similar to electric shuttles, the hydrogen consumption rate of hydrogen shuttles is the hydrogen charging rate when they are in the off-service hydrogen charging state. The hydrogen refueling process of hydrogen shuttles can be described by the hydrogen charging state at the end of period t-1 and the change in hydrogen charging state during period t. The change reflects the change in hydrogen content of the hydrogen shuttle during off-service hydrogen charging or transportation operation. It can be estimated by combining key factors such as passenger capacity, travel distance, and operating time to accurately reflect the hydrogen energy consumption characteristics of the vehicle in various operating states, as shown in formula (53).

[0179]

[0180]

[0181]

[0182]

[0183] In formulas (50)–(53): and They represent the actual energy consumption of the jth shore power station, the mth electric shuttle bus, and the qth hydrogen shuttle bus in period t under multiple uncertainties of logistics operations, in MW or m 3 / h; represents the charging efficiency of the ferry; represents the rated capacity of the battery of the i-th ferry, MWh; and They represent the charging and discharging efficiency of the mth electric shuttle bus respectively; and They represent the no-load transport power and rated operating power of the mth electric shuttle bus, MW; represents the rated capacity of the battery of the mth electric shuttle, MWh; and They represent the charging and discharging efficiency of the qth hydrogen shuttle bus respectively; and They represent the no-load hydrogen consumption rate and rated working hydrogen consumption rate of the qth hydrogen shuttle bus, m 3 / h; represents the rated capacity of the fuel cell of the qth hydrogen shuttle bus, m 3 . and represents the actual electricity and hydrogen demand of the port ferry logistics system during period t, and They represent the charge state of the mth electric shuttle and the hydrogen charge state of the qth hydrogen shuttle at the end of period t, and represents the state of charge of the ferry i battery at time t and time t-1, respectively, Δt L Indicates the unit duration of logistics scheduling.

[0184] The actual energy consumption of the port ferry logistics system under multiple uncertainties in logistics operations needs to be met through the dynamic scheduling of energy production, conversion, and storage equipment in the port integrated energy system, while taking into account the multiple uncertainties in energy production. Therefore, the dynamic coupling relationship between cross-domain multiple uncertainties is expressed as a dynamic balance between the energy consumption of the port ferry logistics system under multiple uncertainties in logistics operations and the energy supply of the port integrated energy system under multiple uncertainties in energy production, as shown in Equation (54), as follows:

[0185]

[0186] In formula (54), Λ and Υ represent the energy conversion and storage matrices, respectively, which are determined by the configuration of energy equipment in the port integrated energy system; and Respectively represent the charging and discharging power of the battery, MW; and Respectively represent the hydrogen storage and release rates of the hydrogen storage tank, m 3 / h; and Respectively represent the storage and release power of the heat storage tank, MW.

[0187] In step I, a robust optimization model for the coordinated scheduling of the port integrated energy system and the ferry logistics system is constructed by introducing a dual incentive mechanism based on the dynamic coupling relationship between multiple uncertainties across domains.

[0188] The port integrated energy system and the ferry logistics system are usually independently scheduled with the goal of maximizing their own interests. It is difficult for the port ferry logistics system to adjust the load demand unconditionally. Supporting the optimization of the port integrated energy system scheduling is necessary to cope with the multiple uncertainties of energy production. To this end, a dual incentive mechanism of "load adjustment incentive-grid demand response incentive" is introduced. The port integrated energy system pays economic subsidies to the ferry logistics system for load adjustment behavior, incentivizing it to support the optimization of the port integrated energy system scheduling, and realizing the effective coordination of the port integrated energy system and the ferry logistics system under cross-domain multi-charge uncertainty. After the dual incentive mechanism is introduced, the port integrated energy system needs to pay the ferry logistics system a load adjustment fee, which is expressed as formula (55):

[0189]

[0190] In formula (55): C la represents the load adjustment cost of a scheduling cycle, yuan; and Respectively represent the load adjustment incentive unit price during peak shaving and valley filling demand periods, RMB / MWh; represents the electricity demand of the initial scheduling plan of the port ferry logistics system under the deterministic scenario, MW. Therefore, the coordinated scheduling objective functions of the port integrated energy system and the ferry logistics system are expressed as Equations (56) and (57), respectively. Furthermore, the robust optimization model for the coordinated scheduling of the port integrated energy system and the ferry logistics system under cross-domain multiple uncertainties can be expressed in a compact form as Equation (58), as follows:

[0191]

[0192]

[0193]

[0194] In formula (58): c τ 、b τ d τ 、r τ 、h τ 、D τ 、R τ 、Fτ 、H τ 、L τ and Y τ U represents the parameter matrix of the port integrated energy system or ferry logistics system scheduling; τ Indicates uncertain scheduling scenario; x τ 、y τ and u τ They represent the scheduling variable matrix composed of binary variables, continuous variables and uncertain variables respectively; Ω represents the scheduling variable matrix composed of x τ and u τ The uncertain scenarios are: PE and PL represent the port integrated energy system and port ferry logistics system respectively. buy 、C om 、C deg 、C ep 、C se and C dr They represent the external energy procurement cost, equipment operation and maintenance cost, energy storage degradation cost, pollution treatment cost, in-port energy sales income and demand response subsidy of a scheduling cycle respectively; C fo represents the joint operation cost of logistics facilities in a scheduling cycle; C ec and C dp They represent the logistics energy cost and delay compensation cost of a scheduling cycle respectively; C PE and C PL They represent the scheduling cost of the port integrated energy system and the port ferry logistics system in one scheduling cycle respectively.

