CN1055026C - On-line optimizing and controlling method of balanced microcomputer for producing processed materials by gas separation unit - Google Patents

Abstract

The present invention discloses an on-line optimizing and controlling method of a balanced microcomputer for producing material by a gas separation device. The existing computer control system is provided with an optimizing machine. Through sampling analysis, material balance optimization calculation and on-line optimized control, the problem that material of a gas separation device can not be automatically balanced in an on-line mode is solved. The present invention has the characteristics of high control precision and little fluctuation of control parameters. The effect of high quality, low consumption and strong control performance for resisting disturbance is obtained. The present invention is suitable for gas fractionation, MTBE and alkylation joint type chemical industry equipment.

Description

The balanced microcomputer for producing processed materials by gas separation unit on-line optimizing and controlling method

The present invention is a kind of microcomputer on-line optimizing and controlling method, is applicable to gas fractionation in the petroleum refining industry, MTBE and alkylation associative form chemical plant installations and similar other chemical plant installations.

In the oil plant task of gas separation unit exactly for MTBE and (or) alkyl plant supplies raw materials.Wherein, MTBE matching requirements isobutene purity reaches more than 18.81% and is advisable; Alkyl plant fresh feed alkane alkene ratio is to be advisable at 1: 1.How to satisfy the technological requirement of stream oriented device simultaneously, make its performance optimization process ability, and the variation of forming according to raw material, carrying out online processing of microcomputer and optimal control in real time is a great problem of original production technology.Analogue instrument control that China is adopted most of refinery and the control of DCS system all can only be handled the flow process control of single material, and can not handle the material balance distribution of package unit.As the analogue instrument control FC system (1987-1989) of Luoyang Petrochemical designing institute design, the Yokogawa DCS system of Japan, the TDC-3000 system of the U.S. etc., material balance all can only lean on the operator artificially to grasp by rule of thumb, therefore, its degree of balance is poor, and imbalance state is serious.Even it is more reasonable that material distributes, adopt the advanced tandem control of industry at present, its material balance control is still undesirable.The fluctuation of tandem principal parameter is generally more than 10%, and second parameter fluctuates more than 30%.Material components etc. are in case to have slightly great fluctuation process just can cause again unbalance, thereby make the control difficulty of material balance big, complicated operation.

The objective of the invention is the problems referred to above at current production existence, seeking the optimization thought of optimizing point down with multi-constraint condition in the optimal control theory is foundation, in conjunction with improvement to tandem control, provide a kind of can with the matching used balanced microcomputer for producing processed materials by gas separation unit on-line optimizing and controlling method of raw material components online gc.Thereby can solve the difficult problem that gas fractionation unit for a long time can't automatic on-line balance material preferably.

For achieving the above object, the present invention is provided with one and finishes the host computer of optimizing computing and control on existing DCS system or other computer control system, and host computer CPU disposal ability requires IBM or the compatible more than 80386.On the former stockline of minute gas by the online gc sampling analysis, if no online gc is artificial off-line setting component also, raw material composition analysis result is delivered to the slave computer of DCS system or other computer control system, pass through the sampled value of raw material composition analysis result and raw material treating capacity last together by slave computer again, the slave computer communication is issued host computer and is carried out material balance optimization computing, material sendout between each tower of branch device of air that calculates is composed to the initial controlled quentity controlled variable of corresponding variable step size controller as corresponding material, pass through the Optimal Control amount that obtains last, slave computer is issued in the slave computer communication, finishes the control function by the field control unit of slave computer; Again the field control result is fed back to host computer, repeat above-mentioned control procedure, to realize the online excellent control of material balance.

Production material on-line optimizing and controlling method of balanced microcomputer of the present invention, simulation is produced in the material optimization that can be used for gas separation unit, but the existing DCS computer control system of coupling apparatus etc. is directly used in online real-time production process control again.Therefore, the design of the system of computer program total two large divisions form, and one is produced simulation for material balance optimization, and this part subsystem can calculate separately, shows, print, and forms the separate analogue production system; It two is that material balance optimizes the production On-line Control.Design philosophy, Mathematical Modeling and the technical characterictic of following division two parts program:

1, material balance optimization (simulation) operation program (algorithm 1).

