CN1139814C - Soft measurement method for conten of p-carboxyl benzaldehyde in product generated by oxidizing reaction of p-xylene - Google Patents
Soft measurement method for conten of p-carboxyl benzaldehyde in product generated by oxidizing reaction of p-xylene Download PDFInfo
- Publication number
- CN1139814C CN1139814C CNB011135174A CN01113517A CN1139814C CN 1139814 C CN1139814 C CN 1139814C CN B011135174 A CNB011135174 A CN B011135174A CN 01113517 A CN01113517 A CN 01113517A CN 1139814 C CN1139814 C CN 1139814C
- Authority
- CN
- China
- Prior art keywords
- value
- content
- cba
- air
- measurement method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Landscapes
- Organic Low-Molecular-Weight Compounds And Preparation Thereof (AREA)
Abstract
The present invention discloses soft measurement technology for the quality of products of a p-xylene oxidation reaction, which comprises the steps that the flow amount of p-xylene, and the flow amount of air and the tail oxygen concentration of reoxidation in a crystallizer are selected to be as quality indexes of the product, namely an input variable of a soft measurement model for the content of p-carboxybenzaldehyde; the trained weight value and the threshold value are introduced into the model to be calculated by real-time, continuous collection of process data by adopting a neural network model by utilizing a PX flow meter, an air flow meter, a tail oxygen concentration analyzer and a DCS system of the existing TA production device, and thereby, the soft measurement of the content of the p-carboxybenzaldehyde in the products of the p-xylene oxidation reaction.
Description
The invention belongs to the Chemical Reaction Engineering field, relate to the pure terephthalic acid (hereinafter to be referred as PTA, be PureTerephthalic Acid) produce in P-xylene (hereinafter to be referred as PX, be P-xylene) the oxidation reaction product quality indicator---to the flexible measurement method of carboxyl benzaldehyde (hereinafter to be referred as 4-CBA, i.e. 4-carboxybenzaldehyde) content.
Soft-measuring technique has obtained a large amount of successful application in industry in recent years, has solved many " measurement " problems that can not survey crucial controlling index.Be successfully applied to devices such as catalytic cracking at home, important parameters such as catalyst recirculation amount, gasoline endpoint, diesel oil solidifying point have wherein been measured with the method for soft measurement respectively; Under need be to situation about controlling with the closely-related significant process variable of product quality, because not only price comparison costliness of on-line analysis instrument, and some industrial processes are also lacked very much, make that some primary variabless in the production run are difficult to or can't be directly obtained by the on-line analysis instrument at all.In the production run of PX oxidation reaction, product crude terephthalic acid (TA, Terephthalic Acid) in to carboxyl benzaldehyde (4-CBA, 4-carboxybenzaldehyde) content is the quality index of a key, it normally by once obtaining every the several hrs manual analysis, does not have in-line analyzer can detect and control this parameter in real time at present.Therefore, the 4-CBA content of measuring in time, exactly in the TA product is the key of the advanced control technology of research and development TA product quality.
In the multiple production technology of TA, the PX oxidation reaction becomes the main flow technology of producing TA gradually with advantages such as its lower acid consumption, material consumptions.Therefore, the 4-CBA flexible measurement method here also is at this production technology.Fig. 1 has shown the process flow diagram of typical PX oxidation reaction process.The PX oxidation reaction process is: raw material (liquid phase P X) and certain catalyzer and solvent are become a kind of acid slurry, enter reactor then, carry out oxidation reaction with airborne oxygen and generate TA under high temperature, high pressure, its main secondary product is 4-CBA.Because 93% TA separates out in reactor, therefore still unoxidized intermediate product can carry out secondary oxidation in first crystallizer, can further reduce the 4-CBA content among the TA like this.By the crystallization of second and third grade crystallizer, the slurry of generation after filtration, drying obtains TA.
Because the 4-CBA in the TA product is the main accessory substance of oxidation reaction, its content directly has influence on the outward appearance and the quality of product, is the important indicator of PTA product quality.The 4-CBA too high levels will make product painted, have a strong impact on product quality; Otherwise, illustrate that combustion reaction is violent in the oxidation reaction process, can increase material consumption, the energy consumption of reaction, and have more secondary product.Therefore each PTA factory commercial city takes various approach to try hard to the 4-CBA stable content a proper level.
