Cracking depth optimization method and system for ethylene cracking furnace group and storage medium

Abstract

The invention provides a cracking depth optimization method for an ethylene cracking furnace group, a cracking depth optimization system for the ethylene cracking furnace group and a computer readable storage medium. The cracking depth optimization method for the ethylene cracking furnace group provided by the invention comprises the following steps: determining a multi-dimensional optimization variable based on each operation variable of each cracking furnace, determining constraint conditions of the optimization variable according to upper and lower limit constraints of the optimization variable, and determining an optimization objective function based on an ethylene yield minimum constraint value of the cracking furnace group; the multi-dimensional optimization variables are input into a deep optimization model, the deep optimization model determines the cracking depth value of the optimized cracking furnace group within the constraint condition range of the optimization variables according to the optimization objective function, and the deep optimization model is constructed by an enhanced exploration mechanism whale optimization algorithm.

Description

Technical Field

[0001] The present invention relates to the field of ethylene production, and in particular to a method for optimizing the cracking depth of an ethylene cracking furnace group, a system for optimizing the cracking depth of an ethylene cracking furnace group, and a computer-readable storage medium. Background Art

[0002] Ethylene is a very important petrochemical product and one of the basic raw materials for manufacturing organic products in the entire petrochemical industry. The technology, output and scale of ethylene indicate the level of development of a country's petrochemical industry. The ethylene cracking furnace is the main device in the thermal cracking process of olefins. Its operation level is directly related to the yield change of various products and affects the economic benefits of the cracking unit. Since the composition and operation process of the cracking furnace are very complex, modeling, simulation and optimization of the cracking process are currently a common and effective means to improve the economic benefits of the ethylene industry.

[0003] Cracking depth is an important indicator to measure the degree of reaction in the cracking furnace. There are many parameters to characterize the cracking depth, such as methane yield, methane to propylene yield ratio, propylene to ethylene yield ratio (propylene to ethylene ratio), cracking furnace cracking gas outlet temperature (Coil Outlet Temperature, COT), etc. But no matter which parameter is used to characterize the cracking depth, the controlled variable is COT. Existing ethylene cracking furnace devices all use COT control to adjust the cracking depth, and the selection of COT set values ​​is often based on the specified value or empirical value under the design raw material conditions provided by the cracking furnace manufacturer. However, faced with factors such as complex cracking raw material sources and large fluctuations in oil product properties, the fixed COT production operation mode cannot timely reflect the changes in cracking depth and cracking product yield caused by changes in cracking raw material composition and cracking furnace operation conditions.

[0004] In order to overcome the above-mentioned defects of the prior art, there is an urgent need in the art for a cracking depth optimization technology for an ethylene cracking furnace group, which can determine the optimal cracking depth in real time according to the real-time operation status of the current cracking furnace. Summary of the invention

[0005] A brief summary of one or more aspects is given below to provide a basic understanding of these aspects. This summary is not an exhaustive overview of all conceived aspects, and is neither intended to identify the key or critical elements of all aspects nor to define the scope of any or all aspects. Its only purpose is to give some concepts of one or more aspects in a simplified form as a prelude to a more detailed description that will be given later.

[0006] In order to overcome the above-mentioned defects in the prior art, the present invention provides a cracking depth optimization technology for an ethylene cracking furnace group, which can determine the optimal cracking depth in real time according to the real-time operation status of the current cracking furnace.

[0007] Specifically, the above-mentioned cracking depth optimization method for an ethylene cracking furnace group provided according to the first aspect of the present invention comprises the steps of: determining a multidimensional optimization variable based on each operating variable of each cracking furnace, determining the constraint conditions of the optimization variable according to the upper and lower limit constraints of the optimization variable, and determining the optimization objective function based on the minimum constraint value of the ethylene production of the cracking furnace group; and inputting the multidimensional optimization variable into a deep optimization model, wherein the deep optimization model determines the optimized cracking depth value of the cracking furnace group within the constraint conditions of the optimization variable according to the optimization objective function, wherein the deep optimization model is constructed by an enhanced exploration mechanism whale optimization algorithm.

[0008] Preferably, in one embodiment of the present invention, the step of determining the optimized cracking depth value of the cracking furnace group includes: within the scope of the constraints, utilizing the enhanced exploration mechanism whale optimization algorithm, by iteratively calculating and adjusting the multidimensional optimization variables, to determine the optimized cracking depth value of the cracking furnace group.

[0009] Preferably, in one embodiment of the present invention, within the scope of the constraints, the step of determining the optimized cracking depth value of the cracking furnace group by iteratively calculating and adjusting the multi-dimensional optimization variables using the enhanced exploration mechanism whale optimization algorithm includes: S1: initializing control parameters and a group of whale populations, and determining the best whale individual according to the fitness function, wherein the whale population includes n whale individuals; S2: updating the control parameters, and determining the first position vector and the fitness function value of each whale individual according to the control parameters, wherein the dimension of the first position vector is within the upper and lower limit constraints of the dimension; S3: using the reverse learning method to determine the second position vector and the fitness function value of each whale individual, wherein the dimension of the second position vector is within the upper and lower limit constraints of the dimension; S4: using the random alternative method to determine The third position vector of each individual whale and its fitness function value, the dimension of the third position vector is within the upper and lower limit constraints of the dimension; S5: using a random allocation method to determine the fourth position vector of each individual whale and its fitness function value, the dimension of the fourth position vector is within the upper and lower limit constraints of the dimension; S6: according to the fitness function value, the first position vector, the second position vector, the third position vector and the fourth position vector of all determined whale individuals are arranged in ascending order, the first n position vectors are selected as the positions of each whale individual in the next generation of whale population, and the position vector of the whale individual with the highest fitness function value is updated as the best whale individual; and S7: repeating S2 to S6 until the number of iterations reaches the maximum number of population iterations or the convergence condition of the deep optimization model is met.

[0010] Preferably, in one embodiment of the present invention, the control parameters include the maximum number of population iterations, the population size, the target prey position, the current whale position, the current number of iterations, the first coefficient vector, the second coefficient vector, the first variable for controlling the position update method, the second variable for controlling the coefficient vector, the logarithmic spiral shape constant, and a random number.

[0011] Preferably, in an embodiment of the present invention, the step of determining the first position vector of each individual whale according to the control parameter includes: in response to the first variable of the control update position mode being less than 0.5 and the first coefficient vector being less than 1, determining the first position vector according to a shrinking and surrounding mechanism with a convergent adaptive weighted strategy; in response to the first variable being less than 0.5 and the first coefficient vector being greater than or equal to 1, determining the first position vector according to a random update mechanism with a convergent adaptive weighted strategy; and in response to the first variable being greater than or equal to 0.5, determining the first position vector according to a spiral update position mechanism.

