CN101261516A - Oil refinery real time intelligent dynamically optimized scheduling modelling approach based on affair logic - Google Patents

Abstract

The invention discloses a modeling method for real-time intelligent dynamic optimization dispatching of an oil refinery plant on the basis of event logic. The modeling method increases the rapidity and the flexibility of the generation of a dynamic dispatching model, improves the practicality of the dispatching optimization theory and solves the difficult problem of modeling of the dynamic dispatching of the petroleum refining industry to a certain extent, which can obtain the better dispatching optimization at the same time of ensuring the real-time property of the dispatching. The method is as follows: 1) the analysis of the process; 2) the definition of the event; 3) the description of the event according to the definition of the event; 4) the generation of the corresponding event logic according to the description of the event; 5) the expression of the generated event logic according to the production process and the description of the event respectively; 6) the respective establishment of a process dispatching model and an event module according to the expression of the event logic of the step 5; 7) and the final constitution of a dynamic dispatching model on the basis of the event logic by combining the process dispatching model and the event model.

Description

Modeling method based on the oil refinery real time intelligent dynamically optimized scheduling of affair logic

Technical field

The invention belongs to advanced the manufacturing and technical field of automation, relate in particular to a kind of modeling method of the oil refinery real time intelligent dynamically optimized scheduling based on affair logic.

Background technology

Oil refining process is the continuous flow procedure of a quasi-representative, compares with other industrial process, have continuously, steadily, high energy consumption and device is many and outstanding feature such as complexity.Though oil refining process is steady substantially, uncertain incident all may take place at any time but plant failure, energy shortage, quality of material are defective etc., caused respectively install production status, material and the energy produce disappear, the continuous variation of supply and demand, often cause the waste of the resource and the energy, cause the company interest loss even threaten production safety.Dynamic dispatching is exactly in time to adjust and to rearrange production control instructions such as the product amount of disappearing of the running status of process units and parameter, material and the energy and tank farm stock after incident takes place as soon as possible, reduce unnecessary waste and loss as far as possible, optimization production operation to greatest extent.As seen the difficulty of dynamic dispatching and importance.

At present, the main dynamic dispatching method that is applied to continuous process comprises the artificial intelligence approach of mathematical optimization method, heuristic and rule-based reasoning.The mathematical optimization method can satisfy the optimization requirement of scheduling decision preferably, but because model scale and the restriction of finding the solution often can not satisfy the real-time requirement of production, can not fully reflect complicacy, randomness and the dynamic of production environment; Though the heuristic real-time is better, do not give full play to the powerful optimization advantage of mathematical model, be more difficult to get the decision scheme that makes the decision maker satisfied; Though the artificial intelligence approach of rule-based reasoning can satisfy the real-time of production, rule is obtained the comparison difficulty, and the content of optimization is less, and optimizing scheduling is relatively poor.Because the existence of the problems referred to above has caused present dynamic dispatching theory to be difficult to practical present situation.Therefore, continuity according to oil refining process, stationarity, global optimization is adjusted the characteristics that combine and to the high real-time requirement of dynamic dispatching with local, how effectively to utilize mathematical programming, heuristic and artificial intelligence approach, carrying out optimizing scheduling in conjunction with the basis of logic planning, find gratifying equilibrium point between property optimized and the real-time, is a feasible and valid approach that solves dynamically optimized scheduling.

The relevant production scheduling method of having applied at present, proposing a kind of as No. 200410021578.1 patents " rule-based special steel industry production dynamic dispatching method " is the rule-based global optimization row product and the dynamic dispatching method of constraint condition with special steel industry key equipment production rule and order contractual delivery phase, this method is not making full use of production rule under the situation by mathematical programming, though had certain scheduling real-time, can't guarantee the optimization of dispatching.200510061523.8 number patent " a kind of modeling method of visible production process description of process industry " proposes a kind of visual modeling method that is applicable to the intermittence or the semi-batch production system of general flow industrial enterprise, this method only realizes having set forth the process of setting up the production system model from software, though comprised a part of logical constraint, do not related to the application therein of optimizing scheduling and logical constraint.

Summary of the invention

In order to solve the problem of above existence, an object of the present invention is to provide a kind of modeling method that is applicable to oil refining process based on the oil refinery real time intelligent dynamically optimized scheduling of affair logic.

For realizing above-mentioned purpose of the present invention, the present invention adopts following technical scheme:

A kind of modeling method of the oil refinery real time intelligent dynamically optimized scheduling based on affair logic, its method be,

1) industrial analysis;

2) event definition;

3) according to event definition, the description incident;

4), produce the events corresponding logic according to event description;

5) respectively the affair logic that produces is expressed according to flow scheduling demand and event description;

6) according to the expression of step 5), set up flow scheduling model and event model respectively to affair logic;

7) the final dynamic dispatching model that constitutes based on affair logic of combined process flow scheduling model and event model.

The modeling method of the oil refinery real time intelligent dynamically optimized scheduling based on affair logic provided by the invention comprises the following steps: 1) industrial analysis; 2) event definition; 3) according to event definition, the description incident; 4), produce the events corresponding logic according to event description; 5) the respectively affair logic that produces is expressed according to flow scheduling demand and event description; 6) according to the expression of step 5 pair affair logic, set up flow scheduling model and event model respectively; 7) the final dynamic dispatching model that constitutes based on affair logic of combined process flow scheduling model and event model.