[0195] The scheduling variable matrix and the uncertain scenario are specifically expressed as Equations (59)–(61):

[0196]

[0197]

[0198]

[0199] In formulas (59)–(61): x PE 、y PE and u PE They represent the decision variable matrices composed of integer variables, continuous variables, and uncertain variables in the robust optimization problem of port integrated energy system scheduling; x PL 、y PL and u PL They represent the decision variable matrices composed of integer variables, continuous variables and uncertain variables in the robust optimization problem of port ferry logistics system scheduling; and Respectively represent the time periods and each time period The matrix of integer variables formed; and Respectively represent each period Each time period Each time period and each time period The integer variable matrix composed of n ESB,m 、n HSB,q 、x i 、f i and l represent the time periods Each time period Each ferry x is , each ferryφ is and various ferries i The matrix of integer variables formed; and Respectively represent each period Each time period Each time period Each time period and each time period The continuous variable matrix formed by t ber,beg , t ω,beg , t ω,end , t dep 、 ζ ESB,m ,ζ HSB,q 、 and Respectively, each ferry Ferries Ferries Ferries Each time period Each time period Each time period Each time period Each time period Each time period Each time period and each time period The continuous variable matrix formed by and Respectively represent each period

[0200] and each time period The uncertain variable matrix formed; and Respectively, each ferry Ferries and ferries The uncertain variable matrix.

[0201] In step J, a distributed collaborative optimization solution framework combining the column and constraint generation algorithm and the adaptive alternating direction multiplier method is used. Based on the characteristics of the robust optimization model for collaborative scheduling of the port integrated energy system and the ferry logistics system, the column and constraint generation algorithm is used to generate uncertain scenarios. The adaptive alternating direction multiplier method is used to ensure the decision independence of each subsystem while realizing distributed collaborative scheduling optimization under limited information interaction.

[0202] The column and constraint generation algorithm has significant flexibility advantages in dealing with large-scale variables and complex constraints in robust optimization problems, and can improve the quality of the solution while ensuring the calculation speed. Based on the column and constraint generation algorithm, the robust optimization model for the coordinated scheduling of the port integrated energy system and the ferry logistics system is reconstructed into a main problem and sub-problems, expressed as Equations (62) and (63). In the first stage, the main problem is solved to obtain the scheduling objectives θ of the port integrated energy system and the ferry logistics system. τ and the matrix x of integer variables τ In the second stage, the KKT condition (Carlo-Kuhn-Tucker condition, which is a necessary and sufficient condition for a nonlinear programming problem to have an optimal solution under certain regular conditions) is used to replace the max-minc τ T y τ Transformed into a maximum value problem and with maxb τ T u τ Combined, and then constitute a sub-problem, by inputting the integer variable matrix x obtained in the first stage τ Calculate the uncertain variable matrix u τ and the continuous variable matrix y τ Then, the lower and upper bounds of the objective function optimization result are updated according to formula (64). Iterate until the residuals of the lower and upper bounds meet the convergence threshold Δ C&CG , see formula (65), and we can get the robust scheduling scheme of the port integrated energy system and ferry logistics system under multiple uncertainties, which is as follows:

[0203]

[0204]

[0205] LB τ =max{LB τ ,θ τ},UB τ =min{UB τ ,f(u τ ,y τ )}; (64)

[0206] |UB τ-LB τ |≤Δ C&CG ; (65)

[0207] In equations (62)–(65): θ τ represents the objective function value of the main problem; κ τ1 , κ τ2 , κ τ3 and κ τ4 Indicates the relationship between y τ Constrained Lagrange multiplier, y τ and κ τ1 The product term can be linearized by introducing binary variables. In addition, the uncertain variable matrix u τ Can be further used Δu τ and B τ Expressed as, where Δu τ Represents the uncertain boundary set, i.e. u τ In the forecast or planned value The maximum fluctuation near B τ is a binary variable matrix indicating whether the fluctuation occurs; [·] represents the length of the matrix; Γ τ It is introduced to limit the occurrence rate of uncertainty reaching the maximum fluctuation, that is, the maximum fluctuation B τ The total number of occurrences and B τ Ratio of matrix lengths; LB τ and UB τ Respectively represent the lower and upper bounds of the optimization results.

