-
The present invention relates to a computer implemented method for determining a (point in) time for decoking steam cracking furnaces of a steam cracking plant and to a method for operating a steam cracking plant as well as to a computing unit, a computer program and a machine-readable storage medium for performing the method.
Background of the invention
-
The present invention is based on the steam cracking technology for the production of olefins and other base chemicals, as e.g. described in the article "Ethylene" in Ullmann's Encyclopedia of Industrial Chemistry, .
-
Presently, the thermal energy required for initiating and maintaining the endothermic cracking reactions in steam cracking is provided by the combustion of fuel gas in a furnace refractory. The process gas initially containing steam and the hydrocarbons to be cracked is passed through so-called cracking coils placed inside the refractory, also called radiant zone or section. On this flow path the process gas is continuously heated, enabling the desired cracking reactions to take place inside the cracking coils, and thus the process gas is continuously enriched in the cracking products. Typical inlet temperatures for the process gas into the cracking coils are between 550 and 750°C, outlet temperatures are typically in the range between 800 and 900°C.
-
In addition to the radiant zone, fired cracking furnaces comprise a so-called convection zone or section and a so-called quench zone or section. The convection zone is usually positioned above the radiant zone and composed of various tube bundles traversing the flue gas duct from the radiant zone. Its main function is to recover as much energy as possible from the hot flue gas leaving the radiant zone. Indeed, only 35 to 50% of the total firing duty is typically transferred within the radiant zone to the process gas passed through the cracking coils. The convection zone therefore plays a central role in the energy management in steam cracking, as it is responsible for the beneficial usage of approximately 40 to 60% of the heat input into a furnace (i.e. of the firing duty).
-
Indeed, when taking the radiant and convection zone together, modern steam cracking plants make use of 90 to 95% of the overall fired duty (based on the fuel's lower heating value or net calorific value). In the convection section, the flue gas is cooled down to temperature levels between 60 and 140°C before leaving the convection section and being released to the atmosphere via stack.
-
The flue gas heat recovered in the convection zone is typically used for process duties such as preheating of boiler feed water and/or hydrocarbon feeds, (partial) vaporization of liquid hydrocarbon feeds (with or without prior process steam injection), and superheating of process steam and high-pressure steam.
-
The quench zone is positioned downstream of the radiant zone along the main process gas route. It is composed of one or more heat exchanger units, having the main functions of quickly cooling the process gas below a maximum temperature level to stop the cracking reactions, to further cool down the process gas for downstream treatment, and to effectively recover sensible heat from the process gas for further energetic usage. In addition, further cooling or quenching can be effected via injection of liquids, e.g. by oil quench cooling when steam cracking liquid feeds.
-
The process gas heat recovered in the quench section is typically used for vaporizing high-pressure (HP) or super-high-pressure (SHP) boiler feed water (typical at a pressure range between 30 and 130 bar absolute pressure), and for preheating the same boiler feed water, before it being fed to a steam drum. Saturated high-pressure or super-high-pressure-steam generated accordingly may be superheated in the convection zone (see above) to form superheated high-pressure or super-high-pressure steam (SHP steam), and from there may be distributed to the central steam system of the plant, providing heat and power for heat exchangers and steam turbines or other rotating equipment. The typical degree of steam superheating achieved in furnace convection zones lies between 150 and 250 K above the saturation temperature (dew point margin). Generally, steam cracking furnaces may operate with high-pressure steam (typically at 30 to 60 bar) or with super-high-pressure-steam (typically at 60 to 130 bar).
-
An important part of the process gas treatment subsequent to quench cooling is compression which is typically performed after further treatment such as the removal of heavy hydrocarbons and process water, in order to condition the process gas for separation. This compression, also called raw gas compression, is typically performed with multistage compressors driven by steam turbines. In the steam turbines, steam at a suitable pressure from the central steam system of the plant mentioned, and thus comprising steam produced using heat from the convection section and from quench cooling, can be used. Typically, in a steam cracking plant of the prior art, heat of the flue gas (in the convection zone) and heat of the process gas (in the quench zone) is well balanced with the heat demand for producing a large part of the steam amounts needed for heating and driving steam turbines. In other words, waste heat may be more or less fully utilized for generating steam which is needed in the plant. Additional heat for steam generation may be provided in a (fired) steam boiler.
