CN112669912B - Ethylene cracking furnace group scheduling method considering average coking quantity and raw material load - Google Patents

Abstract

The invention discloses a dispatching method of an ethylene cracking furnace group considering average coking quantity and raw material load, which comprises the following steps: obtaining average coking amounts and ethylene product amounts of different cracking raw materials put into different furnace type cracking furnaces, obtaining an objective function of a scheduling model, establishing constraint conditions, constructing an MINLP model, and carrying out optimization solving on the MINLP model by using a DICOPT solver. The method comprehensively considers the characteristic of yield decay with time in ethylene production, considers the change of cracking raw material load, and plans the optimal arrangement of batches, batch processing time, decoking sequence and batch raw material feeding quantity in the scheduling time range for each cracking furnace. In addition, the invention can reduce the emission of pollutants in the decoking stage, obtain more considerable environmental benefit at the cost of a small amount of profit, and provide theoretical basis for energy conservation, emission reduction and optimization production of ethylene plants.

Description

Scheduling method of ethylene cracking furnace group considering average coking amount and feedstock load

Technical Field

The invention relates to the technical field of ethylene production, and in particular to a method for dispatching an ethylene cracking furnace group taking into account average coking amount and raw material load.

Background Art

Ethylene is the most important organic compound in the petrochemical/chemical industry, with production volume far exceeding that of other petrochemical products. A wide range of ethylene derivatives play an extremely important role in people's daily lives, such as ethylene oxide, vinyl acetate, ethylene dichloride, and high/low density polyethylene. Steam thermal cracking is the first operating part of an ethylene plant and largely determines the yield of downstream products and the energy consumption of the entire plant. Today, with increasingly tight margins and volatile feedstock markets, ethylene plants are encouraged to diversify their feedstock portfolio and increase their operational resilience in the face of raw material uncertainties.

With the increasingly stringent requirements for energy conservation, emission reduction, and low-carbon environmental protection in the production of ethylene industry in my country, how to effectively reduce the pollutant emissions of ethylene plants needs special attention, and reducing the amount of coking in the cracking furnace can meet the environmental protection requirements from the root. Therefore, the improvement and optimization of the cracking furnace group system has been put on the agenda. Since most of the previous research focused on the operation optimization of a single cracking furnace, in actual industry, multiple cracking furnaces produce ethylene products in parallel. Therefore, it is necessary to schedule the raw materials for the production of the cracking furnace group to achieve the corresponding optimization goals.

Summary of the invention

In order to solve the limitations and defects of the prior art, the present invention provides a method for scheduling an ethylene cracking furnace group taking into account the average coking amount and the raw material load, comprising:

Obtain the average coking amount and ethylene product amount when different cracking raw materials are put into cracking furnaces of different furnace types;

According to the average coking amount and the ethylene product amount, the objective function of the scheduling model is obtained, and the objective function is used to minimize the average coking amount per unit ethylene output. The expression of the objective function is as follows:

Where, F _i,j,k is the raw material feed rate, represents a yield model of ethylene yield changing dynamically with time when raw material i is cracked in cracking furnace j during intermittent operation, wherein a _i,j , b _i,j and c _i,j are fitting parameters of the yield model;

Establishing constraint conditions, wherein the constraint conditions include raw material balance constraint, time constraint, integer constraint, asynchronous coke removal constraint, variable upper and lower limit constraints and additional constraints;

Considering the change of cracking feedstock load, the ethylene yield model under different load levels is obtained through simulation. The cracking feedstock load is only adjusted and optimized within the feed range when the operating conditions change, and remains unchanged within a single scheduling cycle;

According to the objective function, the constraint conditions and the ethylene yield model, a MINLP model for furnace group scheduling is formed;

The DICOPT solver is used to optimize and solve the MINLP model, and the sub-solver CPLEX and the sub-solver CONOPT are used to handle the MILP problem and the NLP problem respectively.

