Low-carbon port distributed energy management method considering computing power network

Abstract

The invention discloses a low-carbon port distributed energy management method considering a computing power network, and relates to the technical field of new energy and energy saving technology / port micro-grid system energy scheduling. According to the method, the energy management problem of the port micro-grid can be effectively solved, the energy consumption calculation model of the geographically dispersed data center is established by considering the data load delay constraint in the port micro-grid and the energy consumption of the cooling system of the data center for the upper geographically dispersed data center, and the optimal scheduling scheme is calculated by using a centralized optimization method. For a lower-layer port micro-grid system, a port refrigeration house is used as a main energy consumption source, a low-carbon port energy management model is constructed based on the power purchase cost, the carbon capture cost, the data center cost and the energy supply equipment operation cost of a main power grid generated by a port, and a distributed energy management strategy is designed based on a multi-agent consistency algorithm; and finally, solving an optimal energy management scheme so as to realize port micro-grid energy management.

Description

Technical Field

[0001] The present invention relates to new energy and energy-saving technologies / energy dispatching of port microgrid systems, and in particular to a low-carbon port distributed energy management method considering a computing power network. Background Art

[0002] As a hub for global trade and logistics, ports are an important source of energy consumption and carbon emissions. By building zero-carbon ports, the carbon footprint of ports can be significantly reduced, reducing the negative impact of climate change. Therefore, optimizing the energy management system of port microgrids is an important guarantee for the sustainable development of ports. However, most port microgrids have limited computing resources, and may not be able to complete distributed economic scheduling relying solely on their own computing resources.

[0003] Due to the special properties of ports, there are few studies on energy management issues of port microgrids that consider computing power requirements. Most existing studies only consider the load requirements of the port's power system. For the upper-level geographically dispersed data centers, due to the large computing power requirements generated by ports and insufficient computing power resources, existing studies have ignored the insufficient computing power resources of ports. For the lower-level port microgrid system, existing studies have mostly used port operating costs as the optimization target, ignoring the additional cost requirements generated by data centers and carbon emissions.

[0004] Therefore, the present invention aims to build a zero-carbon port, reduce the carbon footprint of the port, reduce the negative impact of climate change, and provide a complete two-layer distributed energy management method for low-carbon ports that takes into account the computing power network, thereby achieving safe and reliable operation of the port. Summary of the invention

[0005] Aiming at the energy scheduling problem of the port microgrid system with insufficient computing resources, the present invention provides a two-layer distributed energy management method for low-carbon ports that takes into account the computing network. The present invention proposes an energy management model and a distributed energy management strategy for geographically dispersed data centers and energy supply equipment, respectively. While reducing the port's operating costs and carbon emissions, the present invention meets the port's computing needs through the collaborative cooperation of data centers in multiple geographical locations, so as to achieve low-carbon and stable operation of the port.

[0006] To this end, the present invention provides the following technical solutions:

[0007] A low-carbon port distributed energy management method considering a computing power network, comprising:

[0008] The upper layer takes minimization of energy consumption generated by geographically dispersed data centers for data processing as the first objective function, constructs constraint conditions based on data delay and cooling system energy consumption, establishes an energy consumption calculation model for geographically dispersed data centers, and uses a centralized optimization method to solve the energy consumption calculation model for geographically dispersed data centers to obtain the upper layer optimal scheduling solution; based on the upper layer optimal scheduling solution, obtains the operating cost of the geographically dispersed data centers;

[0009] The lower layer takes the minimization of the port microgrid operation cost, the main grid power purchase or sale cost, the carbon cost and the geographically dispersed data center cost as the second objective function, considers the port operation characteristics, constructs physical constraints, and establishes low-carbon port energy management models for the port microgrid island and grid-connected operation modes respectively;

[0010] A distributed energy management algorithm is designed based on a multi-agent consensus algorithm, and the distributed energy management algorithm is used to solve the low-carbon port energy management model to obtain the optimal energy scheduling solution for the lower layer.

[0011] Furthermore, the energy consumption calculation model of geographically dispersed data centers is:

[0012]

[0013] in, is the power consumption of data center j at time slot t; is the active power of the data center server and air conditioning equipment at time slot t; T = {1, 2, ..., t} represents the set of all time periods in the voyage; Indicates the cooling power of the data center air conditioner; EP n Indicates the energy efficiency coefficient of the data center air conditioning equipment; Indicates the number of air conditioners in the data center; m n,t represents the number of active servers in the data center at time t; They represent the peak power and silent power of a single server in a data center when not processing tasks; u n,t Indicates the average utilization of the servers in the data center.