Because it is more to influence the factor of 4-CBA, every theoretically participation PX oxidation reaction process related technical parameters all can have influence on its content, as: CO in catalyzer composition, oxidation reaction actuator temperature, pressure, air capacity, ratio of solvent, reactant concentration, partial pressure of oxygen, the residence time and the reaction end gas
2Content; The temperature of first crystallizer, pressure, air capacity, tail oxygen concentration etc.
In fact, after the secondary oxidation effect of first crystallizer, the content of 4-CBA is determined.Generally the manual analysis sampling spot of 4-CBA is obtained in discharging place of dryer, needs at least 1 and a half hours from first crystallizer to dryer, adds that the manual analysis time (off-line stratographic analysis) of 4-CBA needs 40 minutes at least.Therefore, in conventional 4-CBA control, current manual analysis value is to be determined by the reaction condition before the several hrs, it is not the real-time measurement of 4-CBA, operating personnel can only by virtue of experience adjust, like this large time delay, have probabilistic adjustment, very easily cause the fluctuation of reactor product quality.Especially when reaction condition is unstable, in order to obtain suitable 4-CBA, whole adjustment process even need reach several days time.
Jap.P. (JP101739) discloses the method for prediction 4-CBA: by unpressed gas concentration lwevel in the detection reaction device waste gas, obtained the concentration value of 4-CBA in the reactor by one-variable linear regression equation Y=A*X+B.
As mentioned above, owing to also will carry out the secondary oxidation reaction in first crystallizer, the content of 4-CBA could final decision after passing through first crystallizer.And if when having the situation of 1 above reactor, numerous course of reaction influence factors intercouple, and also can cause the prediction deviation of the content of 4-CBA.
In order to solve the large time delay control problem of 4-CBA, the online in real time problems of measurement of the 4-CBA that needs to be resolved hurrily.Because no matter process conditions or 4-CBA Determination on content method are all brought lag factor inevitably, therefore can't adopt hardware investment to solve this problem.
The objective of the invention is: the flexible measurement method to carboxyl benzaldehyde content is provided in a kind of p xylene oxidation course of reaction, realizes that the online in real time of 4-CBA is measured.This invention utilizes the computer distributed control system of existing instrument (PX flowmeter, airflow meter, tail oxygen concentration analyser) and existing TA process units (hereinafter to be referred as DCS, be Distributed Control Systems), adopt neural network model to predict, improved precision of prediction and the fault freedom of 4-CBA greatly.For the TA production quality control provides favourable foundation.
The object of the present invention is achieved like this: utilize on the basis of the DCS computer control system of existing PX flowmeter, airflow meter, tail oxygen concentration analyser and existing TA process units in the p xylene oxidation course of reaction, adopt neural network model to predict, real-time, continuous acquisition by DCS systematic procedure data (each input variable data in the finger print type here), the weights and the threshold value that train are brought into and calculated, just can obtain the real-time soft measurement predicted value of 4-CBA content; By anti-normalized, obtain the 4-CBA actual value.
At first, have the air mass flow of the flowmeter of measuring the PX flow, crystallizer flowmeter, measure the in-line analyzer of the tail oxygen content (vo1%) of crystallizer and the manual analysis value of 4-CBA; By measuring PX flow, the air capacity of crystallizer, the tail oxygen content of crystallizer, utilize nerual network technique, set up the soft measuring system of 4-CBA content.This soft measuring system can generally be applicable to the production technology of utilizing the PX oxidation to produce TA.Above-mentioned crystallizer generally is meant first crystallizer.
Soft measuring system of the present invention has been considered the different time of above-mentioned three factors to the 4-CBA influence, according to the reaction condition actual conditions, choose a unit interval before the current time of PX, the flow value (referring generally to the first two unit interval) of preceding another time respectively; Choose the flow value of the previous unit interval of the first crystallizer air mass flow present flow rate, current time; The concentration value of previous unit interval of choosing the current concentration of crystallizer tail oxygen content, current time is as the input variable of neural network model.
