Rolling horizon-based crude oil short-term scheduling optimization method

Abstract

The present invention relates to the technical field of petrochemical engineering, and provides a rolling horizon-based crude oil short-term scheduling optimization method, comprising the following steps: establishing an objective function; storage tank state constraints; material balance constraints; crude oil property constraints; and rolling horizon solving. By means of a complex storage and transportation process, which involves a crude oil tanker arriving at a port to unload crude oil to a proper wharf tank, conveying the crude oil from the wharf tank to a factory tank by means of a pipeline, and finally conveying the crude oil from the factory tank to an atmospheric and vacuum distillation unit by means of a pipeline, the present invention establishes a complete novel crude oil storage and transportation short-term scheduling MINLP model on the basis of a certain coastal refinery in China, and uses a rolling horizon decomposition policy and sets a "safety time slice" and "safety constraint" to ensure the feasibility and optimization of a decomposition model.

Description

A short-term crude oil scheduling optimization method based on rolling horizon Technical Field

[0001] The present invention relates to the technical field of petrochemical industry, and in particular to a crude oil short-term scheduling optimization method based on a rolling time domain. Background Art

[0002] Oil is a non-renewable resource with limited reserves. It is estimated that current oil reserves can only sustain human demand for the next 50 years. Statistics show that in 2018, global oil consumption reached 9.9843*107 barrels per day, a 1.46% increase over the previous year. China's oil consumption alone reached 1.3525*107 barrels per day, a 5.33% increase over the previous year. Faced with the contradiction between huge demand and limited resources, countries around the world, especially China, with its nearly 70% external dependence, need to fully tap the value of low-cost, low-quality, high-sulfur, high-acid crude oil.

[0003] Crude oil blending scheduling (also known as "crude oil storage and transportation") is the first step in the oil processing process, organizing the operational tasks of crude oil unloading, storage, transportation, blending, and distillation for a specific period of time within a refinery. An optimized crude oil blending scheduling scheme can reduce feed switching in oil processing units, ensure smooth operation, avoid tanker demurrage, and reduce costs. It can maximize the processing of low-quality crude oil while meeting unit feed limits, thus generating positive economic benefits for the company. Therefore, crude oil blending scheduling plays a crucial role in the rational utilization of oil resources, and its optimization has attracted considerable attention from scholars.

[0004] With the advancement of computer technology, a number of feasible crude oil scheduling methods have emerged. For example, mixed integer programming (MIP), constraint programming (CP), or event-tree-based optimization methods have been used to guide crude oil transfer. However, these methods require large models and take long to solve. Literature has proposed a crude oil scheduling method based on model predictive control, demonstrating the feasibility of rolling optimization in crude oil scheduling and enhancing the adaptive capabilities of the scheduling model. However, this approach does not take crude oil blending into account. Crude oil scheduling and blending are closely linked and mutually influential in production. On the one hand, when the time, sequence, and quantity of scheduled oil arrivals do not match the blending production requirements, incoming crude oil will occupy the storage tank for a long time, reducing the available tank capacity for turnover. This also makes it difficult to select the appropriate crude oil components for blending, which in turn affects the stable production of the atmospheric and vacuum units. On the other hand, because refined crude oil blending requires consideration of numerous factors, the blending formula often changes according to actual production and processing conditions. This can easily lead to repeated scheduling in existing static scheduling methods, making it difficult for schedulers to cope. Therefore, it is necessary to study how to integrate blending requirements into crude oil scheduling.

[0005] This study focuses on the crude oil pipeline scheduling process of a refinery. By taking blending demand into consideration, lean crude oil scheduling is achieved. A crude oil pipeline scheduling model is established using the MINLP method and solved using the rolling optimization method. This helps the crude oil blending optimization system to dispatch urgently needed crude oil and ensures that blending demand is met from the scheduling supply level. This, while ensuring the continuous and stable production of the refinery, comprehensively reduces crude oil storage and transportation, tanker demurrage costs, and production operating costs, thereby improving the refined management and control level of the enterprise's scheduling.