[0014] Furthermore, the first constraint condition includes: server average utilization constraint, data center server quantity constraint, data load processing delay constraint and data load balance constraint; wherein:

[0015] The average server utilization constraint is:

[0016]

[0017] in, is the total amount of data load processed by the data center at time slot t; Φ is the set of front-end portal websites, and the set of front-end portal website servers is defined as δ(δ∈Φ); μ j is the average service rate of active servers in data center j. When the servers are homogeneous, μ j is the service rate of a single active server, which is a fixed performance parameter provided by the data center and depends on the server performance and data load type; U max Indicates the upper limit of CPU utilization of a single server;

[0018] The number of servers in a data center is constrained as follows:

[0019] 0≤m j,t ≤M j ;

[0020] Among them, M j is the total number of servers in data center j;

[0021] The data payload processing delay constraint is:

[0022]

[0023] Among them, λ δj,t is the data load allocated from the front-end portal server δ to data center j at time slot t; The total data load arriving at data center j at time slot t; is the average residence time of data load in data center j under the M / M / 1 queuing model; D is the delay bound for data load processing;

[0024] The data load balancing constraints are:

[0025]

[0026] Furthermore, physical constraints are constructed, including:

[0027] Taking the minimization of the total energy consumption of the port cold storage system as the third objective function, the power consumption model of the port cold storage equipment is established by considering the refrigeration power consumption and direct power consumption of the cold storage system.

[0028] The power consumption of the cold storage is obtained based on the power consumption model of the port cold storage equipment;

[0029] A physical constraint condition is established based on the power consumption of the cold storage.

[0030] Furthermore, the total energy consumption of the cold storage system is the sum of the refrigeration power consumption and direct power consumption of the cold storage system:

[0031] P TEC =P REC +PDEC ;

[0032] Among them, P TEC is the total energy consumption of the cold storage system, P REC , P DEC They are the refrigeration power consumption and direct power consumption of the port cold storage system;

[0033]

[0034] Among them, Q 1 is the heat flux of the building envelope; Q 2 is the cargo heat flow; Q 3 is the ventilation heat flow; Q 4 is the heat flow of the motor; Q 5 is the operating heat flow; R is the compensation coefficient for cooling loss of refrigeration equipment and pipelines; n 1 is the seasonal correction coefficient of heat flux of the building envelope; n 2 is the mechanical load reduction factor of cargo heat; n 3 is the number of ventilation times in the same period; n 4 is the synchronous operation coefficient of the cold room motor; n 5 is the synchronous operation coefficient of the cold room.

[0035] Furthermore, the second objective function is:

[0036] F=min{F 1 +F 2 +F 3 +F 4};

[0037] Among them, F is the total port cost, F 1 ,F 2 ,F 3 ,F 4 They are the main grid electricity purchase costs, port microgrid operating costs, carbon capture costs and geographically dispersed data center operating costs;

[0038] F 1 =K M P M ;

[0039] Among them, K M is the cost coefficient of the port main power grid purchasing electricity, P M is the output power of the port main power grid;

[0040] F 2 =F w +F pv +F fu +F d ;

[0041] Among them, Fw is the cost of wind power generation, F pv is the photovoltaic power generation cost, F fu is the cost of fossil fuel electricity generation, F d is the battery cost;

[0042]

[0043] Among them, F c For carbon capture costs; F wq is the wind curtailment cost; P c is the power consumption for carbon capture; c The energy consumption required to capture a unit of carbon dioxide; q c is the carbon emission intensity of thermal power units; x c is the capture level; t is the smoke split ratio during period t;

[0044]

[0045] Among them, K p is the cost coefficient of the port processing data; is the output power of the data center.

[0046] Furthermore, the physical constraints include: power balance constraints, power generation equipment capacity constraints, carbon capture equipment constraints and battery capacity constraints; wherein:

[0047] Power balance constraints:

[0048]

[0049] Among them, the number of devices satisfies n 1 +n 2 +n 3 +n 4 =N, N is the total number of power generation equipment in the port; P load is the port load demand; P TEC is the power consumption of the cold storage; P other is other port load;

[0050] Power generation equipment capacity constraints:

[0051]

[0052] in, They are the upper and lower limits of the power of each power generation equipment;

[0053] Carbon capture equipment constraints:

[0054]

[0055] in, They are the upper and lower power limits of the carbon capture equipment;

[0056] Battery capacity constraints:

[0057]

[0058] Among them, P e is the rated power of the battery, SoC max and SoC min They are respectively the maximum and minimum state of charge of the battery at the port.

[0059] Furthermore, the distributed energy management algorithm is used to solve the low-carbon port energy management model under the port microgrid grid-connected mode, and the optimal energy scheduling scheme at the lower level is obtained, including:

[0060] In the grid-connected mode, the iteration method of the electricity price λ follows the leader-follower consistency, and the iteration process is as follows:

[0061]

[0062] Among them, λ 0 The electricity price of the main power grid is is the algorithm step size;

[0063] The power generation capacity of each power generation equipment is:

[0064]

[0065] Among them, P i (k) represents the power generation of the i-th node at the k-th iteration;

[0066] Each node estimates the local mismatch value of all nodes in the system through the leadership consensus algorithm:

[0067]

[0068] Among them, ΔP i (k) is the actual local mismatch value at the kth iteration, is the mismatch value estimated at the kth iteration, ε is the algorithm step size, It means that the neighboring cells of the main grid will directly receive power supplement from the main grid in each iteration;

[0069] The optimal solution for the main grid to purchase electricity is as follows:

[0070]

[0071] Among them, P M (k+1) is the supplementary power value of the main grid to the neighboring node at time k+1.