Then, according to the reasonable time of reaction, unit interval got one hour, and then the application module in computer system (hereinafter to be referred as AM, i.e. APPLICATION MODULE) module realizes obtaining before the PX flow current time 1 hour, 2 hours flow PX (K-1), PX (K-2); Obtain before the air mass flow present flow rate, current time of first crystallizer 1 hour flow value AIR (K), AIR (K-1); Obtain before the current concentration of tail oxygen content, current time of first crystallizer 1 hour concentration value O
2(K), O
2(K-1); If a plurality of measurement instruments are arranged, choose mean value and carry out the mistake diagnosis; And utilize formula (1), above-mentioned each variable is carried out normalized;
In nY=(Y-a)/(b-a) * (0.8-0.2)+0.2 (1) formula, Y represents the actual measured value of input variable, and nY represents the numerical value with input variable after the normalization, [a, b] expression input variable variation range, [0.2,0.8] expression normalization scope, the normalization scope also can be chosen other value, as [0,1], [0.5,0.5], [1,1] etc.
In neural network model, the node number of input layer is i (i=6), and the hidden layer number of plies in middle layer is L (L=1～100), and each the number of hidden nodes is j (2-100), and the output layer node is k (k=1～100).
By historgraphic data recording, choose 100 groups of data, every group of data comprise [PX (K-1), PX (K-2), AIR (K), AIR (K-1), O
2(K), O
2(K-1), 4-CBA manual analysis value (K)], utilize 4-CBA manual analysis value to do desired value, choose corresponding above-mentioned variable constantly as the input of neural network.For example, with 50 groups of neural network trainings,, trained and one group of weights that predicated error is less with 50 groups of prediction neural network generalization abilities; Here, the neural network input variable is 6, and choosing the hidden node number is 6, and output variable is 1.Can choose the different hidden layer numbers of plies and hidden node number according to requirement to model accuracy; In general, the increase hidden layer number of plies in the proper range and the number of node thereof can corresponding raising model accuracies.Variables Selection in the 4-CBA soft-sensing model
The concentration of 4-CBA can be judged by the first crystallizer airload situation and tail oxygen content (VOL%) usually; Promptly, the reaction condition of secondary oxidation can reflect the reaction depth of front reactor (can be separate unit reactor or many reactors) indirectly, and catalyst concentration is corresponding with the PX flow, therefore, the input variable of choosing the 4-CBA model is: PX flow, the air capacity of secondary oxidation and secondary oxidation tail oxygen content; Owing to there is the problem of the residence time in the course of reaction, cause each variable can not in time embody to the influence of 4-CBA, therefore, must be according to concrete reaction condition (as the residence time in material-compound tank, reactor, the haulage time of material etc.), the influence time of each input factor in the model to 4-CBA also taken into account.The 4-C8A soft-sensing model
Because backpropagation (BP, the Back propagation) neural network of structure has the function of any nonlinear function of mapping more than three layers, here, the BP neural network of employing three-decker is set up the soft-sensing model of 4-CBA.Utilize above-mentioned selected model variable, and consider the influence time of model input factor current time (promptly finger) 4-CBA at the first crystallizer place, construction 4-CBA neural network soft sensor model, as shown in Figure 2.The correction of 4-CBA soft-sensing model
Above-mentioned 4-CBA neural network soft sensor model is directly applied to the real-time estimate of commercial plant, because the disturbing factor that exists in the actual industrial production process is a lot, therefore this soft measured value and manual analysis value (off-line stratographic analysis) will produce certain deviation inevitably.For this reason, must be according at set intervals, for example 8 hours manual analysis values are once carried out online " rolling " to soft-sensing model and are optimized correction, make soft-sensing model adapt to the variation of industrial process operating characteristic and the migration of production status.
The soft measuring system of the present invention considers that the variation of catalyzer is corresponding relation with the PX flow, so its prerequisite is control catalyst concentration, makes it to be complementary with the PX flow.