[0006] Summary of the Invention

[0007] In response to the above problems, the present invention provides a crude oil short-term scheduling optimization method based on a rolling time domain, the purpose of which is to achieve lean crude oil scheduling by taking blending demand into consideration. The MIP method is used to establish a crude oil pipeline scheduling model, which is solved by a rolling optimization method. This assists the crude oil blending optimization system in dispatching urgently needed crude oil and ensures that blending demand is met from the scheduling supply level. Therefore, on the premise of ensuring continuous and stable production of oil refining enterprises, the crude oil storage and transportation, tanker demurrage costs and production operation costs are comprehensively reduced, thereby improving the level of refined management and control of enterprise scheduling.

[0008] To achieve the above objectives, the present invention provides the following technical solutions: a rolling horizon-based short-term crude oil scheduling optimization method, establishing a MINLP model for short-term crude oil scheduling. The scheduling process involves the arrival of a crude oil transport cruise ship and the unloading of crude oil into a suitable terminal tank. The crude oil in the terminal tank is then transported via pipeline to a plant tank. Finally, the crude oil in the plant tank is transported via pipeline to a constant-pressure distillation unit for refining. The optimization objective considers cruise ship docking fees, tank inventory costs, and the number of tank oil transfers. The constraints consider tank status, material balance, crude oil properties, and the processing capacity of the constant-pressure distillation unit. The method specifically includes the following steps:

[0009] S1. Establish the objective function: Minimize the cruise ship docking fee, tank inventory cost, and tank fuel switching times as the model optimization objectives, as follows:

[0010] Among them, the set S T 、S C 、S V 、S WT 、S PT 、S NC Respectively represent the scheduling period, oil type, cruise ships arriving at the port within the scheduling period, terminal tanks, plant tanks, and atmospheric and vacuum devices; the letters t, c, v, wt, pt, and nc represent the codes of the scheduling period, oil type, cruise ships arriving at the port within the scheduling period, terminal tanks, plant tanks, and atmospheric and vacuum devices; C V,v VOL represents the cost of crude oil storage per unit volume at the terminal during the cruise ship v period. V,v,c,tC represents the reserves of the cth type of crude oil in the cruise ship v at the end of the tth time period; WT,wt represents the stagnation cost of unit volume of crude oil in the terminal tank wt per unit period, VOL WT,wt,c,t C represents the reserves of the cth type of crude oil in the terminal tank wt at the end of the tth time period; PT,pt represents the stagnation cost of unit volume of crude oil in the plant tank pt per unit period, VOL WT,wt,c,t represents the reserves of type c crude oil in the plant tank pt at the end of the tth time period; BS pt,i,j,t BS represents the binary decision variable for switching from terminal tank i to terminal tank j (i≠j) when delivering oil to the plant tank pt at the end of the tth time period; nc,i,j,t The binary decision variable representing the switch from plant tank i to plant tank j (i≠j) when delivering oil to the atmospheric and vacuum unit nc at the end of the tth time period; the parameter The weight coefficients of cruise ship docking fees, terminal tank inventory costs, plant tank inventory costs, and tank fuel switching times are respectively represented;

[0011] S2. Tank status constraints: The scheduling process involves two types of tanks located in different locations: terminal tanks and plant tanks. The general principle for setting tank status constraints is that a tank cannot receive and deliver oil at the same time, and after receiving oil, the tank must remain stationary for at least T (no less than the unit cycle time Δt) before it can begin delivering oil. Constraint model:

[0012] Among them, the variable expression B v,wt,t 、B wt,pt,t 、B wt,pt,t and B pt,nc,t They all represent the same type of binary variables and can be uniformly expressed as B a,a′,t , where a, a′ represent the complete set of codes for cruise ships arriving at the port, tanks at the terminal, tanks at the plant area, and atmospheric and vacuum devices. a,a′,t =1, it means that the device corresponding to code a pays oil to the device corresponding to code a′ in the tth time period; a,a′,t =0, it means that the device corresponding to code a has not paid oil to the device corresponding to device a′ in the tth time period; ceil(·) indicates the rounding up operation;