-
For reference, and to further illustrate the background of the invention, a conventional fired steam cracking arrangement, steam cracking plant or ethylene plant is illustrated in Figure 1 in a highly simplified, schematic partial representation and is designated 100.
-
The steam cracking arrangement or steam cracking plant 100 illustrated in Figure 1 comprises, as illustrated with a reinforced line, one or more cracking furnaces 10. For conciseness only, "one" cracking furnace 10 is referred to in the following, while typical steam cracking arrangements 100 may comprise a plurality of cracking furnaces 10 which can be operated under the same or different conditions. Furthermore, cracking furnaces 10 may comprise one or more of the components explained below.
-
The cracking furnace 10 comprises a radiant zone 11 and a convection zone 12. In other embodiments than the one shown in Figure 1, also several radiant zones 11 may be associated with a single convection zone 12, etc.
-
In the example illustrated, several heat exchangers 121 to 125 are arranged in the convection zone 12, either in the arrangement or sequence shown or in a different arrangement or sequence. These heat exchangers 121 to 125 are typically provided in the form of tube bundles passing through the convection zone 12 and are positioned in the flue gas stream from the radiant zone 11.
-
In the example illustrated, the radiant zone 11 is heated by means of a plurality of burners 111 arranged on the floor and wall sides of a refractory forming the radiant zone 11, which are only partially designated. In other embodiments, the burners 111 may also be provided solely at the floor side or solely at the wall sides. The latter may e.g. be the case when pure hydrogen is used for firing.
-
In the example illustrated, a gaseous or liquid feed stream 101 containing hydrocarbons is provided to the steam cracking arrangement 100. It is also possible to use several feed streams 101 in the manner shown or in a different manner. The feed stream 101 is preheated in the heat exchanger 121 in the convection zone 12.
-
In addition, a boiler feed water stream 102 is passed through the convection zone 12 or, more precisely, the heat exchanger 122, where it is preheated. The boiler feed water stream 102 is thereafter introduced into a steam drum 13. In the heat exchanger 123 in the convection zone 12, a process steam stream 103, which is typically provided from a process steam generation system located outside the furnace system of the steam cracking arrangement 100, is further heated and, in the example illustrated in Figure 1, thereafter combined with the feed stream 101.
-
A stream 104 of feed and steam formed accordingly is passed through a further heat exchanger 125 in the convection zone 12 and is thereafter passed through the radiant zone 11 in typically several cracking coils 112 to form a cracked gas stream 105. The illustration in Figure 1 is highly simplified. Typically, a corresponding stream 104 is evenly distributed over into a number of cracking coils 112 and a cracked gas formed therein is collected to form the cracked gas stream 105.
-
As further illustrated in Figure 1, a steam stream 106 can be withdrawn from the steam drum 13 and can be (over)heated in a further heat exchanger 124 in the convection zone 12, generating a high-pressure steam stream 107. The high-pressure steam stream 107 can be used in the steam cracking arrangement 100 at any suitable location and for any suitable purpose as not specifically illustrated.
-
The cracked gas stream 105 from the radiant zone 11 or the cracking coils 112 is passed via one or more transfer lines to a quench exchanger 14 where it is rapidly cooled for the reasons mentioned. The quench exchanger 14 illustrated here represents a primary quench (heat) exchanger. In addition to such a primary quench exchanger 14, further quench exchangers may also be present.
-
The cooled cracked gas stream 110 is passed to further process units 15 which are shown here only very schematically. These further process units 15 can, in particular, be process units for scrubbing, compression and fractionation of the cracked gas, and a compressor arrangement including a steam turbine, which may be operated using steam from the steam drum 13, being indicated with 16.
-
In the example shown, the quench exchanger 14 is operated with a water stream 108 from the steam drum 13. A steam stream 109 formed in the quench exchanger 14 is returned to the steam drum 13.
-
Petroleum coke, also referred to as coke or petcoke, a carbon-rich solid material, is a by-product of the cracking process. This coke is derived in the final stage of the cracking process, wherein long chain hydrocarbons are split into shorter chains, and can accumulate in coils, tubes or other components of the cracking furnaces 10. In the course of a decoking process, this accumulated coke is removed in the cracking furnaces 10. With growing amounts of accumulated coke, the efficiency and production yield of cracking furnaces 10 continuously decay. A regular and frequent decoking of the cracking furnaces 10 is therefore necessary.