Step 1) also comprises substep: 1-1) analyze the various a large amount of production incident that causes dynamic dispatching in the oil refining production, obtain the general characteristic of these incidents, as the basis of definition incident; 1-2) analyze the oil refining production procedure, the technology that obtains process units connects (as accompanying drawing 2), the operational factor of process units is (as working ability, yield, energy resource consumption rate and production cost etc.) and the production information of material (as consumption, quantum of output, tank farm stock and storage expenses etc.), as setting up the flow scheduling model based.

In step 2) in, be the unusual condition of process units, material or the energy with event definition, show as the accident in the production.These incidents can cause the running status (as device start and stop, processing capacity etc.) to process units, and the quantum of output of the material or the energy, consumption and tank farm stock etc. carry out the adjustment of the part or the overall situation, to guarantee the smooth production of safety and steady.For embodying the general characteristic of incident, further it is defined as polynary group of form, promptly by state, attribute, behavior becomes the definition content of incident with requirement groups.Wherein whether the state-event presentation of events takes place; Event attribute comprises time of origin, object takes place, influence time scope and influence the device scope, when presentation of events takes place respectively, occur on which device, which kind of material or the energy, occur to the elimination incident for the needed time of influence of producing and the process units of the required adjustment running status of response events from incident; The technology adjustment measure that the dispatcher takes for processing events is represented in the incident behavior, as the side line connection of adjusting gear, and the processing capacity of adjusting gear, the setting device start and stop, the kind of modifier processing or output material changes material purposes etc.; Requirements, event is illustrated in the requirement of interior purchase volume to some material of events affecting time, sales volume, tank farm stock etc.

In step 3) according to step 2) described definition content is described incident.Wherein, describe incident with "Yes" and take place, describe incident with "No" and do not take place for state-event; For event attribute, with concrete scheduling time section Time To Event is described, with the set of concrete device, material or energy numbering description incident generation object, with scheduling time segment limit the events affecting time is described, number description and influence the device scope with influence device scheduling set; For the incident behavior, to describe with the production status (connecting stock material etc. as device technique) of adjusted device or material, the numbering during device of wherein adjusting or material are gathered separately with it is described; For requirements, event, describe with the numbering and the demand thereof of material.

In step 4), produce the events corresponding logic according to the described event description of step 3).Wherein describe and produce Event triggered logic E by state-event _s, if incident s generation, then E _s=True, otherwise E _s=False, wherein s is a Case Number, s incident in the presentation of events S set can trigger its events corresponding model by the Event triggered logic; By event attribute generation event time logic ET is described _t, event device logic EI _iWith incident scope logic ETI _It, if scheduling time section t is in the events affecting time range, i.e. t ∈ T ¹ET then _t=True, otherwise, ET _t=False, if process units i is in events affecting device scope, i.e. i ∈ I ¹, EI then _i=True, otherwise, EI _i=False, incident scope logic ETI _It=EI _iET _t, wherein t is the scheduling time numbering, t time period among the expression scheduling time set T, T ¹The set of presentation of events influence time scope, i is the process units numbering, i device among the indication device set I, I ¹Presentation of events influences the set of device scope; Adjust affair logic EC by incident behavior description generation device technology _ItAnd material/energy balance is adjusted affair logic ES _JtIf at scheduling slot t, i carries out technology adjustment, then EC to device _It=False, i ∈ I ^c, t ∈ T ¹, otherwise, EC _It=True, i ∈ I, t ∈ T, wherein I ^cThe device sets that expression needs technology to adjust; If within the events affecting scope, material or energy j are carried out technology adjust then ES _Jt=False, j ∈ J ^c, t ∈ T ¹, simultaneously the balance outside the events affecting scope is adjusted affair logic and is set to False, be i.e. ES _Jt=False, j ∈ J, $t\∉T^1,$ If material or energy j are not carried out any technology adjustment, then ES _Jt=True, j ∈ J, t ∈ T, wherein J ^cMaterial or energy set that expression needs technology to adjust.

Step 5) also comprises substep: 5-1) express affair logic according to the flow scheduling demand; 5-2) express affair logic according to event description.

Step 5-1) also comprise:

(a) the affair logic expression formula (1) of coupling apparatus production scheduling model:

Wherein, Y _ItBe device i, if the device start and stop logic of time period t is Y _It=True then indication device opens operation, if Y _It=False then indication device is out of service; Ec _ItFor with affair logic EC _ItCorresponding Boolean variable; g _i(x _It, β _IjThe production and processing model of)≤0 indication device i; x _ItBe time period t, the processing capacity of device i, β _IjBe the output capacity of device i material or energy j, α _IjConsumption rate for device i material or energy j; Disjunction expression

Be device production scheduling model.

(b) in conjunction with the affair logic expression formula (2) of material and energy balance: es _JtP _j(x _It, x _It, β _Ij, α _Ij, I _Jt, B _Jt, S _Jt)=0, wherein, es _JtFor with affair logic ES _JtCorresponding Boolean variable; x _ItBe time period t, events affecting device scope is produced the processing capacity of device i outward; I _JtBe time period t, the memory space of material or energy j; B _JtBe time t, the amount of buying of material or energy j; S _JtBe time period t, the sales volume of material or energy j; Equation P _Jt(x _It, x _It, β _Ij, α _Ij, I _Jt, B _Jt, S _Jt)=0 is the balance constraint of time period t material or energy j.