[0013] S3. Material balance constraint: This refers to the material balance between the oil unloading volume of the cruise ship, the oil receiving and delivering volume of the storage tank, and the planned processing volume of the atmospheric and vacuum unit; the constraint model is:

[0014] Among them, VT v,t VT wt,t VT pt,t Respectively represent the volume of crude oil in the cruise ship v, the terminal tank wt, and the plant tank pt during the t-th time period, in m 3 , FV v,t 、FV wt,t 、FV pt,t Respectively represent the oil delivery volume of cruise ship v, terminal tank wt, and plant tank pt in the tth time period, unit: m 3 ;FV v,hi 、FV wt,hi 、FV pt,hi The upper limit of fuel volume paid by cruise ship v, terminal tank wt, and factory tank pt in a unit time period, unit m 3 VT wt,hi VT wt,lo Respectively represent the upper and lower limits of the terminal tank wt storage capacity, unit m 3 VT pt,hi VT pt,lo Respectively represent the upper and lower limits of the plant tank pt storage capacity, unit m 3 ;

[0015] S4. Crude oil property constraints: Based on the known single crude oil information and the requirements of the atmospheric and vacuum unit for the processed crude oil properties; constraint model:

[0016] Among them, the set S M Represents a mixed crude oil collection; m and a represent the codes of the mixed crude oil and the storage tank respectively; the variable VC a,c,t and VM a,m,t Respectively represent the inventory of single crude oil c and mixed crude oil m in storage tank a at the end of time period t, unit: m 3 ; parameter f m,c represents the volume fraction of a single crude oil c in a mixed crude oil m; the set S P represents the crude oil attribute set; p represents the code of crude oil attribute; PRO p,a,t represents the pth property of the mixed crude oil in tank a at the end of the tth time period; Pro p,c represents the pth attribute of a single crude oil c; S5. Constraint on the processing capacity of the atmospheric and vacuum unit: To ensure the stability of the feed rate of the atmospheric and vacuum unit, a maximum of two plant tanks can simultaneously feed oil to a atmospheric and vacuum unit within a unit time period; Constraint model:

[0017] Among them, the collection FNC nc,t It represents the crude oil processing volume of the atmospheric and vacuum unit nc per unit cycle;

[0018] S6. Rolling time domain solution: Divide the long scheduling cycle into multiple consecutive adjacent short cycles and establish multiple scheduling sub-models. Use the measurement results of the previous sub-model as the initial parameters of the next sub-model, solve each sub-model in turn, and set rolling time domain safety constraints based on the simple division of the scheduling cycle.

[0019] The step S6 is a rolling time domain solution: the long scheduling period is divided into multiple consecutive adjacent short periods, and multiple scheduling sub-models are established; the calculation results of the previous sub-model are used as the initial parameters of the next sub-model, each sub-model is solved in turn, and a safe time segment is set based on the simple division of the scheduling period.

[0020] The rolling time domain safety constraint is based on the conventional rolling time domain decomposition processing method: the entire scheduling cycle is evenly divided into n sub-scheduling cycles, namely "rolling time domain segments" (Δt(n)). On this basis, overlapping "safety time segments" Δt'(n) are added for two adjacent sub-scheduling cycles. The value of Δt'(n) is determined based on the fastest unloading time of the tanker and the static time of the tank receiving oil, combined with the computational efficiency of the model.

