CN1465660A - Advanced control method for viscosity-reducing reaction of heavy oil - Google Patents

Abstract

The method for controlling viscosity-reducing reaction of heavy oil is characterized by making real-time on-line analysis of viscosity of oil in the heavy oil viscosity-reducing reactor, and implementing model prediction control of the heavy oil viscosity-reducing reaction, i.e. according to the product brand of finished oil at outlet of heavy oil viscosity-reducing reactor making closed-loop automatic regulation of operation variable of the heavy oil viscosity-reducing reactor so as to implement control of product oil viscosity at outlet of heavy oil viscosity-reducing reactor, and can raise its yield.

Description

A kind of advanced control method of heavy oil visbreaking reaction

Technical field

The present invention relates to the heavy oil production technical field of petrochemical complex, is a kind of advanced control method of heavy oil visbreaking reaction concretely.

Background technology

Visbreaking unit is to adopt the vacuum residuum of various different sorts crude oil to react, mix with wax tailings and catalytic diesel oil to produce 180# and 250# oil fuel through mild thermal cracking.The main purpose of visbreaking unit is to guarantee that the visbreaking product satisfies specification of quality such as viscosity.

At present, by regulating visbreaking temperature of reaction and catalytic diesel oil mixed volume (the decompressed wax oil flow is non-adjustable) viscosity with the control product.Wherein, when producing 180# oil, need conditioned reaction temperature and catalytic diesel oil mixed volume simultaneously; And when producing 250# oil, only need the conditioned reaction temperature to get final product.

In the production process of the 180# of prior art and 250# oil fuel, the control method of used heavy oil visbreaking is: the quality index such as product viscosity that obtain according to assay, the operator by the conditioned reaction temperature and (or) the diesel oil incorporation, in order to control product viscosity.

The control method and the device of heavy oil visbreaking in fact in the past exist a lot of drawbacks: it is a kind of method and device of adjusted open loop for (one).Its control to viscosity rests on the operating experience, and this has just brought bigger personal errors inevitably, will inevitably cause controlling performance not high.So this control method and the device of realizing this method based on artificial experience and operation is not truly closed loop control method and device.(2) need produce switching when the product of producing different oil brands is oily, and the control method of heavy oil visbreaking of the prior art and device need more manual intervention when switching, and the quality product fluctuation is bigger, the switching adjusting time is longer.(3) passing through of adopting at present regulated the diesel oil flow and controls the method for product viscosity and cause product viscosity to be higher than general requirement, thereby consumed the higher relatively diesel oil of more value, do not reach purpose energy-saving and cost-reducing, coordination optimization.

Summary of the invention

The invention provides the advanced control method of a kind of heavy oil visbreaking reaction, make the control of heavy oil visbreaking become the automatic control of closed loop, get, improve product quality for artificial empirical operation; In the production handoff procedure of the product oil of producing different oil brands, reduce the fluctuation of quality product by the automatic control of closed loop, shorten and switch the adjusting time, when improving product quality, can also reach purpose energy-saving and cost-reducing, coordination optimization.

Technical scheme of the present invention is: a kind of advanced control method of heavy oil visbreaking reaction is characterized in that real-time online is analyzed the oil plant viscosity in the heavy oil visbreaking reactor; To heavy oil visbreaking reaction implementation model predictive control, that is: according to the oil brand of heavy oil visbreaking reactor outlet product oil, the operational variable of close-loop automatic adjustment heavy oil visbreaking reactor, realization is to the control of heavy oil visbreaking reactor outlet product oil viscosity, and automatic mutual switching of realizing at least two kinds of trade mark Oil Production.

The oil plant viscosity that described real-time online is analyzed in the heavy oil visbreaking reactor is meant: gather heavy oil visbreaking reactor inlet residual oil viscosity, outlet visbreaking residue viscosity, exported product oil viscosity in real time, and it is carried out the analysis of real-time online viscosity.

Described oil brand according to heavy oil visbreaking reactor outlet product oil, the operational variable of close-loop automatic adjustment heavy oil visbreaking reactor, realization is meant the control of heavy oil visbreaking reactor outlet product oil viscosity: heavy oil visbreaking reactor outlet product oil can be No. 180 oil fuel or No. 250 oil fuel

When product oil is No. 180 oil fuel: heavy oil visbreaking reactor outlet product oil viscosity is adopted regional control;

When product oil was No. 250 oil fuel: employing singly went into singly (SISO) Model Predictive Control to heavy oil visbreaking reactor outlet product oil viscosity;

Switching mutually automatically of at least two kinds of trade mark Oil Production of described realization is meant: switch to No. 250 oil fuel of production from producing No. 180 oil fuel, or switch to No. 180 oil fuel of production from producing No. 250 oil fuel.