Optionally, the ethylene yield model is a functional relationship between time and load, and the expression of the ethylene yield model is as follows:

a _i,j (D _i,j )=pma _i,j ×F _i,j +pna _i,j (2)

b _i,j (D _i,j )=pmb _i,j ×F _i,j +pnb _i,j (3)

c _ij (D _ij )=pmc _ij ×F _ij +pnc _ij (4)

Optionally, the expression of the raw material balance constraint is as follows:

G _i ≤(Dup _i -Dlo _i )H (9)

Flo _i,j gy _i,j,k ≤F _i,j,k ≤Fup _i,j gy _i,j,k (10)

The total feed amount for cracking of all cracking furnaces is within the range of raw material supply, the consumption of feed i within the scheduling time range H, Gi represents the excess of feed i exceeding the lower limit, the feed amount of the cracking raw material is adjusted within the preset range, y _i,j,k is a binary logical variable, if the feed i is not allocated to the kth batch production operation of the jth furnace, the feed amount of the cracking raw material F _i,j,k is 0.

Optionally, the expression of the time constraint is as follows:

The continuous processing time of the feed i is set within a preset minimum and maximum value. If the feed i is not assigned to the kth batch production run of the jth furnace, the processing time is 0;

Ts _j,k is the start time of the cracking batches of all cracking furnaces, Te _j,k is the end time of the cracking batches of all cracking furnaces, the start time of the cracking of the current batch is equal to the end time of the cracking of the previous batch plus the shutdown and decoking time between the two batches, the end time of the current batch is equal to the start time of the current batch plus the batch processing time, the start time of the first batch of all cracking furnaces is 0, if the batch processing start time is greater than the scheduling time range H, the start time Ts _j,k is equal to 0;

A binary variable y _i,j,k is introduced to indicate whether the feed i is allocated to the kth batch production run of the jth furnace. If the cracking feed has been allocated, the binary variable y _i,j,k is 1. If the cracking feed has not been allocated, the binary variable y _i,j,k is 0. If the kth batch has no production run, the start and end times of the kth batch are the same.

Optionally, the expression of the integer constraint is as follows:

The first batch of all cracking furnaces is always used to crack feed, and each batch can only crack one type of feed at most. Within the entire scheduling time range, all types of raw materials are processed at least once. The entire scheduling process uses the previous batch and the next batch in sequence. If a batch of a cracking furnace is not used, the subsequent batches of the cracking furnace will not be used.

Optionally, the expression of the asynchronous defocusing constraint is as follows:

The time intervals for decoking different cracking furnaces are prohibited from having overlapping parts. The asynchronous decoking constraint is described using the time difference between two starting points and the time difference between two end points. The product of the time difference between the two starting points and the time difference between the two end points is less than or equal to 0.

Optionally, the expressions of the upper and lower limit constraints of the variables are as follows:

Tp _i,j,k ≥0,G _i ≥0,y _i,j,k ∈{0,1},H>0 (22)

The upper limit of the start time Ts _j,k is 0, the lower limit of the end time Te _j,k is H, the scheduling time range H is greater than 0, the binary logic variable y _i,j,k is 0 or 1, and other variables are not less than 0.

Optionally, the expression of the additional constraint is as follows:

The additional constraint is the daily average profit of each scheduling scheme.

The present invention has the following beneficial effects:

The present invention takes into account the actual production process of the cracking furnace, abandons the assumption in the original model that all batches of a certain raw material enter a certain cracking furnace at the same time, and introduces a binary variable to indicate whether the raw material enters the kth batch of the cracking furnace for cracking, which is more in line with the actual production situation. At the same time, considering the change in the load of the cracking furnace can better optimize the feed load of each raw material for each cracking furnace, as well as the combination of raw materials put into each cracking furnace. The present invention also fully considers the production characteristics of asynchronous coke clearing in the ethylene production process, thereby ensuring the normal and stable scheduling. The present invention can effectively simulate the scheduling production of ethylene chemical furnace groups, effectively improve environmental benefits, and thus effectively deal with the increasingly serious environmental pollution, and can serve as a reference and guide for actual production.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a scheduling schematic diagram of a cracking furnace group scheduling model provided in Example 1 of the present invention.