[0021] Each of the composite sub-scheduling periods is composed of Δt(n) and Δt'(n), but only the storage and transportation plans within the range of Δt(n) are retained during each measurement.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] A novel MINLP model for the short-term scheduling of crude oil storage and transportation was established using a coastal Chinese refinery as a prototype. This model involved a complex storage and transportation process in which a crude oil tanker arrived at the port and unloaded the crude oil into a suitable terminal tank. The crude oil in the terminal tank was then transported via pipeline to a plant tank, and finally to a atmospheric and vacuum unit. A rolling horizon decomposition strategy was adopted, and the feasibility and optimization of the decomposition model were ensured by setting "safety time segments" and "safety constraints." DETAILED DESCRIPTION

[0024] To facilitate understanding of the present invention, the present invention will be described more comprehensively below. Several embodiments of the present invention are given below. However, the present invention can be implemented in different forms and is not limited to the embodiments described in the text. On the contrary, these embodiments are provided to make the content disclosed in the present invention more thorough and comprehensive.

[0025] The following describes an embodiment of the present invention based on its overall structure.

[0026] In this embodiment, a method for optimizing short-term crude oil scheduling based on a rolling time horizon is described. The crude oil storage and transportation process of a coastal refinery can be described as follows: upon arrival at the port, tankers (Vs) dock at idle berths and unload crude oil into suitable terminal tanks (WTs); a portion of the crude oil in the terminal tanks is directly transported to plant tanks (PTs) via long-distance pipelines; and finally, the crude oil in the CTs is sent to atmospheric and vacuum units (CDUs) for processing.

[0027] After understanding the business process, to handle the crude oil blending scheduling problem, the modeler also needs to collect the following information: the length of the scheduling cycle, the expected arrival information of the tanker, the structural details of the refinery, the inventory information and oil storage information of each storage tank, the range of tanker unloading rates, the range of pipeline transmission rates, the production requirements of the atmospheric tower, and the properties and composition ratio of the oil types.

[0028] A MINLP model for short-term crude oil scheduling was established. The scheduling process involves the arrival of a crude oil tanker and the unloading of crude oil into appropriate terminal tanks. The crude oil in the terminal tanks is then transported via pipelines to plant tanks. Finally, the crude oil in the plant tanks is transported via pipelines to the atmospheric and vacuum unit for refining. The optimization objective considers the cost of ship docking, tank inventory costs, and the number of tank transfers. The constraints consider tank status, material balance, crude oil properties, and the processing capacity of the atmospheric and vacuum unit. The model specifically includes the following steps:

[0029] S1. Establish the objective function: Minimize the cruise ship docking fee, tank inventory cost, and tank fuel switching times as the model optimization objectives, as follows:

[0030] Among them, the set S T 、S C 、S V 、S WT 、S PT 、S NC Respectively represent the scheduling period, oil type, cruise ships arriving at the port within the scheduling period, terminal tanks, plant tanks, and atmospheric and vacuum devices; the letters t, c, v, wt, pt, and nc represent the codes of the scheduling period, oil type, cruise ships arriving at the port within the scheduling period, terminal tanks, plant tanks, and atmospheric and vacuum devices; C V,v VOL represents the cost of crude oil storage per unit volume at the terminal during the cruise ship v period. V,v,c,t C represents the reserves of the cth type of crude oil in the cruise ship v at the end of the tth time period; WT,wt represents the stagnation cost of unit volume of crude oil in the terminal tank wt per unit period, VOL WT,wt,c,t C represents the reserves of the cth type of crude oil in the terminal tank wt at the end of the tth time period; PT,pt VOL represents the stagnation cost of a unit volume of crude oil in the plant tank pt per unit period, WT,wt,c,trepresents the reserves of type c crude oil in the plant tank pt at the end of the tth time period; BS pt,i,j,t BS represents the binary decision variable for switching from terminal tank i to terminal tank j (i≠j) when delivering oil to the plant tank pt at the end of the tth time period; nc,i,j,t The binary decision variable representing the switch from plant tank i to plant tank j (i≠j) when delivering oil to the atmospheric and vacuum unit nc at the end of the tth time period; the parameter The weight coefficients of cruise ship docking fees, terminal tank inventory costs, plant tank inventory costs, and tank fuel switching times are respectively represented;

