CN102436252B - Process industry fault diagnosis method and system based on immune hazard theory - Google Patents
Process industry fault diagnosis method and system based on immune hazard theory Download PDFInfo
- Publication number
- CN102436252B CN102436252B CN201010299142.4A CN201010299142A CN102436252B CN 102436252 B CN102436252 B CN 102436252B CN 201010299142 A CN201010299142 A CN 201010299142A CN 102436252 B CN102436252 B CN 102436252B
- Authority
- CN
- China
- Prior art keywords
- analysis
- fault
- danger signal
- danger
- sdg
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 92
- 230000008569 process Effects 0.000 title claims abstract description 50
- 238000003745 diagnosis Methods 0.000 title claims abstract description 33
- 231100001261 hazardous Toxicity 0.000 claims abstract description 55
- 238000001514 detection method Methods 0.000 claims abstract description 11
- 238000012544 monitoring process Methods 0.000 claims abstract description 4
- 238000004458 analytical method Methods 0.000 claims description 38
- 239000000427 antigen Substances 0.000 claims description 19
- 102000036639 antigens Human genes 0.000 claims description 19
- 108091007433 antigens Proteins 0.000 claims description 19
- 239000000126 substance Substances 0.000 claims description 7
- 238000011156 evaluation Methods 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 235000011470 Adenanthera pavonina Nutrition 0.000 claims 4
- 240000001606 Adenanthera pavonina Species 0.000 claims 4
- 210000000987 immune system Anatomy 0.000 abstract description 14
- 230000007547 defect Effects 0.000 abstract description 3
- 238000004364 calculation method Methods 0.000 abstract description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 27
- 238000003756 stirring Methods 0.000 description 12
- 238000010606 normalization Methods 0.000 description 7
- 239000007788 liquid Substances 0.000 description 5
- 238000004519 manufacturing process Methods 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 5
- 230000036039 immunity Effects 0.000 description 4
- 101150067055 minC gene Proteins 0.000 description 4
- 210000004027 cell Anatomy 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 210000000612 antigen-presenting cell Anatomy 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 210000003719 b-lymphocyte Anatomy 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 229960000074 biopharmaceutical Drugs 0.000 description 1
- 238000001311 chemical methods and process Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 239000000498 cooling water Substances 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000013399 early diagnosis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 230000028993 immune response Effects 0.000 description 1
- 230000007233 immunological mechanism Effects 0.000 description 1
- 239000011344 liquid material Substances 0.000 description 1
- 210000004698 lymphocyte Anatomy 0.000 description 1
- 210000002540 macrophage Anatomy 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 238000011144 upstream manufacturing Methods 0.000 description 1
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Automatic Analysis And Handling Materials Therefor (AREA)
Abstract
The invention aims at using the characteristics of strong adaptability and self-learning capacity of an artificial immune system, overcoming the defects of long calculation time, high incorrect and missed report rate and the like of the existing artificial immune method, and applying an immune hazard theory to the fault diagnosis of a process industry. Compared with the existing artificial immune system, the invention provides a method for determining hazard signals and hazard rules and dividing hazardous areas, and additionally provides a fault diagnosis method based on the immune hazard theory. The process industry fault diagnosis method based on the immune hazard theory comprises a hazardous area dividing step, which is used for dividing the hazardous areas; an antibody generation step, which is used for generating antibodies according to history data in each hazardous area in the original fault samples; and a fault detection step, which is used for monitoring whether a process is faulted or not. The invention additionally provides a process industry fault diagnosis system based on the immune hazard theory.
Description
Invention field
The present invention relates to a kind of method for diagnosing faults and system based on immune danger theory, it is mainly used in chemical process, oil refining process, the fault diagnosis field of the process industries such as bio-pharmaceuticals process.
Background technology
A gordian technique of the stable operation of realization flow industry be exactly under different mode of operations to the unusual service condition occurred can carry out accurately, reliable early diagnosis or early warning.In recent years, along with scientific and technological development, in the modern production process, the system of application is increasingly sophisticated, and the equipment related to is increasingly specialized, for continuous large batch of modern production process, more is necessary to set up supervisory system, finds in time and diagnose the reason that is out of order.From nineteen seventies, the research of process failure diagnosis is just an emphasis of process industry research field always.Fault diagnosis not only can guarantee the stability of production run, the generation of prevention serious accident effectively, timely, can also after occurring, fault provide instruction, help works engineer to be processed, repair fault, effectively reduce the loss that fault causes.
Through years development, now main method for diagnosing faults roughly can be divided three classes from the diagnosis principle: the method based on quantitative model, the method based on qualutative model, the method based on historical data, typical method has neural network, least square method, wavelet analysis method, principal component analysis (PCA) etc.To the fault diagnosis of transient process, more common method is Multivariable Statistical Methods in recent years, its advantage is the generation that fast detecting is abnormal, but not high with the diagnosis accuracy of contribution plot analytical approach, and lack adaptive ability and the self-learning capability of diagnosing new fault.For this present situation, the artificial immune system that has the scholar to propose to have adaptivity and self-learning capability is incorporated into fault diagnosis field.
Artificial immune system is a kind of integrated intelligent system that relates to multidisciplinary field, and it organically combines immunology and engineering science, utilizes the technology such as mathematics, computing machine to set up the immunologic mechanism model, and is applied to the aspects such as design, enforcement of engineering.In recent years, artificial immunity is introduced in fault diagnosis field gradually.But due to traditional artificial immune system algorithm, because calculated amount is very large, the element related to is more, has long problem computing time, the wrong report likely occurred in process fault detection is failed to report.
