Variable-length multi-cycle scheduling method for ethylene cracking process

Abstract

The invention discloses a variable-length multi-cycle scheduling method in an ethylene cracking process, which brings refined chemical engineering device processing requirements into scheduling optimization and solves the problem that only short time, coarse granularity and static scheduling production scheduling are considered. According to the method, variable-length and multi-cycle scheduling can be timely triggered according to an incoming material plan, furnace shutdown and scorching of the device and a change automatic event of component material requirements, a continuous rolling production scheduling scheme from time granularity to hour is achieved, and on the premise that continuous and stable production of chemical enterprises is met, the production efficiency is improved. Raw material storage, frequent device raw material switching and device raw material matching cost are comprehensively reduced, and the refined management and control level of enterprise scheduling is improved.

Description

Technical Field

[0001] The present invention relates to the production scheduling field of chemical enterprises, and specifically to a refined scheduling method for the processing process of chemical plants. Background Art

[0002] Chemical production enterprises led by ethylene cracking plants are characterized by capital intensity, long production and operation cycles, and continuous and uninterrupted full processes. The whole process of their production and operation has the attributes of logistics and chronological characteristics, covering multiple links such as raw material procurement, shipping schedule arrangement, raw material warehousing, intermediate material warehousing, product warehousing and sales. As the core hub connecting upstream cracking raw materials and downstream plant production, it is particularly important to reasonably schedule and optimize the raw material processing of the ethylene cracking plant.

[0003] At present, enterprises mainly rely on manual experience, consider the current total raw material inventory and total demand, and use EXCEL to formulate the production scheduling plan for the ethylene cracking plant in the next week or so. The time granularity is only at the daily level, which is relatively rough. However, in the process of allocating chemical raw material resources suitable for enterprise adaptive processing, it involves determining the types and quantities of raw materials suitable for the actual raw material warehousing conditions of the enterprise, the transportation arrangements of each raw material, as well as the corresponding cracking furnace type structure and feeding arrangements, the switching production of product grades, the warehousing and logistics arrangements of products, etc. It is a dynamic process with a large cycle span and variable influencing factors, and has extremely high requirements for the timeliness and flexibility of configuration optimization. Simply relying on manual experience to complete the implementation and adjustment of this dynamic configuration process is very cumbersome, time-consuming and inefficient. Therefore, there is an urgent need for a new scheduling method that can adapt to the refined device processing requirements in the current ethylene cracking process to improve production efficiency and scheduling management level. Summary of the Invention

[0004] In view of the above problems, the present invention proposes a variable-length multi-cycle scheduling method for the ethylene cracking process. This method not only considers the downstream resource demand problem, but also considers the connection problem of upstream resources in storage and in transit. By establishing a set of event-triggered variable-length and multi-cycle rolling scheduling optimization models, an optimized production scheduling plan is automatically output, which specifically includes the following steps:

[0005] 1) According to the time points of the incoming material plan, the furnace shutdown and decoking plan of the device, and the change of component material demand, construct a set of state switching time points, and then divide the total scheduling period N into a set of variable-length time periods T;

[0006] 2) Based on event triggering, iteratively roll through these T time periods. According to the time period triggered by the event, divide the raw material incoming material plan into these T periods. Whether it is continuous incoming material from the upstream pipeline or batch incoming material through transportation tools, the incoming material i can be divided into the total incoming material volume F within the time period t YL,i,t, map the processing load capacity of device j to the time period t to obtain the maximum processing volume of device j at time t is the maximum processing speed of device j. Map the coking state of device j to the 0-1 decision variable U indicating whether it is coked within the time period t j,t , 0 means coked, 1 means operating;

[0007] 3) If it is the first time period, construct a set of constraint conditions according to the initial inventory status of the total scheduling period; otherwise, use the end-of-period inventory of the previous round as the initial inventory of this period to construct a set of constraint conditions;

[0008] 4) Iterate through T time periods to construct the scheduling model constraints respectively, as follows:

[0009] a) Determine the decision variables, specifically including: the 0-1 decision variable D for raw material i to enter device j for processing at time t Z,i,j,t , 0 means not processed, 1 means processed, the quantity F of material i allocated from the raw material tank to device j at time t Z,i,j,t , the inventory S of raw material i at the end of time t YL,i,t , the inventory S of the component material at the end of time t P,t ;

[0010] b) Determine the set of constraint conditions, including raw material and device matching constraints, raw material inventory dynamic balance constraints, raw material tank capacity upper and lower limits constraints, component material inventory dynamic balance constraints, component material tank capacity upper and lower limits constraints, device production load constraints, raw material demand constraints, and component material consumption constraints, specifically:

[0011] i. Determine the raw material and device matching constraints. Whether a material can be processed in a device is determined by two parts: one is the 0-1 decision variable that plays an algorithm optimization role, and the other is the conflict constraint between the material and the device:

[0012]

[0013] In formula (1), D Z,i,j,t is the 0-1 decision variable for raw material i to be processed in device j at time t, 0 means not processed, 1 means processed, D CON,i,j is whether material i and device j conflict. If there is a conflict, processing cannot be arranged, 0 means conflict, 1 means no conflict;

[0014] ii. Raw material inventory dynamic balance constraints:

[0015]

[0016] In formula (2), S YL,i,t is the inventory of raw material i at the end of time t, which is the sum of the end-of-period inventory of raw material i at time t-1 and the incoming quantity of raw material i at time t SYL,i,t-1 , minus the consumption of raw material i in all processing units for time t, we get, F Z,i,j,t is the processing amount of raw material i in device j at time t;

[0017] iii. Raw material tank capacity upper and lower limit constraints:

[0018]

[0019] In formula (3) The minimum capacity allowed for the raw material tank is The maximum capacity that the tank of raw material i can bear;

[0020] iv. Dynamic balance constraints on component inventory:

[0021]

[0022] In formula (4), S P,t is the inventory of component materials at the end of time t, which is calculated from the ending inventory S of component materials at time t-1. P,t-1 Add the raw materials being produced by all devices at time t and multiply by the raw material yield to get the output of the component material, minus the amount of component material consumed at time t F XH,t , Y i is the component yield of raw material i;

[0023] v. Constraints on upper and lower limits of component tank capacity:

[0024]

[0025] In formula (5) The minimum capacity allowed for the component tank. The maximum capacity that the component tank can bear;

[0026] vi. The production load of raw materials on the device cannot exceed the device capacity constraint:

[0027]

[0028] In formula (6), Z j,t is the maximum production of device j in time t, U j,t Binding according to the shutdown and charring plan, is a 0-1 decision variable for whether device j is charring at time t, 0 for charring and 1 for operation;

[0029] vii. A device can process only one type of raw material at a time t:

[0030]

[0031] viii. The processing volume of raw materials in all cycles of all devices must meet the production demand:

[0032]

[0033] In formula (8), and are the minimum and maximum amounts of raw material i to be processed in the entire cycle;

[0034] ix. The demand for component materials should satisfy:

[0035]

[0036] In formula (9), F XH,t is the consumption of component materials within time t, and are the minimum and maximum amounts of component materials consumed in the entire cycle;

[0037] x. Number of times of material switching in the device:

[0038]

[0039]

[0040] In formulas (10) and (11), D SWT,i,j is an auxiliary variable, a 0-1 variable indicating whether material i is switched in device j at time t. 0 means no switching, 1 means switching. D i,j,t is a 0-1 decision variable indicating whether device j is processing material i at time t. 0 means not processing, 1 means processing. Similarly, D i,j,t-1 indicates whether it was processed at time t - 1. Also consider that material switching is not counted during furnace shutdown and decoking. U j,t indicates whether the device is decoking at time t. 0 means decoking, 1 means running. Similarly, U j,t-1 indicates whether it was decoking at time t - 1;

[0041] 5) Combine the single-cycle constraint conditions constructed above into the total constraint condition group, and determine whether the time cycle iteration of event triggering is completed. If completed, go to step 6); otherwise, go to step 2);

[0042] 6) Determine the objective function:

[0043]

[0044] In formula (12), min(Cost) indicates that the objective of this objective function is to minimize the operating cost. C YL,i represents the storage cost per unit amount of raw material i, S YL,i,T represents the inventory of raw material i at the last time T, C SWT represents the cost incurred by a single material switch in the device, D SWT,i,j,tA 0-1 variable indicating whether material i switches in device j at time t, where 0 means no switch and 1 means switch, C M,i,j Indicating the matching cost per unit quantity of raw material i and device j, D i,j,t A 0-1 decision variable indicating whether material i is processed in device j, where 0 means not processed and 1 means processed.

[0045] 7) Solve the constructed scheduling model with mixed integer programming and output the complete cycle scheduling plan.

[0046] Beneficial effects:

[0047] The present invention discloses a method for scheduling variable-length multi-cycle in the ethylene cracking process. This method automatically rolls the time period for production scheduling according to the incoming material plan, furnace shutdown and decoking of the device, and changes in component material requirements. It can generate an event-triggered variable-length, multi-cycle rolling optimization production scheduling plan with a time granularity up to hours. On the premise of meeting the continuous and stable production of chemical enterprises, it comprehensively reduces the raw material storage, frequent switching of device raw materials, and the matching cost of device raw materials, and improves the refined management and control level of enterprise scheduling. Brief description of the drawings

[0048] Figure 1 It is the process flow chart of the ethylene cracking process of a certain chemical enterprise in the embodiment of the present invention;

[0049] Figure 2 It is the flow chart of a method for scheduling variable-length multi-cycle in the ethylene cracking process in the embodiment of the present invention;

[0050] Figure 3 It is the Gantt chart output of the processing scheduling plan obtained in the embodiment of the present invention. Detailed implementation manners

[0051] The following combines the drawings and specific calculation examples to give a detailed calculation process and specific operation flow to further illustrate the present invention. This implementation case is implemented on the premise of the technical solution of the present invention, but the protection scope of the present invention is not limited to the following implementation cases.