(c) in conjunction with the affair logic expression formula (3) of material and energy demand: R _t(et _t, B _Jt, S _Jt, I _Jt)≤R _t(et _t, B _Jt, S _Jt, I _Jt)≤R _t(et _t, B _Jt, S _Jt, I _Jt), wherein, expression formula R _t(B _Jt, S _Jt, I _Jt) for the buying of time period t material or the energy, sell the Demand Constraint with the stock; Expression formula R _t() and R _t() be demand lower limit function and demand upper limit function respectively, wherein B _Jt, S _Jt, I _JtAnd B _Jt, S _Jt, I _JtBe respectively demand lower limit and higher limit; Et _tFor with affair logic ET _tCorresponding Boolean variable.

(d) the affair logic expression formula (4) that requires in conjunction with flow process: Ω (ETI _It, Y _It)=True, wherein, expression formula Ω (Y _It)=True is flow process requirement constraint, and these technological requirements comprise device start and stop requirement, processing sequence requirement, requirement process time, device collocation requirement etc.

(e) in conjunction with the affair logic expression formula (5) of flow scheduling target: maxZ=OBJ (u _JtS _Jt, τ _JtB _Jt, h _jI _Jt, eti _Itλ _iY _It), i ∈ I, j ∈ J, t ∈ T, wherein, u _JtUnit selling price for time t product or energy j; τ _JtUnit purchasing price for time t raw material or energy j; h _jUnit holding cost for chemical products j; λ _iFixedly processing charges for device i; y _ItFor with device start and stop logic corresponding Boolean variable; OBJ (u _JtS _Jt, τ _JtB _Jt, h _jI _Jt, λ _iY _It) be the flow scheduling objective function.

Step 5-2) also comprise:

(f) according to the affair logic expression formula (6) of event description: f (ET _t, EI _i, ETI _It, EC _It, ES _It)=True, wherein, f ()=Ture is the logical assignment function, the formula that embodies is as described in the step 4).

(g) incident behavior logical expression:

(g-1) the device start and stop are provided with logical expression (7): Y _It=True or False, $\∀i\∈I^1,$ t∈T ¹；

(g-2) device technique is adjusted affair logic expression formula (8):

Wherein, g ' _i(x _It, β _Ij, α _Ij)≤0 is the adjusted production and processing model of device i technology.

(g-3) material/energy balance is adjusted affair logic expression formula (9): $\left[\begin{array}{c}\⫬{\mathrm{ES}}_{\mathrm{jt}}\\ P_{\mathrm{jt}}^c(x_{\mathrm{it}},{\stackrel{\‾}{x}}_{\mathrm{it}},{\mathrm{\β}}_{\mathrm{ij}},{\mathrm{\α}}_{\mathrm{ij}},I_{\mathrm{jt}},B_{\mathrm{jt}},S_{\mathrm{jt}})=0\end{array}\right],$ Wherein, $P_{\mathrm{jt}}^c(x_{\mathrm{it}}{,\stackrel{\‾}{x}}_{\mathrm{it}},{\mathrm{\β}}_{\mathrm{ij}},{\mathrm{\α}}_{\mathrm{ij}},I_{\mathrm{jt}},B_{\mathrm{jt}},S_{\mathrm{jt}})=0$ For adjusting the balance constraint of back time period t material or energy j.

(h) requirements, event logical expression (10): O (I _Jt, x _It, β _Ij, α _Ij)≤0, i ∈ I ¹, t ∈ T ¹, wherein, expression formula O () is to installing processing capacity and stock's Demand Constraint in the events affecting scope.

Step 6) is based on the broad sense planning of extracting, and expresses according to the flow scheduling affair logic and sets up the flow scheduling model, is made up of affair logic expression formula (1)-(5):

$\max Z=\mathrm{OBJ}(u_{\mathrm{jt}}\·S_{\mathrm{jt}},{\mathrm{\τ}}_{\mathrm{jt}}\·B_{\mathrm{jt}},h_j\·I_{\mathrm{jt}},{\mathrm{eti}}_{\mathrm{it}}\·{\mathrm{\λ}}_i\·y_{\mathrm{it}})$

es _jt·P _jt(x _it，x _it，β _ij，α _ij，I _jt，B _jt，S _jt)＝0， $\∀j\∈J,$ t∈T

R _t(et _t， B _jt， S _jt， I _jt)≤R _t(et _t，B _jt，S _jt，I _jt)≤R _t(et _t，B _jt，S _jt，I _jt)， $\∀j\∈J,$ t∈T

Ω(ETI _it，Y _it)＝True， $\∀i\∈I,$ t∈T

Set up event model according to the expression of event description modelling affair logic, form by affair logic expression formula (6)-(10):

$\∀s\∈S$

Step 7) combined process flow scheduling model and event model constitute the dynamic dispatching model based on affair logic.

The present invention also provides a kind of model generating method of the oil refinery real time intelligent dynamic dispatching response accident based on affair logic, comprises step: judge whether that 1) event occurs; 2) trigger the events corresponding model, and it is articulated mutually with flow scheduling; 3) by affair logic the dynamic dispatching model is simplified, reformed, finally responded the optimizing scheduling model of this incident.