Characteristics of the present invention: the flexible measurement method to carboxyl benzaldehyde content is provided in a kind of p xylene oxidation course of reaction, solve the large time delay of 4-CBA content measurement and influence the problem of control, on the basis that utilizes existing instrument (PX flowmeter, airflow meter, tail oxygen concentration analyser), can infer the content of 4-CBA in real time, promptly realize the online in real time measurement, for the TA production quality control provides favourable foundation; Owing to adopt neural network model to predict, improved precision of prediction and the fault freedom of 4-CBA greatly.The soft measuring system of the present invention can generally be applicable to the production technology of utilizing the PX oxidation to produce TA.
The invention will be further described below in conjunction with accompanying drawing and by embodiment:
The block diagram of PX oxidation reaction has been described among Fig. 1,
Wherein: corresponding numeral is 1 P-xylene PX; 2. catalyzer; 3. solvent; 4. air capacity; 5. air capacity.Utilize 1,5 the flowmeter and the tail oxygen concentration assay value at the first crystallizer place, calculate the 4-CBA content in the current time TA product in real time; Shown in Figure 1, the manual analysis value of 4-CBA records in discharging place of TA product usually.Reaction mechanism is basis of the present invention among Fig. 1, P-xylene PX, catalyzer, solvent, under High Temperature High Pressure by air oxidation reaction; Catalyzer then cobalt acetate, manganese acetate is formed, and with certain proportion (as 1: 1) preparation, the concentration of rugged catalyst is between [100ppm, 400ppm]; As control concentration of cobalt ions at 300ppm.
Fig. 2 is the soft measurement block diagram of 4-CBA neural network model, and the K shown in the figure, K-1, K-2 represent current time, current time last hour respectively, the preceding two hours numerical value of current time.The node of input layer is i (i=1-6) b1, and the hidden node in middle layer is j (j=1-6) b2, and the output layer node is k (k=1),
Input variable to model is carried out normalized, and the normalization scope of choosing here is [0.2 0.8].
Fig. 3 is a flow chart, and key is by the program value that whether differentiate outside input point be bad value and calculating goes out 4-CBA according to weights and the threshold calculations of Fig. 2 among the figure, and in the time period.
According to the flowmeter of the air mass flow of the flowmeter of PX flow, first crystallizer, measure the in-line analyzer of the tail oxygen content (vo1%) of crystallizer and the manual analysis value of 4-CBA; AM in computer system (APPLICATION MODULE) module realizes obtaining before the PX flow current time 1 hour, 2 hours flow PX (K-1), PX (K-2); Obtain the air mass flow mean value of two strands of single tubes of first crystallizer: 1 hour flow value AIR (K), AIR (K-1) before present flow rate, the current time; Obtain before the current concentration of tail oxygen content, current time of first crystallizer 1 hour concentration value O
2(K), O
2(K-1); If a plurality of measurement instruments are arranged, choose mean value and carry out the mistake diagnosis; And utilize formula (1), above-mentioned each variable is carried out normalized; The variation range [10,80] of PX (K), AIR (K) variation range [200,1200], O
2(K) variation range [0.5%, 10.5%].
By historgraphic data recording, choose 100 groups of data, every group of data comprise [PX (K-1), PX (K-2), AIR (K), AIR (K-1), O
2(K), O
2(K-1), 4-CBA manual analysis value (K)], utilize 4-CBA manual analysis value to do desired value, choose corresponding above-mentioned variable constantly as the input of neural network.With 50 groups of neural network trainings,, trained and one group of weights that predicated error is less with 50 groups of prediction neural network generalization abilities; Here, the neural network input variable is 6, and choosing the hidden node number is 6, and output variable is 1.Can choose the different hidden layer numbers of plies and hidden node number according to requirement to model accuracy; In general, the increase hidden layer number of plies in the proper range and the number of node thereof can corresponding raising model accuracies.