[0072] Furthermore, the distributed energy management algorithm is used to solve the low-carbon port energy management model under the port microgrid island mode, and the optimal energy scheduling scheme at the lower level is obtained, including:

[0073] In the island mode, the incremental cost is corrected by adding a penalty term to meet the supply and demand balance of the port microgrid. The iterative formula used is as follows:

[0074]

[0075] in, is the step size of the algorithm, ξ(k) is the feedback gain;

[0076] The power generation of each power generation unit is:

[0077]

[0078] Among them, P i (k) represents the power generation of the power generation unit of the i-th node at the k-th iteration;

[0079] Then, according to the known P i (k), estimate the local mismatch value, the formula is as follows:

[0080]

[0081] Among them, ΔP i (k) is the actual local mismatch value at the kth iteration, is the mismatch value estimated at the kth iteration and ε' is the step size of the algorithm.

[0082] Advantages and positive effects of the present invention: The present invention can effectively handle the energy management problem of port microgrids. For the upper geographically dispersed data centers, the data load delay constraints in the port microgrid and the energy consumption of the data center cooling system are considered, and a geographically dispersed data center energy consumption calculation model is established, and the optimal scheduling plan is calculated using a centralized optimization method. For the lower-level port microgrid system, the port cold storage is used as the main energy consumption source, and a low-carbon port energy management model is constructed based on the main power grid electricity purchase cost, carbon capture cost, data center cost and energy supply equipment operation cost generated by the port, and a distributed energy management strategy is designed based on the multi-agent consensus algorithm; finally, the optimal energy management plan is solved, thereby realizing the energy management of the port microgrid. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0084] Figure 1 It is a schematic diagram of the structure framework of the port microgrid energy system in an embodiment of the present invention;

[0085] Figure 2 It is a schematic diagram of the energy management of the port microgrid in the island mode in an embodiment of the present invention;

[0086] Figure 3 This is a schematic diagram of the energy management of a port microgrid in the grid-connected mode in an embodiment of the present invention. DETAILED DESCRIPTION

[0087] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0088] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0089] In order to reduce port carbon emissions, improve operational economy, and achieve safe and reliable operation of ports, the present invention takes the port microgrid system as the research object, and provides a two-layer distributed energy management method for low-carbon ports considering the computing power network for the port microgrid, aiming to reduce port carbon emissions and operating costs, promote the green transformation of ports, and drive ports towards a sustainable future. For the upper geographically dispersed data centers, the present invention considers the data load delay constraints in the port microgrid and the energy consumption of the data center cooling system, establishes a geographically dispersed data center energy consumption calculation model, and uses a centralized optimization method to calculate the optimal scheduling solution. For the lower-level port microgrid system, the port cold storage is used as the main energy consumption source, and a low-carbon port energy management model is constructed based on the main power grid electricity purchase cost, carbon capture cost, data center cost and energy supply equipment operation cost generated by the port. In addition, the present invention proposes a distributed energy management strategy based on the multi-agent consensus algorithm. This strategy can solve the energy management problem and calculate the optimal energy management solution while meeting the port's computing power requirements, thereby ensuring the port's business continuity.

[0090] The structure framework of the port microgrid energy system described in the present invention is as follows: Figure 1 As shown in the figure, the energy management of the port microgrid energy system takes the port cold storage as the main energy consumption source, and can be divided into three parts according to different energy flows: energy supply equipment, carbon capture equipment and data center. Among them, the data center consists of IT equipment such as servers and storage, cooling systems, other equipment and power transmission and distribution systems. When the port is connected to the main power grid for operation, it is a grid-connected mode. The port power system is connected to the main power grid to jointly supply power to the port. The port energy supply equipment includes two photovoltaic power generation equipment, two wind power generation equipment, two fuel generators, a carbon capture device and an energy storage device. Considering the data load delay constraint and the energy consumption of the data center cooling system, a geographically dispersed data center energy consumption calculation model is constructed. The energy supply equipment includes traditional fuel equipment (such as internal combustion units, etc.), new energy equipment (such as photovoltaic units, wind turbine units, etc.), and the energy management model is established with the carbon capture cost, the geographically dispersed data center cost, the energy supply equipment operation cost and the main power grid power purchase or sales cost minimization as the objective function. For the port microgrid island and grid-connected mode, the solution is based on the multi-agent consensus algorithm to obtain the optimal energy management solution.

[0091] In an embodiment of the present invention, a two-layer distributed energy management method for a low-carbon port considering a computing power network is provided, which is mainly divided into energy consumption for geographically dispersed data centers and energy management for a port microgrid system, and specifically includes the following steps:

[0092] Step 1: The upper layer designs geographically dispersed data centers to minimize the energy consumption generated by data processing as the objective function (step 1.1); constructs constraints based on data latency and cooling system energy consumption (step 1.2), and then establishes a geographically dispersed data center energy consumption calculation model, and uses a centralized optimization method to calculate the optimal scheduling solution.