[0031] S2. Tank status constraints: The scheduling process involves two types of tanks located in different locations: terminal tanks and plant tanks. The general principle for setting tank status constraints is that a tank cannot receive and deliver oil at the same time, and after receiving oil, the tank must remain stationary for at least T (no less than the unit cycle time Δt) before it can begin delivering oil. Constraint model:

[0032] Among them, the variable expression B v,wt,t 、B wt,pt,t 、B wt,pt,t and B pt,nc,t They all represent the same type of binary variables and can be uniformly expressed as B a,a′,t , where a, a′ represent the complete set of codes for cruise ships arriving at the port, tanks at the terminal, tanks at the plant area, and atmospheric and vacuum devices. a,a′,t =1, it means that the device corresponding to code a pays oil to the device corresponding to code a′ in the tth time period; a,a′,t =0, it means that the device corresponding to code a has not paid oil to the device corresponding to device a′ in the tth time period; ceil(·) indicates the rounding up operation;

[0033] S3. Material balance constraint: This refers to the material balance between the oil unloading volume of the cruise ship, the oil receiving and delivering volume of the storage tank, and the planned processing volume of the atmospheric and vacuum unit; the constraint model is:

[0034] Among them, VT v,t VT wt,t VT pt,t Respectively represent the volume of crude oil in the cruise ship v, the terminal tank wt, and the plant tank pt during the t-th time period, in m 3 , FVv,t 、FV wt,t 、FV pt,t Respectively represent the oil delivery volume of cruise ship v, terminal tank wt, and plant tank pt in the tth time period, unit: m 3 ;FV v,hi 、FV wt,hi 、FV pt,hi The upper limit of fuel volume paid by cruise ship v, terminal tank wt, and factory tank pt in a unit time period, unit m 3 VT wt,hi VT wt,lo Respectively represent the upper and lower limits of the terminal tank wt storage capacity, unit m 3 VT pt,hi VT pt,lo Respectively represent the upper and lower limits of the plant tank pt storage capacity, unit m 3 ;

[0035] S4. Crude oil property constraints: Based on the known single crude oil information and the requirements of the atmospheric and vacuum unit for the processed crude oil properties; constraint model:

[0036] Among them, the set S M Represents a mixed crude oil collection; m and a represent the codes of the mixed crude oil and the storage tank respectively; the variable VC a,c,t and VM a,m,t Respectively represent the inventory of single crude oil c and mixed crude oil m in storage tank a at the end of time period t, unit: m 3 ; parameter f m,c represents the volume fraction of a single crude oil c in a mixed crude oil m; the set S P represents the crude oil attribute set; p represents the code of crude oil attribute; PRO p,a,t represents the pth property of the mixed crude oil in tank a at the end of the tth time period; Pro p,c represents the pth attribute of a single crude oil c; S5. Constraint on the processing capacity of the atmospheric and vacuum unit: To ensure the stability of the feed rate of the atmospheric and vacuum unit, a maximum of two plant tanks can simultaneously feed oil to a atmospheric and vacuum unit within a unit time period; Constraint model:

[0037] Among them, the collection FNC nc,t It represents the crude oil processing volume of the atmospheric and vacuum unit nc per unit cycle;

[0038] S6. Rolling time domain solution: Divide the long scheduling cycle into multiple consecutive adjacent short cycles and establish multiple scheduling sub-models. Use the measurement results of the previous sub-model as the initial parameters of the next sub-model, solve each sub-model in turn, and set rolling time domain safety constraints based on the simple division of the scheduling cycle.

[0039] Step S6: Rolling time domain solution: The long scheduling cycle is divided into multiple consecutive adjacent short cycles, and multiple scheduling sub-models are established; the measurement results of the previous sub-model are used as the initial parameters of the next sub-model, each sub-model is solved in turn, and a safe time segment is set based on the simple division of the scheduling cycle.