[0012] Preferably, in one embodiment of the present invention, the steps S2 to S5 include: when the dimensions of the first position vector and / or the second position vector and / or the third position vector and / or the fourth position vector violate the upper and lower limit constraints of the dimensions, updating the values ​​of the dimensions.

[0013] Preferably, in an embodiment of the present invention, the steps S2 to S5 further include: converting the first position vector and / or the second position vector and / or the third position vector and / or the fourth position vector into an integer vector according to a coding conversion function.

[0014] Preferably, in one embodiment of the present invention, the optimization objective function includes maximizing ethylene yield and / or maximizing propylene yield and / or maximizing diene yield.

[0015] Preferably, in an embodiment of the present invention, the step of determining the constraint conditions of the optimization variables according to the upper and lower limit constraints of the optimization variables comprises: discretizing the upper and lower limits of the optimization variables to determine the constraint conditions of the optimization variables.

[0016] In addition, the above-mentioned cracking depth optimization system for ethylene cracking furnace group provided according to the second aspect of the present invention includes a memory and a processor. The memory stores computer instructions. The processor is connected to the memory and is configured to execute the computer instructions stored in the memory to implement the cracking depth optimization method for ethylene cracking furnace group provided in any one of the above embodiments.

[0017] In addition, the computer readable storage medium provided according to the third aspect of the present invention stores computer instructions. When the computer instructions are executed by the processor, the method for optimizing the cracking depth of an ethylene cracking furnace group provided in any one of the above embodiments is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above features and advantages of the present invention can be better understood after reading the detailed description of the embodiments of the present disclosure in conjunction with the following drawings. In the drawings, the components are not necessarily drawn to scale, and components with similar related properties or features may have the same or similar reference numerals.

[0019] Figure 1 A flow chart of a method for optimizing cracking depth for an ethylene cracking furnace group according to some embodiments of the present invention is shown;

[0020] Figure 2 A schematic diagram of an ethylene cracking furnace provided according to some embodiments of the present invention is shown;

[0021] Figure 3A flowchart of a whale optimization algorithm of a reinforced exploration mechanism for a deep optimization model provided according to some embodiments of the present invention is shown;

[0022] Figure 4 A schematic diagram showing a spiral update position mechanism provided according to some embodiments of the present invention is shown;

[0023] Figure 5 A rendering showing a method for optimizing cracking depth for an ethylene cracking furnace group; and

[0024] Figure 6 A rendering of the effect of a cracking depth optimization method for an ethylene cracking furnace group is shown.

[0025] Reference numerals:

[0026] 100: Method for optimizing cracking depth of ethylene cracking furnace group;

[0027] S110~S120: steps;

[0028] 200: Ethylene cracking furnace;

[0029] 201: Liquid raw material feeding pipeline;

[0030] 202: Boiler feed water preheater;

[0031] 203: dilution steam distributor;

[0032] 204: Phosphorus-free feedwater preheater;

[0033] 205: Ultra-high pressure steam superheater;

[0034] 206: fuel gas feed line;

[0035] 207: Steam drum;

[0036] 208: sewage device;

[0037] 209: cracking gas outlet;

[0038] S1 to S7: steps; and

[0039] 501, 502, 601, 602: lines. DETAILED DESCRIPTION

[0040] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. Note that the aspects described below in conjunction with the accompanying drawings and specific embodiments are only exemplary and should not be construed as limiting the scope of protection of the present invention in any way.

[0041] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0042] In addition, the terms "upper", "lower", "left", "right", "top", "bottom", "horizontal" and "vertical" used in the following description should be understood as the directions shown in the paragraph and the related drawings. Such relative terms are only used for the convenience of description and do not mean that the device described therein must be manufactured or operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0043] It is understood that although the terms "first", "second", "third", etc. may be used herein to describe various components, regions, layers and / or parts, these components, regions, layers and / or parts should not be limited by these terms, and these terms are only used to distinguish different components, regions, layers and / or parts. Therefore, the first component, region, layer and / or part discussed below may be referred to as a second component, region, layer and / or part without departing from some embodiments of the present invention.

[0044] As mentioned above, existing ethylene cracking furnace devices all use COT control to adjust the cracking depth, and the selection of COT setting values ​​is often based on the specified value or empirical value under the design raw material conditions provided by the cracking furnace manufacturer. However, faced with factors such as the complex source of cracking raw materials and large fluctuations in oil product properties, the fixed COT production operation mode cannot timely reflect the changes in cracking depth and cracking product yield caused by changes in cracking raw material composition and cracking furnace operation conditions.

[0045] In order to overcome the above-mentioned defects in the prior art, the present invention provides a cracking depth optimization technology for an ethylene cracking furnace group, which can determine the optimal cracking depth in real time according to the real-time operation status of the current cracking furnace.

[0046] In some non-limiting embodiments, the above-mentioned cracking depth optimization method for an ethylene cracking furnace group provided by the first aspect of the present invention can be implemented via the above-mentioned cracking depth optimization system for an ethylene cracking furnace group provided by the second aspect of the present invention. Specifically, the cracking depth optimization system for an ethylene cracking furnace group can be configured with a memory and a processor. The memory includes but is not limited to the above-mentioned computer-readable storage medium provided by the third aspect of the present invention, on which computer instructions are stored. The processor is connected to the memory and is configured to execute the computer instructions stored on the memory to implement the cracking depth optimization method for an ethylene cracking furnace group provided by the first aspect of the present invention.

[0047] The following will first describe the working principle of the above-mentioned cracking depth optimization system for ethylene cracking furnace group in conjunction with some embodiments of the cracking depth optimization method for ethylene cracking furnace group. Those skilled in the art will understand that these embodiments of the cracking depth optimization method for ethylene cracking furnace group are only some non-restrictive implementation methods provided by the present invention, which are intended to clearly demonstrate the main concept of the present invention and provide some specific schemes that are convenient for the public to implement, rather than to limit all functions or all working modes of the cracking depth optimization system for ethylene cracking furnace group. Similarly, the cracking depth optimization system for ethylene cracking furnace group is also only a non-restrictive implementation method provided by the present invention, and does not constitute a limitation on the execution subject and execution order of each step in these cracking depth optimization methods for ethylene cracking furnace group.

[0048] Please refer to Figure 1 , Figure 1 A flow chart of a method for optimizing cracking depth of an ethylene cracking furnace group provided according to some embodiments of the present invention is shown.

[0049] like Figure 1 As shown, the cracking depth optimization method for an ethylene cracking furnace group may include step S110: determining a multidimensional optimization variable based on each operating variable of each cracking furnace, determining the constraint conditions of the optimization variable according to the upper and lower limit constraints of the optimization variable, and determining the optimization objective function based on the minimum constraint value of ethylene production of the cracking furnace group.