[0093] Furthermore, the power consumption of different devices in geographically dispersed data centers involved in step 1.1 includes:

[0094] Step 1.1.1: Calculate the power consumption of the data center, which can be expressed as:

[0095]

[0096] in, is the power consumption of data center j at time slot t; is the active power of the data center server and air conditioning equipment at time slot t; T = {1, 2, ..., t} represents the set of all time periods within the voyage. Indicates the cooling power of the data center air conditioner; EP n Indicates the energy efficiency coefficient of the data center air conditioning equipment; Indicates the number of air conditioners in the data center; m n,t represents the number of active servers in the data center at time t; They represent the peak power and silent power of a single server in a data center when not processing tasks; u n,t Indicates the average utilization of the servers in the data center.

[0097] Furthermore, the constraints involved in step 1.2 include:

[0098] Step 1.2.1: Construct the average server utilization constraint, which can be specifically expressed as:

[0099]

[0100] in, is the total amount of data load processed by the data center at time slot t; Φ is the set of front-end portal websites, and the set of front-end portal website servers is defined as δ(δ∈Φ); μ j is the average service rate of active servers in data center j. When the servers are homogeneous, μ j is the service rate of a single active server, which is a fixed performance parameter provided by the data center and depends on the server performance and data load type; U max Indicates the upper limit of CPU usage of a single server.

[0101] Step 1.2.2: Construct the number of servers in the data center, which can be specifically expressed as:

[0102] 0≤m j,t ≤M j ;

[0103] Among them, M j is the total number of servers in data center j.

[0104] Step 1.2.3: Construct the data load processing delay constraint, which can be specifically expressed as:

[0105]

[0106] Among them, λ δj,t is the data load allocated from the front-end portal server δ to data center j at time slot t; The total data load arriving at data center j at time slot t; is the average residence time of data load in data center j under the M / M / 1 queuing model; D is the delay bound for data load processing.

[0107] Step 1.2.4: Construct data load balancing constraints, which can be specifically expressed as:

[0108]

[0109] In the energy management system of the port microgrid, the data center is another important energy consumption source, and its IT equipment energy consumption and cooling system energy consumption account for a considerable proportion. In order to achieve effective management of the energy consumption of the data center, the present invention establishes a data center energy consumption model and adopts a centralized algorithm to solve it. The centralized algorithm can comprehensively consider the energy consumption characteristics, load demand and energy supply of the data center, and formulate the optimal energy allocation strategy, thereby achieving accurate control and optimization of the energy consumption of the data center.

[0110] Step 2: Design the objective function of minimizing the total energy consumption of the port cold storage system (step 2.1); consider the refrigeration energy consumption (REC) and direct energy consumption (DEC) of the cold storage system (step 2.2), and then establish the power consumption model of the port cold storage equipment.

[0111] The cold storage equipment power consumption model is designed for the specific energy demand side of the port cold storage. It takes into account the refrigeration power consumption and direct power consumption, that is, the power consumption required for the cold storage during operation, including the refrigeration system used to maintain the low-temperature environment and the direct power consumption of other equipment in the cold storage. As a sub-model of the low-carbon port energy management model, the cold storage equipment power consumption model provides detailed data on the energy consumption of the cold storage and provides a basis for energy management. The low-carbon port energy management model comprehensively considers the energy consumption and cost of multiple energy demanders and suppliers, including cold storage, and achieves the goal of minimizing the overall cost and energy saving and emission reduction by optimizing the scheduling strategy.

[0112] Furthermore, the objective function involved in step 2.1 includes:

[0113] Step 2.1.1: The total energy consumption (TEC) of the cold storage system is the sum of the refrigeration energy consumption (REC) and direct energy consumption (DEC) of the cold storage system, which can be expressed as:

[0114] P TEC =P REC +P DEC ;

[0115] Among them, P TEC is the total energy consumption of the cold storage system, P REC , P DEC They are respectively the refrigeration power consumption and direct power consumption of the port cold storage system.

[0116] Furthermore, the refrigeration energy consumption (REC) and direct energy consumption (DEC) of the cold storage system involved in step 2.2 include:

[0117] Step 2.2.1: Calculate the refrigeration power consumption and direct power consumption of the cold storage system, which can be expressed as:

[0118]

[0119] Among them, Q 1 is the heat flux of the building envelope; Q 2 is the cargo heat flow; Q 3 is the ventilation heat flow; Q 4 is the heat flow of the motor; Q 5 is the operating heat flow; R is the compensation coefficient for cooling loss of refrigeration equipment and pipelines; in this paper, it is calculated as 1.07 according to the general direct cooling system; n 1 is the seasonal correction coefficient of heat flux of the building envelope, which is set to 1; n 2 is the mechanical load reduction factor of cargo heat, which is 0.6; n 3 is the number of ventilation times in the same period, generally ranging from 0.5 to 1; n 4is the synchronous operation coefficient of the cold room motor, the value is 0.4; n 5 is the synchronous operation coefficient of the cold room, and its value is 0.4.