[0052] The scheduling process of the ethylene cracking device of a certain chemical enterprise involved in the scheduling method of the present invention is as Figure 1 shown. The purchased naphtha is unloaded into the naphtha tank area through the terminal pipeline. The upstream refinery directly transports light naphtha, cracked tail oil, propane, and liquefied gas to the chemical enterprise through the pipeline, and then these raw materials are fed to the ethylene cracking device through the pipeline. The ethylene cracking device will crack to produce main component materials such as ethylene and propylene. Since the relative ratio of propylene yield and ethylene yield is fixed, only ethylene is considered as the component material.

[0053] This example takes the solution of the scheduling plan from 6:00 on January 1, 2025 to 6:00 on January 16, 2025 for a total of 15 days in a certain chemical enterprise. During this period, the enterprise processes a total of 5 raw materials: naphtha, light naphtha, cracked tail oil, propane, and liquefied gas. The initial inventory data of the specific raw materials is shown in Table 1.

[0054] Table 1 Initial inventory of 5 raw materials for processing

[0055]

[0056] The arrival time and quantity of the 5 raw materials planned to arrive within this period are shown in Table 2.

[0057] Table 2 Planned incoming material table

[0058]

[0059] In this example, there are 6 cracking furnaces in the ethylene cracking unit. Each cracking furnace is divided into two hearths, A and B, and each hearth can be processed independently. The upper limit of the processing speed of each hearth is shown in Table 3.

[0060] Table 3 Processing load capacity of cracking furnace hearths

[0061]

[0062]

[0063] The cracking furnace planned to be shut down for decoking during this period is F-101. It will be shut down for decoking from January 10 to January 11. The initial inventory of the cracked component ethylene is 8,800 tons. The lower limit of the ethylene tank is 2,000 tons, and the upper limit is 12,000 tons. The demand for ethylene material has one change during this period, as shown in Table 4.

[0064] Table 4 Planned demand for ethylene material

[0065]

[0066] The process of this embodiment is as Figure 2 shown. The specific implementation steps are as follows:

[0067] 1) According to the incoming material plan, the device shutdown and decoking plan, and the time points of the change in the demand for component materials, construct a set of state switching time points. According to the set of time points, divide the total scheduling period N = 24 * 15 = 360 into 6 time periods of variable lengths. The set of lengths of each time period is T = {72, 48, 96, 48, 24, 72}, with the unit of hours.

[0068] 2) Obtain the feed rate F of various raw materials in each time period YL,i,t , such as the feed rate F of naphtha in the first time periodYL,0,0 = 70000 × 72 / 120 = 42000. The feed rates of each raw material within the period are divided as shown in Table 5.

[0069] Table 5 Feed Rates of Raw Materials in Each Period

[0070]

[0071]

[0072] Obtain the maximum processing capacity of the device for each time period. For example, the maximum processing capacity Z of the cracking furnace F - 101A in the first period j,0 = 35 × 72 = 2520. Determine the decoking variable U during the first period j,0 = 1. The maximum processing capacity and decoking variables of each other furnace in the first period are shown in Table 6. The same applies to other time periods and will not be elaborated further.

[0073] Table 6 Maximum Processing Capacity and Decoking Variables of Each Furnace in the First Period

[0074]

[0075] 3) Establish the constraint condition group in sequence. If it is the first time period, the initial inventory adopts the inventory quantity in Table 1; otherwise, use the end - of - last - period inventory as the beginning - of - this - period inventory.

[0076] 4) Construct the scheduling model constraints as follows: Determine the decision variables, specifically including: the 0 - 1 decision variable D for raw material i entering device j for processing at time t Z,i,j,t , where 0 means not processing and 1 means processing; the quantity F of material i distributed from the raw material tank to device j at time t Z,i,j,t , the inventory quantity S of raw material i at the end of time t YL,i,t , the inventory quantity S of the component material at the end of time t P,t ; Determine the constraint condition group according to Equations (1) to (11).

[0077] 5) Incorporate the constructed single - period constraint conditions into the total constraint condition group. Determine whether the time - period iteration triggered by the event is completed. If completed, go to step 6); otherwise, go to step 2).