Gas divides the raw material composition analysis result such as the table 1 of production line:

Table 1

Sampled point	Gas divides former stockline
										The component title	C 2	C 3 o	C 3 ＝	iC 4 o	iC 4 ＝+C 4-1 ＝	nC 4 o	tC 4-2 ＝	cC 4-2 ＝	C5
Raw material is formed	0.2	10	33	20	16	5	8.5	6	1.3

If optimizing the decision-making effectiveness factor is X (representing deisobutanizer top normal butane and suitable, anti-butylene virtual component sum), the optimal control target function is Q (X).Guarantee that gas branch material balance has three constraintss, is respectively:

1, guarantees that gas separation unit alkane alkene ratio is 1: 1 (± 0.05);

2, guarantee that the isobutene purity requirement meets or exceeds design objective 18.81%;

3, the deisobutanizer isobutane purity requires 50% (± 5%).

Setting the sub-goal function respectively is J1 (X), J2 (X), J3 (X).So, the optimum control function is: $Q\left(X\right)=\min \underset{L}{\overset{H}{\∫}}\left[(J2\left(X\right)+J3\left(X\right))\right]\mathrm{dx}$

Making the condition of J1 (X) satisfied 1 ± 0.05 is H≤20, L 〉=4; Wherein: $J3\left(X\right)=|{\mathrm{iC}}_4^o\&CenterDot;100/({\mathrm{iC}}_4^o+\mathrm{<figure-callout\; id="4"\; label="i\; C"\; filenames="C9511193000201.png"\; state="\{\{state\}\}">i</figure-callout>}{C}_{}^{}{}_4^{=}+C_{4-1}^{=})\&CenterDot;(1+X/100)-50|$

J2(X)＝|(100-J3(X)-X-50+2)/200-18.81|

The online gc composition was gathered every 2 minutes, and computing is once optimized the also corresponding result of calculating simultaneously and produced material balance schemes.Find the solution optimizing decision effectiveness factor X by the microcomputer on-line operation, associative list 1 also calculates by gas separation unit material flow shown in Figure 2 and to try to achieve material sendout and iso-butane between each tower, isobutene purity.

Gas divides deisobutanizer isobutane purity J=FiC ₄ ^o/ B

Gas divides deisobutanizer isobutene purity I=(100-J-X+2)/200

Wherein F is that gas divides former stockline gas divisional processing amount, can be by the online gc collection; B is a deisobutanizer top output quantity, B=F (iC ₄ ^o+ iC ₄ ⁼+ C _4-1 ⁼) % (1+X%).

2, material balance is optimized production On-line Control program (algorithm 2)

With liquid level---the flow cascade table is an example:

If liquid level---main contral parameter is L0, and allowing the fluctuation deviation is ± Δ L0 to allow fluctuation

Codomain is L0 ± Δ L0

If flow---secondary control parameter is F0, and allowing upper and lower limit is that F0 is big, F0 is little; Initial controlled quentity controlled variable F0 is obtained by algorithm 1 or F0=(F0 big+F0 is little)/2 by formula

Fluctuating range be Δ F=|F0 big-F0 is little |/2

Initial step length is K * Δ F * Δ L/ Δ L0; Wherein, K is a hysteresis factors, can be according to the adjustment of table type, and suitable value is 0.1--10; Δ L is the poor of main contral parameter measured value and L0.

Linear relationship is: Δ L--F+K. Δ F. Δ L/ Δ L0

Secondary control restrained boundary is: and F0 is little≤and L0≤F0 is big for F+K. Δ F. Δ L/ Δ

Control procedure:

If actual measurement principal parameter changing value is LPV, tight back changing value is LPV ', and the fluctuation deviation is Δ L=LPV-L0

At first judge Δ L:

When | Δ L|≤| during Δ L0|, then work as the corresponding at this moment subloop flow measured value in LPV '＞L0 and LPV '-LPV＜0 and be designated as F ' greatly; To be designated as F ' little when corresponding subloop flow actual measurement LPV '＜L0 and LPV '-LPV＞0 this moment.F ' is big, F ' little in pairs after, following carrying out:

F '=(F ' big+F ' is little)/2; Δ F '=| F ' is big-and F ' is little | and/2;

Control action is: Δ L--F+K ' Δ F ' Δ L/ Δ L0;

Otherwise when | Δ L|＞| the unconditional tandem of carrying out is controlled during Δ L0|.