FIG. 2a is a schematic diagram showing the variation of ethylene cracking yield over time provided in Example 1 of the present invention.

FIG2 b is another schematic diagram of the variation of ethylene cracking yield over time provided in Example 1 of the present invention.

FIG3a is a trend diagram showing the change of the coking amount with the operating time provided in the first embodiment of the present invention.

FIG3 b is another trend diagram of the coking amount versus operating time provided in Example 1 of the present invention.

FIG. 3 c is another trend diagram of the coking amount changing with the operating time provided in the first embodiment of the present invention.

FIG. 4a is a schematic diagram of a basic model scheduling solution provided in Embodiment 1 of the present invention.

FIG. 4 b is another schematic diagram of the basic model scheduling solution provided in the first embodiment of the present invention.

FIG. 4c is another schematic diagram of the basic model scheduling solution provided in the first embodiment of the present invention.

FIG. 5 a is a schematic diagram of an improved model scheduling solution provided in Embodiment 1 of the present invention.

FIG5 b is another schematic diagram of the improved model scheduling solution provided in the first embodiment of the present invention.

FIG. 5c is another schematic diagram of the improved model scheduling solution provided in the first embodiment of the present invention.

DETAILED DESCRIPTION

In order to enable those skilled in the art to better understand the technical solution of the present invention, the method for scheduling an ethylene cracking furnace group taking into account the average coking amount and the feedstock load provided by the present invention is described in detail below with reference to the accompanying drawings.

Embodiment 1

The purpose of this embodiment is to take the average coking amount per unit of ethylene product as the objective function, take into account the change in raw material load, construct an improved ethylene cracking furnace cluster scheduling model, and provide a reliable and efficient furnace cluster scheduling solution after solving it, so as to provide technical support and technical reference for ethylene enterprises to improve production efficiency and reduce environmental pollution. This embodiment establishes the minimization of the average coking amount per unit of ethylene product as the objective function, establishes corresponding constraints through the actual production process of ethylene, takes into account the change in raw material load, models the furnace cluster scheduling problem, and optimizes and solves the constructed model.

The objective function provided in this embodiment needs to introduce the concept of average coking amount. This embodiment defines the average coking amount avecoke as the mass of coke generated on the inner wall of the furnace tube for every ton of ethylene produced by the cracking furnace. Here, the variable output is used to represent the total output of ethylene products in the cracking furnace system within the scheduling time range, and the variable cokemass represents the total coke mass generated in the cracking furnace system within the scheduling time range. The fitting function relationship of the coking amount of the cracking furnace tube obtained by COILSIM1D simulation over time, wherein _P1 , _P2 , _P3 , and _P4 are all fitting parameters. This embodiment sums the ethylene production and the coking amount of the furnace tube of all cracking furnaces and all batches in the cracking furnace group system, thereby calculating the total average coking amount of the furnace group system, and further obtaining the average coking amount.

The constraints provided in this embodiment include raw material balance constraints, time constraints, integer constraints, asynchronous coke removal constraints, upper and lower limit constraints of variables, and additional constraints. This embodiment takes into account the change of raw material load, and the feed amount of cracking raw material (raw material load) can be adjusted within a certain range according to the actual working conditions. The ethylene yield model under different load levels is obtained by simulation. The raw material load is adjusted and optimized within the feed range only when the operating conditions change, and remains unchanged within a scheduling cycle.