[0040] The rolling time domain safety constraint is based on the conventional rolling time domain decomposition processing method: the entire scheduling cycle is divided into n sub-scheduling cycles, namely "rolling time domain segments" (Δt(n), evenly divided according to the arrival time of the tanker and the scheduling cycle. On this basis, an overlapping "safety time segment" Δt'(n) is added for two adjacent sub-scheduling cycles. The value of Δt'(n) is determined based on the fastest unloading time of the tanker and the static time of the oil storage tank, combined with the computational efficiency of the model.

[0041] Each composite sub-scheduling period consists of Δt(n) and Δt'(n), but only the storage and transportation plans within the range of Δt(n) are retained in each measurement.

[0042] The above is an exemplary description of the present invention. Obviously, the specific implementation of the present invention is not limited to the above-mentioned method. As long as such non-substantial improvements are made by adopting the method concept and technical solution of the present invention, or the concept and technical solution of the present invention are directly applied to other occasions without improvement, they are all within the scope of protection of the present invention.

Claims

1. A crude oil short-term scheduling optimization method based on rolling time domain, characterized in that A MINLP model for short-term crude oil scheduling is established. The scheduling process involves the crude oil transport cruise ship arriving at the port to unload the crude oil into a suitable terminal tank, the crude oil in the terminal tank is transported to the plant tank through a pipeline, and finally the crude oil in the plant tank is transported to the atmospheric and vacuum unit through a pipeline for refining production; its optimization goal takes into account the cruise ship berthing fee, tank inventory cost and tank oil switching times, and the constraints take into account the tank status, material balance, crude oil properties, and atmospheric and vacuum unit processing volume. Specifically, it includes the following steps: S1. Establish the objective function: Minimize the cruise ship berthing fee, tank inventory cost, and tank oil switching times as the model optimization objectives, as follows: Among them, the set S T , S C , S V , S WT , S PT , S NC They represent the dispatching period, oil type, cruise ship arriving at the port within the dispatching period, tank at the terminal, tank at the plant area, and the collection of atmospheric and vacuum devices; the letters t, c, v, wt, pt, and nc represent the codes of the dispatching period, oil type, cruise ship arriving at the port within the dispatching period, tank at the terminal, tank at the plant area, and atmospheric and vacuum devices; C V,v represents the cost of stagnation of a unit volume of crude oil at the terminal within a unit period of cruise ship v, VOL V,v,c,t represents the reserves of the cth type of crude oil in the cruise ship v at the end of the tth time period; C WT,wt represents the stagnation cost of unit volume of crude oil in the terminal tank wt per unit period, VOL WT,wt,c,t represents the reserves of the cth crude oil in the terminal tank wt at the end of the tth time period; C PT,pt It represents the stagnation cost of unit volume of crude oil in the plant tank pt per unit period, VOL WT,wt,c,t represents the reserves of the cth type of crude oil in the plant tank pt at the end of the tth time period; BS pt,i,j,t Indicates that the oil delivery to the plant tank pt at the end of the tth time period is switched from the terminal tank i to the terminal tank j (i≠j) Binary decision variables of BS nc,i,j,t The binary decision variable representing the switch from plant tank i to plant tank j (i≠j) when delivering oil to the atmospheric and vacuum unit nc at the end of the tth time period; parameter They represent the weight coefficients of cruise ship berthing fee, terminal tank inventory cost, plant tank inventory cost and tank oil delivery switching times respectively; S2. Tank status constraint: The scheduling process involves two types of tanks located in different locations: terminal tanks and plant tanks. The general principle for setting tank status constraints is that the tank cannot receive and deliver oil at the same time, and the tank must be stationary for at least T (not less than the unit cycle time Δt) after receiving oil before it can start delivering oil; constraint model: Among them, the variable expression B v,wt,t , B wt,pt,t , B wt,pt,t and B pt,nc,t They all represent the same type of binary variables and can be uniformly expressed as B a,a′,t , where a, a′ represent the complete set of codes for cruise ships arriving at the port, tanks at the terminal, tanks in the plant area, and atmospheric and vacuum devices. a,a′,t =1, it means that the device corresponding to code a pays oil to the device corresponding to code a′ in the tth time period; a,a′,t = 0, it means that the device corresponding to code a has not paid oil to the device corresponding to device a′ in the tth time period; ceil(·) means rounding up; S3. Material balance constraint: that is, consider the material balance between the oil unloading volume of the cruise ship, the oil receiving and paying volume of the storage tank, and the planned processing volume of the atmospheric and vacuum device; constraint model: Among them, VT v,t VT wt,t VT pt,t FV represents the volume of crude oil in the cruise ship v, terminal tank wt, and plant tank pt in the tth time period, in m3. v,t 、FV wt,t 、FV pt,t They represent the oil delivery of cruise ship v, terminal tank wt, and plant tank pt in the tth time period, in m3; FV v,hi 、FV wt,hi 、FV pt,hi The upper limit of the oil volume paid by the cruise ship v, the terminal tank wt, and the factory tank pt in a unit time period, unit m 3 VT wt,hi VT wt,lo Respectively represent the upper and lower limits of the terminal tank wt storage, unit m 3 VT pt,hi VT pt,lo Respectively represent the upper and lower limits of the plant tank pt storage capacity, unit m 3 ; S4. Crude oil property constraints: Based on the known single crude oil information and the requirements of the atmospheric and vacuum distillation unit for the properties of the processed crude oil; Constraint model: Among them, the set S M represents the mixed crude oil set; m and a represent the codes of the mixed crude oil and the storage tank respectively; the variable VC a,c,t and VM a,m,t They represent the inventory of single crude oil c and mixed crude oil m in storage tank a at the end of the tth time period, in m3; parameter f m,c represents the volume fraction of a single crude oil c in a mixed crude oil m; the set S P represents the crude oil attribute set; p represents the code of crude oil attribute; PRO p,a,t represents the pth property of the mixed crude oil in tank a at the end of the tth time period; Pro p,c represents the pth property of a single crude oil c; S5. Constraints on the processing volume of the atmospheric and vacuum unit: To ensure the stability of the feed volume of the atmospheric and vacuum unit, at most two plant tanks can simultaneously feed oil to one atmospheric and vacuum unit within a unit time period; Constraint model: Among them, the collection FNC nc,t It represents the crude oil processing volume of the atmospheric and vacuum unit nc in a unit cycle; S6. Rolling time domain solution: The long scheduling cycle is divided into multiple consecutive adjacent short cycles, and multiple scheduling sub-models are established; the calculation results of the previous sub-model are used as the initial parameters of the next sub-model, and each sub-model is solved in turn. And the rolling time domain safety constraints are set based on the simple division of the scheduling cycle.