The advanced control method of described a kind of heavy oil visbreaking reaction, its step comprises:

Gather heavy oil visbreaking reactor inlet residual oil viscosity, outlet visbreaking residue viscosity, exported product oil viscosity in real time, and it is carried out the analysis of real-time online viscosity;

Heavy oil visbreaking reactor outlet product oil can be No. 180 oil fuel or No. 250 oil fuel,

When product oil is No. 180 oil fuel: heavy oil visbreaking reactor outlet product oil viscosity is adopted regional control;

When product oil was No. 250 oil fuel: employing singly went into singly (SISO) Model Predictive Control to heavy oil visbreaking reactor outlet product oil viscosity;

Can realize automatically switching to No. 250 oil fuel of production, or automatically switch to No. 180 oil fuel of production from producing No. 250 oil fuel from producing No. 180 oil fuel.

Described when product oil is No. 180 oil fuel: as to adopt regional control to be meant to heavy oil visbreaking reactor outlet product oil viscosity:

Set the meadow, when visbreaking reactor outlet viscosity measurement value is outside the meadow, that is: outlet viscosity utilizes corresponding ratio control to regulate diesel oil mediation amount (F2) greater than the set(ting)value upper limit or less than the set(ting)value lower limit; When visbreaking reactor outlet viscosity measurement value is within the meadow, adopt Model Predictive Control.

Described when product oil is No. 250 oil fuel: as heavy oil visbreaking reactor outlet product oil viscosity to be adopted singly to go into singly (SISO) Model Predictive Control and be meant: when No. 250 oil fuel are carried out visbreaking control, only need carry out close-loop automatic adjustment to visbreaking reactor inlet temperature (TI).

The described realization automatically switches to and produces No. 250 oil fuel and be meant from producing No. 180 oil fuel: transfer to and produce No. 250 oil fuel and can adopt direct switching from producing No. 180 oil fuel.

Describedly automatically switch to and produce No. 180 oil fuel and be meant: transfer to and produce No. 180 oil fuel from producing No. 250 oil fuel from producing No. 250 oil fuel, can keep under the constant condition of TI, one step F2 increases, that is: increase the average of F2 when measuring No. 180 oily ordinary production, the system that makes promptly enters the meadow at short notice, treat that it enters after the meadow, cancellation TI remains unchanged, and implements control by the general control of No. 180 oil.

Described regional control is meant the product viscosity multivariable prediction control based on mechanism model; Its with product viscosity as main controlled variable, and for guaranteeing the material balance of separation column, liquid level at the bottom of the tower is kept existing conventional control, promptly utilize liquid level at the bottom of the bottom product discharge control tower, with temperature of reaction and diesel oil incorporation as main operational variable, by regulating furnace outlet temperature and diesel oil incorporation control product viscosity; When control, should change temperature of reaction, and the diesel oil incorporation is in less important manipulated variable as far as possible, and the least possible;

In the mechanism model of setting up, include in and can survey interference, as the influence of unit capacity, wax oil incorporation, material viscosity, material density, raw material carbon residue, diesel oil viscosity and wax oil viscosity, can utilize feed-forward and feedback control to reduce their disturbances to whole process control.

Described Model Predictive Control is meant: the multivariable prediction based on mechanism model is coordinated control, it can set up the dynamic mathematical models that can reflect production process according to the material balance and the energy balance of reactor and separation column, the multivariate monodrome predictive control of feedback of status is predicted and had to implementation model, controls according to the deviation and the overrun condition of the controlled variable of predicting.

Describedly transfer to and produce No. 250 oil fuel and can adopt the step of direct switching to be from producing No. 180 oil fuel:

In m control cycle, divide the m step to carry out the transition to No. 250 production,

Get last steadily produce No. 250 when oily reactor inlet temperature and the programmed values SP of diesel oil flow;

SP and this programmed values SP that switches preceding reactor inlet temperature and diesel oil flow is poor, difference is done the average of m, as the manipulated variable of switching controls;

Manipulated variable is added on the programmed values SP of current reactor inlet temperature and diesel oil flow, as SP output, control of output in every n minute promptly automaticallyed switch to go in the production decision of No. 250 oil after m*n minute.

Describedly automatically switch to the step of producing No. 180 oil fuel and comprise from producing No. 250 oil fuel:

In m control cycle, divide the m step to carry out the transition to No. 180 production,

Get last steadily produce No. 180 when oily reactor inlet temperature and the programmed values SP of diesel oil flow;

SP before SP and this switching is poor, difference is done the average of m, as the manipulated variable of switching controls;

Manipulated variable is added on the programmed values SP of current reactor inlet temperature and diesel oil flow, as SP output, control of output in every n minute promptly automaticallyed switch to go in the production decision of No. 180 oil after m*n minute.