[0120] Step 2.2.2: The heat flow is expanded as follows:

[0121]

[0122] Where A is the heat transfer area; Δt is the temperature difference between the indoor and outdoor sides adopted in the design; Q 21 is the food calorie; Q 22 The heat of packaging materials and carriers; h 1 is the enthalpy value of the goods before entering the cold room for freezing. The enthalpy value corresponding to the cherries entering the 0-10℃ warehouse at a temperature of 29℃ is 339.8kJ / kg; h 2 is the enthalpy value of the goods after freezing in the cold room, and the enthalpy value corresponding to the freezing temperature of cherries in the 0-10℃ warehouse is 236.3kJ / kg; G is the calculated tonnage of the cold storage design; G′ is the daily purchase volume of the cold room, and its limit is limited according to the different functions of the cold storage. According to the conservative calculation, regardless of the type of goods, the median value is taken, which is 6.5% of the calculated tonnage, that is, 0.065G; G is calculated in accordance with GB50072-2010 "Cold Storage Design Code" 3.0.2; T is the cold processing time of the goods, and the cold storage room is taken as 24h; B is the weight coefficient of the goods packaging material or carrier, and the most conservative value is 0.6; t 1 The temperature of the packaging material or vehicle when it enters the cold room is 29.6℃; t 2 is the temperature of the packaging material or vehicle when the cooling in the cold room ends, and the lowest storage temperature in the design storage temperature is taken; the cold storage for storing vegetables and fruits in the cargo port in this paper is not equipped with ventilation and ventilation operation is not required, so Q 3 The value of P is 0; u is the rated power of the motor; ζ is the heat conversion coefficient, which is 1 for motors in cold rooms; s is the motor operation time coefficient, which is 1 for motors equipped with air coolers, and 8 / 24 for motors of other equipment, calculated based on 8 hours of operation per day and night; Q is the lighting heat, 52 is the door opening heat, Q 53 is the operator's heat; q a The lighting heat per square meter of floor area is 2.3W / m 2 Value; A is the cold room floor area; h w is the specific enthalpy of the air outside the cold room, and the air enthalpy value corresponding to the outdoor calculated temperature and the outdoor calculated relative humidity in summer ventilation in Dalian is 239.79; h n is the specific enthalpy of the air in the cold room, and the air enthalpy value is taken as 8.02 according to the temperature and humidity under the design conditions of the cold storage; n is the number of ventilation times, and the value is taken as 2; F is the net size of the warehouse, and the storage capacity is directly taken as a conservative estimate; ρn is the air density in the warehouse, which is the air density corresponding to the lowest warehouse temperature in the design warehouse temperature; n x is the number of operators; q r is the operator’s heat, which is taken as 410W / person; M is the air curtain efficiency correction factor, which is calculated as if no air curtain is installed, and the value is 1.

[0123] The research background of the present invention focuses on the port microgrid, which is a microgrid system that integrates multiple energy types, complex network structures and diverse load demands. As an important hub of the global economy, the operational efficiency and sustainability of the port are of great significance to global trade and environmental protection. Therefore, the present invention proposes a low-carbon port distributed energy management method considering the computing power network, aiming to reduce carbon emissions in the port operation process and improve energy utilization efficiency by optimizing energy distribution and utilization. In the port microgrid, the cold storage is an important load demand side, and its energy consumption accounts for a large proportion and has certain volatility. In order to more accurately describe the energy consumption characteristics of the cold storage, the present invention establishes a cold storage energy consumption model and incorporates it into the energy management system as part of the load demand side. By real-time monitoring of the energy consumption data of the cold storage, the system can dynamically adjust the energy allocation strategy to ensure that the cold storage minimizes energy consumption while maintaining a constant temperature. At the same time, in view of the carbon emission problem generated during the operation of the port, the present invention uses a method for processing using carbon capture technology. Carbon capture technology is an effective emission reduction method that can capture, separate and store carbon dioxide emitted by the port, thereby reducing the carbon emissions of the port. By integrating carbon capture technology into the energy management system, precise control and optimization of the port's carbon emissions are achieved, further promoting the low-carbon development of the port.

[0124] Step 3: The lower layer takes the minimization of the port microgrid operating cost, the main grid power purchase or sales cost, the carbon cost and the geographically dispersed data center cost as the objective function (step 3.1), considers the port operation characteristics, constructs physical constraints, and establishes low-carbon port energy management models for the port microgrid island and grid-connected operation modes (step 3.2).

[0125] Furthermore, the objective function involved in step 3.1 includes:

[0126] Step 3.1.1: Based on the main grid electricity purchase cost, port microgrid operation cost, carbon capture cost and geographically dispersed data center cost function, establish the objective function, including:

[0127] F=min{F 1 +F 2 +F 3 +F 4};

[0128] Among them, F is the total port cost, F1 ,F 2 ,F 3 ,F 4 They are the electricity purchase cost of the main power grid, the operating cost of the port microgrid, the carbon capture cost and the operating cost of geographically dispersed data centers.