Here, the node of input layer is i (i=1-6), and the hidden node in middle layer is j (j=1-6), and the output layer node is k (k=1), then through after the training, obtains following one group of weights: w
11=-9.9996 w
12=4.2762 w
13=12.1089 w
14=-7.1424 w
15=4.0033 w
16=6.6466w
21=-8.9448 w
22=-3.3336 w
23=10.2651 w
24=-4.3202 w
25=-3.1763 w
26=12.1868w
31=4.0490 w
32=9.6067 w
33=1.6904 w
34=-1.2490 w
35=7.1705 w
36=14.3591w
41=-3.8343 w
42=11.2557 w
43=9.7094 w
44=5.3649 w
45=-0.9249 w
46=0.2625w
51=2.4643 w
52=4.6943 w
53=7.0411 w
54=-7.5121 w
55=-3.1356 w
56=16.7473w
61=-10.5346 w
62=-2.2789 w
63=-3.4760 w
64=11.8624 w
65=9.9663 w
66=-7.7713ww
1=-4.2749 ww
2=1.0098 ww
3=-1.8173 ww
4=1.9705 ww
5=0.9672ww
6=1.3803b
1=-0.5942 b
2=-2.5289 b
3=-23.1472 b
4=-11.6570 b
5=-7.8892 b
6=0.2932bpb
1=1.4523 (w wherein
IjBe the weights of i node to j node; Ww
jBe the weights of j node of middle hidden layer to k node of output layer; b
jThreshold value for j node of middle hidden layer; Bpb
kThreshold value for the output layer node)
It more than is the off-line simulation process, in the real time execution of device, then need go up the establishment that realizes the said process control language at the AM/APM of the DCS of TA process units system (APPLICATION MODULE/ADVANCED PROCESSMANAGER), real-time, continuous acquisition by DCS systematic procedure data (each input variable data in the finger print type here), the weights and the threshold value that train are brought into and calculated, just can obtain the real-time soft measurement predicted value of 4-CBA content; The 4-CBA numerical value that obtain this moment utilizes (2) formula between [0.2,0.8], carry out anti-normalized, obtains the 4-CBA actual value.
Y＝(nY-0.2)/(0.8-0.2)*(b-a)+a (2)
In addition, for guarantee forecast model accurately, operation effectively, utilize the manual analysis value to carry out " rollings " correction.
As model input variable PX (the K-1)=45m that chooses
3/ h, PX (K-2)=45m
3/ h, AIR (K)=700m
3/ h, AIR (K-1)=700m
3/ h, O
2(K)=5.5%, O
2(K-1)=5.5% after the normalization, numerical value is respectively 0.5,0.5,0.5,0.5,0.5,0.5, then calculates by neural network model: net1=w
11* PX (k-1)+w
12* PX (k-2)+w
13* AIR (k)+w
14* AIR (k-1)+w
15* O
2(k)+w
16* O
2(k-1)+b
1(3) net2=w
21* PX (k-1)+w
22* PX (k-2)+w
23* AIR (k)+W
24* AIR (k-1)+w
25* O
2(k)+w
26* O
2(k-1)+b
2(4) net3=w
31* PX (k-1)+w
32* PX (k-2)+w
33* AIR (k)+w
34* AIR (k-1)+w
35* O
2(k)+w
36* O
2(k-1)+b
3(5) net4=w
41* PX (k-1)+w
42* PX (k-2)+w
43* AIR (k)+w
44* AIR (k-1)+w
45* O
2(k)+w
46* O
2(k-1)+b
4(6) net5=w
51* PX (k-1)+w
52* PX (k-2)+w
53* AIR (k)+w
54* AIR (k-1)+w
55* O
2(k)+w
56* O
2(k-1)+b
5(7) net6=w
61* PX (k-1)+w
62* PX (k-2)+w
63* AIR (k)+w
64* AIR (k-1)+w
65* O
2(k)+w
66* O
2(k-1)+b
6(8)out1＝1/ (1+exp (-net1)) (9)out2＝1/ (1+exp (-net2)) (10)out3＝1/ (1+exp (-net3)) (11)out4＝1/ (1+exp (-net4)) (12)out5＝1/ (1+exp (-net5)) (13)out6＝1/ (1+exp (-net6)) (14)net7＝ww
1* PX (k-1)+ww
2* PX (k-2)+ww
3* AIR (k)+ww
4* AIR (k-1)+ww
5* O
2(k)+ww
6* O
2(k-1)+bpb
1(15) out7=1/ (1+exp (net7)) (16) out8=(out7-0.2) * 0.2167+0.2 (17)
When bringing weights, threshold value to (3)～(17) into, out8 is the 4-CBA predicted value that neural network model calculates, and error is within ± 80ppm between 4-CBA measured value that obtains and the manual analysis value.