Described product viscosity multivariable prediction control based on mechanism model need be satisfied following condition:

The output condition of No. 180 oil is: $\mathrm{MU}108\left(k\right)=C_{180}*T0(k-\mathrm{tao}3)+D_{180}\left[\begin{array}{c}F1(k-\mathrm{tao}4)\\ F2(k-\mathrm{tao}5)\\ F3(k-\mathrm{tao}6)\\ \mathrm{MU}1(k-\mathrm{tao}1)\end{array}\right]$

The modeling condition of No. 180 oil of controlled plant is:

The control condition of No. 180 oil of controlled plant is:

Described multivariable prediction based on mechanism model is coordinated control and need be satisfied following condition:

The output condition of No. 250 oil of controlled plant is: $\mathrm{MU}250\left(k\right)=C_{250}*T0(k-\mathrm{tao}3)+D_{250}*\left[\begin{array}{c}F1(k-\mathrm{tao}4)\\ \mathrm{MU}1(k-\mathrm{tao}1)\end{array}\right]$

The modeling condition of No. 180 oil of controlled plant is:

The control condition of No. 250 oil of controlled plant is:

Beneficial effect of the present invention is: realized the multivariable prediction control of reactor product viscosity, reduced the fluctuation of product viscosity, undulating quantity is no more than 5% of controlling valu, guarantees that product viscosity is steadily qualified.Guarantee that other controlled variable, operational variable do not exceed its given upper and lower limit, guarantee whole device safety and steady operation.Reduced the incorporation of blended diesel, the reduction amplitude is no less than 1%, has improved economic benefit.The two kinds of schemes of 180# and 250# that realized are convenient to be switched, and has shortened product switching time.

As seen, compared with prior art, reactor smooth operation of the present invention, it is big to have changed adjusted open loop quality product fluctuation in the past, and temperature in is regulated frequent, the situation that the amplitude of accommodation is big; And diesel oil mediation amount obviously reduces the quality product edge; Particularly product oil is produced to switch and has been realized automatic control, and the adjusting time obviously reduces.Thereby the closed loop that has realized the heavy oil visbreaking is controlled automatically, has improved productive rate, has optimized the running status of reactor, has received good result energy-saving and cost-reducing, coordination optimization, has created economic benefit.

Description of drawings

Fig. 1 is the synoptic diagram of the heavy oil visbreaking control device of prior art;

Fig. 2 is regional control curve and flow process;

Fig. 3 is a functional block diagram of the present invention.

Embodiment

With reference to Fig. 1, can be according to online viscosity analyser to import residual oil, reactor outlet visbreaking residue and the real-time analysis that goes out the visbreaking product viscosity of device, adopt model predictive control technique, visbreaking unit is implemented advanced control: the viscosity to the reactor outlet visbreaking residue realizes regional control, the visbreaking product viscosity is realized set point control, in closed-loop control visbreaking temperature of reaction and diesel oil infiltration capacity with when reaching controlled target, realize the switching mutually automatically of the safe ready of two kinds of control of 180# oil product and 250# oil product scheme, when replacing manually regulating measure, shorten switching time, and reduce the incorporation of catalytic diesel oil as much as possible, reach the purpose of increasing economic efficiency.

Control divides two portions: control of 180# oil product and the control of 250# oil product, and it specifically implements as follows:

(1) viscosity controller strategy:

As shown in table 1 to No. 180 oil fuel: table 1:

Controlled variable	Operational variable	Perturbing variables
			Device export oil viscosity	Visbreaking reactor inlet temperature diesel oil mediation flow	The inlet concentration of acetylene inlet ethylene concentration inlet density of hydrogen inlet raw material inlet amount inlet raw hydrogen inlet amount pure hydrogen inlet amount that enters the mouth

Comprehensive advanced control advantage and optimization strategy when guaranteeing steadily control, make that diesel oil mediation amount is as much as possible little, reach the purpose of control.

To adopt the method for regional control when viscosity controller, set certain control area, promptly as " meadow " of figure, and to the enforcement logic control (as shown in Figure 2) of diesel oil mediation amount:

1. when visbreaking reactor outlet viscosity measurement value is outside the meadow (outlet viscosity is greater than the set(ting)value upper limit or less than the set(ting)value lower limit) utilizes corresponding ratio control to regulate diesel oil mediation amount.Regulate formula as shown in Figure 2.

2. when visbreaking reactor outlet viscosity measurement value is within the meadow, adopt model predictive control technique.Wherein, SP is the MU180 set(ting)value, SP _HI, SP _LOHigh lower bound for set(ting)value

Δ F2 is the diesel oil manipulated variable, in the control, it as interference

K is a proportional control factor

Δ Fd is the F2 optimal control in the plateau region (in the diagram between the high lower bound)

α is the half-breadth of plateau region

2. to No. 250 oil fuel, be not in harmonious proportion, have only a manipulated variable TI, press SISO predictive control strategy and get final product because of it has.

3. the steady switching of viscosity controller production decision:

(1) transfers No. 250 oil fuel of production to from producing No. 180 oil fuel: directly switch.

(2) transfer No. 180 oil fuel of production to from producing No. 250 oil fuel:

Keeping under the constant condition of TI, one step F2 increases (F2's was average when No. 180 oily ordinary production were measured in increase), and the system that makes promptly enters the meadow at short notice, treats that it enters after the meadow, cancellation TI remains unchanged, by the general control scheme implementation control of No. 180 oil.