Step 1) based on the expansion the Event triggered logical expression:

$E_s\left(e{g}_{\mathrm{ijst}}\right)=\left[\begin{array}{c}\mathrm{True},{\mathrm{eg}}_{\mathrm{ijst}}>{\stackrel{\&OverBar;}{\mathrm{eg}}}_{\mathrm{ij}}\mathrm{or\; e}{g}_{\mathrm{ijst}}<{\underset{\&OverBar;}{\mathrm{eg}}}_{\mathrm{ij}}\\ \mathrm{False},{\underset{\&OverBar;}{\mathrm{eg}}}_{\mathrm{ij}}\&le;{\mathrm{eg}}_{\mathrm{ijst}}{\&le;\stackrel{\&OverBar;}{\mathrm{eg}}}_{\mathrm{ij}}\end{array}\right],$ Eg wherein _IjstBe the real-time production control index (as the quality index of the material or the energy, the side line outlet temperature of process units, pressure, the output of process units, material or the energy, consumption etc.) of time t device i or material, energy j incident s, eg _IjWith Eg _IjBe respectively the bound of production control index.If the production control index is greater than the index upper limit or less than the index lower limit in real time, then incident s takes place, E _s=True, the time t of this moment, device i, material or energy j are the time of origin of incident s and object take place; If the production control index is in indication range in real time, then incident s does not take place, E _s=False.

Step 2) triggers the event model of corresponding event based on the value of step 1) Event triggered logic, and articulate, simultaneously with the time of origin of incident initial time as dynamic dispatching with the flow scheduling model.

The logical value of the affair logic in the event model that the step 3) utilization triggers is chosen or is discharged based on the respective logic disjunction expression in the dynamic dispatching model of affair logic, and with the corresponding Boolean variable of affair logic also can be to the model parameter of dynamic dispatching model, the variable peace treaty does up the effect that disappears unit and simplify, the disjunction expression of simplifying the back identity logic is integrated, the only optimizing scheduling model in the events affecting scope of this accident meets with a response, thereby simplified the scale of dynamic dispatching model greatly, in the real-time that improves dynamic dispatching, guaranteed certain optimization again.

In sum, the present invention is directed to the problem that present petrochemical industry dynamic dispatching exists, the logical relation that makes full use of the oil refining production technology characteristic and in production scheduling, embodied, consider basic inducement---the incident of dynamic dispatching, proposed the notion of affair logic, and designed a kind of modeling method of real time intelligent dynamically optimized scheduling based on this.The present invention has considered contingent incident in the oil refining production, production technology characteristic and production scheduling logical relation have been made full use of, handle dynamic dispatching based on affair logic, in based on strict mathematical model optimizing, demonstrated fully artificial intelligence again, when taking place, incident utilize the present invention to respond rapidly, scheduling scheme by solving model can be optimized in time adjusts production.The present invention is simple and practical, when guaranteeing the scheduling real-time, obtain optimizing scheduling preferably, reduce the waste of the resource and the energy to greatest extent, reduce producing cost, improve productivity effect, satisfy oil refining and produce under the complex environment for the Real-time and Dynamic scheduling requirement.

Description of drawings

Fig. 1 is based on the FB(flow block) of the real time intelligent dynamically optimized scheduling modeling method of affair logic;

Fig. 2 is certain refinery production procedure artwork;

Fig. 3 is based on the dynamic dispatching model of affair logic, the FB(flow block) of the optimizing scheduling model generating method of response accident.

Embodiment

Below in conjunction with accompanying drawing and embodiment invention is described further.

Among Fig. 1, modeling method of the present invention is:

1) industrial analysis;

2) event definition;

3) according to event definition, the description incident;

4), produce the events corresponding logic according to event description;

5) respectively the affair logic that produces is expressed according to flow scheduling demand and event description;

6) according to the expression of step 5), set up flow scheduling model and event model respectively to affair logic;

7) the final dynamic dispatching model that constitutes based on affair logic of combined process flow scheduling model and event model.

Now describe the processing procedure of this dynamic dispatching example in detail according to Fig. 2.

Find that by industrial analysis this refinery has two cover atmospheric and vacuum distillation units, the material of output identical type, the continuous reformer of one cover, one cover catalytic cracking unit, one cover delayed coking unit, an one cover gasoline hydrogenation device and a cover diesel hydrotreating unit, this flow process has 7 covering devices, set numbering 1-7 is set represents said apparatus, be I={1...7}, relate to the 21 kinds of materials or the energy in the production altogether, comprise the material of reforming, diesel oil, gasoline, fuel gas,, steam etc. are provided with set numbering 1-21 and represent said apparatus, be J={1...21}, adopt rolling scheduling on the five, promptly scheduling time section t has 5 sections, constitutes scheduling time set T={1...5}.

Obtain the production data α of manufacturing technique requirent and device, material and the energy by industrial analysis _Ij, β _Ij, B _Jt, S _Jt, I _Jt, B _Jt, S _Jt, I _Jt, u _Jt, τ _Jt, h _j, λ _jDeng, each meaning of parameters as previously mentioned, no longer explanation herein.

Further find often to have in the oil refining process following incident to take place by industrial analysis:

Incident 1: the 2 reformation material that often reduce pressure are off quality.

As follows according to event definition to event description:

Incident 1: state-event the unknown, attribute (time of origin the unknown, it is material 4 that object takes place, influence time scope 2 days, influencing the device scope is device 1,2,3), behavior (the reformation material of device 2 is cut circulation, and device 3 reduces turnout), demand does not have.