[0129] Step 3.1.2: Calculate the cost of purchasing electricity from the main power grid, which can be expressed as:

[0130] F 1 =K M P M ;

[0131] Among them, K M is the cost coefficient of the port main power grid purchasing electricity, P M is the output power of the port's main power grid.

[0132] Step 3.1.3: Calculate the operating cost of the port microgrid, which can be expressed as:

[0133] F 2 =F w +F pv +F fu +F d ;

[0134] Among them, F w is the cost of wind power generation, F pv is the photovoltaic power generation cost, F fu is the cost of fossil fuel electricity generation, F d It is the battery cost.

[0135] Table 1

[0136]

[0137] Step 3.1.4: Calculate the cost of carbon capture, which can be expressed as:

[0138]

[0139] Among them, F c For carbon capture costs; F wq is the wind curtailment cost; P c is the power consumption for carbon capture; c The energy consumption required to capture a unit of carbon dioxide; q c is the carbon emission intensity of thermal power units; x c is the capture level; t is the flue gas split ratio during period t.

[0140] Step 3.1.5: Calculate the cost of operating geographically dispersed data centers, which can be expressed as:

[0141]

[0142] Among them, K p is the cost coefficient of port processing data, which is [0.6127, 0.7247, 0.8521]; is the output power of the data center.

[0143] Furthermore, the physical constraints involved in step 3.2 include:

[0144] Step 3.2.1: Construct the power balance constraint as follows:

[0145]

[0146] Among them, the number of devices satisfies n 1 +n 2 +n 3 +n 4 =N, where N is the total number of power generation equipment in the port. load is the port load demand; P TEC is the power consumption of the cold storage; P other It is the other load of the port.

[0147] Step 3.2.2: Construct the capacity constraints of power generation equipment as follows:

[0148]

[0149] in, They are the upper and lower limits of the power of each power generation equipment.

[0150] Step 3.2.3: Construct the carbon capture equipment constraints as follows:

[0151]

[0152] in, They are the upper and lower power limits of the carbon capture equipment.

[0153] Step 3.2.4: Construct the battery capacity constraint as follows:

[0154]

[0155] Among them, P e is the rated power of the battery, SoC max and SoC min They are respectively the maximum and minimum state of charge of the battery at the port.

[0156] At this point, the distributed energy management model of the port microgrid in step 3 is established.

[0157] Step 4: Design a distributed energy management algorithm based on the multi-agent consensus algorithm, and use the distributed energy management algorithm to solve the low-carbon port energy management model to obtain the optimal energy scheduling solution for the lower layer.

[0158] In the multi-agent consensus algorithm, each intelligent control unit iteratively estimates the electricity price of the distribution system and the average power mismatch of the entire microgrid through the leader-follower algorithm and the average consensus algorithm. In each iteration, the energy router calculates the incremental active power to be exchanged with the distribution system in the next period in a distributed manner. While ensuring convergence, it can reduce the fluctuation of the exchanged power.

[0159] Furthermore, the distributed energy management algorithm involved in step 4 includes:

[0160] Step 4.1: The port microgrid operation cost in the energy management model is rewritten as follows:

[0161]

[0162] Among them, F i (P i ) is the cost of the i-th device; α i ,β i , is the equivalent transformation coefficient of the cost function.

[0163] Step 4.2: The energy management model is rewritten as follows:

[0164]

[0165] Step 4.3: The objective function of the energy management model is rewritten as follows:

[0166]

[0167] Step 4.4: Design the Lagrangian function as follows:

[0168]

[0169] Among them, λ, ν are all Lagrange multipliers, and the carbon capture load can be expressed as P c =B i P i .

[0170] Step 4.5: Analyze its KKT (Karush Kuhn Tucker) conditions as follows:

[0171]

[0172] In the grid-connected mode, the iteration method of the electricity price λ follows the leader-follower consistency, and the iteration process is as follows:

[0173]

[0174] Among them, λ 0 The electricity price of the main power grid is is the algorithm step size.

[0175] The power generation capacity of each power generation equipment is:

[0176]

[0177] Among them, P i (k) represents the power generation of the i-th node at the k-th iteration.

[0178] Each node estimates the local mismatch value of all nodes in the system through the leadership consensus algorithm as follows:

[0179]

[0180] Among them, ΔP i (k) is the actual local mismatch value at the kth iteration, is the mismatch value estimated at the kth iteration, ε is the algorithm step size, It means that the neighboring cells of the main grid will directly receive power supplement from the main grid in each iteration.

[0181] The optimal solution for the main grid to purchase electricity is as follows:

[0182]

[0183] Among them, P M (k+1) is the supplementary power value of the main grid to the neighboring node at time k+1.