[0208] Based on the multiple uncertainty scenarios of energy production and logistics operations generated by the column and constraint generation algorithm, the uncertainty variable matrix obtained by the column and constraint generation algorithm is used as a parameter. The adaptive alternating direction multiplier method is used to perform distributed collaborative scheduling optimization based on the dynamic balance consensus of energy supply and demand between the port integrated energy system and the ferry logistics system, see formula (66). By introducing dynamic adaptive parameters, this method effectively overcomes the residual oscillation problem that is common in the traditional alternating direction multiplier method when solving complex large-scale problems, significantly improves the convergence speed and stability, and ensures the efficient solution of the collaborative scheduling optimization problem. The distributed collaborative scheduling optimization model is shown in formula (67). In each iteration, the port integrated energy system is based on the energy consumption of the ferry logistics system obtained by the previous iteration. Optimize self-scheduling and update energy supply plans And send it to the port ferry logistics system; the port ferry logistics system optimizes its own scheduling according to the latest energy supply situation and updates the logistics energy consumption And the demand is transmitted back to the port integrated energy system. Subsequently, the penalty operator matrix and the Lagrange multiplier matrix are updated according to the original residual and the dual residual, as shown in Equation (68). The iteration stops when the convergence condition is met, as shown in Equation (69), and the final coordinated robust scheduling scheme of the port integrated energy system and the ferry logistics system can be obtained. This method alleviates the oscillation caused by the sharp fluctuation of the original residual and the dual residual by introducing the residual norm smoothing parameter, effectively reduces the convergence fluctuation, and adopts parameter adaptive adjustment to accelerate the iterative convergence. The specific formula is as follows:

[0209]

[0210]

[0211]

[0212]

[0213] In formulas (66)–(69): A T y PE B represents the energy variable matrix of the port integrated energy system supplying the ferry logistics system; T y PL represents the energy demand matrix of the port ferry logistics system; and Respectively represent The residuals of energy supply and demand variables (i.e., original residuals) and dual residuals of the iteration; and denote the smoothing parameters of the original residual norm and the dual residual norm respectively; and Respectively represent The sub-penalty operator matrix and the Lagrange multiplier matrix, and Respectively represent The sub-penalty operator matrix and the Lagrange multiplier matrix, u τ represents the scheduling variable matrix composed of uncertain variables; ε AADMM,1 and ε AADMM,2 represent the iterative convergence thresholds of the primal residual norm and the dual residual norm respectively. and Respectively represent The smoothed values ​​of residual norm and dual residual norm of energy supply and demand variables of the iteration; Indicates the The smoothed value of the residual norm of the energy supply and demand variables of the iteration, Indicates the The energy variables of the port integrated energy system supplying the ferry logistics system are obtained by the iteration. and Respectively represent Second and The energy demand variables of the ferry logistics system obtained by the iteration are: ε1 and ε2 represent the parameters of the penalty operator matrix adjustment, ρ and Represent the lower limit and upper limit matrices of the penalty operator respectively; the iteration stops when the convergence condition is met, and the final coordinated robust scheduling scheme of the port integrated energy system and the ferry logistics system is obtained. and Respectively represent The energy scheduling continuous variables and logistics scheduling continuous variables of the iteration, c PE T y PE and b PE T u PE Respectively represent the functions related to continuous variables and uncertain variables in the energy scheduling sub-problem, T represents the transpose operation, c PL T y PL and b PL T u PL Respectively represent the functions related to continuous variables and uncertain variables in the logistics scheduling sub-problem, Indicates the use of The energy demand variables of the ferry logistics system obtained by the iteration are used to solve the penalty term when the port integrated energy system supplies the energy variables of the ferry logistics system. Indicates the use of The penalty term when solving the energy demand variable of the ferry logistics system by using the energy variable of the port integrated energy system supplied by the iteration is used. and Respectively represent The residual norm and dual residual norm of energy supply and demand variables at the iteration.

[0214] like Figure 4 As shown, in step K, the models in steps B, C, E, F, H, and I are solved to obtain a robust scheduling solution for the coordinated port integrated energy system and ferry logistics system. The complete solution steps are as follows:

[0215] 1) Input the parameters and data of the port’s integrated energy system equipment and ferry logistics system facilities.

[0216] 2) Establish a multiple uncertainty representation model for port energy production and logistics operations, and determine the uncertainty set boundary based on the uncertainty fluctuation interval definition method of convolutional neural network and bidirectional long short-term memory network.

[0217] 3) Calculate the scheduling optimization model of the port integrated energy system and ferry logistics system under the deterministic scenario as the initial scheduling plan of the port integrated energy system and the initial scheduling plan of the ferry logistics system.

[0218] 4) Establish the main problem model and sub-problem model of the two-stage robust optimization of the coordinated scheduling of the port integrated energy system and the ferry logistics system based on the column and constraint generation algorithm.

[0219] 5) Initialize the two-stage robust optimization variables, Lagrange multipliers, and iterative convergence domain values ​​for the coordinated scheduling of the port integrated energy system and the ferry logistics system.

[0220] 6) Based on the main problem model of two-stage robust optimization of coordinated scheduling of port integrated energy system and ferry logistics system, calculate θ τ and the integer variable matrix x τ , and input it into the sub-problem model to calculate the uncertain variable matrix u τ and the continuous variable matrix y τ .

[0221] 7) Update the lower bound LB τ and upper bound UB τ If the residual meets the convergence condition, the iteration stops, and the output result is the robust scheduling plan of the port integrated energy system and ferry logistics system under multiple uncertainties, and then goes to step 8); otherwise, returns to step 6).

[0222] 8) will u τ As known quantities, a distributed collaborative scheduling optimization model of the port integrated energy system and ferry logistics system based on the adaptive alternating direction multiplier method is established.