[0078] 6) Considering the minimum raw material storage cost, the minimum material switching cost in the device, and the minimum raw material - device matching cost, determine the minimization objective function as:

[0079]

[0080] Where the storage cost C of any raw material YL,i is taken as 100 yuan / ton, and the cost C incurred by a single material switch in the device SWTTake 2,000 yuan / ton, the matching cost C of naphtha, light naphtha and hydrocracking tail oil in cracking furnace F-106 M,i,j Take 300 yuan / ton, the matching cost C of propane and liquefied gas in F-101~F-105 M,i,j Take 300 yuan / ton.

[0081] 7) Solve the above objective function. In this example, the established mixed-integer programming model is optimized and solved in the Python 3.8 environment, and the total scheduling period production plan is output as Figure 3 shown.

[0082] It can be seen that this method realizes an ethylene cracking scheduling and production plan with event-triggered variable-length multi-period rolling for up to 15 days. On the premise of meeting the continuous and stable production of chemical enterprises, it comprehensively reduces raw material storage, frequent switching of raw materials in the device and the matching cost of raw materials in the device, meets the requirements of fine production, and improves production efficiency and scheduling management level.

Claims

1. A method for scheduling ethylene cracking process with variable length and multiple cycles, characterized in that By establishing a scheduling model with mixed integer mathematical programming and variable length and multi-period continuous rolling, the optimized production scheduling plan is output, which specifically includes the following steps: 1) According to the incoming material plan, the shutdown and coking plan of the device, and the time point of the change in the component material demand, a set of state switching time points is constructed, and then the total scheduling period N is divided into a set of indefinite length time periods T; 2) Based on the event trigger, the T time periods are rolled and iterated, and the incoming material quantity, device processing quantity and device charring status are mapped to the T time periods according to the time period triggered by the event; 3) If it is the first time period, the constraint condition group is constructed according to the initial inventory status of the total scheduling period, otherwise the constraint condition group is constructed by taking the final inventory of the previous round as the initial inventory of this period; 4) Iterate T time periods to construct model constraints respectively, including raw material and device matching constraints, raw material inventory dynamic balance constraints, raw material tank capacity upper and lower limit constraints, component material inventory dynamic balance constraints, component material tank capacity upper and lower limit constraints, device production load constraints, raw material demand constraints and component material consumption constraints; 5) Merge the single-cycle constraint condition group constructed above into the total constraint condition group, and determine whether the time cycle iteration triggered by the event is completed. If completed, go to step 6), otherwise go to step 2); 6) Determine the objective function in the model; 7) Solve the constructed model and output the production scheduling plan for the complete cycle.

2. The method for scheduling an ethylene cracking process in an indefinite length and multiple periods according to claim 1, characterized in that In step 1), the time period is cut in an event-driven manner according to the incoming material plan, the device shutdown and coking plan, and the component material demand changes.

3. The method for scheduling an ethylene cracking process in an indefinite length and multiple cycles according to claim 1, characterized in that Map the incoming amount of raw material i into time period t, map the processing amount of device j into time period t, and map the charring state of device j into time period t.

4. The method for scheduling an ethylene cracking process in an indefinite length and multiple periods according to claim 1, characterized in that The dynamic balance constraints of raw material inventory and component material inventory in the established model are as follows: (1) The inventory of raw material i at the end of any time period t is obtained by adding the inventory of raw material i at the end of time period t-1 to the incoming material quantity of raw material i in time period t, minus the consumption of raw material i by all devices in time period t; (2) The inventory of component materials at the end of any time period t is obtained by adding the inventory of component materials at the end of time period t-1 to the amount of component materials produced by all devices in time period t, minus the amount of component materials consumed in time period t.

5. The method for scheduling an ethylene cracking process with variable length and multiple cycles according to claim 1, characterized in that The matching constraints of raw materials and equipment in the established model are as follows: (1) Whether any raw material i and any device j can be processed together is determined by two parts: one is the 0-1 decision variable that plays a role in algorithm optimization, and the other is the conflict constraints between materials and devices; (2) Any device j can process only one type of raw material in the same time period t.

6. The method for scheduling an ethylene cracking process with variable length and multiple cycles according to claim 1, characterized in that The objective function in the established model is: Where min(Cost) indicates that the objective function is to minimize the operating cost, C YL,i represents the storage cost per unit of raw material i, S YL,i,T represents the inventory of raw material i in the last time period T, C SWT Indicates the cost of a single switch of equipment materials, D SWT,i,j,t A 0-1 variable indicating whether material i in device j is switched during time period t, 0 means no switch, 1 means switch, C M,i,j represents the cost of matching the unit quantity of raw material i and device j, D i,j,t A 0-1 decision variable indicating whether material i is processed in device j, 0 for not processing and 1 for processing.

Images