(2) based on the viscosity multivariable predictive controller of mechanism model

(1) based on the multivariable prediction tuning controller of mechanism model

In the visbreaking unit with closely-related reactor and the separation column of mainly containing of product viscosity.Material balance and energy balance according to reactor and separation column are set up the dynamic mathematical models that can reflect production process, adopt predictive controller UPCC strategy, the multivariate monodrome Prediction and Control Technology of feedback of status is predicted and had to implementation model, deviation and overrun condition according to the controlled variable of predicting are controlled, and have stronger robustness.

(2) product viscosity multivariable predictive controller

As main controlled variable, for guaranteeing the material balance of separation column, liquid level at the bottom of the tower is kept existing conventional control with product viscosity simultaneously, promptly utilize liquid level at the bottom of the bottom product discharge control tower.

With temperature of reaction and diesel oil incorporation as main operational variable, by regulating furnace outlet temperature and diesel oil incorporation control product viscosity.In order to improve the economic benefit of device, guaranteeing under the product quality premise, change temperature of reaction and the diesel oil incorporation is in less important manipulated variable at first as far as possible, and the least possible.

In the mechanism model of setting up, include in and can survey interference, as the influence of unit capacity, wax oil incorporation, material viscosity, material density, raw material carbon residue, diesel oil viscosity and wax oil viscosity, can utilize feed-forward and feedback control to reduce their disturbances to whole process control.

(3) the dynamic mechanism mathematical model of visbreaking

1. the derivation of process dynamic model:

The consideration of (1) system's hysteresis, discrete time and prediction time domain:

Pre-when determining retardation time, dynamic mathematical models discrete time and control prediction according to measured data and processing parameter.

Interior time lag: 20-60 minute of reactor

Time lag: 3-15 minute (as table 2) after reactor extremely is in harmonious proportion

Table 2

Data	MU1	?TI	?T0	?F1	?F2	?F3	Reactor outlet viscosity	The finished product viscosity
									Pure hysteresis (minute)	0-10	?20-60	?3-15	?20-60	?0-6	?0-6	3-15	0-2

The visbreaking reaction is a process that reaction is slower, and retardation time is bigger, needs the choose reasonable discrete time, discrete time is too small, will increase data processing amount and calculated amount, increases the DCS system loading, discrete time is excessive, will be not enough to reflect real systematic procedure state.Here, determine that discrete time is 3-15 minute, so as shown in table 3;

Table 3:

Data	MU1	?TI	?T0	?F1	?F2	?F3	Reactor outlet viscosity	The finished product viscosity
									Pure hysteresis (step-length)	tao1	?tao2	?tao3	?tao4	?tao5	?tao6	tao7	tao8

Prediction step is taken as the P=5-15 step.In the table:

MU1 is (100 ℃ of material viscosities

);

TI be reactor inlet temperature (℃);

T0 be the reactor outlet temperature (℃);

F1 is material inlet flow (t/h);

F2 is blended diesel flow (t/h);

F3 is for being in harmonious proportion wax oil flow (t/h);

Tao is technology retardation time (step/minute);

Reactor outlet product viscosity: for No. 250 oil fuel, be not in harmonious proportion, so item is consistent with the finished product viscosity because of having;

(100 ℃ of the finished product viscosity: MU250 (No. 250 oil fuel product viscosity), MU180 (product viscosity after No. 180 oil fuel is in harmonious proportion) ).

(2) derivation of output equation:

Because the visbreaking system feeding is formed changeable, visbreaking reaction height complexity, at present both at home and abroad chemical does not still have quantitative mechanism research to it, the accurate relational expression of visbreaking system product viscosity is unknown, but, we are considering reactor self conduction time lag, reactor to being in harmonious proportion under the situation of section time lag, and carrying out match according to the historical data that a series of and product viscosity of collection in worksite is relevant, to draw the proximate calculation expression formula of product viscosity as follows:

1. No. 250 oil fuel product viscosity $\mathrm{MU}250\left(k\right)=C_{250}*T0(k-\mathrm{tao}3)+D_{250}*\left[\begin{array}{c}F1(k-\mathrm{tao}4)\\ \mathrm{MU}1(k-\mathrm{tao}1)\end{array}\right]$

2. No. 180 oil fuel product viscosity: $\mathrm{MU}108\left(k\right)=C_{180}*T0(k-\mathrm{tao}3)+D_{180}\left[\begin{array}{c}F1(k-\mathrm{tao}4)\\ F2(k-\mathrm{tao}5)\\ F3(k-\mathrm{tao}6)\\ \mathrm{MU}1(k-\mathrm{tao}1)\end{array}\right]$ Wherein:

T0 reactor outlet temperature (℃);

F1 raw material inlet amount (t/h);

F2 blended diesel inlet amount (t/h);

F3 is in harmonious proportion wax oil inlet amount (t/h);

MU1 material viscosity (100 ℃, motion);

K discrete step (7 minutes per steps);

Tao is technology retardation time (step/minute);

C ₂₅₀, C ₁₈₀, D ₂₅₀, D ₁₈₀Be the corresponding coefficient battle array.Annotate: the kinematic viscosity when the formula medium viscosity all has been converted into 100 ℃

(3) derivation of state equation

Adopted based on the predictive control principle on the model basis in the product viscosity control, so at first will determine reactor model, predictive control is not very high to the requirement of model, and this is one of predictive control biggest advantage.

1. No. 250 oil fuel product viscosity control

(1) model is derived

To reactor R01,, have according to energy balance: $C_{p}W\frac{\mathrm{dT}}{\mathrm{dt}}=F_1\[C_p(\mathrm{TI}-T0)-\mathrm{\ΔH}\]-q$ Wherein, C _pMean specific heat for residual oil in the reactor;

W is the reactor storage;

F ₁Be the reaction feed amount;

TI, T0 are respectively reactor inlet and temperature of reaction (≈ temperature out);

Q is a thermosteresis, can ignore;

Δ H is reaction heat (≈ 0), so get=C _P* (MU1-MUQ), MU1, MUQ are respectively raw material residual oil and reactor outlet residual oil viscosity.Steady state relation: at a certain steady state point, Δ H=Cp* (MU1-MUQ), the linearizing following formula, and consider time lag, discretize is handled to be had $T0\left(k\right)=A_{250}*T0(k-\mathrm{tao}3)+B_{250}*\mathrm{TI}(k-\mathrm{tao}2)+V_{250}\left[\begin{array}{c}F1(k-\mathrm{tao}4)\\ \mathrm{MU}1(k-\mathrm{tao}1)\end{array}\right]$

2. No. 180 oil fuel product viscosity control

(1) model is derived

Equally, to reactor R01,, have according to energy balance: $C_{p}W\frac{\mathrm{dT}}{\mathrm{dt}}=F_1\[C_p(\mathrm{TI}-T0)-\mathrm{\ΔH}\]-q$

Linearizing following formula: wherein:

C _pMean specific heat for residual oil in the reactor;

W is the reactor storage;

F ₁Be the reaction feed amount;

TI, T0 are respectively reactor inlet and temperature of reaction (=temperature out);

Q is a thermosteresis, can ignore;

Δ H is reaction heat (≈ 0), so get=C _P* (MU1-MU180Q), MU1, MU180Q are respectively raw material residual oil and reactor outlet residual oil viscosity.

Δ H steady state relation: at a certain steady state point, Δ H=Cp* (MU1-MU180Q),

Consider time lag, and through discretize handle state equation: $T0(k+1)=A_{180}*T0\left(k\right)+B_{180}*\mathrm{TI}(k-\mathrm{tao}2+\mathrm{tao}3)+V_{180}*\left[\begin{array}{c}F1(k-\mathrm{tao}4+\mathrm{tao}3)\\ F2(k-\mathrm{tao}5+\mathrm{tao}3)\\ F3(k-\mathrm{tao}6+\mathrm{tao}3)\\ \mathrm{MU}1(k-\mathrm{tao}1+\mathrm{tao}3)\end{array}\right]$

(4) discrete back dynamic mathematical models:

Simultaneous output equation and state equation promptly draw the dynamic mathematical models of visbreaking reactor:

1. controlled plant 250# oil fuel mathematical model is expressed as:

2. controlled plant 180# oil fuel mathematical model is expressed as:

(4) viscosity multivariable prediction zone controller control law

Suppose that pure hysteresis is as follows: (referring to table 4)

Table 4

Data	MU1	?F1	?TI	?T0	?F2	?F3	Reactor outlet viscosity	The finished product viscosity
									Pure hysteresis (step-length)	0	?7	?7	?1	?0	?0	1	0

1. 250# oil

Wherein:

X(k)＝T0(k)????U(k)＝TI(k)?? $V\left(k\right)=\left[\begin{array}{c}F1\left(k\right)\\ \mathrm{MU}1\left(k\right)\end{array}\right]$

Y(k)＝MU250(k)

X: control model state variable, promptly the reactor reaction temperature is one 1 * 1 vector in this example;

U: control model manipulation variable, i.e. reactor inlet temperature, 1 * 1 vector;

V: control model perturbing variables, i.e. reactor feed flow, charging viscosity.2 * 1 vectors;

Y: control model controlled variable (model output variable), i.e. visbreaking system outlet product viscosity, 1 * 1 vector;

A, B, C, F1, F2: model coefficient battle array;

Feedback modifiers is carried out in prediction output, and adopts the monodrome predictive control algorithm, obtains the optimal control law of incremental form:

Δ u (k)=S ^-1YS-Y-KX-SU-SF1-SF2} wherein: $S^{-1}=\underset{i=1}{\overset{P-7}{\mathrm{\Σ}}}{\mathrm{CA}}^{i-1}B$

YS＝Y ^S(k+P)

Y＝Y(k)

KX＝CA ^P-1[X(k)-X(k-P)] $\mathrm{SU}=\underset{i=1}{\overset{P-7}{\mathrm{\Σ}}}{\mathrm{CA}}^{i-1}B\[u(k-1)-u(k-P)\]+\underset{i=p-6}{\overset{P-1}{\mathrm{\Σ}}}{\mathrm{CA}}^{i-1}B\[u(k+P-i-7)-u(k-i-7)\]$ $\mathrm{SF}1=\underset{i=1}{\overset{P-7}{\mathrm{\Σ}}}{\mathrm{CA}}^{i-1}{F}_1\left(1\right)\[F1(k-1)-F1(k-P)\]+\underset{i=p-6}{\overset{P-1}{\mathrm{\Σ}}}{\mathrm{CA}}^{i-1}{F}_1\left(1\right)\[F1(k+P-i-7)-F1(k-i-7)\]$ $-\underset{i=1}{\overset{P-1}{\mathrm{\Σ}}}{\mathrm{CA}}^{i-1}{F}_1\left(2\right)\[\mathrm{MU}1(k-1)-\mathrm{MU}1(k-P)\]$

SF2＝F2(1)*[F1(k-1)-F1(k-7)]

Wherein:

Δ u: manipulated variable control increment;

Ys: output variable (controlled variable) set(ting)value;

Y: output variable (controlled variable) observed value;

S ^-1: prediction time domain step response battle array;

P: prediction time domain;

F ₁(1): matrix F 1 first column element;

F ₁(2): matrix F 1 secondary series element;

F ₂(1): matrix F 2 first column elements.

2. 180# oil

Model specificationization obtains:

Wherein:

X(k)＝T0(k)????U(k)＝TI(k)?? $V\left(k\right)=J\left[\begin{array}{c}F1\left(k\right)\\ F2\left(k\right)\\ F3\left(k\right)\\ \mathrm{MU}1\left(k\right)\end{array}\right]$

Y(k)＝MU180(k)

X: control model state variable, promptly the reactor reaction temperature is one 1 * 1 vector in this example;

U: control model manipulation variable, i.e. reactor inlet temperature, 1 * 1 vector;

V: control model perturbing variables, i.e. reactor feed flow, charging viscosity.2 * 1 vectors;

Y: control model controlled variable (model output variable), i.e. visbreaking system outlet product viscosity, 1 * 1 vector;

A, B, C, F1, F2, F3: model coefficient battle array;

Prediction output is also carried out feedback modifiers, and adopt the monodrome predictive control algorithm, obtains the optimal control law of incremental form:

Δ u (k)=S ^-1YS-Y-KX-SU-SF1-SF2} wherein: $S^{-1}=\underset{i=1}{\overset{P-7}{\mathrm{\Σ}}}{\mathrm{CA}}^{i-1}B$

YS＝Y ^S(k+P)

Y＝Y(k)

KX＝CAP ^-1[X(k)-X(k-P)] $\mathrm{SU}=\underset{i=1}{\overset{P-7}{\mathrm{\Σ}}}{\mathrm{CA}}^{i-1}B\[u(k-1)-u(k-P)\]+\underset{i=p-6}{\overset{P-1}{\mathrm{\Σ}}}{\mathrm{CA}}^{i-1}B\[u(k+P-i-7)-u(k-i-7)\]$ $\mathrm{SF}1=\underset{i=1}{\overset{P-7}{\mathrm{\Σ}}}{\mathrm{CA}}^{i-1}F1\[V(k-1)-V(k-P)\]+\underset{i=p-6}{\overset{P-1}{\mathrm{\Σ}}}{\mathrm{CA}}^{i-1}F1\[V(k+P-i-7)-V(k-i-7)\]$

SF2＝F2*[V(k-1)-V(k-7)]

Wherein:

Δ u: manipulated variable control increment;

Ys: output variable (controlled variable) set(ting)value;

Y: output variable (controlled variable) observed value;

S ^-1: prediction time domain step response battle array;

P: prediction time domain.

(5) different Oil Production schemes automatically switch tactful

1. switch No. 180 oil of production by producing No. 250 oil

In m control cycle (every n minute program is sent a SP, i.e. a control cycle), divide the m step to carry out the transition to production decision No. 180.(m, n decides on field condition)

(1) obtain last produce No. 180 when oily reactor inlet temperature and the programmed values SP of diesel oil flow

(2) SP before SP and this switching is poor, difference is done the average of m, as the manipulated variable of switching controls

(3) manipulated variable is added on the programmed values SP of current reactor inlet temperature and diesel oil flow, as SP output, control of output in every n minute promptly automaticallyed switch to go in the production decision of No. 180 oil after m*n minute

2. switch No. 250 oil of production by producing No. 180 oil

In m control cycle (every n minute program is sent a SP, i.e. a control cycle), divide the m step to carry out the transition to production decision No. 250.

(1) obtain last produce No. 250 when oily reactor inlet temperature and the programmed values SP of diesel oil flow

(2) SP and this programmed values SP that switches preceding reactor inlet temperature and diesel oil flow is poor, difference is done the average of m, as the manipulated variable of switching controls

(3) manipulated variable is added on the programmed values SP of current reactor inlet temperature and diesel oil flow, as SP output, control of output in every n minute promptly automaticallyed switch to go in the production decision of No. 250 oil after m*n minute.