The condition of above-mentioned requirements all can satisfy, so this invention has universality in most TA process units.
Claims (6)
1, in a kind of p xylene oxidation reactor product to the flexible measurement method of carboxyl benzaldehyde content, it is characterized in that utilizing the DCS system of existing PX flowmeter, airflow meter, tail oxygen concentration analyser and existing TA process units, adopt the BP neural network model, set up the soft measuring system of 4-CBA content, real-time, continuous acquisition by process data in the DCS system, bring the weights that train and the value of cutting off from into model and calculate, just can obtain the real-time soft measurement predicted value of 4-CBA content.
2, by in the described p xylene oxidation reactor product of claim 1 to the flexible measurement method of carboxyl benzaldehyde content, it is characterized in that choosing unit interval before the current time of PX and the flow value of the first two unit interval of current time; Choose the flow value of the previous unit interval of the present flow rate of the first crystallizer air mass flow and current time; The concentration value of previous unit interval of choosing current concentration of crystallizer tail oxygen content and current time is as the input variable of neural network model.
3, by in the described p xylene oxidation reactor product of claim 2 to the flexible measurement method of carboxyl benzaldehyde content, it is characterized in that the unit interval got one hour, realize obtaining before the PX flow current time 1 hour, 2 hours flow PX (K-1), PX (K-2); Obtain before the air mass flow present flow rate, current time of first crystallizer 1 hour flow value AIR (K), AIR (K-1); Obtain before the current concentration of tail oxygen content, current time of first crystallizer 1 hour concentration value 0
2(K), 0
2(K-1); If a plurality of measurement instruments are arranged, choose mean value and carry out the mistake diagnosis.
4, by in the described p xylene oxidation reactor product of claim 3 to the flexible measurement method of carboxyl benzaldehyde content, it is characterized in that above-mentioned each variable is carried out normalized; The variation range [10,80] of PX (K), AIR (K) variation range [200,1200], 0
2(K) variation range [0.5%, 10.5%].
5, by in the described p xylene oxidation reactor product of claim 3 to the flexible measurement method of carboxyl benzaldehyde content, it is characterized in that in the described neural network model, the node number of input layer is i (I=6), the hidden layer number of plies in middle layer is L (L=1-100), each the number of hidden nodes is j (2-100), and the output layer node is k (k=1～100).
6, by in the described p xylene oxidation reactor product of claim 1 to the flexible measurement method of carboxyl benzaldehyde content, it is characterized in that soft measured value carries out online " rolling " with the manual analysis value at set intervals and optimize to proofread and correct to soft-sensing model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB011135174A CN1139814C (en) | 2001-04-11 | 2001-04-11 | Soft measurement method for conten of p-carboxyl benzaldehyde in product generated by oxidizing reaction of p-xylene |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB011135174A CN1139814C (en) | 2001-04-11 | 2001-04-11 | Soft measurement method for conten of p-carboxyl benzaldehyde in product generated by oxidizing reaction of p-xylene |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1316647A CN1316647A (en) | 2001-10-10 |
| CN1139814C true CN1139814C (en) | 2004-02-25 |
Family
ID=4660238
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB011135174A Expired - Fee Related CN1139814C (en) | 2001-04-11 | 2001-04-11 | Soft measurement method for conten of p-carboxyl benzaldehyde in product generated by oxidizing reaction of p-xylene |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1139814C (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101173952B (en) * | 2007-10-25 | 2012-05-30 | 华东理工大学 | Soft measuring method for confirming 4-carboxybenzaldehyde content |
| CN106485094A (en) * | 2016-11-30 | 2017-03-08 | 华东理工大学 | A kind of PX oxidation reaction production process agent model modeling method |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100444182C (en) * | 2004-06-03 | 2008-12-17 | 华东理工大学 | Online soft measuring method for cobalt, manganese catalyst consumption during PX oxidation reaction process |