Along with flow steadily, codomain narrows down, corresponding step-length also diminishes thereupon, finally reaches the purpose of steady adjusting.When treating capacity changes or other factors will recover tandem control when causing the main contral parameter measured value not restrain to jump out Δ L0 scope automatically, corresponding material balance optimal value is chosen in the first control of secondary simultaneously control parameter again.After treating that the main contral parameter measured value enters predefined ranges of value, re-execute the advanced control of variable step size again, to guarantee the stable and continuous of whole excellent control process.

After adopting this method to carry out the material optimal control, can make gas separation unit production regulate material balance automatically, and control accuracy is improved greatly, the variable step size controller is compared with former tandem control, the principal parameter fluctuation is not higher than 5%, second parameter fluctuation is not higher than 20%, received high-quality, low consumption, the effect that the control vulnerability to jamming is strong.

The invention will be further described below in conjunction with drawings and embodiments.

Fig. 1 is the sub-goal function curve diagram of algorithm 1;

Fig. 2 is gas separation unit material flow figure;

Fig. 3 is the control procedure block diagram;

Fig. 4 optimizes computing tutorial program flow chart for material balance;

Fig. 5 optimizes operational flowchart for material balance;

Fig. 6 is the advanced control of variable step relevant parameter setting process figure;

Fig. 7 is the advanced control flow chart of variable step;

F ' is big, F ' interlude flow chart in order to seek for Fig. 8;

Fig. 9 is a gas separation unit material balance optimal control design concept block diagram.

Among Fig. 1, J3 (X) is that isobutane purity is with optimizing the factor variations function; J2 (X) is that isobutene purity is with optimizing the factor variations function.

Among Fig. 2, F is a gas divisional processing amount, and its composition sees Table 1, represents with percentage by weight; A, B, C, D, E, G, H, L are material input and output amount between each tower; T101 is a depropanizing tower, and T102 is a dethanizer, and T103 is a depropenizer, and T104 is a deisobutanizer, and T105 is a depentanizer; T2J is the dethanizer inlet amount, T2D for dethanizer at the bottom of output, T3J is the depropenizer inlet amount, T3D is a propone output, T3C is a propane output; M is a MTBE output, and N is an alkylate oil output, and Q is the ethene input quantity, and W is for removing storage tank along anti-butylene; V=is that gas divides deisobutanizer to alkylation C ₄ ⁼Amount of alkene, V ⁼=(B-D) [(100-J)-(X/2+2)]/100; Vi ^oFor gas divides deisobutanizer to alkylation of isobutane amount, V _i ^o=(B-D) J/100; M ⁼For the MTBE extraction tower to baseization C ₄ ⁼Amount of alkene, M ⁼=H[D (100-J-X/2-2)/100-0.98DI]/(H+Q); M _i ^oFor the MTBE extraction tower to alkylation of isobutane amount, M _i ^o=H (DJ/100-M0.05/100)/(H+Q); Tc is that gas divides the depentanize cat head along anti-butylene purity, tc=(FC _4-2 ⁼/ 100-B (X/2-2)/100)/(C-FC ₅/ 100); When V is the alkylation fresh feed by certain alkane alkene than control, need gas divide the depentanize cat head to mend, V=[(Vi into amount ^o+ Mi ^o)-Vb (V ⁼+ M ⁼)]/(Vbtc); Vb is an alkylation fresh feed alkane alkene ratio, and Vb=1; Ed is that gas divides propylene tower purified propylene, ed=0.93--0.98.

Among Fig. 3, T101 is a depropanizing tower; T104 is a deisobutanizer; SP is a gas chromatograph; F is that gas divides throughput per hour, i.e. the depropanizing tower inlet amount; LT is a fluid level transmitter; FT is a flow transmitter; FV is a control valve; XWJ is a slave computer; KBT is variable step size controller, i.e. host computer; DQ is an electropneumatic transducer; FS is that T101 feed composition or flow are formed; On behalf of algorithm 1,02,01 represent algorithm 2.