2. The crude oil short-term scheduling optimization method based on rolling time domain according to claim 1 is characterized in that: The step S6 is a rolling time domain solution: the long scheduling cycle is divided into multiple consecutive adjacent short cycles, and multiple scheduling sub-models are established; the calculation result of the previous sub-model is used as the initial parameter of the next sub-model, each sub-model is solved in turn, and a safe time segment is set based on a simple division of the scheduling cycle.

3. The crude oil short-term scheduling optimization method based on rolling time domain according to claim 2 is characterized in that: The rolling time domain safety constraint is based on the conventional rolling time domain decomposition processing method: the entire scheduling cycle is divided into n sub-scheduling cycles, namely "rolling time domain segments" (Δt(n) on the basis of evenly dividing the oil tanker arrival time and the scheduling cycle, and adding overlapping "safety time segments" Δt′(n) for two adjacent sub-scheduling cycles. The value of Δt′(n) is determined based on the fastest oil unloading time of the oil tanker and the standing time of the oil storage tank, combined with the calculation efficiency of the model.

4. The crude oil short-term scheduling optimization method based on rolling time domain according to claim 3 is characterized in that: Each of the composite sub-scheduling cycles is composed of Δt(n) and Δt′(n), but only the storage and transportation schemes within the range of Δt(n) are retained in each calculation.

Images