| CN100444181C (en) * | 2004-06-03 | 2008-12-17 | 华东理工大学 | Online soft measuring method for promoter consumption during PX oxidation reaction process of industrial equipment |
| CN100352912C (en) * | 2004-12-13 | 2007-12-05 | 东南大学 | Contrary flexible method in neural network for measuring procedure variable of fermenting erythromycin and structure of flexible instrument |
| CN101598737B (en) * | 2009-06-26 | 2014-04-23 | 华东理工大学 | Soft-measurement method for 4-carboxybenzaldehyde (4-CBA) in pure terephthalic acid (PTA) |
| CN101963785B (en) * | 2010-09-17 | 2012-06-27 | 南京工业大学 | On-line control method for oxidation mother liquor filter process in purified terephthalic acid production |
| CN104504232B (en) * | 2014-10-25 | 2018-02-13 | 南京邮电大学 | A kind of flexible measurement method of the 4 CBA contents based on Sparse least squares support vector machine |
| CN108227649A (en) * | 2017-12-13 | 2018-06-29 | 华东理工大学 | The method of industrial p xylene oxidation reaction member monitoring running state |
-
2001
- 2001-04-11 CN CNB011135174A patent/CN1139814C/en not_active Expired - Fee Related
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101173952B (en) * | 2007-10-25 | 2012-05-30 | 华东理工大学 | Soft measuring method for confirming 4-carboxybenzaldehyde content |
| CN106485094A (en) * | 2016-11-30 | 2017-03-08 | 华东理工大学 | A kind of PX oxidation reaction production process agent model modeling method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1316647A (en) | 2001-10-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1139814C (en) | Soft measurement method for conten of p-carboxyl benzaldehyde in product generated by oxidizing reaction of p-xylene | |
| Tong | The evaluation of fuzzy models derived from experimental data | |
| CN1693892A (en) | On-line measuring method for one section outlet acetylene concentration of acetylene hydrogenation reaction chamber | |
| CN102880905B (en) | Online soft sensor method done by a kind of normal top oil | |
| Xuefeng | Hybrid artificial neural network based on BP-PLSR and its application in development of soft sensors | |
| CN108088974B (en) | Soft measurement method for effluent nitrate nitrogen in anaerobic simultaneous denitrification methanogenesis process | |
| Joo et al. | The effects of data preprocessing in the determination of coagulant dosing rate | |
| CN112861350B (en) | Temperature overheating defect early warning method for stator winding of water-cooled steam turbine generator | |
| CN103245759A (en) | Product quality monitoring method based on autoregression total projection to latent structures (T-PLS) | |
| CN111142494B (en) | Intelligent control method and system for amine liquid regeneration device | |
| CN101995891B (en) | Method for online analysis of water content of solid master batch recovery system in aromatic acid production | |
| CN111783354A (en) | Flow-induced corrosion characteristic prediction and service life evaluation method of RBF neural network model | |
| De Magalhães et al. | Joint economic model for totally adaptive X and R charts | |
| CN108204997A (en) | Normal line oil flash point online soft sensor method | |
| WO2018177651A1 (en) | Method for determining the methane index of a hydrocarbon-containing combustion gas mixture | |
| CN113515893A (en) | Establishment method, prediction method and prediction device of real-time prediction model in rare earth extraction process | |
| CN101173952B (en) | Soft measuring method for confirming 4-carboxybenzaldehyde content | |
| CN1373118A (en) | Intelligent control method for reaction temp in liquid-phase oxidizing reactor of p-xylene | |
| CN1417192A (en) | Establishing method of dynamic model of liquid phase catalytic paraxylene oxidizing reaction | |
| CN101591240A (en) | The modeling method of crude terephthalic acid hydrofining reaction industrial device model | |
| CN113191614B (en) | Polypropylene product quality real-time monitoring method based on slow characteristic regression analysis | |
| WO2023012007A1 (en) | Method for monitoring and/or controlling a chemical plant using hybrid models | |
| CN1382672A (en) | Intelligent control method of propene rectifying tower in ethylene equipment | |
| CN102486632A (en) | On-line analyzing method of terephthalic acid crystal particle diameter in P-xylene oxidation process | |
| CN101251747B (en) | Modelling method for industrial device model for dimethylbenzene oxidation reaction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C06 | Publication | ||
| PB01 | Publication | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20040225 Termination date: 20170411 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |