Cracking furnace operation optimization method and device, storage medium and processor

Abstract

The embodiment of the invention provides a cracking furnace operation optimization method and device, a storage medium and a processor, and belongs to the technical field of chemical engineering automatic control. The method comprises the following steps: continuously acquiring operation parameters of the cracking furnace, inputting a first prediction model of deep learning to obtain a first prediction value of the cracking depth of the cracking furnace at a moment T + T0 after a current moment T, and forming a first curve reflecting the change of the operation effect of the cracking furnace within a preset time period; continuously obtaining reaction condition parameters of the cracking furnace, inputting a second prediction model of deep learning to obtain a second prediction value of the cracking depth of the cracking furnace at the moment T + T0 after the current moment T, and forming a second curve reflecting the operation effect change of the cracking furnace within the preset time period; if the similarity of the first curve and the second curve exceeds a preset threshold value, the cracking furnace operation is corrected. Through the method, the real-time operation of the cracking furnace can be quickly and effectively optimized, and the operation optimization direction of the cracking furnace is consistent with the optimization direction of the reaction conditions.

Description

Technical Field

[0001] The invention relates to the technical field of chemical automatic control, and in particular to a method, device, storage medium and processor for optimizing the operation of a cracking furnace. Background Art

[0002] In the cracking production process, the operating conditions of the cracking furnace will have a significant impact on its yield. In the actual production process, there are two main ways to adjust the operating conditions of the cracking furnace: one is based on the reference value and manual experience provided by the cracking furnace manufacturer, but in the actual production process, the source of cracking oil products is usually not fixed, and different oil products will be switched as needed during the production process. Therefore, the current manual experience cannot achieve the optimization and adjustment of the optimal operating conditions of the cracking furnace; the other way is to use an online component analyzer to measure the relative content of key components in the cracking gas, and use this as a reference to adjust the production operating conditions. However, it takes a certain amount of time for the online analyzer to analyze a sample, which has a large hysteresis and randomness, which is not conducive to the stable operation of the cracking furnace. At the same time, if the online analyzer fails, this operation mode will fail, and in severe cases, it will even cause the emergency shutdown of the cracking furnace, resulting in serious economic losses. Therefore, in order to achieve rapid optimization and adjustment of the operating conditions of the cracking furnace, a fast and effective real-time operation optimization method for the cracking furnace is urgently needed. Summary of the invention

[0003] The purpose of the embodiment of the present invention is to provide a method for optimizing the operation of a cracking furnace, by which the real-time operation of the cracking furnace can be optimized quickly and effectively, and the operation optimization of the cracking furnace is consistent with the optimization direction of the reaction conditions of the cracking furnace.

[0004] In order to achieve the above object, an embodiment of the present invention provides a method for optimizing the operation of a cracking furnace, the method comprising:

[0005] Continuously obtain the operating parameters of the cracking furnace, and input the operating parameters into the first prediction model of deep learning to obtain a first predicted value of the cracking depth of the cracking furnace at time T+T0 after the current time T, and form a first curve reflecting the change of the operating effect of the cracking furnace within a preset time period;

[0006] Continuously acquiring reaction condition parameters of the cracking furnace, and inputting the reaction condition parameters into a second prediction model of deep learning to obtain a second prediction value of the cracking depth of the cracking furnace at a time T+T0 after the current time T, and forming a second curve reflecting the change of the operation effect of the cracking furnace within a preset time period; and

[0007] If the similarity between the first curve and the second curve exceeds a preset threshold, the cracking furnace operation is corrected.

[0008] Preferably, the cracking severity includes any one of the following: feed conversion, methane yield, propylene / ethylene yield ratio, methane / propylene yield ratio, hydrogen content of cracked liquid products, outlet temperature, kinetic severity, or yield of C3 and lighter components.

[0009] Preferably, the operating parameters are the cracking furnace operation control conditions determined according to the cracking furnace tube type and cracking furnace model, and the operating parameters include one or more of the following: total steam feed ratio, total feed amount, COT, target outlet temperature, feed amount of each furnace tube group, feed water flow rate, bottom fuel gas flow rate, side wall fuel flow rate, drum liquid level, radiation section furnace pressure, and feed cross section pressure.

[0010] Furthermore, the first prediction model is a trained long short-term memory network model, and the adopted training data set includes a group of training samples, and the training samples are historical operating data of the current cracking furnace in cracking the current raw material, including operating parameter sample values ​​and corresponding cracking depth sample values. The training process of the first prediction model includes:

[0011] Obtaining an initialized long short-term memory network model, the input of the long short-term memory network model is the sample value of the operation parameter, and the output is the predicted value of the cracking depth;

[0012] The sample values ​​of the operating parameters are input into the cracking depth prediction model, and the deviation between the cracking depth prediction value and the cracking depth sample value is calculated by the error loss function;

[0013] When the deviation is less than a preset threshold or the training reaches a preset number of times, the training is terminated; otherwise, the long short-term memory network model is corrected and the next round of training is continued; and

[0014] The one with the smallest deviation is selected as the final first prediction model.

[0015] Preferably, the reaction condition parameters are the monitoring values ​​of the feed ratio of the cracking reaction, the content of cracking gas and intermediate products, and the temperature and pressure conditions of the cracking reaction determined according to the reaction mechanism of the cracking raw material. The reaction condition parameters include one or more of the following: cracking gas pressure, bottom fuel gas pressure, temperature deviation and COT deviation of each furnace tube group, feed cross-section temperature of each furnace tube group, cracking gas waste heat boiler temperature, furnace temperature, each group of feed furnace outlet temperature, and each group of feed mixing preheating outlet temperature.

[0016] Furthermore, the second prediction model is a trained long short-term memory network model, and the training data set used includes a group of training samples, and the training samples are historical reaction condition data of the current cracking furnace in cracking the current raw material, including reaction condition parameter sample values ​​and corresponding cracking depth sample values, and the training process includes:

[0017] Obtaining an initialized long short-term memory network model, the input of the long short-term memory network model is a sample value of a reaction condition parameter, and the output is a predicted value of a cracking depth;

[0018] The sample values ​​of the reaction condition parameters are input into the long short-term memory network model, and the deviation between the predicted value of the cracking depth and the sample value of the cracking depth is calculated by the error loss function;

[0019] When the deviation is less than a second preset threshold or the training reaches a second preset number of times, the training is terminated; otherwise, the long short-term memory network model is corrected and the next round of training is continued; and

[0020] The one with the smallest deviation is selected as the final second prediction model.

[0021] Furthermore, the training data set is obtained by collecting historical operating data of the cracking furnace, and performing data classification and screening, wherein:

[0022] Historical operating data include: cracking feedstock oil properties, cracking furnace tube types, DCS monitoring related parameters and corresponding ethylene yield and propylene yield;

[0023] The classification of historical operating data includes: classification of cracking properties according to the properties of cracking feedstock, and the classification is based on one or more of the following performance indicators: relative density, distillation range, group composition, viscosity, hydrogen content, and average molecular weight of the oil;

[0024] The screening of historical operating condition data includes deleting operating condition data that meets the following conditions: the error of any one of the relevant parameters, ethylene yield, and propylene yield exceeds the corresponding preset range, or there is an abnormal operating condition.

[0025] Furthermore, the training data set includes historical operating data of the cracking furnace in the last 0.1-10 years.

[0026] Preferably, the preset time period is n1 minutes, T0 is n2 minutes, and n1 and n2 are values ​​between 1-30.

[0027] In another aspect, the present invention provides an optimization device for cracking furnace operation, comprising:

[0028] A first prediction module continuously obtains the operating parameters of the cracking furnace and inputs the operating parameters into a first prediction model of deep learning to obtain a first predicted value of the cracking depth of the cracking furnace at a time T+T0 after the current time T, and forms a first curve reflecting the change of the operating effect of the cracking furnace within a preset time period;

[0029] A second prediction module continuously obtains reaction condition parameters of the cracking furnace, and inputs the reaction condition parameters into a second prediction model of deep learning to obtain a second predicted value of the cracking depth of the cracking furnace at a time T+T0 after the current time T, and forms a second curve reflecting the change of the operation effect of the cracking furnace within a preset time period; and

[0030] The optimization module is run, and if the similarity between the first curve and the second curve exceeds a preset threshold, the cracking furnace operation is corrected.

[0031] On the other hand, the present invention provides a machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the method for optimizing the operation of a cracking furnace of the present application.

[0032] On the other hand, the present invention provides a processor, characterized in that it is used to run a program, wherein the program is used to execute the optimization method of cracking furnace operation of the present application when it is run.

[0033] Through the above technical scheme, on the one hand, a first curve reflecting the change of the operating effect of the cracking furnace in a certain time period is formed through the first prediction model of deep learning and the operating parameters of the cracking furnace, representing the change trend of the first prediction value of the effect of the cracking furnace in the time period; on the other hand, a second curve reflecting the change of the operating effect of the cracking furnace in the same time period is formed through the second prediction model of deep learning and the reaction condition parameters of the cracking furnace, representing the change trend of the second prediction value of the effect of the cracking furnace in the time period, and then the change trends of the two curves in the time period are compared, and when the change trends are inconsistent, the operation of the cracking furnace is corrected, which can optimize the operation of the cracking furnace in real time and make the operation optimization of the cracking furnace consistent with the optimization direction of the reaction conditions of the cracking furnace.

[0034] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0036] Figure 1 It is a flow chart of an embodiment of a method for optimizing cracking furnace operation of the present application;

[0037] Figure 2 yes Figure 1 A schematic diagram of the training process of the first prediction model in the embodiment; and

[0038] Figure 3 It is a composition structure diagram of an embodiment of an optimization device for cracking furnace operation of the present application. DETAILED DESCRIPTION

[0039] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0040] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0041] In addition, it should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0042] The embodiment of the present invention provides a method for optimizing the operation of a cracking furnace, by which the real-time operation of the cracking furnace can be optimized quickly and effectively. The process of the first embodiment of the method for optimizing the operation of the cracking furnace of the present application is as follows: Figure 1 As shown, including:

[0043] Step 1: Continuously obtain the operating parameters of the cracking furnace, and input the operating parameters into the first prediction model of deep learning to obtain a first predicted value of the cracking depth of the cracking furnace at time T+T0 after the current time T, and form a first curve reflecting the change of the operating effect of the cracking furnace within a preset time period;

[0044] Step 2: continuously acquiring reaction condition parameters of the cracking furnace, and inputting the reaction condition parameters into a second prediction model of deep learning to obtain a second prediction value of the cracking depth of the cracking furnace at time T+T0 after the current time T, and forming a second curve reflecting the change of the operation effect of the cracking furnace within a preset time period; and

[0045] Step 3: If the similarity between the first curve and the second curve exceeds a preset threshold, correct the cracking furnace operation.

[0046] In actual production, the cracking furnace operator can refer to the vector composed of the operating variables of the cracking furnace and the first and second predicted values ​​of the cracking depth, and select the optimal operating conditions of the cracking furnace according to the actual production needs on site, and adjust and optimize the operating state of the cracking furnace.

[0047] It should be noted that the above steps 1 and 2 are all continuous detection of the relevant parameters of the cracking furnace. In practice, periodic detection, regular or irregular detection, simultaneous or alternating detection can be performed as needed, and it is not necessary to be limited to the execution order of the steps disclosed in the present application. The similarity between the first curve and the second curve can be Euclidean distance, Manhattan distance, Chebysh distance, Minkowski distance, standardized Euclidean distance, Mahalanobis distance, angle cosine, etc.

[0048] In this embodiment, the cracking furnace can be 1, 1-1, 2-1, 4-1, 4-1-1-1, 2-1-1-1, 1-1-1-1, or 8-4-2-1 furnace type, and the cracking raw materials of the cracking furnace are ethane, propane, LPG, naphtha, diesel, aviation kerosene, or hydrogenated tail oil.

[0049] It should also be noted that the cracking furnace is usually in a continuous working state, and according to the present application, its operating parameters and reaction condition parameters are also continuously obtained, so the first and second predicted values ​​obtained according to the operating parameters and reaction condition parameters can form a change curve equivalent to the operation time of the cracking furnace. The first and second curves of the change of the cracking furnace operation effect in the preset time period in this embodiment are partial curves of the preset time period intercepted from the above-mentioned change curve according to actual needs, and the time period can be arbitrarily set, preferably the nearest 1-30 minutes, so as to obtain whether the trend between the first and second predicted values ​​in this period is consistent. In addition, the value of T0 can be determined reasonably according to the training process of the first and second prediction models, so the value can be the same or different from the length of the preset time period.

[0050] The cracking depth can be determined according to the cracking raw material and the cracking reaction mechanism. For example, if the cracking raw material is naphtha, the cracking depth can be the propylene-ethylene content ratio in the cracking gas, i.e. the cracking reaction depth. In other words, the cracking depth in the present application is a parameter that reflects the operating effect of the cracking furnace, and can also be the ethylene or propylene yield in actual operation. Depending on the cracking raw materials, the cracking depth can also be the conversion rate of the raw material, the methane yield, the methane / propylene yield ratio, the hydrogen content of the cracked liquid product, the outlet temperature, the kinetic depth, or the yield of C3 and lighter components, etc.

[0051] Taking naphtha as the cracking raw material, its cracking product composition is restricted by many factors, such as cracking raw material quality, residence time, cracking furnace tube configuration, hydrocarbon partial pressure, reaction temperature, etc. In steps 1 and 2, the operating parameters and reaction condition parameters of the cracking furnace can be selected from the detection parameters of the DCS system based on experience to reflect the operating conditions of the cracking furnace, or all parameters corresponding to the operation and reaction conditions of the cracking furnace are selected from the detection parameters of the DCS system, or they can also be important parameters first selected according to an artificial intelligence algorithm.

[0052] In step 1, the first prediction model is a trained long short-term memory network model based on time series analysis, and the training sample is the historical operation data of the current cracking furnace in cracking the current raw material, that is, the first prediction model is a model that reflects the corresponding relationship between the cracking furnace operation and the cracking depth. By analyzing the historical operating condition data of the cracking furnace, a prediction model for the cracking depth is established.

[0053] In step 2, the second prediction model is a trained long short-term memory network model, and the training samples are the historical reaction condition data of the current cracking furnace in cracking the current raw material, that is, the second prediction model is a prediction model based on the reaction mechanism. By analyzing the historical reaction condition data of the cracking furnace, a prediction model of cracking depth is established to reflect the corresponding relationship between the reaction conditions in the cracking furnace and the cracking depth.

[0054] It should also be noted that the present application can not only monitor and predict the cracking depth of the cracking furnace, but also monitor and predict any one of the following according to different cracking production purposes: the maximum values ​​of ethylene, propylene, ethylene and propylene in the cracking products, the maximum yield of trienes, or the maximum yield of the value of the cracking products, that is, replace the "cracking depth" in this embodiment with any one of the above monitoring objects to allow the cracking reaction to proceed in the desired direction.

[0055] Compared with the prior art, the technical advantages of this embodiment include:

[0056] (1) A second prediction model is established based on the cracking reaction mechanism, which can be cross-referenced and verified with the prediction results of the first prediction model, thereby reducing the dependence on online analytical instruments, improving the real-time performance of the cracking furnace operation optimization method, and making the operation optimization of the cracking furnace consistent with the optimization direction of the cracking furnace reaction conditions;

[0057] (2) A multi-objective real-time operation optimization model for the cracking furnace was established to provide a variety of control schemes for the cracking furnace operators, which could be flexibly selected according to the on-site production needs, effectively improving the overall production efficiency and economic benefits of the ethylene plant.

[0058] In some embodiments, the training data set used by the first prediction model is a group of training samples, and the training samples are historical operation data of the current cracking furnace in cracking the current raw material, and the training samples include operating parameter sample values ​​and corresponding cracking depth sample values. Wherein, the training samples are obtained from the offline operating data of the cracking furnace in the past 0.1-10 years according to the time series collection, and the collected data types include the properties of the cracking raw material oil, the type of cracking furnace tube, all relevant operating parameters monitored on the DCS and the corresponding diene yield. The above data are used to establish a cracking furnace cracking depth database, and then, according to experience, the important parameters that can reflect the operating conditions of the cracking furnace are selected from the detection parameters of the DCS system. Wherein, the operating parameters are the cracking furnace operation control conditions determined according to the cracking furnace tube type and the cracking furnace model, and the operating parameters include one or more of the following: total steam feed ratio, total feed amount, COT, target outlet temperature, feed amount of each furnace tube group, feed water flow, bottom fuel gas flow, side wall fuel flow, drum liquid level, radiation section furnace pressure, and feed cross section pressure.

[0059] In some embodiments, considering that the composition of naphtha cracking products is restricted by various factors, such as cracking feed quality, residence time, cracking furnace tube configuration, hydrocarbon partial pressure, reaction temperature, etc. The operating parameters include cracking feed feed flow rate, cracking feed feed temperature and pressure, dilution steam feed flow rate, dilution steam feed temperature, cracking furnace cross section temperature, cracking furnace cross section pressure, cracking furnace tube outlet temperature and pressure, waste heat boiler outlet temperature, cracking furnace furnace temperature, and burner fuel flow rate.

[0060] In some embodiments, for different furnace tube configurations of the cracking furnace, such as single-pass furnace tubes, 1-1, 2-1 and 4-1 two-pass furnace tubes, and 1-1-1-1 and 2-1-1-1 four-pass furnace tubes, for example, ethylene cracking furnaces using naphtha as cracking raw materials mostly use 1-1 two-pass furnace tubes or 2-1 two-pass furnace tubes. Therefore, when obtaining the historical operating condition data of the cracking furnace, in addition to collecting the cracking furnace operating parameter data, the corresponding cracking furnace furnace tube configuration parameters also need to be included in the cracking depth database of the cracking furnace.

[0061] Before training the first prediction model, the training data set or the cracking furnace cracking depth database needs to be classified and screened.

[0062] In some embodiments, the cracking properties of naphtha oil products are classified according to their properties, and the naphtha cracking oil products are divided into L categories with similar cracking performance. The classification is based on their main performance indicators, namely, the relative density, distillation range, group composition (PONA value), viscosity, hydrogen content, and average molecular weight of the oil products.

[0063] On the basis of classification, the relevant data in each oil product classification is corrected, the collected data is cleaned, and the cracking furnace operating condition data and diene yield data containing obvious errors are deleted; data cleaning means that according to the range of relevant operating condition parameters and diene yield range of the cracking furnace in the operation process of naphtha cracking to ethylene, the extracted cracking furnace historical operating condition data is screened to remove abnormal operating condition values ​​with obvious errors;

[0064] Furthermore, the corrected data are integrated, and 50%-95% of the total data are used as training set data, and the remaining 50%-5% are used as test set data.

[0065] refer to Figure 2 As shown, during the training process of the first prediction model, the input is the operating parameter sample value of training sample n, wherein training sample n is the nth training sample selected from a group of training samples in the above-mentioned training data set, and the operating parameter sample value should include all the selected operating parameters and the cracking furnace operating conditions at time t; the output is the cracking depth prediction value, and the cracking depth prediction value is compared with the cracking depth sample value in training sample n through the error loss function, so as to determine the prediction deviation of the current first prediction model.

[0066] It should be noted that the input operating parameter sample value is the cracking furnace operating condition at time t, and accordingly, the cracking depth sample value used for comparison with the predicted value should be the actual value of the cracking depth at time t+n (n≥1), and the output is the predicted value of the cracking depth at time t+n (n≥1). Wherein, the value of n can be any time before the cracking production of the current raw material is completed, such as 1min, 5min, 10min, 13min, 15min, 24min, 35min, 42min, 48min, 54min, 60min, 2 hours, or even 3 hours, 5 hours, etc.

[0067] The training process of the first prediction model is as follows: first, an initialized long short-term memory network model is obtained, the input of the long short-term memory network model is the operation parameter sample value, and the output is the cracking depth prediction value; then multiple trainings are performed according to a preset number of training times and a preset deviation threshold, wherein each training includes inputting the operation parameter sample value into the long short-term memory network model, and calculating the deviation between the cracking depth prediction value and the cracking depth sample value through an error loss function; when the deviation is less than a first preset threshold or the training reaches a first preset number of times, the training is terminated; otherwise, the long short-term memory network model is corrected and the next round of training continues; finally, the one with the smallest deviation is selected as the final first prediction model.

[0068] In some embodiments, the root mean square error loss is used to calculate the deviation between the two, and the models saved after multiple trainings are screened, and the one with the smallest root mean square error is selected as the final prediction model.

[0069] Compared with the prior art, the technical advantage of this embodiment is that the historical operating condition data of the cracking furnace is analyzed by a long short-term memory algorithm based on time series analysis to establish a prediction model for cracking depth. Compared with the traditional time series analysis method, this method has obvious advantages in processing problems with longer historical data sequences. While ensuring that the information of neighboring time nodes is fully utilized, it can also ensure that the information of nodes with more distant similarity will not be forgotten, thereby improving the utilization rate of historical data and the accuracy of the prediction model.

[0070] The second prediction model is a trained long short-term memory network model. The training data set used can be a group of training samples selected from the cracking furnace cracking depth database established above. The training samples are the historical reaction condition data of the current cracking furnace in cracking the current raw materials. The training samples include reaction condition parameter sample values ​​and corresponding cracking depth sample values.

[0071] It should be noted that the input operating parameter sample value is the cracking furnace operating condition at time t, and accordingly, the cracking depth sample value used for comparison with the predicted value should be the actual value of the cracking depth at time t+n (n≥1), and the output is the predicted value of the cracking depth at time t+n (n≥1). Wherein, the value of n can be any time before the cracking production of the current raw material ends, such as 1min, 5min, 10min, 13min, 15min, 24min, 35min, 42min, 48min, 54min, 60min, etc. Preferably, n is 1-30 minutes.

[0072] In some embodiments, the reaction condition parameters are monitoring values ​​of the feed ratio of the cracking reaction, the content of cracking gas and intermediate products, and the temperature and pressure conditions of the cracking reaction determined according to the reaction mechanism of the cracking raw material. The reaction condition parameters include one or more of the following: cracking gas pressure, bottom fuel gas pressure, temperature deviation and COT deviation of each furnace tube group, feed cross-section temperature of each furnace tube group, cracking gas waste heat boiler temperature, furnace temperature, each group of feed furnace outlet temperature, and each group of feed mixing preheating outlet temperature.

[0073] The training process of the second prediction model is as follows: first, an initialized long short-term memory network model is obtained, the input of the long short-term memory network model is the sample value of the reaction condition parameter, and the output is the prediction value of the cleavage depth; then, multiple trainings are performed according to a preset number of training times and a preset deviation threshold, wherein each training includes inputting the sample value of the reaction condition parameter into the long short-term memory network model, and calculating the deviation between the prediction value of the cleavage depth and the sample value of the cleavage depth through an error loss function; when the deviation is less than a second preset threshold or the training reaches a second preset number of times, the training is terminated; otherwise, the long short-term memory network model is corrected and the next round of training continues; finally, the one with the smallest deviation is selected as the final second prediction model.

[0074] In some embodiments, a virtual component model of the cracking raw material is established with the main performance index parameters as a guide, and the second prediction model established on this basis is a cracking reaction mechanism model (hereinafter referred to as the second prediction model as the cracking reaction mechanism model), and the cracking furnace operating conditions at time t are used as input variables to calculate the cracking depth prediction value corresponding to the operating conditions, and the prediction value is compared with the cracking depth operating value obtained by the cracking gas online analyzer to ensure that the relative error of the two is less than a given threshold value of 2%, otherwise, the cracking reaction mechanism model is corrected, and finally the preferred cracking reaction mechanism model is established.

[0075] In some embodiments, the cracking raw material is naphtha. Since naphtha is a mixture of different hydrocarbons, its specific hydrocarbon composition is difficult to obtain by measurement. In actual industrial operation, its cracking performance is mainly characterized according to its main performance indicators. However, when establishing a specific cracking reaction mechanism model, its detailed hydrocarbon composition needs to be constructed. For example, for naphtha, a special cracking raw material, a virtual component model corresponding to it is established according to its main performance indicators, mainly including C5-C12 alkanes and their isomers. Based on this, a cracking reaction mechanism model corresponding to this hydrocarbon composition is established. During the cracking process, the calculation results of the mechanism model are compared in real time with the measurement data of the cracking gas online analyzer, and the operation of the cracking furnace is optimized accordingly.

[0076] In some embodiments, the maximum and minimum allowable values ​​of each operating variable of the cracking furnace are given as constraints with the goal of maximizing the yield of ethylene and propylene, and a multi-objective real-time operation optimization model for the cracking furnace is established. Ultimately, the optimized set values ​​of each operating variable of the cracking furnace are obtained, and the optimal solution geometry of each control variable and its corresponding ethylene and propylene yield are provided to the on-site operator. According to the current actual production needs, the corresponding optimal operating conditions are selected to optimize and adjust the operating conditions of the cracking furnace in real time.

[0077] The current actual production needs mainly include the following situations:

[0078] ·Determine the specific value of the output variable based on the current market prices of ethylene and propylene and the maximum profit;

[0079] According to the different demands for ethylene and propylene in the downstream process of the cracking furnace, the one with the highest demand is selected to determine the specific value of the output variable;

[0080] Based on the range of the ratio of ethylene and propylene yields on site, we hope to obtain the highest comprehensive yield of ethylene and propylene, thereby determining the specific value of the output variable;

[0081] When there is no specific requirement for the yield of other products during the production process and only the yield of ethylene is concerned,

[0082] Select the one with the highest ethylene yield and determine the specific values ​​of the output variables.

[0083] The embodiment of the present invention provides a device for optimizing the operation of a cracking furnace, the composition structure of which is as follows: Figure 3 As shown, it includes a first prediction module, a second prediction module and an operation optimization module, wherein the first prediction module continuously obtains the operating parameters of the cracking furnace, and inputs the operating parameters into a first prediction model of deep learning to obtain a first prediction value of the cracking depth of the cracking furnace at a moment T+T0 after the current moment T, and forms a first curve reflecting the change of the operation effect of the cracking furnace within a preset time period; the second prediction module continuously obtains the reaction condition parameters of the cracking furnace, and inputs the reaction condition parameters into a second prediction model of deep learning to obtain a second prediction value of the cracking depth of the cracking furnace at a moment T+T0 after the current moment T, and forms a second curve reflecting the change of the operation effect of the cracking furnace within a preset time period; the operation optimization module corrects the cracking furnace operation if the similarity between the first curve and the second curve exceeds a preset threshold.

[0084] In this embodiment, the first prediction module and the second prediction module obtain the first and second predicted values ​​of cracking depth based on the operating parameters and reaction condition parameters of the cracking furnace respectively. When the change trends of the monitored value of the cracking depth and the first and second predicted values ​​are inconsistent, it is considered that the current operation of the cracking furnace needs to be optimized, and the vector composed of the operating variables of the cracking furnace and its corresponding cracking depth, the first and second predicted values ​​of the cracking depth are provided to the cracking furnace operator, so that the operator can select the optimal operating conditions of the cracking furnace according to the actual production needs on site, adjust and optimize the operating state of the cracking furnace, and make the operation optimization of the cracking furnace consistent with the optimization direction of the reaction conditions of the cracking furnace.

[0085] In this embodiment, the training methods of the first prediction model and the second prediction model and the training data sets used can refer to the embodiment of the optimization method for cracking furnace operation, and will not be described in detail here.

[0086] An embodiment of the present invention also provides an optimization device for cracking furnace operation, including a processor and a memory. The above-mentioned first prediction module, second prediction module and operation optimization module are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.

[0087] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the kernel parameters can be adjusted to determine whether there is a deviation in the cracking furnace and make corresponding adjustments.

[0088] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0089] An embodiment of the present invention provides a storage medium on which a program is stored. When the program is executed by a processor, the method for optimizing the operation of a cracking furnace of the present application is implemented.

[0090] An embodiment of the present invention provides a processor, which is used to run a program, wherein the program executes the method for optimizing the operation of a cracking furnace of the present application when it is run.

[0091] The embodiment of the present invention provides a device, which includes a processor, a memory, and a program stored in the memory and executable on the processor, and the processor implements the steps of the optimization method for cracking furnace operation of the present application when executing the program. The device in this article can be a server, a PC, a PAD, a mobile phone, etc.

[0092] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program for initiating the steps of the optimization method for cracking furnace operation of the present application.

[0093] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0094] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0095] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0096] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0097] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0098] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0099] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0100] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0101] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for optimizing cracking furnace operation, comprising: Continuously acquiring the operating parameters of the cracking furnace, and inputting the operating parameters into a first prediction model of deep learning to obtain a first prediction value of the cracking depth of the cracking furnace at a time T+T0 after a current time T, and forming a first curve reflecting the change of the operating effect of the cracking furnace within a preset time period; Continuously acquiring the reaction condition parameters of the cracking furnace, and inputting the reaction condition parameters into a second prediction model of deep learning to obtain a second prediction value of the cracking depth of the cracking furnace at a time T+T0 after the current time T, and forming a second curve reflecting the change of the operation effect of the cracking furnace within the preset time period; as well as If the similarity between the first curve and the second curve exceeds a preset threshold, the cracking furnace operation is corrected.

2. The method according to claim 1, characterized in that The cracking severity includes any one of the following: feedstock conversion rate, methane yield, propylene / ethylene yield ratio, methane / propylene yield ratio, hydrogen content of cracked liquid products, outlet temperature, kinetic severity, or yield of C3 and lighter components.

3. The method according to claim 1, characterized in that The operating parameters are the cracking furnace operation control conditions determined according to the cracking furnace tube type and cracking furnace model, and the operating parameters include one or more of the following: total steam feed ratio, total feed amount, COT, target outlet temperature, feed amount of each furnace tube group, feed water flow, bottom fuel gas flow, side wall fuel flow, drum liquid level, radiation section furnace pressure, and feed cross section pressure.

4. The method for optimizing cracking furnace operation according to claim 3, characterized in that: The first prediction model is a trained long short-term memory network model, and the adopted training data set includes a group of training samples, and the training samples are historical operating data of the current cracking furnace in cracking the current raw material, including operating parameter sample values ​​and corresponding cracking depth sample values. The training process of the first prediction model includes: Obtaining an initialized long short-term memory network model, wherein the input of the long short-term memory network model is the sample value of the operation parameter, and the output is the predicted value of the cracking depth; Inputting the sample value of the operating parameter into the cracking depth prediction model, and calculating the deviation between the cracking depth prediction value and the cracking depth sample value by using an error loss function; When the deviation is less than a preset threshold or the training reaches a preset number of times, the training is terminated; otherwise, the long short-term memory network model is corrected and the next round of training is continued; and The one with the smallest deviation is selected as the final first prediction model.

5. The method for optimizing cracking furnace operation according to claim 1, characterized in that: The reaction condition parameters are monitoring values ​​of the feed ratio of the cracking reaction, the content of cracking gas and intermediate products, and the temperature and pressure conditions of the cracking reaction determined according to the reaction mechanism of the cracking raw materials. The reaction condition parameters include one or more of the following: cracking gas pressure, bottom fuel gas pressure, temperature deviation and COT deviation of each furnace tube group, feed cross section temperature of each furnace tube group, cracking gas waste heat boiler temperature, furnace temperature, each group of feed furnace outlet temperature, and each group of feed mixing preheating outlet temperature.

6. The method for optimizing cracking furnace operation according to claim 5, characterized in that: The second prediction model is a trained long short-term memory network model, and the training data set used includes a group of training samples, and the training samples are historical reaction condition data of the current cracking furnace in cracking the current raw material, including reaction condition parameter sample values ​​and corresponding cracking depth sample values, and the training process includes: Obtaining an initialized long short-term memory network model, wherein the input of the long short-term memory network model is the sample value of the reaction condition parameter, and the output is the predicted value of the cracking depth; Inputting the reaction condition parameter sample value into the long short-term memory network model, and calculating the deviation between the cracking depth prediction value and the cracking depth sample value through an error loss function; When the deviation is less than a second preset threshold or the training reaches a second preset number of times, the training is terminated; otherwise, the long short-term memory network model is corrected and the next round of training is continued; and The one with the smallest deviation is selected as the final second prediction model.

7. The method for optimizing the operation of a cracking furnace according to claim 4 or 6, characterized in that: The training data set is obtained by collecting the historical operating data of the cracking furnace, and performing data classification and screening, wherein: The historical operating data include: properties of cracking raw oil, types of cracking furnace tubes, relevant parameters monitored by the DCS, and corresponding ethylene yields and propylene yields; The classification of the historical operating condition data includes: classifying the cracking properties according to the properties of the cracking feedstock, and the classification is based on one or more of the following performance indicators: relative density, distillation range, group composition, viscosity, hydrogen content, and average molecular weight of the oil; The screening of the historical operating condition data includes deleting the operating condition data that meets the following conditions: the error of any one of the relevant parameters, the ethylene yield, and the propylene yield exceeds the corresponding preset range, or there is an abnormal operating condition.

8. The method according to claim 4 or 6, characterized in that: The training data set includes historical operating data of the cracking furnace within the last 0.1-10 years.

9. The method according to claim 1, characterized in that: The preset time period is n1 minutes, T0 is n2 minutes, and n1 and n2 are values ​​between 1-30.

10. An optimization device for cracking furnace operation, characterized in that: include: A first prediction module continuously acquires the operating parameters of the cracking furnace, and inputs the operating parameters into a first prediction model of deep learning to obtain a first predicted value of the cracking depth of the cracking furnace at a time T+T0 after a current time T, and forms a first curve reflecting the change of the operating effect of the cracking furnace within a preset time period; A second prediction module continuously obtains the reaction condition parameters of the cracking furnace, and inputs the reaction condition parameters into a second prediction model of deep learning to obtain a second predicted value of the cracking depth of the cracking furnace at a time T+T0 after the current time T, and forms a second curve reflecting the change of the operation effect of the cracking furnace within the preset time period; as well as The optimization module is run, and if the similarity between the first curve and the second curve exceeds a preset threshold, the cracking furnace operation is corrected.

11. A machine-readable storage medium having instructions stored thereon, the instructions being used to cause a machine to execute: the method for optimizing the operation of a cracking furnace as claimed in any one of claims 1 to 9.

12. A processor, characterized in that: Used to run a program, wherein the program, when run, is used to execute: the method for optimizing the operation of a cracking furnace as described in any one of claims 1-9.

Images