[0223] 9) Initialize the distributed collaborative scheduling optimization variables of the port integrated energy system and ferry logistics system, energy supply and demand coupling variables, Lagrange multipliers, penalty operator matrix, and iterative convergence domain value.

[0224] 10) Based on the distributed collaborative scheduling optimization model of the port integrated energy system and the ferry logistics system, calculate the energy supply A of the port integrated energy system T y ESO Based on this, the ferry logistics system scheduling is optimized and the logistics energy consumption B is obtained. T y LSO .

[0225] 11) Calculate the original residual Δ1 and the dual residual Δ2, update the penalty operator matrix ρ and the Lagrange multiplier matrix λ, and if the residual meets the convergence condition, the iteration stops and the output result is the final coordinated robust scheduling plan for the port integrated energy system and the ferry logistics system; otherwise, return to step 10).

[0226] The following provides a specific embodiment to further illustrate the implementation effect of the collaborative scheduling method of the port integrated energy system and the ferry logistics system under cross-domain multiple uncertainties provided by the present invention.

[0227] For example, a ferry port has 5 shore power units, 10 electric shuttles, and 10 hydrogen shuttles. The electricity prices in this area are based on peak-flat-valley time-of-use electricity prices, which are 1.35 yuan / kWh, 0.90 yuan / kWh, and 0.48 yuan / kWh, respectively. The natural gas prices are 2.70 yuan / m 3 . The port integrated energy system sells electricity, heat and cooling energy to the terminal at prices of 0.6 yuan / kWh, 0.25 yuan / kWh and 0.28 yuan / kWh respectively. The demand response incentive and load adjustment incentive prices during peak and valley load periods are 4.0 yuan / kWh, 1.2 yuan / kWh and 2.0 yuan / kWh, 0.6 yuan / kWh respectively. The ferry port operates from 6:00 to 24:00 every day. Based on the data of planned ferries arriving at the port on a certain day and the data related to electricity, heat and cooling loads of the infrastructure, the method provided by the present invention is used to solve the coordinated robust scheduling plan of the port integrated energy system and the ferry logistics system on that day, and compare it with the comparative example based on the prior art.

[0228] 1) Comparison of the results of the definition of uncertain fluctuation range

[0229] Comparison of the definition of uncertain fluctuation range between the prior art and the method of the present invention, such as Figure 5 – Figure 9 As shown, the uncertainty variable fluctuation range obtained by the method provided by the present invention significantly alleviates the problem of over-conservatism in the existing technology. Since it is closer to the actual fluctuation situation, it effectively improves the robustness of the system and thus improves the economy of scheduling optimization.

[0230] 2) Comparison of scheduling costs between the port integrated energy system and the ferry logistics system

[0231] Table 1 Comparison of the dispatching costs of the port integrated energy system and the port ferry logistics system of the prior art and the method of the present invention

[0232]

[0233] Comparing the scheduling cost results of the existing technology and the method of the present invention, the method provided by the present invention reduces the scheduling costs of the port integrated energy system and the ferry logistics system by 15.60% and 3.16% respectively. For the port integrated energy system, although the existing technology requires a load adjustment subsidy for the ferry logistics system, the demand response incentive income increases significantly, the external energy purchase cost and the energy equipment operation and maintenance cost decrease, and the income from energy sales in the port increases, which reduces the scheduling cost of the port integrated energy system. For the port ferry logistics system, although the logistics energy cost increases, the ferry logistics system scheduling cost is reduced due to the load adjustment subsidy and the fact that the logistics facility operation cost and delay compensation cost do not change significantly. It can be seen that the method of the present invention can improve the economic efficiency of the operation of the port integrated energy system and the ferry logistics system under cross-domain multiple uncertainties.

[0234] 3) Comparison of scheduling schemes for port integrated energy systems and ferry logistics systems

[0235] The comparison of the port integrated energy system and ferry logistics system scheduling schemes of the existing technology and the method of the present invention is as follows: Figure 10 – Figure 17 and Figure 18 – Figure 20 As shown. Based on the scheduling scheme of the port integrated energy system and ferry logistics system solved by the method of the present invention, during the period of high electricity prices, the utilization rate of hydrogen shuttle buses is increased, and the charging load of part of the ferries and electric shuttle buses is transferred to the valley period. The energy stored in the hydrogen storage tank is released to meet the demand for hydrogen in logistics, thereby reducing the purchase of electricity from the external power grid to respond to the grid demand response incentive. During the valley electricity price period, the port integrated energy system increases the power purchased from the external power grid to meet the growing demand for electricity in logistics. The heat and cold demands of the port infrastructure load are mainly met by increasing the output of electric boilers and electric refrigeration units. As a result, the power consumption of electric boilers and electric refrigeration units increases, the output of the combined heat, cold and power units decreases, and the excess power is stored in batteries or used for hydrogen production in electrolyzers and stored in hydrogen storage tanks. It can be seen that collaborative scheduling significantly improves the efficiency of the coordinated utilization of energy and logistics resources and the robustness of the operation of the port integrated energy system and ferry logistics system by flexibly adjusting the energy production and logistics operation plans, and effectively responds to the impact of multiple cross-domain uncertainties.