Disclosure of the invention
-
The present invention relates to a computer implemented method for determining a time, i.e. point in time, for decoking steam cracking furnaces of a steam cracking plant and to a method for operating a steam cracking plant as well as to a computing unit, a computer program and a machine-readable storage medium for performing the method with the features of the independent claims. Further advantages and embodiments of the invention will become apparent from the description and the appended figures.
-
The steam cracking plant or ethylene plant comprises at least two steam cracking furnaces. The present invention allows to computationally simulate the operation of the entire steam cracking plant and to determine optimised individual times for decoking each of the cracking furnaces using computer implemented optimisation. The corresponding actual steam cracking plant can then be operated according to the theoretical simulation and the cracking furnaces can actually be decoked according to the computationally determined optimised decoking times.
-
For this purpose, a computer implemented model is provided for computationally simulating an operation of each cracking furnace of the steam cracking plant during a predetermined time interval. By means of this model, a production yield, e.g. an ethylene yield, of each cracking furnace and an accumulation of coke in each cracking furnace, e.g. in one or several coils of each cracking furnace, during this predetermined time interval are determined. For example, operation of the cracking furnace during a predetermined time interval of one day, one week, one month, etc. can be simulated.
-
This model is particularly a theoretical, mathematical, thermodynamic model of the steam cracking plant. Particularly, the model can be a composition of several individual sub-models, wherein each sub-model simulates at least a part or a section of one of the cracking furnace. For example, several stand-alone models can be provided, wherein each of these models characterises a specific part of one of the cracking furnaces, e.g. a radiant section, a convection section, etc. These various stand-alone sub-models can particularly be combined to the common, overall model of the entire steam cracking plant.
-
The model can further for example be used for scenario simulation, especially for studying effects on the production in case of changing various operating conditions. Further, the model can e.g. be used as an online monitor for supervising and evaluating the performance of the current operation of the plant.
-
Further, at least two scenarios or operating scenarios are defined, predetermined or specified for each cracking furnace of the steam cracking plant, wherein each operating scenario specifies or defines individual operating conditions for the operation of the corresponding cracking furnace. These operating scenarios particularly characterise different possibilities, how the individual furnaces can be operated and how the cracking process in the individual furnaces can be executed. The model can expediently simulate the operation of each furnace with the corresponding individual operating scenarios. The different operating conditions of each scenario can e.g. define set points for temperature, pressure, flowrate, etc.
-
According to the present invention, the model or data generated via the model is used for a computer implemented optimisation, expediently for formulating and solving an optimisation problem. Utilising the model, it is formulated as an optimisation problem to maximise the production yield of each cracking furnace in dependence of an individual (point in) time of decoking of each cracking furnace, at which during the predetermined time interval the accumulated coke in the corresponding cracking furnace is at least partially removed. The optimisation problem is solved and an optimised time of decoking for each cracking furnace is determined.
-
Solving the optimisation problem comprises selecting one of the corresponding at least two scenarios for each of the cracking furnaces. Expediently, a set of representative operating scenarios for each furnace is defined and the corresponding optimisation algorithm chooses one of the defined scenarios for each furnace for the operation during the predetermined time interval. The optimisation process thus especially determines not only the optimum times for decoking but also the specific operating conditions under which the various furnaces are to be operated.
-
Particularly, for solving the optimisation problem, the individual decoking times of each furnace for different operating scenarios during the time interval are varied until a set of decoking times is found, with which a maximum overall production yield of the steam cracking plant during the time interval can be achieved. Thus, different decoking times for each furnace are weighted against each other until an optimised set of times is found. For example, it can be weighted against each other whether individual furnaces are to be decoked once or several times or even not at all during the predetermined time interval, when the various furnaces are operated under certain operating conditions.
-
Operating a corresponding furnace in different scenarios can result in different coking behaviour of the furnace and can therefore influence, when the furnace needs to be decoked. For example, the use of different feeds with different coking behaviour can also impact the yield, such that change of feed may also be considered over the optimization time. Therefore, for determining the decoking schedule it particularly needs to be considered under which operating conditions the furnaces are operated. Thus, the defined operating scenarios provide an additional degree of freedom or parameter, which can be adjusted in order to determine the optimised decoking times of each furnace. Since the steam cracking plant can comprise a multitude of different furnaces and since each furnace can be operated under a multitude of possible operating conditions, finding an optimum combination of furnace operating conditions for an optimised decoking schedule can be a very cumbersome process and might not be performed manually or only with very great effort. The use of the computational optimisation process is therefore particularly expedient, since it can automatically consider the various operating conditions and find an optimum combination of operating scenarios for an optimum decoking schedule.
-
For example, a decoking condition can be predetermined to decoke a corresponding cracking furnace when the corresponding amount of accumulated coke reaches a threshold or is in a certain range around a threshold. Further, individual properties of each furnace during the time interval can be taken into account, e.g. a velocity or rate of coke accumulation or for example if a furnace has to be shut down because of different reasons, e.g. because an operating temperature reaches a threshold.
-
The formation of coke within the coils of the cracking furnaces is an undesired but unavoidable by-product of the steam cracking process. With an increasing amount of accumulated coke whiting the furnace coils, the efficiency and production yield of the corresponding furnace decreases. Regular and frequent decoking of each cracking furnace is therefore necessary, e.g. when the amount of accumulated coke in the corresponding cracking furnace reaches a predetermined threshold. The coke formation of the individual furnaces therefore influences a runtime of the furnaces until they need to be shut down for decoking in order to increase their efficiency once again. However, it is for example undesirable that several or even all furnaces of the plant have to be decoked simultaneously, which could lead to a decrease in production capacity and to unnecessary losses in production yield. In order to avoid such simultaneous downtimes of several furnaces due to decoking, the present invention allows to plan the downtimes for decoking of each furnace in dependence of each other and coordinated with each other, such that maximum production efficiency can be provided. Determining the optimised times for decoking of each cracking furnace expediently allows to determine an optimised decoking schedule and to optimally plan the individual downtimes of the furnaces for coke removal.
-
Therefore, the present invention allows an efficient operation of the steam cracking plant and to reasonably plan the downtimes for decoking of each furnace, expediently avoiding unnecessary and simultaneous downtimes of several furnaces. The invention particularly provides a possibility to systematically setup steam cracking furnace models and a corresponding optimisation framework including thermodynamic simulation models. The invention particularly provides a tool that can predict the plant operation and automatically optimise the furnace decoking schedule with respect to production yield over a planning horizon. Particularly, an optimised turnaround management and an optimised recurrent cycle between production and following decoking can be provided. The cracking furnace operation can particularly be optimised with respect to individual products of each furnace. A transparent and reproducible setup of furnace scheduling can be provided.
-
Expediently, various components of the steam cracking plant can be considered and can especially be simulated by the model and taken into account for determining the decoking times, e.g. supplying equipment like a product separation section or the furnace convection section. Furthermore, the optimisation algorithm can be used to optimise the overall site production especially with respect to product market prices, e.g. optimising a product that has currently a high market price.
-
The requirement that each furnace needs to be switched off regularly for decoking particularly yields in a binary problem (i.e. the furnace is either switched on or off), which can expediently be addressed using a mathematical solution or optimisation strategy. Further, different furnace operating conditions can result in different coking behaviour and consequently in different runtimes achievable for each furnace. In order to find optimised decoking times, a multitude of different scenarios need to be considered in order to find an optimal combination of furnace operation conditions yielding an optimum production yield and avoiding capacity limitations. The use of the computational optimisation is therefore particularly expedient.
-
In general, the optimisation problem is formulated in that a solution space Ω, i.e. a set of possible solutions or variables, as well as an objective function fare determined. To solve this optimisation problem, a set of values of variables or solutions x ∈ Ω is sought such that f(x) satisfies a given criterion, for example, becomes maximum or minimum. Often, constraints can be specified, whereby admissible solutions x must fulfil these specified constraints. Such constraints or boundary conditions can be given in the form of equations or inequalities or describe an explicit set, e.g. only integer values of process parameters. The set of all solutions that satisfy all given boundary conditions is called the admissible set.
-
The present invention therefore improves common simulation and decoking scheduling strategies of steam cracking plants according to the prior art by expediently formulating and solving the corresponding optimisation problem. For a detailed discussion of common simulation and decoking scheduling strategies, it is e.g. referred to the following articles:
- " Kunjie Yu, Lyndon While, Mark Reynolds, Xin Wang, Zhenlei Wang, Cyclic scheduling for an ethylene cracking furnace system using diversity learning teaching-learning-based optimization, Computers & Chemical Engineering, Volume 99, 2017, Pages 314-324, ISSN 0098-1354, https://doi.org/10.1016/j.compchemeng.2017.01.024 "
- " Development of Optimal Decoking Scheduling Strategies for an Industrial Naphtha Cracking Furnace System, Heejin Lim, Jaein Choi, Matthew Realff, Jay H. Lee, and Sunwon Park, Industrial & Engineering Chemistry Research 2006 45 (16), 5738-5747, DOI: 10.1021/ie050129n "
- " Dynamic Scheduling for Ethylene Cracking Furnace System, Chuanyu Zhao, Chaowei Liu, and Qiang Xu, Industrial & Engineering Chemistry Research 2011 50 (21), 12026-12040, DOI: 10.1021/ie200318p "
- " Lijie Su, Lixin Tang, Ignacio E. Grossmann, Scheduling of cracking production process with feedstocks and energy constraints, Computers & Chemical Engineering, Volume 94, 2016, Pages 92-103, ISSN 0098-1354, https://doi.org/10.1016/j.compchemeng.2016.07.023 "
-
Advantageously, defining a corresponding operating scenario for a respective cracking furnace comprises evaluating effects of changing operating conditions on the production yield of the respective cracking furnace. Expediently, a set of operating conditions can be defined as corresponding operating scenario, with which the production yield of the respective cracking furnace lies within a desired interval or within an interval around a desired value. It can therefore particularly be evaluated under which operating conditions the furnaces can technically reasonably be operated in order to achieve reasonable production yields. The operating scenarios can e.g. be defined in advance by process experts and can cover a wide range of possible operating conditions. The scenarios expediently follow guidelines in terms of plausibility, operability, profitability. During process design various scenarios can be calculated (e.g. 100%, 90%, 105% cases, but also different severities and feedstocks). This knowledge can particular be used to specify realistic operating scenarios, which then expediently also fulfil full thermodynamic requirements of the overall process. For example, previous periods of operating of the various furnaces can be evaluated in order to extract reasonable operation conditions, with which reasonable production yields can be achieved. Alternatively or additionally, underlying technical principles of the furnaces and process sequences of furnace operation can expediently be evaluated. For example, the specific components installed in the furnace can be considered, e.g. specific heat exchangers, burners, etc., as well as the specific materials of these components. It can further e.g. be considered how these components and their materials are affected by decoking. Thus, reasonable possibilities how the different furnaces can be operated are specified for the optimisation algorithm to choose from. For example, for different operating scenarios, different feeds with different coking behaviour can be considered, which impact the yield in a different way, such that a change of feed can be considered over the optimisation time. The algorithm can then weight the different operating scenarios and when to decoke the furnaces during these different operating scenarios against each other. Thus, the algorithm can not only determine the best time for decoking, but also the best possible operating scenario.
-
Preferably, a so called mixed integer linear programming (MILP) is used for formulating and solving the objective problem, wherein the objective function is linear and the constraints can be represented by a system of piecewise linear equations and inequalities that also contain integer decisions. It is preferably also possible to use non-linear mixed integer programming.
-
Advantageously, formulating the optimisation problem further comprises defining constraints or boundary conditions regarding the operation of each cracking furnace, especially regarding properties or physical properties of each cracking furnace.
-
Accordingly, the optimisation problem is particularly solved in dependence of these defined constraints. Expediently, these constraints characterise actual physical properties or measurement values of physical quantities of the various furnaces during their operation.
-
Advantageously, defining the constraints comprises defining threshold values for properties or physical properties of each cracking furnace, especially threshold values for measurement values of physical quantities of the furnaces during their operation. Expediently, maximum values for physical quantities can be defended as the constraints, which are not to be exceeded during furnace operation. For example, threshold values for key performance indicators (KPI) of each furnace can be defined, e.g. threshold or maximum values for a temperature and/or a pressure of a component of each cracking furnace. Expediently, a maximum tube metal temperature and/or a maximum coil inlet pressure can be defined as constraints. When a corresponding threshold value is reached during operation of one of the furnaces, e.g. when a tube metal temperature or a coil inlet pressure of a furnace exceeds its defined maximum value, a shutdown of the corresponding furnace might be necessary. The optimisation algorithm can then consider, whether it is reasonable to perform a decoking of the corresponding furnace during this downtime. If so, the optimisation algorithm can accordingly adjust the decoking times of the remaining furnaces.
-
According to a preferred embodiment, providing the model comprises providing a thermodynamic model of the cracking furnaces, particularly at least one thermodynamic model of each cracking furnace. These thermodynamic models can expediently be presented by non-linear equations like cubic spline etc. Particularly, non-linear equations in the thermodynamic model are linearized. Expediently, these non-linear equations are approximated by means of linearisation. For example, when using mixed integer linear programming (MILP) for solving the optimisation problem, non-linearities, which occur e.g. in a full-blown thermodynamic model of the radiant section of furnace, need to be linearised. Although linearization might result in a loss of accuracy, it yields the advantages of handling and solving the optimisation problem efficiently. The thermodynamic model of the steam cracking furnaces can show a strong non-linear behaviour e.g. in temperature evolvement over coil length, chemical kinetics and also in coke formation velocity prediction overtime. Thus, the non-linear equations in the model can e.g. be approximated using a piecewise linear representation with one or multiple independent variables. However, each piecewise linearization generates integer variables that need be chosen by the optimisation algorithm. If the linearization is directly applied to all non-linearities of the thermodynamic model, the combinatoric of binary variables can grow exponentially, resulting in very long solving times. By defining the corresponding set of operating scenarios for each cracking furnaces and by having the optimisation algorithm chose one scenario for each furnace, the solving time can significantly be decreased. Having selected the corresponding scenarios, the production yield, coke formation and physical properties of each furnace can expediently be predicted by a minimal set of piecewise linear equations.
-
The invention further relates to a method of operating a corresponding steam cracking plant, which comprises least one cracking furnace. For this purpose, an optimised time of decoking is determined for each cracking furnace according to the above description using the computer implemented model simulating the operation of each furnace and applying the computational optimisation. The actual steam cracking plant is then operated over an operating time period corresponding to the predetermined time interval of the corresponding simulation. Decoking of the various cracking furnaces is performed at the correspondingly determined optimised times. Further, each furnace is operated during this operating time period according to the corresponding operating scenario selected by the optimisation algorithm. The model and the optimisation can therefore be used to determine an operating schedule for the steam cracking plant, particularly to determine how to operate and when to decoke each cracking furnace to achieve a maximum production yield.
-
A computing unit according to the invention is configured, in particular by a computer program, to carry out an inventive method.
-
The implementation of the invention in the form of software is advantageous because this allows particularly low costs, especially if an executing processing unit is still being used for other tasks and therefore is present anyway. Suitable media for providing the computer program are in particular discettes, hard drives, flash memory, EEPROM, CD-ROMs, DVDs etc. Downloading a program via computer networks (Internet, intranet, etc.) is possible.
-
Further advantages and developments of the invention are specified in the description and the associated drawings.
-
It should be noted that the previously mentioned features and the features to be further described in the following are usable not only in the respectively indicated combination, but also in further combinations or taken alone, without departing from the scope of the present invention.
-
The present invention will now be described further, by way of example, with reference to the accompanying drawings, in which
- Figure 1
- schematically shows a steam cracking plant with a cracking furnace, wherein a time for decoking the cracking furnace can be determined in the course of a preferred embodiment of the method according to the present invention.
- Figure 2
- schematically shows a preferred embodiment of the method according to the present invention as a block diagram.
Detailed description
-
Figure 1 schematically shows a steam cracking arrangement or steam cracking plant 100 comprising a cracking furnace 10 and was already described at the outset.
-
As explained above, the steam cracking plant 100 can expediently comprise several cracking furnaces 10, which can e.g. be identical constructed. The radiant zone 11 of each furnace can be heated by means of a plurality of burners 111 arranged on the floor and wall sides of a refractory forming the radiant zone 11. The process gas comprising a mixture of hydrocarbons and steam is passed through the radiant zone 11 in typically several cracking coils 112 of the corresponding furnace 10 to form the cracked gas stream 105.
-
Petroleum coke or coke or is derived as a by-product of the cracking process and accumulates in the cracking coils 112 of each furnace 10. Efficiency and production yield of the corresponding cracking furnaces 10 continuously decay with growing amounts of accumulated coke in the cracking coils 112, such that frequent decoking of each furnace 10 is necessary.
-
In order to avoid that several or all of these furnaces 10 are decoked simultaneously, leading to an undesirable decrease in production capacity, individual decoking times for decoking of each cracking furnace 10 are determined in the course of a preferred embodiment of the method according to the present invention, as shall be explained hereafter with reference to the figures 2 and 3.
-
Figure 2 schematically shows a preferred embodiment of the method according to the present invention as a block diagram.
-
In step 201, a computer implemented model is provided for computationally simulating the operation of each cracking furnace during a predetermined time interval. For example, the model can be a composition of several stand-alone sub-models, wherein different sections of each cracking furnace can be emulated by individual sub-models. In particular, a composite computational model of the entire steam cracking plant is provided, which can determine a production yield of each cracking furnace and an accumulation of coke in each cracking furnace during this predetermined time interval. For example, the model can further be used e.g. for studying effects on the production yield when changing furnace operating conditions or e.g. for online monitoring and supervising performance of the plant operation.
-
In step 202, a set of different operation scenarios is defined for each cracking furnace. Each of these operating scenarios specifies individual operating conditions for the operation of the corresponding cracking furnace. For example, these operating conditions can e.g. define set points for temperature, pressure, flowrate, etc. The operating scenarios are e.g. defined by a process expert and strongly depend on the given plant configuration going to be optimised in terms of the decoking schedule. The scenarios can e.g. refer to a plant operation at workloads between 90% and 105%.
-
In step 203, utilising the model an optimisation problem is formulated, in particular to maximise the production yield of each cracking furnace in dependence of a time or decoking time of decoking of each cracking furnace, at which during the predetermined time interval the accumulated coke is in the corresponding cracking furnace is at least partially removed. In order to solve the optimisation problem, the individual furnace decoking times are to be varied until an optimum set of decoking times can be found, with which a maximum possible production yield can be achieved.
-
In step 204, constraints for the optimisation problem are defined, for example regarding physical properties of each furnace. For example, threshold values for key performance indicators KPI of each furnace can be defined, expediently threshold values for shutdown KPIs. For exorable, a maximum tube metal temperature and a maximum coil inlet pressure for each cracking furnace can be defined as constraints.
-
In step 205, the optimisation problem is computationally solved by a corresponding optimisation algorithm, e.g. by using mixed integer linear programming (MILP). For this purpose, the corresponding algorithm selects in step 206 for each cracking furnace one of the operating scenarios as defined in step 202. In step 207, the algorithm performs a linearisation and approximates non-linear equations in the thermodynamic model of the various cracking furnaces. By defining the sets of operating scenarios and selecting one scenario for each furnace, the time for solving the optimisation problem can significantly be decreased, since only non-linear equations in the models for the selected scenario are approximated by means of the corresponding linearisiation. In step 208, the algorithm determines as a result of solving the optimisation problem optimised decoking times for each furnace.
-
Therefore, the optimisation algorithm can adjust as degrees of freedom over the predetermined time interval or prediction horizon a status of each furnace (i.e. being activated and performing the cracking process or being deactivated for decoking), as well as a production rate and further a coil outlet temperature (COT) and conversion. The optimisation algorithm adjusts these parameters in order to maximize the production yield or profit of the steam cracking plant in dependent of given and predetermined constraints, e.g. in dependence of site topology, production constraints (e.g. furnace capacity and demand) and coke formation as well as decoking or shutdown constraints (e.g. maximum tube metal temperature and maximum coil inlet pressure) and the selected predefined operation scenarios.
-
In step 209, the actual steam cracking plant is then operated over an operating time period corresponding to the simulated predetermined time period of the thermodynamic model and decoking of the various cracking furnaces is performed at the optimised decoking times as determined in step 208.
-
In the following, exemplary values of physical quantities are given, which can be determined in the course of solving the optimisation problem for an exemplary steam cracking plant with two cracking furnaces (#1 and #2), wherein a specific operating scenario was selected for each of these two furnaces.
-
As constraints, for each of the furnaces #1 and #2 a maximum tube metal temperature of e.g. 995°C was defined. As a further constraints, a maximum coil inlet pressure of e.g. 3.6 bara was defined for the first furnace #1 and a maximum coil inlet pressure of e.g. 4.1 bara was defined for the second furnace #2. With a corresponding thermodynamic model, operation of each furnace over a predetermined time interval of eight days was simulated. For each physical quantity, one simulated value per day was simulated.
-
Hereafter, exemplary sets of values of various physical quantities are given for each of the furnaces #1 and #2, which were determined as a result of solving the underlying optimisation problem. Each set comprises eight simulated values, wherein each value represents one day of operation. The first two sets of values represents a status of the corresponding furnace, wherein the value "1.0" characterises that the corresponding furnace is to be activated on the respective day, and wherein the value "0.0" characterises that the corresponding furnace is to be deactivated. The remaining sets represent an average coil outlet temperature, a total ethylene yield, an average velocity of coke accumulation, an average height of accumulated coke, an average tube metal temperature and an average coil inlet pressure to be expected for the corresponding furnace on the respective day.
-
Furnace status:
- #1: ["1.00", "1.00", "1.00", "1.00", "1.00", "1.00", "1.00", "1.00"]
- #2: ["1.00", "1.00", "1.00", "0.00", "1.00", "1.00", "1.00", "1.00"]
-
Coil outlet temperature:
- #1: ["855.00", "853.00", "853.00", "853.00", "853.00", "853.00", "853.00", "861.00"]
- #2: ["861.00", "861.00", "861.00", "0.00", "853.00", "855.00", "853.00", "861.00"]
-
Ethylene yield:
- #1: ["319.00", "315.00", "315.00", "315.00", "315.00", "315.00", "315.00", "328.00"]
- #2: ["328.00", "328.00", "328.00", "0.00", "319.00", "320.00", "319.00", "320.00"]
-
Velocity of coke accumulation:
- #1: ["0.80", "0.70", "0.70", "0.70", "0.70", "0.70", "0.70", "1.30"]
- #2: ["1.20", "1.20", "1.20", "0.00", "0.80", "0.90", "0.80", "1.20"]
-
Height of accumulated coke:
- #1: ["0.00", "0.80", "1.50", "2.20", "2.90", "3.60", "4.30", "5.00"]
- #2: ["0.00", "1.20", "2.40", "0.00", "0.00", "0.80", "1.70", "2.50"]
-
Tube metal temperature:
- #1: ["982.00", "983.60", "985.00", "986.40", "987.80", "982.20", "990.60", "992.00"]
- #2: ["990.00", "992.40", "994.80", "0.00", "990.00", "991.60", "993.40", "995.00"]
-
Coil inlet pressure:
- #1: ["3.10", "3.18", "3.25", "3.32", "3.39", "3.46", "3.53", "3.60"]
- #2: ["3.50", "3.62", "3.74", "0.00", "3.50", "3.58", "3.67", "3.75"]
-
According to the above example, the tube metal temperature of the second cracking furnace #2 is expected to reach the corresponding threshold on the third day of operation and the second furnace #2 is to be deactivated during the fourth day. In the first three day of operation, the second furnace #2 produced a higher ethylene yield than the first furnace #1. Accordingly, the amount of accumulated coke grows quicker in the second furnace #2 than in the first furnace #1 during these first three days. Therefore, decoking of the second cracking furnace #2 is to be performed on the fourth day. Since coke is accumulated less rapidly in the first furnace #1, no decoking is necessary in the first furnace #1 during the simulated eight day operation. Therefore, the fourth day of operation is determined as the optimised decoking time for the second furnace #2 whereas the optimised decoking time for the first furnace #1 is determined to be after the eight day of operation.
-
For the above example, a very simplified linear representation of the corresponding thermodynamic furnace model was used for the sake of simplicity. However, more complex models can be applied, which can simulate furnaces with higher accuracy by defining more operation scenarios and by extending piecewise linearisation models to more independent variables to achieve a description of full furnace operating range.