The function of this control strategy, as shown in Figure 3:

The concrete step of implementing is (describing in the example that is embodied as of certain factory with the present invention)

The through engineering approaches of model before implementing;

Engineering construction comprises:

(1) DCS configuration

This part is assisted to finish by manufacturer, and main task is to provide necessary condition for writing of controller software bag.

(2) the controller software bag writing and debugging.

Implement to require:

(1) conventional instrument is working properly;

(2) operating mode is steady;

(3) the DCS system is normal.

Beneficial effect of the present invention is: realized the multivariable prediction control of reactor product viscosity, reduced the fluctuation of product viscosity, undulating quantity is no more than 5% of controlling valu, guarantees that product viscosity is steadily qualified.Guarantee that other controlled variable, operational variable do not exceed its given upper and lower limit, guarantee whole device safety and steady operation.Reduced the incorporation of blended diesel, the reduction amplitude is no less than 1%, has improved economic benefit.The two kinds of schemes of 180# and 250# that realized are convenient to be switched, and has shortened product switching time.

As seen, compared with prior art, reactor smooth operation of the present invention, it is big to have changed adjusted open loop quality product fluctuation in the past, and temperature in is regulated frequent, the situation that the amplitude of accommodation is big; And diesel oil mediation amount obviously reduces the quality product edge; Particularly product oil is produced to switch and has been realized automatic control, and the adjusting time obviously reduces.Thereby the closed loop that has realized the heavy oil visbreaking is controlled automatically, has improved productive rate, has optimized the running status of reactor, has received good result energy-saving and cost-reducing, coordination optimization, has created economic benefit.

Above embodiment only is used to illustrate the present invention, but not is used to limit the present invention.

Claims (15)

1. the advanced control method of a heavy oil visbreaking reaction is characterized in that, real-time online is analyzed the viscosity of oil plant in the heavy oil visbreaking reactor; To heavy oil visbreaking reaction implementation model predictive control, that is: according to the oil brand of heavy oil visbreaking reactor outlet product oil, the operational variable of close-loop automatic adjustment heavy oil visbreaking reactor, realization is to the control of heavy oil visbreaking reactor outlet product oil viscosity, and automatic mutual switching of realizing at least two kinds of trade mark Oil Production.

2. method according to claim 1, it is characterized in that, the viscosity that described real-time online is analyzed oil plant in the heavy oil visbreaking reactor is meant: gather heavy oil visbreaking reactor inlet residual oil viscosity, outlet visbreaking residue viscosity, exported product oil viscosity in real time, and it is carried out the analysis of real-time online viscosity.

3. method according to claim 1, it is characterized in that, described oil brand according to heavy oil visbreaking reactor outlet product oil, the operational variable of close-loop automatic adjustment heavy oil visbreaking reactor, realize the control of heavy oil visbreaking reactor outlet product oil viscosity is meant: heavy oil visbreaking reactor outlet product oil can be No. 180 oil fuel or No. 250 oil fuel;

When product oil is No. 180 oil fuel: heavy oil visbreaking reactor outlet product oil viscosity is adopted regional control;

When product oil is No. 250 oil fuel: heavy oil visbreaking reactor outlet product oil viscosity is adopted singly to go into singly to go out Model Predictive Control.

4.. method according to claim 1, it is characterized in that, switching mutually automatically of at least two kinds of trade mark Oil Production of described realization is meant: switch to No. 250 oil fuel of production from producing No. 180 oil fuel, or switch to No. 180 oil fuel of production from producing No. 250 oil fuel.

5. method according to claim 1, its step comprises:

Gather heavy oil visbreaking reactor inlet residual oil viscosity, outlet visbreaking residue viscosity, exported product oil viscosity in real time, and it is carried out the analysis of real-time online viscosity;

Heavy oil visbreaking reactor outlet product oil can be No. 180 oil fuel or No. 250 oil fuel,

When product oil is No. 180 oil fuel: heavy oil visbreaking reactor outlet product oil viscosity is adopted regional control;

When product oil is No. 250 oil fuel: heavy oil visbreaking reactor outlet product oil viscosity is adopted singly to go into singly to go out Model Predictive Control;

Can realize automatically switching to No. 250 oil fuel of production, or automatically switch to No. 180 oil fuel of production from producing No. 250 oil fuel from producing No. 180 oil fuel.

6. method according to claim 5 is characterized in that, and is described when product oil is No. 180 oil fuel: adopt regional control to be meant to heavy oil visbreaking reactor outlet product oil viscosity:

Set the meadow, when visbreaking reactor outlet viscosity measurement value is outside the meadow, that is: outlet viscosity utilizes corresponding ratio control to regulate diesel oil mediation amount F2 greater than the set(ting)value upper limit or less than the set(ting)value lower limit; When visbreaking reactor outlet viscosity measurement value is within the meadow, adopt Model Predictive Control.