[0050] Please refer to Figure 2 , Figure 2 A schematic diagram of an ethylene cracking furnace provided according to some embodiments of the present invention is shown.

[0051] like Figure 2As shown, in a non-limiting embodiment, the ethylene cracking furnace 200 is a STR-V type cracking furnace, including a liquid raw material feed pipeline 201, a boiler feed water preheater 202, a dilution steam distributor 203, a phosphorus-free feed water preheater 204, an ultra-high pressure steam superheater 205, a fuel gas feed pipeline 206, a steam drum 207, a blowdown device 208, a cracking gas outlet 209 and a quench heat exchanger 210. The hydrocarbon raw material can be fed into the ethylene cracking furnace 200 through the liquid raw material feed pipeline 201, mixed with the dilution steam after preheating, and then cracked under high temperature and oxygen-free conditions to generate small molecular olefins. After the cracking is completed, the high-temperature cracked gas can be quickly cooled by the quench heat exchanger 210 to recover the heat of the cracked gas to generate ultra-high pressure steam, and finally the cracked gas is discharged from the cracked gas outlet 209 to enter the subsequent treatment.

[0052] The operating variables of a single cracking furnace may include operating conditions and raw material properties. The operating conditions may include information such as feed load, dilution steam ratio, feed temperature, feed pressure, outlet temperature, outlet pressure, etc. The raw material properties may be oil product properties, including raw material density, mass percentage of alkanes and olefins, etc., real-time data of cracking gas online analysis instruments, etc.

[0053] Based on each operating variable of each cracking furnace, the cracking depth optimization system can construct the input elements of multidimensional optimization variables, and the number of rows of the matrix of the input elements of the multidimensional optimization variables is the number of cracking furnaces in the cracking furnace group, and the number of columns is the number of each operating variable selected. In an example, each column can be outlet temperature, dilution steam ratio, feed flow rate, outlet pressure, cross section feed temperature, and cross section feed pressure. The input elements of each cracking furnace can constitute a multidimensional optimization variable.

[0054] The optimization variable of each cracking furnace can be the outlet temperature (COT) in the operating conditions. According to the upper and lower limits of the optimization variable COT, the constraints of the optimization variable can be determined. Here, the cracking depth optimization system can discretize the original continuous optimization problem into a combinatorial optimization problem, that is, discretize the upper and lower limits of the optimization variable COT, and the feasible set formed after discretization is the constraints of the optimization variable.

[0055] Preferably, the cracking depth optimization system can establish a cracking furnace mechanism model to predict products based on the furnace type information of the currently operating cracking furnace, thereby determining the yields of key components such as ethylene and propylene under the current working conditions. The cracking furnace mechanism model can be a mechanism model established based on chemical engineering principles and chemical reaction engineering, in which the cracking reaction mechanism is a free radical reaction mechanism.

[0056] According to the working condition demand on site, the cracking depth optimization system can determine the ethylene output minimum constraint value of the cracking furnace group, and further select the optimization target function.Here, the optimization target function can include ethylene yield maximization, propylene yield maximization or diene yield maximization, and in a preferred embodiment, the optimization target function is ethylene yield maximization or diene yield maximization.According to the ethylene actual output situation, the cracking depth optimization system can choose to improve ethylene output as the optimization target, and can also be used as the optimization target according to the selected optimization target function.For example, when the ethylene actual output was lower than the ethylene output minimum constraint value, the cracking depth optimization system selected to improve ethylene output as the optimization target.

[0057] like Figure 1 As shown, the cracking depth optimization method for an ethylene cracking furnace group may include step S120: inputting multidimensional optimization variables into a depth optimization model, the depth optimization model determining the cracking depth value of the optimized cracking furnace group within the constraint range of the optimization variables according to the optimization objective function, wherein the depth optimization model is constructed by an enhanced exploration mechanism whale optimization algorithm.

[0058] Specifically, the cracking depth optimization system can use the enhanced exploration mechanism whale optimization algorithm within the scope of constraints to determine the cracking depth value of the optimized cracking furnace group through iterative calculation and adjustment of multi-dimensional optimization variables.

[0059] Please refer to Figure 3 , Figure 3 A flowchart of a whale optimization algorithm using an enhanced exploration mechanism for a deep optimization model provided according to some embodiments of the present invention is shown.

[0060] like Figure 3 As shown, the deep optimization model constructed by the whale optimization algorithm with enhanced exploration mechanism first executes step S1: initializing control parameters and a set of whale populations, and determining the best whale individual according to the fitness function, wherein the whale population includes n whale individuals.

[0061] Here, the control parameters include the maximum number of population iterations Max_iters, the population size population_size, the target prey position X best (i.e., the current best whale individual), the current whale position X(t), the current number of iterations t, the first coefficient vector A, the second coefficient vector C, the first variable p that controls the method of updating the position, the second variable h that controls the range of the coefficient vector A, the logarithmic spiral shape constant b, and the random number l. Among them, the first variable p controls the method of updating the position, including shrinking the encirclement and spiral updating, etc., the value of the second variable h decreases linearly from 2 to 0, the logarithmic spiral shape constant b is generally set to 0, and the random number l can be a random number in [0, 1].

[0062] In an initialized group of whale populations, the position vector of the i-th whale individual is:

[0063] X i (t) = [x 0，i (t), x 1，i (t), ..., x N-1，i (t)],

[0064] Each dimension corresponds to the cracking depth value of a cracking furnace in the cracking furnace group, and thus each dimension is constrained by the upper and lower limits of the cracking depth of the corresponding cracking furnace. Here, the upper and lower limits of the cracking depth of the corresponding cracking furnace can be:

[0065] x j，min ≤x j，i (t)≤x j，max j=0,1,2…,N-1,

[0066] Among them, the subscript j represents the number of dimensions, that is, the jth cracking furnace; the subscript i represents the ith individual in the whale population. The number of individuals in the whale population can be determined by the population size population_size. In this example, the number of individuals in the whale population is n. Therefore, the formula for randomly initializing the upper and lower limits of each cracking furnace is:

[0067] x j，i (0) = x j，min +rand·(x j，max -x j，min ),

[0068] Where rand is a random number between 0 and 1. Repeat the above random initialization steps to generate the initial position of each whale individual in the whale population. By calculating the fitness function value of each whale individual, the whale individual with the best fitness function value is selected and marked as the best whale individual X. best .

[0069] Afterwards, when the number of iterations iter is less than the maximum number of population iterations Max_iters, step S2 is executed: the control parameters are updated, and the first position vector and the fitness function value of each individual whale are determined according to the control parameters, and the dimension of the first position vector is within the upper and lower limit constraints of the dimension.