Fig. 9 is to be the design block diagram of body general procedure design, has designed nine modules altogether, and wherein the data communication module is the data communication structure that connects material optimization system and existing computer control system.Its communication modes can be selected the communications protocol identical with this computer control system network configuration, also can be by the computer intercommunication data of RS-232 interface and original control system; Native system is to adopt U.S. ARCNET procotol and original control system to share data.

The optimization system design is divided into two lines according to material balance optimization simulation and online excellent control two large divisions, promptly selects A and selects B.Select A to produce simulation program for material balance optimization, it comprises inlet 1--inlet 6, and the content of each module defines in Fig. 4.Wherein enter the mouth 3, inlet 4, the program of inlet 5 for compiling and editing and carry out, other inlet all can be abridged as required.Selecting B is that material balance is optimized production On-line Control program, it comprises inlet 1, inlet 5, inlet 4, inlet 7 and online advanced variable step module, wherein enter the mouth 1, inlet 5 and entering the mouth 4 can be ready in selecting A, 7 will prepare the desired parameters of variable step size control and enter the mouth, Fig. 6 is this module parameter setting process figure.

The key modules of configuration program is the advanced control module of inlet 4, inlet 7 and online variable step.Fig. 5 is the program flow diagram of inlet 4 modules.Calculate the absolute value of the difference of isobutene purity and design objective 18.81% along with the circulation (40--200) of variable i; Calculate the absolute value of the difference of isobutane purity and design objective 50%; Seek two absolute values and minimum of a value, and with the cyclic variable of this point divided by after 10, compose and cheat variable Point.The value of point is best decision effectiveness factor X so, at this moment.

Inlet 7 is variable step size controller desired parameters setting modules, and Fig. 6 is its flow chart.Material balance is optimized operation result if will be dropped in line traffic control, need to realize through the variable step size controller, and realize that variable step control needs to set the relevant parameter value of each cascade table, be the initial optimization point of master meter, excellent control alarm point and subtabulation excellent control alarm point, F0 is big, F0 is little and the hysteresis factors K0 when calculating excellent control output, its setting range is: the initial controlled quentity controlled variable L0=35--75% of main contral parameter liquid level; Permissible variation amount Δ L0=3--15% high alarm setting is L0+ Δ L0, and low alarm setting is L0-Δ L0; The constraint upper limit F0 of secondary control parameter flow is big≤the instrument measurement upper limit, and constraint lower limit F0 is little 〉=the instrument measurement lower limit; Initial controlled quentity controlled variable F0 is for optimizing operation result or F0=(F0 big+F0 is little)/2; Hysteresis factors K=0.1--10.After setting finishes, can throw optimal control from the whole cascade tables of procedure Selection, or select current certain cascade table to throw optimal control, in case selected, system will control the optimization result of this cascade table application material balance.

The advanced control module of online variable step, its flow chart such as Fig. 7, Fig. 8, this program is to produce with interrupt mode, thereby operation has real-time, move a suboptimum control program each second, be not subjected to other program whether to move the influence that finishes, so just guaranteed the real-time and the continuity of control.

The control of material balance microcomputer on-line optimization is exactly to realize material level and corresponding material optimal value balance Control Scheme at each tower of gas fractionation unit or return tank material level storeroom corresponding with it on the tandem basis.Now wherein one the tunnel to be controlled to be example, promptly depropanizing tower bed material position and deisobutanizer charging cascade table are explained.Suppose that gas divides 30 tons of throughput per hours, at first on host computer depropanizing tower bed material position is arranged on 60% by technology, allowing the fluctuation deviation setting is 5%, and allowing the fluctuation codomain according to algorithm 2 is 55--65%, the depropanizing tower tower bottom flow amount upper limit is made as 24 tons/hour, and lower limit set is 13 tons/hour.Hysteresis factors K is set to 1.0, throw excellent control, at this moment, raw material composition analysis data are delivered to the slave computer of DCS system or other control system by chromatograph, by slave computer the sampled value (or setting value) of raw material composition analysis result and raw material treating capacity is issued host computer by the communication of upper and lower machine together again, and carry out computing by algorithm 1; Again the optimal value of inventory at the bottom of the depropanizing tower that calculates is composed to the variable step size controller, as the beginning control amount of depropanizing tower substrate material; To issue slave computer by the communication of upper and lower machine through the Optimal Control amount that the variable step size controller obtains again, by the field control unit of slave computer, promptly adjuster is finished near the balance adjustment that depropanization Tata bed material is positioned at this bottoms material optimal value or the material upper and lower limit is set again.Again the field control result is fed back to host computer by LT, FT, repeat above-mentioned control procedure then.When treating capacity changes or other factors causes the main contral parameter measured value not restrain, will recover tandem control when jumping out the 55--65% scope automatically.The first control amount of secondary simultaneously control parameter is chosen corresponding material balance optimal value again.After treating that the main contral parameter measured value enters predefined ranges of value, re-execute the advanced control of variable step size again, to guarantee the stable and continuous of whole excellent control process.

Claims (7)

1, a kind of balanced microcomputer for producing processed materials by gas separation unit on-line optimizing and controlling method is to divide at gas to adopt the analysis of online gc sample on the production line, and the result delivers to computer control system with the raw material composition analysis; It is characterized in that on DCS system or other existing computer control system, being provided with one and finish the upper calculation machine of optimizing computing and control, and carry out as follows:

A, provide the Mathematical Modeling that the gas separation unit material is optimized computing;

B, according to optimizing Mathematical Modeling establishment material balance optimizing operation program;

C, be equilibrated at line optimal control program by optimizing operation result establishment material;

D, by operation optimize computing, control program is realized online excellent control.

2, optimal control method according to claim 1, the Mathematical Modeling that it is characterized in that said gas separation unit material optimization computing, be under given constraints, press the optimal design index of gas separation unit material flow, ask for the mathematical operation formula of the optimization decision-making effectiveness factor X that represents deisobutanizer top normal butane and suitable, anti-butylene virtual component sum, that is: $Q\left(X\right)=\min \underset{L}{\overset{H}{\&Integral;}}(J2\left(X\right)+J3\left(X\right))\mathrm{dx}$

Wherein: H≤20, L 〉=4; $J3\left(X\right)=|i{C}_4^o\&CenterDot;100/(i{C}_4^o-i{C}_4^{=}+C_{4-1}^{=})\&CenterDot;(1+X/100)-50|$ $J2\left(X\right)=|(100-J3\left(X\right)-X-50+2)/200-18.81|.$

3, optimal control method according to claim 1 is characterized in that said material balance optimization operation program thes contents are as follows:

A, along with the circulation of variable i, calculate the absolute value of the difference of isobutene purity and design objective 18.81%;

B. calculate the absolute value of the difference of isobutane purity and design objective 50%;

C, seek two absolute values and minimum of a value.And with the cyclic variable of this point divided by after 10, compose and give variable Point, this moment, the value of Point was best decision effectiveness factor X.

4, optimal control method according to claim 3, the span that it is characterized in that cyclic variable i is 40--200.

5, optimal control method according to claim 1 is characterized in that said material balance on-line optimization control program is online real-time variable step control program, and it comprises:

A, online excellent control relevant parameter are set;

B, initial step length are K Δ F Δ L/ Δ L0

C, principal parameter fluctuation deviation delta L=LPV-L0;

D, call and seek that F ' is big, F ' interlude;

E, control action are: Δ L-F+K Δ F Δ L/ Δ L0.

6, optimal control method according to claim 5 is characterized in that said searching F ' content big, F ' interlude is:

A, judgement | Δ L|≤| whether Δ L0| sets up, if, then:

B, when LPV '＞L0 and LPV '-LPV＜0, this moment corresponding subloop flow measured value to be designated as F ' big;

C, when LPV＜L0 and LPV '-LPV＞0, this moment corresponding subloop flow measured value to be designated as F ' little;

D, if | Δ L|＞| Δ L0|, unconditionally carry out tandem control.

7, optimal control method according to claim 5 is characterized in that the content that online excellent control relevant parameter is set is:

The initial controlled quentity controlled variable L0=35--75% of main contral parameter liquid level;

Permissible variation amount Δ L0=3--15%;

High alarm setting is L0+ Δ L0, and low alarm setting is L0-Δ L0;

The constraint upper limit F0 of secondary control parameter flow is big≤the instrument measurement upper limit, and constraint lower limit F0 is little 〉=the instrument measurement lower limit;

Initial controlled quentity controlled variable F0 is for optimizing operation result or F0=(F0 big+F0 is little)/2

Hysteresis factors K=0.1--10.

Images