7. method according to claim 5, it is characterized in that, described when product oil is No. 250 oil fuel: as heavy oil visbreaking reactor outlet product oil viscosity to be adopted singly to go into singly to go out Model Predictive Control be meant: when No. 250 oil fuel are carried out visbreaking control, only need carry out close-loop automatic adjustment to visbreaking reactor inlet temperature TI.

8. method according to claim 5 is characterized in that, the described realization automatically switches to and produce No. 250 oil fuel and be meant from producing No. 180 oil fuel: transfer to and produce No. 250 oil fuel and can adopt direct switching from producing No. 180 oil fuel.

9. method according to claim 5, it is characterized in that, describedly automatically switch to and produce No. 180 oil fuel and be meant: transfer to and produce No. 180 oil fuel from producing No. 250 oil fuel from producing No. 250 oil fuel, can keep under the constant condition of TI, a step F2 increases, and increases the average of F2 when measuring No. 180 oily ordinary production that is:, the system that makes promptly enters the meadow at short notice, treat that it enters after the meadow, cancellation TI remains unchanged, and implements control by the general control of No. 180 oil.

10. method according to claim 6 is characterized in that, described regional control is meant the product viscosity multivariable prediction control based on mechanism model; Its with product viscosity as main controlled variable, and for guaranteeing the material balance of separation column, liquid level at the bottom of the tower is kept existing conventional control, promptly utilize liquid level at the bottom of the bottom product discharge control tower, with temperature of reaction and diesel oil incorporation as main operational variable, by regulating furnace outlet temperature and diesel oil incorporation control product viscosity; When control, should change temperature of reaction, and the diesel oil incorporation is in less important manipulated variable as far as possible, and the least possible;

In the mechanism model of setting up, include in and can survey interference, as the influence of unit capacity, wax oil incorporation, material viscosity, material density, raw material carbon residue, diesel oil viscosity and wax oil viscosity, can utilize feed-forward and feedback control to reduce their disturbances to whole process control.

11. method according to claim 7, it is characterized in that, described Model Predictive Control is meant: the multivariable prediction based on mechanism model is coordinated control, it can set up the dynamic mathematical models that can reflect production process according to the material balance and the energy balance of reactor and separation column, the multivariate monodrome predictive control of feedback of status is predicted and had to implementation model, controls according to the deviation and the overrun condition of the controlled variable of predicting.

12. method according to claim 8 is characterized in that, describedly transfers to and produces No. 250 oil fuel and can adopt the step of direct switching to be from producing No. 180 oil fuel:

In m control cycle, divide the m step to carry out the transition to No. 250 production,

Obtain last produce No. 250 when oily reactor inlet temperature and the programmed values SP of diesel oil flow;

SP and this programmed values SP that switches preceding reactor inlet temperature and diesel oil flow is poor, difference is done the average of m, as the manipulated variable of switching controls;

Manipulated variable is added on the programmed values SP of current reactor inlet temperature and diesel oil flow, as SP output, control of output in every n minute promptly automaticallyed switch to go in the production decision of No. 250 oil after m*n minute.

13. method according to claim 9 is characterized in that, describedly automatically switches to the step of producing No. 180 oil fuel and comprises from producing No. 250 oil fuel:

In m control cycle, divide the m step to carry out the transition to No. 180 production,

Obtain last produce No. 180 when oily reactor inlet temperature and the programmed values SP of diesel oil flow;

SP before SP and this switching is poor, difference is done the average of m, as the manipulated variable of switching controls;

Manipulated variable is added on the programmed values SP of current reactor inlet temperature and diesel oil flow, as SP output, control of output in every n minute promptly automaticallyed switch to go in the production decision of No. 180 oil after m*n minute.

14. method according to claim 10 is characterized in that, described product viscosity multivariable prediction control based on mechanism model need be satisfied following condition:

The output condition of No. 180 oil is: $\mathrm{MU}108\left(k\right)=C_{180}*T0(k-\mathrm{tao}3)+D_{180}\left[\begin{array}{c}F1(k-\mathrm{tao}4)\\ F2(k-\mathrm{tao}5)\\ F3(k-\mathrm{tao}6)\\ \mathrm{MU}1(k-\mathrm{tao}1)\end{array}\right]$

The modeling condition of No. 180 oil of controlled plant is:

The control condition of No. 180 oil of controlled plant is:

15. method according to claim 11 is characterized in that, described multivariable prediction based on mechanism model is coordinated control and need be satisfied following condition:

The output condition of No. 250 oil of controlled plant is: $\mathrm{MU}250\left(k\right)=C_{250}*T0(k-\mathrm{tao}3)+D_{250}*\left[\begin{array}{c}F1(k-\mathrm{tao}4)\\ \mathrm{MU}1(k-\mathrm{tao}1)\end{array}\right]$

The modeling condition of No. 180 oil of controlled plant is:

The control condition of No. 250 oil of controlled plant is:

Images