Produce affair logic E according to event description ₁, EI _i, ET _t, ETI _It, EC _It, ES _Jt $\&ForAll;i\&Element;I,$ J ∈ J, t ∈ T and corresponding Boolean variable et thereof _t, ei _i, eti _It, ec _It, es _JtProductive set I ¹=1,2,3}, T ¹=1,2}, I ^c=2,3}, J ^c={ 4}.

Express the dynamic dispatching model of setting up following (the affair logic expression provides as previously mentioned no longer step by step) herein by affair logic based on affair logic:

Wherein incident 1 model is as follows:

Wherein the flow scheduling model is as follows:

$\max Z=\mathrm{OBJ}(u_{\mathrm{jt}}\&CenterDot;S_{\mathrm{jt}},{\mathrm{\&tau;}}_{\mathrm{jt}}\&CenterDot;B_{\mathrm{jt}},h_j\&CenterDot;I_{\mathrm{jt}},{\mathrm{eti}}_{\mathrm{it}}\&CenterDot;{\mathrm{\&lambda;}}_i\&CenterDot;y_{\mathrm{it}})$

es _jt·P _jt(x _it，x _it，β _ij，α _ij，I _jt，B _jt，S _jt)＝0， $\&ForAll;j\&Element;J,$ t∈T

R _t(et _t， B _jt， S _jt， I _jt)≤R _t(et _t，B _jt，S _jt，I _jt)≤R _t(et _t，B _jt，S _jt，I _jt)， $\&ForAll;j\&Element;J,$ t∈T

Ω(ETI _it，Y _it)＝True， $\&ForAll;i\&Element;I,$ t∈T

The above-mentioned dynamic dispatching model of setting up based on the real time intelligent dynamically optimized scheduling modeling method of affair logic, comprised the accident in the production run, model can respond rapidly when incident takes place, and by finding the solution to the dynamic dispatching model, can in time adjust, when guaranteeing the scheduling real-time, obtain optimizing scheduling preferably, reduce the waste of the resource and the energy to greatest extent production, reduce producing cost, improve productivity effect.

The modeling method of above-mentioned real time intelligent dynamically optimized scheduling based on affair logic also can be applied to flow industry enterprises such as other continuous process industries or discrete, batch processing.Needn't make amendment to the method body when being applied to other industry, only need incident to be redefined according to different flow process, obtain being more suitable for the event description and the affair logic of the sector, the model structure of this method proposition simultaneously need not change yet, the production models and the parameter that only need to introduce the sector device get final product, so this method has compatibility and dirigibility preferably.

Among Fig. 3, provide the generation method of the optimizing scheduling model of response accident, comprised step: judge whether that 1) event occurs; 2) trigger the events corresponding model, and it is articulated mutually with flow scheduling; 3) by affair logic the dynamic dispatching model is simplified, reformed, finally responded the optimizing scheduling model of this incident.

The generation method of the optimizing scheduling model of response accident is described according to the embodiment of the invention now.

If 3 decompression routines 2 are reformed the real-time quality index of material less than its index lower limit, i.e. eg ₂₄₁₃＜ Eg ₂₄, then E takes place in incident 1 ₁=True, and carry out reschedule as the initial time of dynamic dispatching with Time To Event.

E ₁=True, trigger event 1 model, and articulate mutually with the flow scheduling model, obtain as drag:

maxZ＝OBJ(u _jt·S _jt，τ _jt·B _jt，h _j·I _jt，eti _it·λ _i·y _it)

ET _t＝False，

t＝3，4，5

ET _t＝True， t＝1，2

EI _i＝False，

i＝4，5，6，7

EI _i＝True，

i＝1，2，3

ETI _it＝False，

i＝4，5，6，7?or?t＝3，4，5

ETI _it＝True，

i＝1，2，3，t＝1，2

EC _it＝False，

i＝2，3，t＝1，2

ES _jt＝True，

$j\&NotElement;J^c,$ t＝1，2

ES _jt＝False，

j＝4，t＝1，2?or?j∈J，t＝3，4，5

i＝1，2，3，t＝1，2

$\left[\begin{array}{c}\&Not;{\mathrm{ES}}_{\mathrm{jt}}\\ P_{\mathrm{jt}}^c(x_{\mathrm{it}},{\stackrel{\&OverBar;}{x}}_{\mathrm{it}},{\mathrm{\&beta;}}_{\mathrm{ij}},{\mathrm{\&alpha;}}_{\mathrm{ij}},I_{\mathrm{jt}},B_{\mathrm{jt}},S_{\mathrm{jt}})=0\end{array}\right],$

j＝4，t＝1，2

es _jt·P _jt(x _it，x _it，β _ij，α _ij，I _jt，B _jt，S _jt)＝0， $\&ForAll;j\&Element;J,$ t∈T

R _t(et _t， B _jt， S _jt， I _jt)≤R _t(et _t，B _jt，S _jt，I _jt)≤R _t(et _t，B _jt，S _jt，I _jt)， $\&ForAll;j\&Element;J,$ t∈T

Ω(ETI _it，Y _it)＝True， $\&ForAll;i\&Element;I,$ t∈T

The application affairs logic is simplified, is reformed above-mentioned model, and explains that by following two examples affair logic is for whole simplified models process.

Example one: ETI ₁₁=True then can trigger

I=1, t=1 makes

Set up; ETI ₄₃=False does not then trigger

I=4, t=3, and with its cancellation from model.

Example two: EC ₃₁=False then can trigger

I=3, t=1 makes

Set up, again because ec ₃₁=0, cancellation simultaneously

I=3, the g among the t=1 _i(x _It, β _Ij, α _Ij)≤0, ETI again ₃₁=True reforms the disjunction expression of finally being set up to installing 3 device production scheduling disjunction expression

Other affair logics are the same for the simplified models process, no longer narration.

The final dynamic dispatching model that forms is as follows:

maxZ＝OBJ(u _jt·S _jt，τ _jt·B _jt，h _j·I _jt，eti _it·λ _i·y _it)

j∈J，i∈I，t＝1，2

i＝1，t＝1，2

i＝2，3，t＝1，2

P _j(x _it，x _it，β _ij，α _ij，I _jt，B _jt，S _jt)＝0，

$j\&NotElement;J^c,$ t＝1，2

$P_{\mathrm{jt}}^c(x_{\mathrm{it}}{\stackrel{\&OverBar;}{x}}_{\mathrm{it}},{\mathrm{\&beta;}}_{\mathrm{ij}},{\mathrm{\&alpha;}}_{\mathrm{ij}},I_{\mathrm{jt}},B_{\mathrm{jt}},S_{\mathrm{jt}})=0,$

j＝4，t＝1，2

R _t(et _t， B _jt， S _jt， I _jt)≤R _t(et _t，B _jt，S _jt，I _jt)≤R _t(et _t，B _jt，S _jt，I _jt)，

t＝1，2

Ω(Y _it)＝True， i＝1，2，3，t＝1，2

As seen, the dynamic dispatching model does not comprise seven covering devices, five time periods before simplifying, model is larger, use above-mentioned short-cut method simplification back model and only comprise three devices and two time periods, the model scale has been dwindled greatly, and be the simple broad sense planning form of extracting, then use existing optimisation technique and can find the solution rapidly fully.Therefore, the method that the present invention proposes can guarantee certain optimization can reduce the waste of the Enterprise Resource and the energy to greatest extent when satisfying the dynamic dispatching real-time, reduces producing cost, improves productivity effect.

The method that the present invention proposes has bigger dirigibility, can realize dynamic dispatching and static scheduling easily simultaneously.Because when incident 1 does not take place, E ₁=False, the trigger event model $\left[\begin{array}{c}\&Not;E_1\\ {\mathrm{ET}}_t=\mathrm{True}\\ {\mathrm{EI}}_i=\mathrm{True}\\ {\mathrm{ETI}}_{\mathrm{it}}=\mathrm{True},\&ForAll;i,j,t\\ {\mathrm{EC}}_{\mathrm{it}}=\mathrm{True}\\ {\mathrm{ES}}_{\mathrm{jt}}=\mathrm{True}\end{array}\right],$ And it is articulated mutually with the flow scheduling model, use above-mentioned model simplification process, obtain following 5 days static scheduling models:

$\max Z=\mathrm{OBJ}(u_{\mathrm{jt}}\&CenterDot;S_{\mathrm{jt}},{\mathrm{\&tau;}}_{\mathrm{jt}}\&CenterDot;B_{\mathrm{jt}},h_j\&CenterDot;I_{\mathrm{jt}},{\mathrm{\&lambda;}}_i\&CenterDot;y_{\mathrm{it}})$

P _jt(x _it，x _it，β _ij，α _ij，I _jt，B _jt，S _jt)＝0， $\&ForAll;j\&Element;J,$ t∈T

R _t( B _jt， S _jt， I _jt)≤R _t(B _jt，S _jt，I _jt)≤R _t(B _jt，S _jt，I _jt)， $\&ForAll;j\&Element;J,$ t∈T

Ω(Y _it)＝True， $\&ForAll;i\&Element;I,$ t∈T

In addition, generation method according to the optimizing scheduling model of the modeling method of the real time intelligent dynamically optimized scheduling based on affair logic of the present invention and response accident can be write as the dynamic dispatching Optimization Software, and can use corresponding production field, improve the production efficiency and the benefit in this field.

As mentioned above, according to the present invention, formerly system is made and technical field of automation, can effectively handle the oil refining process dynamic scheduling problem, when satisfying the dynamic dispatching real-time, can guarantee certain optimization again, reduce the waste of the Enterprise Resource and the energy to greatest extent, reduce producing cost, improve productivity effect.

Claims (9)

1, a kind of modeling method of the oil refinery real time intelligent dynamically optimized scheduling based on affair logic is characterized in that: its method is,

1) industrial analysis;

2) event definition;

3) according to event definition, the description incident;

4), produce the events corresponding logic according to event description;

5) respectively the affair logic that produces is expressed according to flow scheduling demand and event description;

6) according to the expression of step 5), set up flow scheduling model and event model respectively to affair logic;

7) the final dynamic dispatching model that constitutes based on affair logic of combined process flow scheduling model and event model.

2, the modeling method of the oil refinery real time intelligent dynamically optimized scheduling based on affair logic as claimed in claim 1 is characterized in that: also comprise following substep in the described step 1):

1-1) analyze the various a large amount of production incident that causes dynamic dispatching in the oil refining production, obtain the general characteristic of these incidents, as the basis of definition incident;

1-2) analyze the oil refining production procedure, the technology that obtains process units connects, and the operational factor of process units and the production information of material are as setting up the flow scheduling model based.