This embodiment establishes a scheduling optimization model, and comprehensively considers the above-mentioned objective function and constraint variables to establish a basic MINLP mathematical model for the scheduling of a cracking furnace group. The model provided in this embodiment is solved using GAMS, and the solver for solving the MINLP problem is DICOPT, and the sub-solvers CPLEX and CONOPT are used to handle the MILP and NLP problems respectively. This embodiment provides a scheduling modeling method for an ethylene cracking furnace group that takes into account the average coking amount and load changes, which is used to optimize the furnace group scheduling process. By improving the objective function and constraint conditions of the model, the accuracy of the model can be improved, and better optimization results can be obtained. This embodiment can achieve the acquisition of a scheduling plan under the optimal target value, thereby playing an optimization role and providing a guiding plan.

This embodiment can be effectively applied to a parallel multi-feed, multi-product ethylene cracking furnace group scheduling system. Figure 1 is a scheduling schematic diagram of the cracking furnace group scheduling model provided in Example 1 of the present invention. As shown in Figure 1. Through the optimal scheduling of the system, the average coking amount per unit product can be effectively reduced at the cost of sacrificing a small amount of profit. This embodiment proposes a new MINLP model to obtain a scheduling strategy for maximizing the environmental benefits of the cracking furnace system under the constraint that the product yield decays exponentially over time. Figure 2a is a schematic diagram of the change of ethylene cracking yield over time provided in Example 1 of the present invention. Figure 2b is another schematic diagram of the change of ethylene cracking yield over time provided in Example 1 of the present invention. The relationship between ethylene yield and time is shown in Figures 2a and 2b.

The following information is given in this embodiment: upper and lower limits of raw material load; product yield models of different cracking furnaces cracking different raw materials; decoking time of each cracking furnace when cracking different feeds; upper and lower limits of batch processing time; decoking and configuration costs; product price index; given scheduling time range; fitting function of coking rate of the inner wall of each cracking furnace tube over time. Information that can be determined by optimal scheduling includes: the number of batches assigned to each cracking furnace; the type of feed processed in each batch operation; the starting time point Ts _j,k and the ending time point Te _j,k of each batch cracking operation; and the specific decoking operation sequence of the entire cracking furnace system.

This embodiment establishes an objective function. The objective function of the scheduling model is to minimize the average coking amount per unit of ethylene product. The expression is as follows:

In this embodiment, taking into account the change in raw material load, the functional relationship between ethylene yield and time and load is as follows:

a _i,j (D _i,j )=pma _i,j ×F _i,j +pna _i,j (2)

b _i,j (D _i,j )=pmb _i,j ×F _i,j +pnb _i,j (3)

c _ij (D _ij )=pmc _ij ×F _ij +pnc _ij (4)

This embodiment establishes constraint conditions, which include raw material balance constraints, time constraints, integer constraints, asynchronous coke clearing constraints, upper and lower limit constraints of variables, and additional constraints.

For the raw material balance constraint, the total feed amount of all cracking furnaces should be within the range of raw material supply. The consumption of feed i within the scheduling time range H, where Gi represents the excess of feed i exceeding the lower limit. In this embodiment, the feed amount (raw material load) of the cracking raw material can be adjusted within a certain range according to the actual working conditions. _Yi,j,k is a binary logical variable. If feed i is not allocated to the kth batch production operation of the jth furnace, the feed amount of the cracking raw material F _i,j,k is 0, and the expression is as follows:

G _i ≤(Dup _i -Dlo _i )H (9)

Flo _i,j gy _i,j,k ≤F _i,j,k ≤Fup _i,j gy _i,j,k (10)

For time constraints, from the perspective of management and operation, the continuous processing time of feed i is within the minimum and maximum values actually allowed. If feed i is not assigned to the kth batch production run of the jth furnace, the processing time of the batch is 0. The start time Ts _j,k and end time Te _j,k of the cracking batches of all cracking furnaces are defined. The cracking start time point of the current batch is equal to the cracking end time point of the previous batch plus the shutdown and decoking time between the two batches. The end time point of the current batch is equal to its start time point plus the batch processing time.

Assume that all cracking units are clean at the beginning of the time range and the start time of the first batch of all cracking furnaces is 0. If the batch start time point is greater than the scheduling time range H, it means that the batch is not actually used, that is, Tp _j,k is equal to 0. Introduce a binary variable _yi,j,k to express whether feed i is allocated to the kth batch production run of the jth furnace. If the feed is allocated, then the binary variable _yi,j,k is 1, otherwise it is 0. It should be noted that the start and end times of a batch are arranged in this way: if the kth batch is not actually used, then its start and end time points will be the same, and the expression is as follows:

For integer constraints, in actual production, the first batch of all cracking furnaces must always be used to crack the feed. For all cracking furnaces, each batch can only crack at most one feed. All types of feed must be processed at least once within the entire scheduling time range. During the entire scheduling process, the previous batch must be used before the next batch is used. If a batch of a cracking furnace is not used, the subsequent batches of the cracking furnace will not be used. The expression is as follows:

For the non-synchronous decoking constraint, the time intervals of decoking of different cracking furnaces should not overlap, as shown in Figure 1. The constraint can be described by the time difference between the two starting points and the time difference between the two end points. The product of these two types of time differences should be less than or equal to 0. Note that this constraint cannot be applied to the last cleaning of each furnace, that is, it is feasible to decoke simultaneously in the last production run of all devices. The expression is as follows:

For the upper and lower limit constraints of variables, the upper and lower limits of Ts _j,k and Te _j,k are 0 and H respectively; the scheduling time domain H should be greater than 0; the binary variable y _i,j,k can only be 0 or 1; other variables need to be not less than 0, and the expressions are as follows:

Tp _i,j,k ≥0,G _i ≥0,y _i,j,k ∈{0,1},H>0 (22)

For additional constraints, the average daily profit under each scheduling scheme is used as an additional constraint of the scheduling model, and the expression is as follows:

Through the above steps, a new MINLP optimization model can be established with formula (1) as the objective function and formulas (2)-(23) as constraints. In order to highlight the effect of this embodiment, the old basic model that does not consider the change of raw material load is compared with the new model.

The case provided in this embodiment comes from a domestic ethylene plant. Three types of cracking furnaces are studied, namely GK-VI, GK-V and GK-III, represented by 1, 2 and 3 respectively; three cracking raw materials are processed, light naphtha (NAP), light naphtha (LNAP) and liquefied petroleum gas (LPG), represented by A, B and C respectively. The scheduling time domain H is 200 days. The relevant parameters required in the scheduling model are shown in Table 1:

Table 1 Related parameter values of cracking furnace group system

FIG. 3a is a trend diagram of the coking amount with the running time provided in the first embodiment of the present invention. FIG. 3b is another trend diagram of the coking amount with the running time provided in the first embodiment of the present invention. FIG. 3c is another trend diagram of the coking amount with the running time provided in the first embodiment of the present invention. Wherein, NAP is naphtha, LNAP is light naphtha, LPG is liquefied petroleum gas, GK-VI represents the GK-VI model cracking furnace, GK-V represents the GK-V model cracking furnace, and CBL-III represents the CBL-III model cracking furnace. FIG. 4a is a schematic diagram of the basic model scheduling scheme provided in the first embodiment of the present invention. FIG. 4b is another schematic diagram of the basic model scheduling scheme provided in the first embodiment of the present invention. FIG. 4c is another schematic diagram of the basic model scheduling scheme provided in the first embodiment of the present invention. The raw material batch processing order is: No. 1 cracking furnace: A-B-C; No. 2 cracking furnace: A-B-C; No. 3 cracking furnace: A-B-C. FIG. 5a is a schematic diagram of the improved model scheduling scheme provided in the first embodiment of the present invention. FIG. 5b is another schematic diagram of the improved model scheduling scheme provided in the first embodiment of the present invention. Figure 5c is another schematic diagram of the improved model scheduling scheme provided in Example 1 of the present invention. The raw material batch processing order is: No. 1 cracking furnace: C-A-C-B; No. 2 cracking furnace: B-C-A; No. 3 cracking furnace: A-B-A-C.