[0070] Here, the cracking depth optimization can update the first coefficient vector A, the second coefficient vector C, the first variable p that controls the update position, the second variable h that controls the range of the coefficient vector A, and the random number l in the control parameters.

[0071] The cracking depth optimization system first updates the second variable h in the range of the control coefficient vector A:

[0072]

[0073] Afterwards, the first coefficient vector A is updated according to the updated second variable h, and the second coefficient vector C, the first variable p and the random number l that control the update position are continued to be updated:

[0074] A=2h·rand-h,

[0075] C = 2 rand,

[0076] p = rand(0, 1),

[0077]

[0078] Afterwards, in a preferred embodiment, based on the first variable p and the first coefficient vector A in the updated control parameters, the lysis depth optimization system can determine a method for updating the position, and determine the first position vector of each individual whale based on the determined method for updating the position.

[0079] Specifically, when the first variable p is less than 0.5, the first coefficient vector A is judged. When the first coefficient vector A is less than 1, the cracking depth optimization system can determine the first position vector X according to the shrinking and encircling mechanism with a convergent adaptive weighting strategy. local (t+1).

[0080] The update formula of the shrinking and encircling mechanism with convergent adaptive weighted strategy is as follows:

[0081] X local (t+1)=r×X best (t)-A·B,

[0082] The calculation formula of weight r is:

[0083]

[0084] The calculation formula of B is:

[0085] B=|C·X best (t)-X local (t)|,

[0086] Among them, X local (t) is the position vector X(t) of the individual whale selected to update its position in the current population.

[0087] In response to the first variable p being less than 0.5 and the first coefficient vector A being greater than or equal to 1, the cracking depth optimization system may determine the first position vector X according to a random update mechanism with a convergent adaptive weighting strategy. local (t+1).

[0088] The update formula of the random update mechanism with convergent adaptive weighted strategy is as follows:

[0089] X local (t+1)=r×X best (t)-A·B,

[0090] The calculation formula of the weight r is the same as the shrinking and encircling mechanism with convergent adaptive weighting strategy:

[0091]

[0092] The calculation formula of B is:

[0093] B=|C·X rand (t)-X local (t)|,

[0094] Among them, X rand (t) is the position vector of a randomly selected whale individual in the current population.

[0095] In addition, when the first variable p is greater than or equal to 0.5, the cracking depth optimization system can determine the first position vector X according to the spiral update position mechanism. local (t+1).

[0096] Please refer to Figure 4 , Figure 4 A schematic diagram of a spiral update position mechanism provided according to some embodiments of the present invention is shown.

[0097] like Figure 4 As shown in the figure, the horizontal axis 1 represents the parameters in the spiral update position mechanism, and the vertical axis x is the position of the whale. The spiral update position mechanism can imitate the spiral movement of humpback whales when hunting. First, the whale at (X, Y) and the whale at (X * , Y * ) and then create a spiral equation between the positions of the whale and the prey to simulate the spiral movement of the whale.

[0098] The equation for this spiral is as follows:

[0099] X local (t+1)=X best (t)+B p ·e bl cos(2πl),

[0100] Among them, B p The calculation formula is as follows:

[0101] B p =|X best (t)-X local (t)|.

[0102] In this way, the first position vector of each individual whale in the whale population is updated according to the control parameters to complete the update of the first position vectors of all individual whales in the whale population.

[0103] Furthermore, each dimension of the first position vector of the determined whale individual is within the upper and lower limit constraints of the corresponding dimension, that is, within the upper and lower limit constraints of the operating variable corresponding to the dimension. When the dimension of the first position vector violates the upper and lower limit constraints of the corresponding dimension, the value of the dimension needs to be updated so that each dimension is within the upper and lower limit constraints of the corresponding dimension. Specifically, the value of the dimension can be updated by the following formula:

[0104]

[0105]

[0106] Among them, x j，i (t) is the jth dimension of the i-th whale individual in the t-th iteration of the whale population, x j，i new is the jth dimension in the position vector of the i-th whale individual to be updated. Here, the position vector is the first position vector X local (t+1).

[0107] Afterwards, the fitness function value of each whale individual is calculated according to the updated first position vectors of all whale individuals, so as to be used for the subsequent determination of the best whale individual X for the next generation of whale population. best (t+1).

[0108] Please continue to refer to Figure 3 , perform steps S3 to S5 respectively:

[0109] S3: Use the reverse learning method to determine the second position vector and fitness function value of each individual whale, and the dimension of the second position vector is within the upper and lower limit constraints of the dimension;

[0110] S4: Use the random selection method to determine the third position vector and fitness function value of each individual whale, and the dimension of the third position vector is within the upper and lower limit constraints of the dimension;

[0111] S5: Use the random allocation method to determine the fourth position vector and fitness function value of each individual whale, and the dimension of the fourth position vector is within the upper and lower limit constraints of the dimension.

[0112] For step S3, the reverse solution set of the solution set of the whale population can be obtained by using the reverse learning method. Specifically, in the multidimensional space, assuming that the solution set of the whale population is: IP = (X1, X2, .., X n), that is, the set of all current whale positions X(t), so the reverse solution set of the whale population is: OP = (X 1op , X 2op , .., X nop ), the search range of the solution space is: [Lb, Ub], so the reverse solution set OP of the solution set IP of the whale population can be realized by the following equation:

[0113] X iop =Lb+Ub-X i ,

[0114] Among them, X iop is the second position vector of the determined whale individual, X i is the current whale position X(t).

[0115] Like the first position vector, the second position vector X iop Each dimension of must also be within the upper and lower limits of the corresponding dimension. iop When the dimension of violates the upper and lower limit constraints of the dimension, the value of the dimension needs to be updated. Specifically, the second position vector X iop The dimension can be updated by the following formula:

[0116]

[0117]

[0118] Among them, x j,i new is the jth dimension in the position vector of the i-th whale individual to be updated. Here, the position vector is the second position vector X iop .

[0119] In this way, according to the reverse learning method, the cracking depth optimization system can update the second position vector of each individual whale in the whale population to complete the update of the second position vectors of all whale individuals in the whale population, thereby forming a new solution set X by the second position vectors of all whale individuals. op And, according to the updated second position vector X of each individual whale iop , calculate the fitness function value of each whale individual.

[0120] In addition, the splitting deep optimization system can use the random selection method to gather the whales after each iteration to the best position, thereby better protecting the position of the best whale. Moreover, the random selection method can accelerate the convergence behavior of the deep optimization model, but at the same time, the random selection method is also prone to the risk of the deep optimization model falling into the local optimum.