3, the modeling method of the oil refinery real time intelligent dynamically optimized scheduling based on affair logic as claimed in claim 1, it is characterized in that: described step 2), event definition is the unusual condition of process units, material or the energy, show as the accident in the production, these incidents can cause the running status to process units, the quantum of output of the material or the energy, consumption and tank farm stock etc. carry out the adjustment of the part or the overall situation, to guarantee the smooth production of safety and steady.

4, the modeling method of the oil refinery real time intelligent dynamically optimized scheduling based on affair logic as claimed in claim 3, it is characterized in that: described incident also is defined as polynary group of form, promptly by state, attribute, behavior becomes the definition content of incident with requirement groups; Wherein, whether the state representation incident takes place; Attribute comprises time of origin, object takes place, influence time scope and influence the device scope, when presentation of events takes place respectively, occur on which device, which kind of material or the energy, occur to the elimination incident for the needed time of influence of producing and the process units of the required adjustment running status of response events from incident; The technology adjustment measure that the dispatcher takes for processing events is represented in behavior; Demand schedule is shown in the requirement of interior purchase volume to some material of events affecting time, sales volume, tank farm stock.

5, the modeling method of the oil refinery real time intelligent dynamically optimized scheduling based on affair logic as claimed in claim 4 is characterized in that: described state-event, and describe incident with "Yes" and take place, describe incident with "No" and do not take place; For event attribute, with concrete scheduling time section Time To Event is described, with the set of concrete device, material or energy numbering description incident generation object, with scheduling time segment limit the events affecting time is described, number description and influence the device scope with influence device scheduling set; For the incident behavior, to describe with the production status of adjusted device or material, the numbering during device of wherein adjusting or material are gathered separately with it is described; For requirements, event, describe with the numbering and the demand thereof of material.

6, the modeling method of the oil refinery real time intelligent dynamically optimized scheduling based on affair logic as claimed in claim 5 is characterized in that: the described description by state-event produces Event triggered logic E _s, if incident s generation, then E _s=True, otherwise E _s=False, wherein s is a Case Number, s incident in the presentation of events S set can trigger its events corresponding model by the Event triggered logic; By event attribute generation event time logic ET is described _t, event device logic EI _iWith incident scope logic ETI _It, if scheduling time section t is in the events affecting time range, i.e. t ∈ T ¹ET then _t=True, otherwise, ET _t=False, if process units i is in events affecting device scope, i.e. i ∈ I ¹, EI then _i=True, otherwise, EI _i=False, incident scope logic ETI _It=EI _iET _t, wherein t is the scheduling time numbering, t time period among the expression scheduling time set T, T ¹The set of presentation of events influence time scope, i is the process units numbering, i device among the indication device set I, I ¹Presentation of events influences the set of device scope; Adjust affair logic EC by incident behavior description generation device technology _ItAnd material/energy balance is adjusted affair logic ES _JtIf at scheduling slot t, i carries out technology adjustment, then EC to device _It=False, i ∈ I ^c, t ∈ T ¹, otherwise, EC _It=True, i ∈ I, t ∈ T, wherein I ^cThe device sets that expression needs technology to adjust; If within the events affecting scope, material or energy j are carried out technology adjust then ES _Jt=False, j ∈ J ^c, t ∈ T ¹, simultaneously the balance outside the events affecting scope is adjusted affair logic and is set to False, be i.e. ES _Jt=False, j ∈ J, $t\&NotElement;T^1,$ If material or energy j are not carried out any technology adjustment, then ES _Jt=True, j ∈ J, t ∈ T, wherein J ^cMaterial or energy set that expression needs technology to adjust.

7, the modeling method of the oil refinery real time intelligent dynamically optimized scheduling based on affair logic as claimed in claim 1 is characterized in that: in the described step 5) affair logic is expressed for the logic-based disjunction expression, the steps include:

5-1) express affair logic according to the flow scheduling demand;

5-2) express affair logic according to event description;

Step 5-1 wherein) concrete grammar is:

(a) the affair logic expression formula (1) of coupling apparatus production scheduling model:

Wherein, Y _ItBe device i, if the device start and stop logic of time period t is Y _It=True then indication device opens operation, if Y _It=False then indication device is out of service; Ec _ItFor with affair logic EC _ItCorresponding Boolean variable; g _i(x _It, β _IjThe production and processing model of)≤0 indication device i; x _ItBe time period t, the processing capacity of device i, β _IjBe the output capacity of device i material or energy j, α _IjConsumption rate for device i material or energy j; Disjunction expression Be device production scheduling model;

(b) in conjunction with the affair logic expression formula (2) of material and energy balance: es _JtP _j(x _It, x _It, β _Ij, α _Ij, I _Jt, B _Jt, S _Jt)=0, wherein, es _JtFor with affair logic ES _JtCorresponding Boolean variable; x _ItBe time period t, events affecting device scope is produced the processing capacity of device i outward; I _JtBe time period t, the memory space of material or energy j; B _JtBe time t, the amount of buying of material or energy j; S _JtBe time period t, the sales volume of material or energy j; Equation P _Jt(x _It, x _It, β _Ij, α _Ij, I _Jt, B _Jt, S _Jt)=0 is the balance constraint of time period t material or energy j;

(c) in conjunction with the affair logic expression formula (3) of material and energy demand: R _t(et _t, B _Jt, S _Jt, I _Jt)≤R _t(et _t, B _Jt, S _Jt, I _Jt)≤R _t(et _t, B _Jt, S _Jt, I _Jt), wherein, expression formula R _t(B _Jt, S _Jt, I _Jt) for the buying of time period t material or the energy, sell the Demand Constraint with the stock; Expression formula R _t() and R _t() be demand lower limit function and demand upper limit function respectively, wherein B _Jt, S _Jt, I _JtAnd B _Jt, S _Jt, I _JtBe respectively demand lower limit and higher limit; Et _tFor with affair logic ET _tCorresponding Boolean variable;

(d) the affair logic expression formula (4) that requires in conjunction with flow process: Ω (ETI _It, Y _It)=True, wherein, expression formula Ω (Y _It)=True is flow process requirement constraint, and these technological requirements comprise device start-stop requirement, processing sequence requirement, requirement process time, device collocation requirement etc.;

(e) in conjunction with the affair logic expression formula (5) of flow scheduling target: maxZ=OBJ (u _JtS _Jt, τ _JtB _Jt, h _jI _Jt, eti _Itλ _iY _It), i ∈ I, j ∈ J, t ∈ T, wherein, u _JtUnit selling price for time t product or energy j; τ _JtUnit purchasing price for time t raw material or energy j; h _jUnit holding cost for chemical products j; λ _iFixedly processing charges for device i; y _ItFor with device start and stop logic corresponding Boolean variable; OBJ (u _JtS _Jt, τ _JtB _Jt, h _jI _Jt, λ _iY _It) be the flow scheduling objective function;

Step 5-2) also comprises

(f) according to the affair logic expression formula (6) of event description: f (ET _t, EI _i, ETI _It, EC _It, ES _It)=True, wherein, f ()=Ture is the logical assignment function;

(g) incident behavior logical expression:

(g-1) the device start and stop are provided with logical expression (7): Y _It=True or False, $\&ForAll;i\&Element;I^1,$ t∈T ¹；

(g-2) device technique is adjusted affair logic expression formula (8):

Wherein, g ' _i(x _It, β _Ij, α _Ij)≤0 is the adjusted production and processing model of device i technology;

(g-3) material/energy balance is adjusted affair logic expression formula (9): $\left[\begin{array}{c}\&Not;{\mathrm{ES}}_{\mathrm{jt}}\\ P_{\mathrm{jt}}^c(x_{\mathrm{it}},{\stackrel{\&OverBar;}{x}}_{\mathrm{it}},{\mathrm{\&beta;}}_{\mathrm{ij}},{\mathrm{\&alpha;}}_{\mathrm{ij}},I_{\mathrm{jt}},B_{\mathrm{jt}},S_{\mathrm{jt}})=0\end{array}\right],$ Wherein, $P_{\mathrm{jt}}^c(x_{\mathrm{it}}{,\stackrel{\&OverBar;}{x}}_{\mathrm{it}},{\mathrm{\&beta;}}_{\mathrm{ij}},{\mathrm{\&alpha;}}_{\mathrm{ij}},I_{\mathrm{jt}},B_{\mathrm{jt}},S_{\mathrm{jt}})=0$ For adjusting the balance constraint of back time period t material or energy j;

(h) requirements, event logical expression (10): O (I _Jt, x _It, β _Ij, α _Ij)≤0, i ∈ I ¹, t ∈ T ¹, wherein, expression formula O () is to installing processing capacity and stock's Demand Constraint in the events affecting scope.

8, the modeling method of the oil refinery real time intelligent dynamically optimized scheduling based on affair logic as claimed in claim 1, it is characterized in that: described step 6) is based on the broad sense planning of extracting, it is expressed according to the flow scheduling affair logic and sets up the flow scheduling model, is made up of affair logic expression formula (1)-(5):

$\max Z=\mathrm{OBJ}(u_{\mathrm{jt}}\&CenterDot;S_{\mathrm{jt}},{\mathrm{\&tau;}}_{\mathrm{jt}}\&CenterDot;B_{\mathrm{jt}},h_j\&CenterDot;I_{\mathrm{jt}},{\mathrm{eti}}_{\mathrm{it}}\&CenterDot;{\mathrm{\&lambda;}}_i\&CenterDot;y_{\mathrm{it}})$

es _jt·P _jt(x _it，x _it，β _ij，α _ij，I _jt，B _jt，S _jt)＝0， $\&ForAll;j\&Element;J,$ t∈T

R _t(et _t， B _jt， S _jt， I _jt)≤R _t(et _t，B _jt，S _jt，I _jt)≤R _t(et _t，B _jt，S _jt，I _jt)， $\&ForAll;j\&Element;J,$ t∈T

Ω(ETI _it，Y _it)＝True， $\&ForAll;i\&Element;I,$ t∈T

Set up event model according to the expression of event description modelling affair logic, form by affair logic expression formula (6)-(10):

$\&ForAll;s\&Element;S.$

9, the modeling method of the oil refinery real time intelligent dynamically optimized scheduling based on affair logic as claimed in claim 1 is characterized in that: described step 7) combined process flow scheduling model and event model constitute the oil refinery real time intelligent dynamically optimized scheduling model based on affair logic.

Images