[0236] From the above analysis, it can be seen that compared with the existing technology, the method provided by the present invention can effectively improve the economy and robustness of the operation of the port integrated energy system and the ferry logistics system, reducing the scheduling costs of the port integrated energy system and the ferry logistics system by 15.60% and 3.16% respectively, and increasing the grid demand response incentive income by about 12.40%.

[0237] The above embodiments are only for illustrating the technical ideas of the present invention and cannot be used to limit the scope of protection of the present invention. Any changes made on the basis of the technical solutions in accordance with the technical ideas proposed by the present invention fall within the scope of protection of the present invention; any technologies not involved in the present invention can be implemented by existing technologies.

Claims

1. A collaborative scheduling method for port integrated energy systems and ferry logistics systems under cross-domain and multiple uncertainties, characterized by: include: Construct a multiple uncertainty representation model for port energy production and logistics operations; optimize the accuracy of the multiple uncertainty representation model for port energy production and logistics operations using an uncertainty fluctuation interval definition method based on convolutional neural networks and bidirectional long short-term memory networks; With the goal of economic optimization, determine the scheduling optimization model under multiple uncertainty scenarios; Scheduling constraint optimization is performed based on interval modeling; based on the dynamic coupling correlation between multiple uncertainties across multiple domains, a dynamic balance is achieved between the energy consumption of the port ferry logistics system under multiple uncertainties in logistics operations and the energy supply of the port integrated energy system under multiple uncertainties in energy production; A robust optimization model for the coordinated scheduling of the port's integrated energy system and the ferry logistics system is established; through the construction of a distributed collaborative optimization solution framework that combines the column and constraint generation algorithm and the adaptive alternating direction multiplier method, a robust scheduling solution for the coordinated scheduling of the port's integrated energy system and the ferry logistics system is obtained.

2. The collaborative scheduling method of a port integrated energy system and a ferry logistics system under cross-domain multiple uncertainties according to claim 1 is characterized in that: The construction of the multiple uncertainty characterization model for port energy production and logistics operations includes: Based on the box uncertainty set and considering the volatility, the uncertain output characteristics of renewable energy units and the uncertain load characteristics of infrastructure, a multiple uncertainty representation model for port energy production is constructed. The formula of the multiple uncertainty representation model for port energy production is: Where, and They represent the renewable energy equipment χ in period t respectively RES The actual output and predicted output; represents the renewable energy equipment χ in period t RES Output fluctuation range; Ω RES represents a collection of renewable energy devices; and They represent the actual and predicted values ​​of energy type load σ in period t respectively; represents the fluctuation amplitude of energy type load σ during period t; e, h, and c represent electricity, heat, and cooling, respectively. Based on the changes in ferry arrival time, disembarkation time, and power consumption when the ferry arrives at the port, the formula for the multiple uncertainty characterization model of logistics operations is: Where, and They represent the actual arrival time and the planned arrival time of the i-th ferry respectively; represents the maximum fluctuation of the actual arrival time of the i-th ferry; N FV represents the number of arriving ferries within the scheduling period; U represents the fluctuation range; Indicates the deviation between the actual arrival time and the planned arrival time; and They represent the actual disembarkation time and planned disembarkation time of passengers on the i-th ferry respectively; represents the maximum fluctuation of the time it takes for passengers to disembark from the i-th ferry; Indicates the deviation between the actual disembarkation time and the planned time; and represents the starting time of passenger ferrying for the i-th ferry; and They represent the actual state of charge and planned state of charge of the i-th ferry when it arrives at the port; represents the maximum fluctuation value of the state of charge when the i-th ferry arrives at the port; Indicates the deviation of the actual state of charge from the planned state of charge; and denote the charging start and end time of the i-th ferry respectively; represents the departure time of the i-th ferry.

3. The collaborative scheduling method of a port integrated energy system and a ferry logistics system under cross-domain multiple uncertainties according to claim 1 is characterized in that: The method of defining uncertainty fluctuation intervals based on convolutional neural networks and bidirectional long short-term memory networks is used to optimize the accuracy of multiple uncertainty characterization models for port energy production and logistics operations, including: Key features are extracted from historical data through convolutional neural networks to provide input for the prediction of bidirectional long short-term memory networks. Collaborative management information based on historical data is collected to obtain the uncertainty fluctuation range of the multiple uncertainty characterization model of port energy production and logistics operations, and the box boundaries are dynamically adjusted to form a generalized box uncertainty set.

4. The collaborative scheduling method of a port integrated energy system and a ferry logistics system under cross-domain multiple uncertainties according to claim 1 is characterized in that: The scheduling optimization model for determining multiple uncertainty scenarios with the goal of economic optimization includes: Based on the external energy procurement cost, equipment operation and maintenance cost, energy storage degradation cost, pollution treatment cost, in-port energy sales income and demand response subsidy, and with the principle of minimizing the port comprehensive energy scheduling cost under multiple uncertainties in energy production, the objective function of the port comprehensive energy system scheduling optimization model under multiple uncertainties in energy production is constructed: Where C buy 、C om 、C deg 、C ep 、C se and C dr They represent the external energy procurement cost, equipment operation and maintenance cost, energy storage degradation cost, pollution treatment cost, in-port energy sales income, and demand response subsidy during the dispatch period; and They represent the prices of electricity purchased from and sold to the grid during period t; Indicates the unit price of natural gas purchased from the gas grid; and They represent the power purchased and sold from the power grid during period t respectively; They represent the gas purchase rate from the gas network during period t; Δt E Indicates the unit duration of energy dispatch; c om,χ represents the unit operation and maintenance price of energy equipment χ; Indicates that the output of renewable energy units is uncertain; represents the output power of energy device χ; Ω represents the set of energy devices; represents the energy storage device χ ESS Unit degradation cost of and Represent the energy storage device χ ESS Charging and discharging power; Ω ESS Represents a collection of energy storage devices; represents the emission tax rate or emission right unit price for pollutant τ; ε e,τ and ε g,τ They represent the grid emission factor and the gas pollutant τ emission factor respectively; Q τ Conversion factor indicating the pollution equivalent value of τ; CO2, SO2 and NO x represent carbon dioxide, sulfur dioxide, and nitrogen oxide gases respectively; Indicates the unit price of energy sold within the port's integrated energy system; represents the energy consumption of the port ferry logistics system under multiple uncertainties; represents the energy consumption of port infrastructure under multiple uncertainties; T p and T v They represent the peak shaving and valley filling demand time periods published by the power grid; and They represent the unit price of demand response subsidies corresponding to peak shaving and valley filling periods respectively; The power purchased from external sources in the initial dispatch plan of the port integrated energy system under a deterministic scenario without considering demand response; Indicates the output power of energy equipment other than renewable energy units, Δt E Indicates the unit duration of energy dispatch; Based on the operating costs of all shore power, electric shuttles, and hydrogen shuttles within the scheduling period, and taking minimizing the port ferry logistics scheduling cost under multiple uncertainties of logistics operations as the principle, the objective function of the port ferry logistics system scheduling optimization model under multiple uncertainties of logistics operations is constructed: C fo =C CI +C ESB +C HSB ; Where C fo represents the joint operation cost of logistics facilities during the scheduling period; C CI 、C ESB and C HSB are the operating costs of all shore power, electric shuttles, and hydrogen shuttles during the dispatch period; c CI,on 、c ESB,on and c HSB,on are the basic operating costs of a single shore power, electric shuttle, and hydrogen shuttle respectively; c ESB,w and c HSB,w represent the unit cost of ferrying operations of electric shuttle and hydrogen shuttle respectively; The state variable representing the j-th shore power station serving the i-th ferry during period t; and They represent the state variables of the mth electric shuttle and the qth hydrogen shuttle serving the i-th ferry during period t; Δt L Indicates the unit duration of logistics scheduling; C ec and C dp They represent the logistics energy cost and delay compensation cost of the scheduling period respectively; c da 、c ds and c dd denote the unit penalty costs of ferry delayed arrival, delayed departure and delayed departure respectively; and They represent the actual departure time and the latest scheduled departure time of the i-th ferry respectively; represents the energy consumption of the port ferry logistics system under multiple uncertainties; Indicates the deviation between the actual arrival time and the planned arrival time; Indicates the deviation between the actual disembarkation time and the planned time; It indicates the unit price of energy sold within the port's integrated energy system.

5. The collaborative scheduling method of a port integrated energy system and a ferry logistics system under cross-domain multiple uncertainties according to claim 4 is characterized in that: The constraints of the port integrated energy system scheduling optimization model under multiple uncertainties of energy production include: multi-energy supply and demand balance, external energy network supply constraints and energy equipment operation constraints; Among them, the power generation, hydrogen production, heat production and cooling production of the port's integrated energy system at various times meet the power, hydrogen, heat and cooling energy needs of the port's ferry logistics system and infrastructure at various times; The supply constraints of the external energy network constrain the energy purchase and sales interaction between the port's integrated energy system and the external energy network. These constraints include the power supply capacity of the external grid, the power returned to the grid being limited by the maximum transmission power of the power cables, and the gas supply capacity of the external natural gas network being limited by the maximum gas supply rate. Ensure that the energy storage equipment operates within a safe range and that the energy level of the energy storage equipment at the end of the dispatch period is consistent with that at the beginning of the dispatch period; The constraints of the port ferry logistics system scheduling optimization model under multiple uncertainties of logistics operations include: ferry berthing constraints, ferry logistics facility allocation constraints and ferry logistics facility operation constraints; The port applies a continuous berth allocation method to meet the first-come, first-served principle; the ferry berthing time is later than the arrival time to ensure the feasibility of the scheduling plan; the berthing position and berthing time of any two ferries in the same period are determined based on the principle of avoiding each other; Ferry logistics facility allocation constraints include independent allocation of a single facility and joint allocation constraints between facilities; ferry logistics facility operation constraints include work efficiency, operating power and battery capacity constraints.

6. The collaborative scheduling method of a port integrated energy system and a ferry logistics system under cross-domain multiple uncertainties according to claim 1 is characterized in that: The scheduling constraint optimization based on interval modeling includes: The interval modeling method is used to reduce the dimensionality of three-dimensional integer variables in the port ferry logistics system scheduling optimization model to optimize the scheduling optimization model. Among them, by introducing a two-dimensional binary time variable to define the start time, processing time, and end time of each ferry berthing, charging, and ferrying, and optimizing the constraints of the port ferry logistics system scheduling optimization model under multiple uncertainties of logistics operations, the constraint condition formula of the optimized port ferry logistics system scheduling optimization model is obtained: Where, Indicates whether the i-th ferry starts the facility ω service phase during time period t; Indicates whether the i-th ferry has completed the facility ω service phase during time period t; represents the set of positive integers; and presence(ω i ) indicate whether the i-th ferry accepts the logistics facility ω service during period t; and Indicates whether the i-th ferry accepts the logistics facility ω service during the period t1+1, t1, t3, t2-1 and t2 respectively; and Indicates whether the i-th ferry berths during period t2, period t2-1 and period t respectively; Indicates whether the sth ferry is berthing during time period t; Indicates whether the i-th ferry accepts the logistics facility ω service during time period t'; represents the state variable of the zth facility ω serving the i-th ferry at time t'; and They represent whether the i-th ferry is in the charging, berthing, electric ferrying and hydrogen ferrying stages during time period t; M represents an infinite constant, l i represents the starting berth position of ferry i, ι i represents the length of ferry i, l max Indicates the length of the dock shoreline, and They represent the maximum and minimum number of electric shuttle buses and hydrogen shuttle buses that can be allocated to the i-th ferry, and denote the time when the logistics facility ω starts and ends serving the i-th ferry, T represents the time that the logistics facility ω serves the i-th ferry, L represents the logistics scheduling cycle, and Indicates whether the i-th ferry is berthed at berth b during period t2 and period t2-1 respectively; The state variable representing the zth facility ω serving the i-th ferry at time t; The state variable representing the jth shore power station serving the i-th ferry during period t; and They represent the state variables of the mth electric shuttle and the qth hydrogen shuttle serving the i-th ferry in period t respectively.

7. The collaborative scheduling method of a port integrated energy system and a ferry logistics system under cross-domain multiple uncertainties according to claim 1 is characterized in that: The method of achieving a dynamic balance between the energy consumption of the port ferry logistics system under multiple uncertainties in logistics operations and the energy supply of the port integrated energy system under multiple uncertainties in energy production based on the dynamic coupling association between multiple uncertainties across multiple domains includes: Energy production, conversion, and storage equipment are dynamically dispatched through the energy production of the port integrated energy system to meet the actual energy consumption of the port ferry logistics system under multiple uncertainties in logistics operations. Based on the multiple uncertainties in energy production, the dynamic coupling correlation form between multiple uncertainties across multiple fields is determined, and the expression of the dynamic coupling correlation form between multiple uncertainties across multiple fields is: Where Λ and Υ represent the energy conversion and storage matrices, respectively; and Respectively represent the charging and discharging power of the battery; and Respectively represent the hydrogen storage and release rates of the hydrogen storage tank; and Respectively represent the storage and release power of the heat storage tank; and They represent the power purchased and sold from the power grid during period t, represents the gas purchase rate from the gas network during period t, and represents the actual electricity and hydrogen demand of the port ferry logistics system during period t, and represent the actual electricity, heating and cooling loads of the port infrastructure during period t, represents the transmission power loss from the port integrated energy system to the ferry logistics system during period t, represents the hydrogen loss from the port integrated energy system to the ferry logistics system during period t, and They represent the losses in the process of transmitting electricity, heat and cold from the port integrated energy system to the logistics system during period t.

8. The collaborative scheduling method of a port integrated energy system and a ferry logistics system under cross-domain multiple uncertainties according to claim 1 is characterized in that: The establishment of a robust optimization model for coordinated scheduling of the port integrated energy system and the ferry logistics system includes: Introducing a dual incentive mechanism consisting of load adjustment incentives and grid demand response incentives, so that the port integrated energy system pays the ferry logistics system for load adjustment costs; The objective function of the coordinated scheduling of the port integrated energy system and the ferry logistics system is constructed, and based on the parameter matrix of the port integrated energy system or the ferry logistics system scheduling, a robust optimization model for the coordinated scheduling of the port integrated energy system and the ferry logistics system under cross-domain multiple uncertainties is constructed. The formula of the robust optimization model for the coordinated scheduling of the port integrated energy system and the ferry logistics system is: Where c τ 、b τ d τ 、r τ 、h τ 、D τ 、R τ 、F τ 、H τ , L τ and Y τ Represents the parameter matrix of different port integrated energy systems or ferry logistics system scheduling; U τ Indicates uncertain scheduling scenario; x τ 、y τ and u τ They represent the scheduling variable matrix composed of binary variables, continuous variables and uncertain variables respectively; Ω represents the scheduling variable matrix composed of x τ and u τ The uncertain scenarios constituted by PE and PL represent the port integrated energy system and port ferry logistics system respectively.

9. The collaborative scheduling method of a port integrated energy system and a ferry logistics system under cross-domain multiple uncertainties according to claim 8 is characterized in that: The expressions of scheduling variable matrix and uncertain scenarios are: Where x PE 、y PE and u PE They represent the decision variable matrices composed of integer variables, continuous variables, and uncertain variables in the robust optimization problem of port integrated energy system scheduling; x PL 、y PL and u PL They represent the decision variable matrices composed of integer variables, continuous variables and uncertain variables in the robust optimization problem of port ferry logistics system scheduling; and Respectively represent the time periods and each time period The matrix of integer variables formed; and Respectively represent the time periods Each time period Each time period and each time period The integer variable matrix composed of n ESB,m 、n HSB,q 、x i 、f i and l represent the periods of time Each time period Each ferry x is , each ferryφ is and ferries i The matrix of integer variables formed; and Respectively represent each period Each time period Each time period Each time period and each time period The continuous variable matrix formed by t ber,beg , t ω,beg , t ω,end , t dep 、 ζ ESB,m ,ζ HSB,q 、 and Respectively, each ferry Ferries Ferries Ferries Each time period Each time period Each time period Each time period Each time period Each time period Each time period and each time period The continuous variable matrix formed; and Respectively represent each period and each time period The uncertain variable matrix formed; and Respectively, each ferry Ferries and ferries The uncertain variable matrix formed; in, represents the energy storage device χ ESS The charging status of Indicates the electricity purchasing status; The state variable representing the berthing work of the i-th ferry at berth b during period t; The state variable representing the jth shore power station serving the i-th ferry during period t; and They represent the state variables of the ferry operation of the mth electric shuttle and the qth hydrogen shuttle serving the i-th ferry during period t; represents the actual passenger capacity of the mth electric shuttle bus during period t; represents the passenger capacity of the qth hydrogen shuttle bus in period t; x is The state variables of the berthing positions of any ferry i and ferry s; φ is The state variable of the berthing time of any ferry i and s; l i represents the starting berth position of ferry i; represents the output power of energy equipment χ; and Represent the energy storage device χ ESS The charging and discharging power; Indicates the power or rate of energy purchased from the energy network; represents the time when the i-th ferry starts berthing; represents the starting time of the phase when the i-th ferry is served by facility ω; represents the end time of the period in which the i-th ferry is served by facility ω; represents the actual departure time of the i-th ferry; represents the state of charge of the battery of the i-th ferry during period t; and They represent the charge state of the mth electric shuttle and the hydrogen charge state of the qth hydrogen shuttle at the end of period t respectively; and They represent the actual energy consumption of the jth shore power station, the mth electric shuttle bus, and the qth hydrogen shuttle bus during period t respectively; and represents the actual electricity and hydrogen demand of the port ferry logistics system during period t; Indicates that the output of renewable energy units is uncertain; It represents the energy consumption of port infrastructure under multiple uncertainties; represents the actual arrival time of the i-th ferry; represents the actual disembarkation time for passengers of the i-th ferry; represents the actual state of charge of the i-th ferry when it arrives at the port.

10. The collaborative scheduling method of a port integrated energy system and a ferry logistics system under cross-domain multiple uncertainties according to claim 1 is characterized in that: The distributed collaborative optimization solution framework constructed by combining the column and constraint generation algorithm and the adaptive alternating direction multiplier method is used to solve the coordinated robust scheduling scheme of the port integrated energy system and the ferry logistics system, including: Based on the column and constraint generation algorithm, the robust optimization model of the coordinated scheduling of the port integrated energy system and the ferry logistics system is reconstructed into the main problem and sub-problems; In the first stage, the main problem is solved using a deterministic matrix to obtain the scheduling objectives and integer variable matrices of the port integrated energy system and ferry logistics system. In the second stage, the continuous variable correlation function in the objective function is transformed into a maximum value problem using the KKT condition and combined with the uncertain variable correlation function to form a sub-problem. Based on the integer variable matrix obtained in the first stage, the uncertain variable matrix and the continuous variable matrix are calculated. The uncertain variable matrix is ​​used as a parameter, and the adaptive alternating direction multiplier method is used to perform distributed collaborative scheduling optimization based on the consensus on the dynamic balance of energy supply and demand of the port integrated energy system and the ferry logistics system. In each iteration, the port integrated energy system optimizes its own scheduling based on the energy consumption of the ferry logistics system obtained in the previous iteration, and sends the updated energy supply plan to the port ferry logistics system; The port ferry logistics system optimizes its own scheduling based on the latest energy supply plan and transmits the updated logistics energy demand back to the port integrated energy system. Based on the original residual and the dual residual, the penalty operator matrix and the Lagrange multiplier matrix are updated: Where, and Respectively represent The residuals and dual residuals of energy supply and demand variables at the iteration; and denote the smoothing parameters of the original residual norm and the dual residual norm respectively; and Respectively represent The sub-penalty operator matrix and the Lagrange multiplier matrix, and Respectively represent The sub-penalty operator matrix and the Lagrange multiplier matrix, u τ represents the scheduling variable matrix composed of uncertain variables; and Respectively represent The smoothed values ​​of residual norm and dual residual norm of energy supply and demand variables of the iteration; Indicates the The smoothed value of the residual norm of the energy supply and demand variables of the iteration, Indicates the The energy variables of the port integrated energy system supplying the ferry logistics system are obtained by the iteration. and Respectively represent Second and The energy demand variables of the ferry logistics system obtained by the iteration are: ε1 and ε2 represent the parameters involved in the penalty operator matrix adjustment. ρ and Respectively represent the lower limit and upper limit matrices of the penalty operator; The iteration stops when the convergence condition is met, and the final robust scheduling scheme for the coordinated port integrated energy system and ferry logistics system is obtained.

Images