[0121] Specifically, the random selection method can replace the spatial position of the individual whale with the spatial position of the best individual whale in each iteration of the whale population when a certain probability condition is met to form a third position vector. Here, the probability condition is:

[0122]

[0123] Each dimension of the third position vector must also be within the upper and lower limit constraints of the corresponding dimension. When the dimension of the third position vector violates the upper and lower limit constraints of the dimension, the value of the dimension needs to be updated. Specifically, the dimension of the third position vector can be updated by the following formula:

[0124]

[0125]

[0126] Among them, x j,i new is the j-th dimension in the position vector to be updated for the i-th whale individual. Here, the position vector is the third position vector.

[0127] Thus, according to the random selection method, the cracking depth optimization system can update the third position vector of each individual whale in the whale population to complete the update of the third position vectors of all whale individuals in the whale population, thereby forming a new solution set X by the third position vectors of all whale individuals. rs And, according to the updated third position vector of each individual whale, the fitness function value of each individual whale is calculated.

[0128] In addition, the cracking depth optimization system can use the random allocation method to enhance the global search capability of the depth optimization model based on the whale optimization algorithm of the enhanced exploration mechanism. Specifically, after each iteration, the position of each whale individual is randomly redistributed in the area of ​​radius k:

[0129] X ird =X i +k×rand×sgin(rand-0.5),

[0130]

[0131] Among them, X ird is the fourth position vector, X i is the current whale position X(t), [Lb, Ub] is the search range of the solution space, and the sgin function is a sign function, which is a function that returns the corresponding result according to the positive or negative value of the input value. Specifically, when the input value is greater than 0, the sgin function returns 1; when the input value is equal to 0, the sgin function returns 0; when the input value is less than 0, the sgin function returns -1.

[0132] Fourth position vector X ird Each dimension of must also be within the upper and lower limits of the corresponding dimension. ird When the dimension of violates the upper and lower limit constraints of the dimension, the value of the dimension needs to be updated. Specifically, the fourth position vector X ird The dimension can be updated by the following formula:

[0133]

[0134]

[0135] Among them, x j，i new is the jth dimension in the position vector of the i-th whale individual to be updated. Here, the position vector is the fourth position vector X ird .

[0136] Thus, according to the random selection method, the split depth optimization system can update the fourth position vector X of each individual whale in the whale population. ird , to complete the fourth position vector X for all whale individuals in the whale population ird The update of the fourth position vector X of all whale individuals is ird Form a new solution set X rd And, according to the updated fourth position vector X of each individual whale ird , calculate the fitness function value of each whale individual.

[0137] Therefore, after determining the first position vector, the second position vector, the third position vector and the fourth position vector of each individual whale in the initial whale population, the cracking depth optimization system obtains 4n position vectors (n is the number of whale individuals in the whale population).

[0138] More preferably, after determining the first position vector, the second position vector, the third position vector and the fourth position vector that do not violate the upper and lower limit constraints of the dimension, the cracking depth optimization system can also convert the first position vector, the second position vector, the third position vector and the fourth position vector into integer vectors according to the encoding conversion function.

[0139] Afterwards, if Figure 3 As shown, the cracking depth optimization system continues to execute step S6 of the offline training phase: according to the fitness function value, the first position vectors, the second position vectors, the third position vectors and the fourth position vectors of all the determined whale individuals are arranged in ascending order, and the first n position vectors are selected as the positions of each whale individual in the next generation of whale population, and the position vector of the whale individual with the highest fitness function value is updated as the best whale individual.

[0140] Please continue to refer to Figure 3 , the cracking depth optimization system continues to execute step S7 of the offline training phase: repeat the above steps S2 to S6 until the number of iterations reaches the maximum number of population iterations Max_iters or the convergence condition of the deep optimization model is met.

[0141] By executing Figure 3 The enhanced exploration mechanism of the whale optimization algorithm shown in the figure can obtain a set of values ​​that meet the constraints: This set of values That is, the cracking depth value of the cracking furnace group is determined by the depth optimization model according to the optimization objective function within the constraint range of the optimization variables.

[0142] The following is a specific non-limiting preferred embodiment, based on which the cracking depth optimization system for an ethylene cracking furnace group proposed by the present invention is further described.

[0143] In this embodiment, the cracking depth optimization system selects the cracking depth index optimized by the depth optimization model as the propylene-ethylene ratio (the yield ratio of propylene to ethylene), and provides a set value of cracking depth for the cracking depth controller of each cracking furnace in the cracking furnace group according to the optimized cracking depth value determined by the cracking depth optimization system.

[0144] First, the cracking depth optimization system determines the optimization objective function according to the selected optimization index to improve the flexibility of the depth optimization model. In the present embodiment, the optimization index is the propylene-ethyl ratio. The optimization objective function includes maximizing ethylene yield, maximizing propylene yield and maximizing diene yield.

[0145] Here, the mathematical expression of the optimization objective function can be shown as follows:

[0146]

[0147] Where, f1(x), f2(x) and f3(x) represent the maximization of ethylene yield, propylene yield and diene yield respectively, X0 is the current feed amount of the cracking furnace, W C2H4 is the mass yield of the product C2H4, W C3H6 is the product mass yield of product C3H6.

[0148] The current whale position X(t) is the cracking depth value of each cracking furnace. Here, the cracking depth X of the cracking furnaces F1 to F18 is i The upper limit is The lower limit is The minimum constraint value of ethylene production of the cracking furnace group is MIN C2H4 , the actual ethylene production is REAL C2H4, through the cracking furnace mechanism model, the ethylene production can be predicted as MODEL C2H4 In the optimization process of the deep optimization model, the ethylene production can be predicted by the cracking furnace mechanism model as MODEL′ C2H4 .

[0149] In this embodiment, the cracking depth optimization system can select to optimize the jth cracking furnace, so that the upper and lower limits of the cracking depth of the cracking furnace can be maintained as follows: j = 0, 1, 2, ..., 18; the cracking depth optimization system can also choose not to optimize the jth cracking furnace, and the upper and lower limits of the cracking depth optimization range of the cracking furnace are the current cracking depth Right now This indicates that the cracking furnace does not need to be optimized.

[0150] Therefore, when the actual ethylene production situation REAL C2H4 Greater than the minimum constraint value MIN of ethylene production of the cracking furnace group C2H4 When the ethylene production MODEL′ calculated by the cracking furnace mechanism model during the optimization process is satisfied C2H4 Compared with the ethylene production model before optimization C2H4 The decline is smaller than the actual production of ethylene C2H4 Compared with the minimum constraint value MINC of ethylene production of the cracking furnace group 2H4 When the surplus value is constrained, the cracking depth optimization system seeks the best result in the direction of maximizing the optimization objective function set by the user:

[0151]

[0152] in, and These are f1(x) and f2(x) in the above optimization objective function.