[0184] In the island mode, the incremental cost is corrected by adding a penalty term to meet the supply and demand balance of the port microgrid. The iterative formula used is as follows:

[0185]

[0186] in, is the step size of the algorithm, ξ(k) is the feedback gain;

[0187] The power generation of each power generation unit is:

[0188]

[0189] Among them, P i (k) represents the power generation of the power generation unit of the i-th node at the k-th iteration;

[0190] Then, according to the known P i (k), estimate the local mismatch value, the formula is as follows:

[0191]

[0192] Among them, ΔP i (k) is the actual local mismatch value at the kth iteration, is the mismatch value estimated at the kth iteration and ε' is the step size of the algorithm.

[0193] For the two operating modes of port microgrid, island and grid-connected, the present invention adopts the distributed average consensus algorithm and the distributed leader-follower consensus algorithm for solution respectively. In the island mode, since the port microgrid is disconnected from the main grid, it needs to rely on its own energy supply and regulation capabilities to maintain stable operation. At this time, the distributed average consistency algorithm can utilize the information interaction and synergy of each distributed energy node to achieve balanced and stable energy distribution. In the grid-connected mode, the port microgrid is connected to the main grid, and it is necessary to consider the main grid's electricity purchase or sales costs and the port's own energy needs at the same time. At this time, the distributed leader-follower consensus algorithm can achieve coordinated optimization and efficient energy utilization with the main grid while ensuring the stable operation of the port microgrid.

[0194] Based on the distributed energy management model of the port microgrid in step 3 and the distributed energy management algorithm in step 4, the optimal energy management solution is calculated and executed.

[0195] At this point, a low-carbon port microgrid energy management dispatch is completed. The output power and λ iterations of the power supply equipment in the isolated and grid-connected modes, the mismatch value iterations and the total power supply iterations are shown in the attached figure. Figure 2-3 Specifically, the computing power requirements of the port are met, and the port load and carbon dioxide emission treatment based on the optimal energy management strategy meet the expected settings, which can achieve the green development of the port. According to the above simulation diagram, the effectiveness of the energy management method proposed in the present invention can be seen.

[0196] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A low-carbon port distributed energy management method considering computing power network, characterized in that: include: The upper layer takes minimizing the energy consumption of geographically dispersed data centers for data processing as the first objective function, constructs constraints based on data delay and cooling system energy consumption, establishes a geographically dispersed data center energy consumption calculation model, and uses a centralized optimization method to solve the geographically dispersed data center energy consumption calculation model to obtain the upper layer optimal scheduling solution; Based on the upper-layer optimal scheduling solution, the operation cost of geographically dispersed data centers is obtained; The lower layer takes the minimization of the port microgrid operation cost, the main grid power purchase or sale cost, the carbon cost and the geographically dispersed data center cost as the second objective function, considers the port operation characteristics, constructs physical constraints, and establishes low-carbon port energy management models for the port microgrid island and grid-connected operation modes respectively; A distributed energy management algorithm is designed based on a multi-agent consensus algorithm, and the distributed energy management algorithm is used to solve the low-carbon port energy management model to obtain the optimal energy scheduling solution for the lower layer.

2. A low-carbon port distributed energy management method considering computing power network according to claim 1, characterized in that: The energy consumption calculation model for geographically dispersed data centers is: in, is the power consumption of data center j at time slot t; is the active power of the data center server and air conditioning equipment at time slot t; T = {1, 2, ..., t} represents the set of all time periods in the voyage; Indicates the cooling power of the data center air conditioner; EP n Indicates the energy efficiency coefficient of the data center air conditioning equipment; Indicates the number of air conditioners in the data center; m n,t represents the number of active servers in the data center at time t; They represent the peak power and silent power of a single server in a data center when not processing tasks; u n,t Indicates the average utilization of the servers in the data center.

3. A low-carbon port distributed energy management method considering computing power network according to claim 2, characterized in that: The first constraint condition includes: server average utilization constraint, data center server quantity constraint, data load processing delay constraint and data load balance constraint; wherein: The average server utilization constraint is: Among them, L δ t is the total amount of data load processed by the data center at time slot t; Φ is the set of front-end portal websites, and the set of front-end portal website servers is defined as δ(δ∈Φ); μ j is the average service rate of active servers in data center j. When the servers are homogeneous, μ j is the service rate of a single active server, which is a fixed performance parameter provided by the data center and depends on the server performance and data load type; U max Indicates the upper limit of CPU utilization of a single server; The number of servers in a data center is constrained as follows: 0≤m j ,t≤M j ; Among them, M j is the total number of servers in data center j; The data payload processing delay constraint is: Among them, λ δj,t is the data load allocated from the front-end portal server δ to data center j at time slot t; The total data load arriving at data center j at time slot t; is the average residence time of data load in data center j under the M / M / 1 queuing model; D is the delay bound for data load processing; The data load balancing constraints are:

4. A low-carbon port distributed energy management method considering computing power network according to claim 1, characterized in that: Construct physical constraints, including: Taking the minimization of the total energy consumption of the port cold storage system as the third objective function, the power consumption model of the port cold storage equipment is established by considering the refrigeration power consumption and direct power consumption of the cold storage system. The power consumption of the cold storage is obtained based on the power consumption model of the port cold storage equipment; A physical constraint condition is established based on the power consumption of the cold storage.