As shown in Figures 4a, 4b, 4c and 5a, 5b, 5c, both the basic model and the improved model are solved by GAMS, and the solver is DICOPT. The MINLP problem is decomposed into MILP and NLP sub-problems, which are solved by CPLEX solver and CONOPT solver respectively. The total cycle time is 200 days. The average coking amount of the basic model is 0.173 kg/t ethylene, and the average daily profit is 735,070 yuan/day. The average coking amount of the new improved model is 0.164 kg/t ethylene, and the average daily profit is 764,890 yuan/day. Both the average daily profit and the average coking amount can be improved.

Comparison of the two models in Figure 3a, Figure 3b, Figure 3c and Figure 4a, Figure 4b, Figure 4c shows that the order of raw material processing in the improved model is very different from that before the improvement. The actual production raw materials are not added in batches according to the prescribed order, and the processing time of different batches of the same raw materials is not exactly the same. Therefore, the arrangement of raw material batches is more reasonable, and its average profit is also improved compared with the original model.

It can be seen from the calculation results of Table 2 that the cracking furnace group scheduling modeling optimization method considering the average coking amount can achieve a higher coking amount reduction at the expense of a small amount of profit, thereby achieving better environmental benefits. The present invention fully considers the production characteristics of asynchronous coking during ethylene production, thereby ensuring the normal and stable scheduling. The present invention can effectively simulate the scheduling production of ethylene chemical furnace groups, effectively improve environmental benefits, and effectively respond to the increasingly serious environmental pollution, which can serve as a reference and guidance for actual production.

Table 2 Calculation results of old model and new model

The present embodiment provides a method for scheduling a group of ethylene cracking furnaces taking into account the average coking amount and the raw material load, including: obtaining the average coking amount and the amount of ethylene products of different cracking raw materials put into cracking furnaces of different furnace types, obtaining the objective function of the scheduling model, establishing constraints, constructing a MINLP model, and using a DICOPT solver to optimize and solve the MINLP model. The present embodiment comprehensively considers the characteristics of the yield decaying over time in ethylene production, takes into account the changes in the cracking raw material load, and plans the best arrangement of batches, batch processing time, decoking sequence, and batch raw material feed amount within the scheduling time range for each cracking furnace. In addition, the present embodiment can reduce the emission of pollutants in the decoking stage, obtain more considerable environmental benefits at the cost of sacrificing a small amount of profit, and provide a theoretical basis for energy conservation, emission reduction, and optimized production scheduling of ethylene plants.

It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present invention, but the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims (5)

1. A method for scheduling an ethylene cracking furnace group taking into account average coking amount and feedstock load, characterized in that it comprises: Obtain the average coking amount and ethylene product amount when different cracking raw materials are put into cracking furnaces of different furnace types; According to the average coking amount and the ethylene product amount, the objective function of the scheduling model is obtained, and the objective function is used to minimize the average coking amount per unit ethylene output. The expression of the objective function is as follows: Where, F _i,j,k is the raw material feed rate, represents a yield model of ethylene yield changing dynamically with time when raw material i is cracked in cracking furnace j during intermittent operation, wherein a _i,j , b _i,j and c _i,j are fitting parameters of the yield model; Establishing constraint conditions, wherein the constraint conditions include raw material balance constraint, time constraint, integer constraint, asynchronous coke removal constraint, variable upper and lower limit constraints and additional constraints; Considering the change of cracking feedstock load, the ethylene yield model under different load levels is obtained through simulation. The cracking feedstock load is only adjusted and optimized within the feed range when the operating conditions change, and remains unchanged within a single scheduling cycle; According to the objective function, the constraint conditions and the ethylene yield model, a MINLP model for furnace group scheduling is formed; The MINLP model is optimized and solved using the DICOPT solver, and the MILP problem and the NLP problem are processed using the sub-solver CPLEX and the sub-solver CONOPT respectively; The ethylene yield model is a functional relationship between time and load. The expression of the ethylene yield model is as follows: a _i,j (D _i,j )=pma _i,j ×F _i,j +pna _i,j (2) b _i,j (D _i,j )=pmb _i,j ×F _i,j +pnb _i,j (3) c _ij (D _ij )=pmc _ij ×F _ij +pnc _ij (4) The expression of the raw material balance constraint is as follows: G _i ≤(Dup _i -Dlo _i )H (9) Flo _i,j gy _i,j,k ≤F _i,j,k ≤Fup _i,j gy _i,j,k (10) The total feed amount of all cracking furnaces is within the range of raw material supply, the consumption of feed i within the scheduling time range H, Gi represents the excess of the feed i exceeding the lower limit, the feed amount of the cracking raw material is adjusted within the preset range, _yi,j,k is a binary logical variable, if the feed i is not allocated to the kth batch production operation of the jth furnace, the feed amount of the cracking raw material F _i,j,k is 0; The expression of the time constraint is as follows: The continuous processing time of the feed i is set within a preset minimum and maximum value. If the feed i is not assigned to the kth batch production run of the jth furnace, the processing time is 0; Ts _j,k is the start time of the cracking batches of all cracking furnaces, Te _j,k is the end time of the cracking batches of all cracking furnaces, the cracking start time of the current batch is equal to the cracking end time of the previous batch plus the shutdown and decoking time between the two batches, the end time of the current batch is equal to the start time of the current batch plus the batch processing time, the start time of the first batch of all cracking furnaces is 0, if the batch processing start time is greater than the scheduling time range H, the start time Ts _j,k is equal to 0; A binary variable y _i,j,k is introduced to indicate whether the feed i is allocated to the kth batch production run of the jth furnace. If the cracking feed has been allocated, the binary variable y _i,j,k is 1. If the cracking feed has not been allocated, the binary variable y _i,j,k is 0. If the kth batch has no production run, the start and end times of the kth batch are the same. 2. The method for scheduling an ethylene cracking furnace group considering average coking amount and feedstock load according to claim 1, wherein the expression of the integer constraint is as follows: The first batch of all cracking furnaces is always used to crack feed, and each batch can only crack one type of feed at most. Within the entire scheduling time range, all types of raw materials are processed at least once. The entire scheduling process uses the previous batch and the next batch in sequence. If a batch of a cracking furnace is not used, the subsequent batches of the cracking furnace will not be used. 3. The method for scheduling an ethylene cracking furnace group considering average coking amount and feedstock load according to claim 2, characterized in that the expression of the asynchronous coking constraint is as follows: The time intervals for decoking different cracking furnaces are prohibited from having overlapping parts. The asynchronous decoking constraint is described using the time difference between two starting points and the time difference between two end points. The product of the time difference between the two starting points and the time difference between the two end points is less than or equal to 0. 4. The method for scheduling an ethylene cracking furnace group considering average coking amount and feedstock load according to claim 3, characterized in that the expressions of the upper and lower limit constraints of the variables are as follows: Tp _i,j,k ≥0,G _i ≥0,y _i,j,k ∈{0,1},H>0 (22) The upper limit of the start time Ts _j,k is 0, the lower limit of the end time Te _j,k is H, the scheduling time range H is greater than 0, the binary logic variable y _i,j,k is 0 or 1, and other variables are not less than 0. 5. The method for scheduling an ethylene cracking furnace group considering average coking amount and feedstock load according to claim 4, characterized in that the expression of the additional constraint is as follows: The additional constraint is the daily average profit of each scheduling scheme.