[0153] In addition, when the actual ethylene production situation REAL C2H4 Less than the minimum constraint value MIN of ethylene production of the cracking furnace group C2H4 When , the cracking depth optimization system takes the ethylene yield maximization f1(x) as the optimization objective function, and in the optimization process, the ethylene yield MODEL′ calculated by the cracking furnace mechanism model C2H4 Compared with the ethylene production model before optimization C2H4 The increase cannot exceed the REAL ethylene production C2H4 Compared with the minimum constraint value MIN of ethylene production of the cracking furnace group C2H4 The missing values ​​of are multiplied by some empirical coefficient α:

[0154]

[0155] Afterwards, based on the operating variables of each cracking furnace, the cracking depth optimization system can construct the input elements of the multidimensional optimization variables. The number of rows of the matrix of the input elements of the multidimensional optimization variables is the number of cracking furnaces in the cracking furnace group, and the number of columns is the number of each selected operating variable. Here, the input element a of each cracking furnace is composed of the outlet temperature cot, the dilution steam ratio dor, the feed flow rate feed, the outlet pressure pout, the cross-section feed temperature tlec, and the cross-section feed pressure pin. j :

[0156] a j =[cot j ,dor j , feed j , pout j ,tlec j , pin j ]j=1,2,…,18。

[0157] According to the input element a of each cracking furnace j , we can construct the multi-dimensional optimization variables of the cracking furnace group:

[0158] X=[a1,a2,...a N ],

[0159] Wherein, N is the number of cracking furnaces (in this embodiment, N=18), and the optimization variable is the COT of each cracking furnace.

[0160] In the application scenario of ethylene cracking furnaces, the whale population size is often hundreds, and when calculating the position vector and objective function value of a whale individual, the number of model calls is consistent with the number of cracking furnaces. Therefore, in the process of continuous optimization, one iteration may need to call the mechanism model thousands of times, and when exiting the iteration, the mechanism model may need to be called tens of thousands of times. Therefore, the time and computing power spent on using the continuous optimization algorithm in the continuous feasible domain are too high.

[0161] Preferably, in order to reduce the number of calls to the cracking furnace mechanism model and improve the timeliness of the whale optimization algorithm in the deep optimization model, the cracking depth optimization system discretizes the original continuous optimization problem into a combinatorial optimization problem, that is, the upper and lower limits of the COT of the jth cracking furnace [m j ,h j ]m j ,h j ∈Z is discretized into a feasible set {m j , m j +1, ..., h j -1,h j}, find and determine the cracking depth value of the cracking furnace group in the constraint space of the feasible set. Here, the constraint space of the feasible set is the constraint condition of the optimization variable COT.

[0162] Thus, according to the above-mentioned multi-dimensional optimization variables, the determined optimization objective function and the constraints of the optimization variables, the depth optimization model is utilized to determine the cracking depth value of the optimized cracking furnace group within the scope of the constraints.

[0163] Specifically, when determining the cracking depth value of the optimized cracking furnace group, the cracking depth optimization system first initializes the control parameters and a group of whale populations. Here, the cracking depth optimization system can set the maximum number of population iterations Max_iters in the control parameters to 100 times, the population size population_size to 150, the logarithmic spiral shape constant b to 1.0, and the current best whale individual X best The number of unchange_iters that does not change during the iteration process is set to 50 times as the convergence condition of the deep optimization model.

[0164] The initial whale population is {X1(0), X2(0), X3(0), ..., X population_size-1 (0)}, the whale population is a uniformly distributed randomly generated integer. In an initialized set of whale populations, the position vector of the i-th whale individual is:

[0165] X i (0) = [x 0，i (0), x 1，i (0), ..., x N-1，i (0)],

[0166] Each dimension corresponds to the cracking depth value of a cracking furnace in the cracking furnace group, and thus each dimension is constrained by the upper and lower limits of the cracking depth of the corresponding cracking furnace. Here, the upper and lower limits of the cracking depth of the corresponding cracking furnace can be:

[0167] x j，min ≤x j，i (0)≤x j，max j=0,1,2…,N-1,

[0168] Among them, the subscript j represents the dimension number, that is, the j-th cracking furnace; the subscript i represents the i-th individual in the whale population, and N is the number of cracking furnaces.

[0169] Afterwards, the first coefficient vector A, the second coefficient vector C, the first variable p for controlling the way of updating the position, the second variable h for controlling the range of the coefficient vector A and the random number l in the control parameters are initialized (ie, the current iteration number t is 0).

[0170] The cracking depth optimization system first updates the second variable h in the range of the control coefficient vector A:

[0171]

[0172] Afterwards, the first coefficient vector A is updated according to the updated second variable h, and the second coefficient vector C, the first variable p and the random number l that control the update position are continued to be updated:

[0173] A=2h·rand-h,

[0174] C = 2 rand,

[0175] p = rand(0, 1),

[0176]

[0177] Afterwards, based on the first variable p and the first coefficient vector A in the updated control parameters, the fragmentation depth optimization system can determine a method for updating the position, and determine the first position vector of each individual whale based on the determined method for updating the position.

[0178] Specifically, when the first variable p is less than 0.5, the first coefficient vector A is judged. When the first coefficient vector A is less than 1, the cracking depth optimization system can determine the first position vector X according to the shrinking and encircling mechanism with a convergent adaptive weighting strategy. local (t+1).

[0179] The update formula of the shrinking and encircling mechanism with convergent adaptive weighted strategy is as follows:

[0180] X local (t+1)=r×X best (t)-A·B,

[0181] The calculation formula of weight r is:

[0182]

[0183] The calculation formula of B is:

[0184] B=|C·X best (t)-X local (t)|,

[0185] Among them, X local (t) is the position vector X(t) of the individual whale selected to update its position in the current population.

[0186] In response to the first variable p being less than 0.5 and the first coefficient vector A being greater than or equal to 1, the cracking depth optimization system may determine the first position vector X according to a random update mechanism with a convergent adaptive weighting strategy. local (t+1).

[0187] The update formula of the random update mechanism with convergent adaptive weighted strategy is as follows:

[0188] X local (t+1)=r×X best (t)-A·B,

[0189] The calculation formula of the weight r is the same as the shrinking and encircling mechanism with convergent adaptive weighting strategy:

[0190]

[0191] The calculation formula of B is:

[0192] B=|C·X rand (t)-X local (t)|,

[0193] Among them, X rand (t) is the position vector of a randomly selected whale individual in the current population.

[0194] In addition, when the first variable p is greater than or equal to 0.5, the cracking depth optimization system can determine the first position vector X according to the spiral update position mechanism. local (t+1).

[0195] The spiral update position mechanism first calculates the position of the whale at (X, Y) and the position of the whale at (X * , Y * ) and then create a spiral equation between the positions of the whale and the prey to simulate the whale's spiral movement.

[0196] The equation for this spiral is as follows:

[0197] X local (t+1)=X best (t)+B p ·e bl cos(2πl),

[0198] Among them, B p The calculation formula is as follows:

[0199] B p =|X best (t)-X local (t)|.

[0200] In this way, the first position vector of each individual whale in the whale population is updated according to the control parameters to complete the update of the first position vectors of all individual whales in the whale population.

[0201] Furthermore, each dimension of the first position vector of the determined whale individual is within the upper and lower limit constraints of the corresponding dimension. When the dimension of the first position vector violates the upper and lower limit constraints of the corresponding dimension, the value of the dimension needs to be updated so that each dimension is within the upper and lower limit constraints of the corresponding dimension. Specifically, the value of the dimension can be updated by the following formula:

[0202]

[0203]

[0204] Among them, x j，i (t) is the jth dimension of the i-th whale individual in the t-th iteration of the whale population, x j，i new is the jth dimension in the position vector of the i-th whale individual to be updated. Here, the position vector is the first position vector X local (t+1).

[0205] Preferably, in order to make the value of the updated position vector within the discrete integer feasible set, the following encoding conversion function is used to convert the real vector into an integer vector:

[0206]

[0207] Among them, N j The jth cracking furnace is within the upper and lower limits [m j , n j ]m, the number of integer feasible sets in n∈Z, w is [1, N j -2] integer, y j，i is the value of the j-th dimension of the i-th whale individual after the real number is converted into an integer through the encoding conversion function.

[0208] After the first position vectors of all whale individuals in the whale population are converted by the above encoding conversion function, their fitness function values ​​are calculated.

[0209] Afterwards, the splitting depth optimization system uses the reverse learning method, the random selection method and the random allocation method to determine the second position vector, the third position vector and the fourth position vector of each individual whale and their corresponding fitness function values.

[0210] The reverse learning method is used for each individual whale in the whale population to form a second position vector X of each individual whale. iop :

[0211] X iop=Lb+Ub-X i ,

[0212] Among them, [Lb, Ub] is the search range of the solution space. Then, the second position vector X of all whale individuals is iop Form a new reverse solution set X op .

[0213] Preferably, the cracking depth optimization system can generate the reverse solution set X op The second position vector X of each whale individual in iop After conversion using the above encoding conversion function, the fitness function value is calculated.

[0214] Similarly, a random selection method is used for each individual whale in the whale population. When the position vector of a whale individual satisfies the following probability conditions, its position is updated to the position of the current best whale individual, that is, the third position vector of the whale individual, forming a new solution set X rs :

[0215]

[0216] The cleavage depth optimization system can also use random allocation to distribute the position of each individual whale in the whale population. According to the following formula, the cleavage depth optimization system randomly redistributes the position of each individual whale in the area of ​​radius k to form a new fourth position vector X ird :

[0217] X ird =X i +k×rand×sgin(rand-0.5),

[0218]

[0219] Then, the cracking depth optimization system uses the above encoding conversion function to convert the fourth position vector X ird After converting to an integer vector, calculate its fitness function value.

[0220] Therefore, after each whale individual has determined the first position vector, the second position vector, the third position vector and the fourth position vector, the first position vector, the second position vector, the third position vector and the fourth position vector of all determined whale individuals are arranged in ascending order according to the fitness function value, and the whale individuals with the first population_size position vectors are selected as the whale individuals of the next generation of whale population, and the position vector of the whale individual with the highest fitness function value is updated as the best whale individual.

[0221] Furthermore, when updating the best whale individual, if there is a whale individual with a better position vector than the current best whale individual, the best whale individual and the corresponding best fitness function value are updated, and the number of unchange_iters of the best whale individual that has not changed is set to 0. Conversely, if there is no whale individual with a better position vector than the current best whale individual, the best whale individual and the corresponding best fitness function value are not updated, and the number of unchange_iters of the best whale individual that has not changed is increased by 1.

[0222] After that, the number of iterations iters is increased by 1, and the next round of iterations is started until the deep optimization model meets the convergence condition (in this example, when the number of unchange_iters reaches 50, it is considered to meet the convergence condition) or the maximum number of population iterations Max_iters is reached.

[0223] Therefore, the cracking depth optimization system obtains a set of values ​​that meet the constraints: That is, the cracking depth value of the cracking furnace group is determined by the depth optimization model according to the optimization objective function within the constraint range of the optimization variables.

[0224] Please refer to Figure 5 and Figure 6 , Figure 5 and Figure 6 A rendering of the effect of a cracking depth optimization method for an ethylene cracking furnace group is shown.

[0225] like Figure 5 , Figure 6 As shown in the figure, two optimization indicators were determined before and after the cracking depth optimization system was put into use. The cracking depth optimization system first selected the maximization of diene yield as the optimization objective function to optimize the cracking depth value of the cracking furnace group. Figure 5 As shown, the vertical axis on the right is the COT value, and the vertical axis on the left is the cracking depth (SEV) value. It can be seen from the figure that when maximizing the diolefin yield is selected as the optimization objective function, line 501 will decrease, that is, the cracking depth value decreases; line 502 will increase, that is, the COT temperature increases. Based on the changes in the above operating conditions, that is, the cracking depth value decreases and the COT temperature increases, the diolefin yield will increase. Then, the cracking depth optimization system selects the maximization of propylene yield as the optimization objective function to optimize the cracking depth value of the cracking furnace group. As shown Figure 6 As shown, when maximizing the propylene yield is selected as the optimization objective function, line 601 rises, that is, the cracking depth value rises; line 602 falls, that is, the COT temperature falls. In this way, the propylene yield increases by increasing the cracking depth value and decreasing the COT temperature. The above results show that the cracking depth optimization method for an ethylene cracking furnace group provided by the present invention can effectively improve the operating efficiency of the cracking furnace.

[0226] In summary, the cracking depth optimization technology for an ethylene cracking furnace group provided according to the present invention has a faster convergence speed and better accuracy, and can determine the optimal cracking depth in real time according to the real-time operation status of the current cracking furnace, thereby significantly improving the yield of the product while meeting the overall constraints.

[0227] Although the above methods are illustrated and described as a series of actions for simplicity of explanation, it should be understood and appreciated that these methods are not limited by the order of the actions, because according to one or more embodiments, some actions may occur in a different order and / or concurrently with other actions from those illustrated and described herein or not illustrated and described herein but understandable to those skilled in the art.

[0228] Those skilled in the art will appreciate that information, signals, and data may be represented using any of a variety of different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols, and chips cited throughout the above description may be represented by voltage, current, electromagnetic waves, magnetic fields or magnetic particles, optical fields or optical particles, or any combination thereof.

[0229] Those skilled in the art will further appreciate that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or a combination of the two. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps are generally described above in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. The technician may implement the described functionality in different ways for each specific application, but such implementation decisions should not be interpreted as resulting in a departure from the scope of the present invention.

[0230] The various illustrative logic modules and circuits described in conjunction with the embodiments disclosed herein may be implemented or executed with a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in cooperation with a DSP core, or any other such configuration.

[0231] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software module may reside in a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor so that the processor can read and write information from / to the storage medium. In an alternative, a storage medium may be integrated into a processor. The processor and the storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and the storage medium may reside in a user terminal as discrete components.

[0232] In one or more exemplary embodiments, the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented as a computer program product in software, each function may be stored on or transmitted by a computer-readable medium as one or more instructions or codes. Computer-readable media include both computer storage media and communication media, including any medium that facilitates the transfer of a computer program from one place to another. Storage media may be any available medium that can be accessed by a computer. As an example and not limitation, such a computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer. Any connection is also properly referred to as a computer-readable medium. For example, if the software is transmitted from a website, a server, or other remote source using a coaxial cable, a fiber optic cable, a twisted pair, a digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwaves, the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwaves are included in the definition of the medium. Disk and disc as used herein include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and Blu-ray disc, wherein disk often reproduces data magnetically, while disc reproduces data optically with lasers. Combinations of the above should also be included within the scope of computer-readable media.

[0233] The previous description of the disclosure is provided to enable any person skilled in the art to make or use the disclosure. Various modifications to the disclosure will be apparent to those skilled in the art, and the general principles defined herein may be applied to other variations without departing from the spirit or scope of the disclosure. Thus, the disclosure is not intended to be limited to the examples and designs described herein, but should be granted the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for optimizing the cracking depth of an ethylene cracking furnace group, characterized in that: Includes steps: Determining a multidimensional optimization variable based on each operating variable of each cracking furnace, determining the constraint conditions of the optimization variable according to the upper and lower limit constraints of the optimization variable, and determining the optimization objective function based on the minimum constraint value of ethylene production of the cracking furnace group; as well as The multidimensional optimization variables are input into a deep optimization model, and the deep optimization model determines the optimized cracking depth value of the cracking furnace group within the constraint range of the optimization variables according to the optimization objective function, wherein the deep optimization model is constructed by the whale optimization algorithm of the enhanced exploration mechanism.

2. The method for optimizing cracking depth according to claim 1, wherein: The step of determining the optimized cracking depth value of the cracking furnace group comprises: Within the scope of the constraints, the enhanced exploration mechanism whale optimization algorithm is used to iteratively calculate and adjust the multi-dimensional optimization variables to determine the optimized cracking depth value of the cracking furnace group.

3. The method for optimizing cracking depth according to claim 2, wherein: The step of determining the optimized cracking depth value of the cracking furnace group by iteratively calculating and adjusting the multi-dimensional optimization variables using the enhanced exploration mechanism whale optimization algorithm within the scope of the constraint conditions comprises: S1: Initialize control parameters and a group of whale populations, and determine the best whale individual according to the fitness function, wherein the whale population includes n whale individuals; S2: updating the control parameters, and determining the first position vector and fitness function value of each individual whale according to the control parameters, wherein the dimension of the first position vector is within the upper and lower limit constraints of the dimension; S3: Determine the second position vector of each individual whale and its fitness function value using a reverse learning method, wherein the dimension of the second position vector is within the upper and lower limit constraints of the dimension; S4: using a random selection method to determine the third position vector and fitness function value of each individual whale, wherein the dimension of the third position vector is within the upper and lower limit constraints of the dimension; S5: Determine the fourth position vector and fitness function value of each individual whale using a random allocation method, wherein the dimension of the fourth position vector is within the upper and lower limit constraints of the dimension; S6: according to the fitness function value, the first position vector, the second position vector, the third position vector and the fourth position vector of all determined whale individuals are arranged in ascending order, the first n position vectors are selected as the positions of each whale individual of the next generation of whale population, and the position vector of the whale individual with the highest fitness function value is updated as the best whale individual; and S7: Repeat S2 to S6 until the number of iterations reaches the maximum number of population iterations or the convergence condition of the deep optimization model is met.

4. The method for optimizing cracking depth according to claim 3, wherein: The control parameters include the maximum number of population iterations, the population size, the target prey position, the current whale position, the current number of iterations, the first coefficient vector, the second coefficient vector, the first variable controlling the position update method, the second variable controlling the coefficient vector, the logarithmic spiral shape constant, and a random number.

5. The method for optimizing cracking depth according to claim 3, characterized in that: The step of determining the first position vector of each individual whale according to the control parameter comprises: In response to the first variable controlling the updating position mode being less than 0.5 and the first coefficient vector being less than 1, determining the first position vector according to a shrinking and surrounding mechanism with a convergent adaptive weighting strategy; In response to the first variable being less than 0.5 and the first coefficient vector being greater than or equal to 1, determining the first position vector according to a random update mechanism with a convergent adaptive weighting strategy; and In response to the first variable being greater than or equal to 0.5, the first position vector is determined according to a spiral update position mechanism.

6. The method for optimizing cracking depth according to claim 3, characterized in that: The steps S2 to S5 include: When the dimensions of the first position vector and / or the second position vector and / or the third position vector and / or the fourth position vector violate the upper and lower limit constraints of the dimensions, the values ​​of the dimensions are updated.

7. The method for optimizing cracking depth according to claim 6, wherein: The steps S2 to S5 also include: The first position vector and / or the second position vector and / or the third position vector and / or the fourth position vector are converted into integer vectors according to a coding conversion function.

8. The method for optimizing cracking depth according to claim 1, wherein: The optimization objective function includes maximizing ethylene yield and / or maximizing propylene yield and / or maximizing diene yield.

9. The method for optimizing cracking depth according to claim 1, wherein: The step of determining the constraint conditions of the optimization variables according to the upper and lower limit constraints of the optimization variables comprises: The upper and lower limits of the optimization variables are discretized to determine the constraint conditions of the optimization variables.

10. A cracking depth optimization system for an ethylene cracking furnace group, characterized in that: include: a memory having computer instructions stored thereon; as well as A processor is connected to the memory and is configured to execute computer instructions stored in the memory to implement the cracking depth optimization method for an ethylene cracking furnace group as described in any one of claims 1 to 9.

11. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the computer instructions are executed by the processor, the method for optimizing the cracking depth of an ethylene cracking furnace group as claimed in any one of claims 1 to 9 is implemented.

Images