5. A low-carbon port distributed energy management method considering computing power network according to claim 4, characterized in that: The total energy consumption of the cold storage system is the sum of the refrigeration power consumption and direct power consumption of the cold storage system: P TEC =P REC +P DEC ; Among them, P TEC is the total energy consumption of the cold storage system, P REC , P DEC They are the refrigeration power consumption and direct power consumption of the port cold storage system; P REC =Q1+pQ2+Q3+Q4+Q5 P DEC =Z(n1∑Q1+n2∑Q2+n3∑Q3+n4∑Q4+n5∑Q5); Among them, Q1 is the heat flow of the enclosure structure; Q2 is the heat flow of the cargo; Q3 is the ventilation heat flow; Q4 is the heat flow of the motor operation; Q5 is the operation heat flow; R is the compensation coefficient for the cold loss of refrigeration equipment and pipelines; n1 is the seasonal correction coefficient for the heat flow of the enclosure structure; n2 is the mechanical load reduction coefficient of the cargo heat; n3 is the number of ventilation times in the same period; n4 is the synchronous operation coefficient of the cold room motor; n5 is the synchronous operation coefficient of the cold room.

6. A low-carbon port distributed energy management method considering computing power network according to claim 4, characterized in that: The second objective function is: F = min{F1+F2+F3+F4}; Among them, F is the total cost of the port, F1, F2, F3, and F4 are the electricity purchase cost of the main power grid, the operating cost of the port microgrid, the carbon capture cost, and the operating cost of geographically dispersed data centers; F1=K M P M ; Among them, K M is the cost coefficient of the port main power grid purchasing electricity, P M is the output power of the port main power grid; F2=F w +F pv +F fu +F d ; Among them, F w is the cost of wind power generation, F pv is the photovoltaic power generation cost, F fu is the cost of fossil fuel electricity generation, F d is the battery cost; Among them, F c For carbon capture costs; F wq is the wind curtailment cost; P c is the power consumption for carbon capture; c The energy consumption required to capture a unit of carbon dioxide; q c is the carbon emission intensity of thermal power units; x c is the capture level; t is the smoke split ratio during period t; Among them, K p is the cost coefficient of the port processing data; is the output power of the data center.

7. A low-carbon port distributed energy management method considering computing power network according to claim 6, characterized in that: Physical constraints, including: power balance constraints, power generation equipment capacity constraints, carbon capture equipment constraints and battery capacity constraints; among which: Power balance constraints: P load =P TEC +P other Among them, the number of equipment satisfies n1+n2+n3+n4=N, where N is the total number of power generation equipment in the port; P load is the port load demand; P TEC is the power consumption of the cold storage; P other is other port load; Power generation equipment capacity constraints: in, They are the upper and lower limits of the power of each power generation equipment; Carbon capture equipment constraints: in, They are the upper and lower power limits of the carbon capture equipment; Battery capacity constraints: Among them, P e is the rated power of the battery, SoC max and SoC min They are respectively the maximum and minimum state of charge of the battery at the port.

8. A low-carbon port distributed energy management method considering computing power network according to claim 1, characterized in that: The distributed energy management algorithm is used to solve the low-carbon port energy management model under the port microgrid grid-connected mode, and the optimal energy scheduling scheme at the lower level is obtained, including: In the grid-connected mode, the iteration method of the electricity price λ follows the leader-follower consistency, and the iteration process is as follows: Among them, λ0 is the electricity price of the main grid, is the algorithm step size; The power generation capacity of each power generation equipment is: Among them, P i (k) represents the power generation of the i-th node at the k-th iteration; Each node estimates the local mismatch value of all nodes in the system through the leadership consensus algorithm: Where ΔP i (k) is the actual local mismatch value at the kth iteration, is the mismatch value estimated at the kth iteration, ε is the algorithm step size, It means that the neighboring cells of the main grid will directly receive power supplement from the main grid in each iteration; The optimal solution for the main grid to purchase electricity is as follows: Among them, P M (k+1) is the supplementary power value of the main grid to the neighboring node at time k+1.

9. A low-carbon port distributed energy management method considering computing power network according to claim 1, characterized in that: The distributed energy management algorithm is used to solve the low-carbon port energy management model under the port microgrid island mode, and the optimal energy scheduling scheme at the lower level is obtained, including: In the island mode, the incremental cost is corrected by adding a penalty term to meet the supply and demand balance of the port microgrid. The iterative formula used is as follows: in, is the step size of the algorithm, ξ(k) is the feedback gain; The power generation of each power generation unit is: Among them, P i (k) represents the power generation of the power generation unit of the i-th node at the k-th iteration; Then, according to the known P i (k), estimate the local mismatch value, the formula is as follows: Where ΔP i (k) is the actual local mismatch value at the kth iteration, is the mismatch value estimated at the kth iteration and ε' is the step size of the algorithm.

Images