For these problems, other method can be combined with artificial immune system on the one hand, set up the Hybrid fault diagnosis system of multi-method.On the other hand, artificial immune system itself is improved, the artificial immune system based on danger theory is exactly a kind of effective method wherein.
The immunity danger theory is for the immune many deficiencies of Traditional Man, a new immune model of proposition.This model thinks, immune system is not to the nonego reaction, but to hazardous reaction.After the cell that is upset produces alarm signal, this class antigen presenting cell of macrophage is by near antigen capture and offer to lymphocyte.This signal itself can be set a hazardous location around oneself, and only the B cell in this scope just likely produces the antibody of coupling antigen, experience immune response process.Do not mate or not the cell in hazardous location can not be upset.Danger theory, as a kind of brand-new theoretical model, no longer needs to distinguish the concept of oneself and nonego, thereby and only need judge whether to bring loss to produce danger signal to body cell.It provides the thought of more effective deal with data for artificial immune system, concentrate on dangerous, on significant data.
Summary of the invention
The object of the invention is to utilize the strong characteristics of artificial immune system adaptability and self-learning capability simultaneously, overcome the defects such as existing artificial immunity method long wrong report rate of failing to report computing time is higher, immune danger theory is applied to the fault diagnosis of process industry, and proposes the division methods of the definite and hazardous location of danger signal and dangerous rule.More existing artificial immune system, definite and division methods hazardous location of the danger signal that the present invention proposes and dangerous rule has represented a kind of brand-new method for diagnosing faults based on immune danger theory.
According to an aspect of the present invention, provide a kind of method for diagnosing faults based on immune danger theory, it is characterized in that comprising:
The accompanying drawing explanation
Fig. 1 has shown the process flow diagram of method according to an embodiment of the invention.
Fig. 2 has shown the process flow diagram of the risk zontation that method according to an embodiment of the invention adopts.
Fig. 3 has shown the secondary continuous stir reactor schematic flow sheet of method application according to an embodiment of the invention.
Fig. 4 has shown the SDG figure of the secondary continuous stir reactor flow process of method application according to an embodiment of the invention.
Embodiment
The object of the invention is to utilize the strong characteristics of artificial immune system adaptability and self-learning capability simultaneously, overcome the defects such as existing artificial immunity method long wrong report rate of failing to report computing time is higher, immune danger theory is applied to the fault diagnosis of process industry.More existing artificial immune system, the present invention proposes the division methods that reaches hazardous location of determining of danger signal and dangerous rule, and proposed a kind of method for diagnosing faults based on immune danger theory.
For the fault diagnosis of process industry, as shown in Figure 1, method according to an embodiment of the invention comprises:
-SDG and HAZOP analytical procedure
Immune danger theory diagnostic system according to an embodiment of the invention can directly carry out SDG (signed digraph) and HAZOP (hazard and operability analysis) analysis at step S101 to the original process flow process, at first by SDG, analyze, obtain the mutual relationship between the flow process variable, the hazard level of preliminary judgment variable, and obtain SDG figure.Analyze by HAZOP again, find the dangerous hidden danger existed in flow process, and analyze the judgement danger classes by calculating Dow Chemical index and LOPA (layer of protection analysis).In this step, HAZOP analyzes and can use other safe evaluation methods such as safety checklist method, fault tree analysis to replace.
Definite step of-danger signal and dangerous rule
The Dow Chemical index obtained according to step S101 and LOPA analysis result, can determine the dangerous hidden danger of technique and obtain the danger classes of variable at step S102, the variable that danger classes is higher is decided to be danger signal, and according to characteristics and the HAZOP analysis result of variable, determine dangerous rule, reach the variable requirement that danger signal must be satisfied, otherwise just think to have occurred fault.
-risk zontation step
Once S102 has determined danger signal through step, enter step S103, divide hazardous location according to the SDG analysis result centered by described danger signal.
According to a specific embodiment of the present invention as shown in Figure 2, this risk zontation step further comprises:
1 makes the neighbouring relations matrix by SDG figure.The numerical value of each row of each row is done and is added and obtain:
C
i=d
1i+d
2i+...d
ni,R
j=d
j1+d
j2+...+d
jn (1)
D wherein
ijfor the element of the capable j row of adjacency matrix i, C
ifor row element and, R
jfor column element and.
2.1 determine the hazardous location under each variable i.At first, make i=1.
2.2 judge whether danger signal of i, if i is the dangerous center using it as a hazardous location of danger signal, to 2.3.Otherwise, to 3.1, calculate the affiliated hazardous location of i, according to SDG figure, with D
irecord variable i calculates immediate hazardous location, SD downstream
irecord variable i is to the distance of calculating downstream center, immediate hazardous location.
2.3 make i=i+1, to the next point of 2.2 inspections.
3.1 whether judgment variable i has link variable downwards, if do not have, i.e. and R
i=0, explanatory variable i does not have the hazardous location connected downwards, makes D
i=0, SD
i=0, to 4.1 steps.Otherwise explanatory variable i is at least 1 to the distance of calculating downstream center, immediate hazardous location, makes SD
i=1, to 3.2 steps, calculate the hazardous location that variable i connects downwards.
3.2.1 judgment variable i has several link variables downwards, if only have link variable, i.e. a R
i=1, make i*=j, wherein j meets d
ij=1, walk to 3.2.2.Otherwise link variable is greater than 1, to 3.3.1.
3.2.2 judge whether i* is danger signal, if danger signal, i* is the center, hazardous location met the demands, and makes D
i=i*, to 4.1 steps.Otherwise, enter the 3.2.3 step.
3.2.3 judge whether that i forms loop to i*, or i* is downwards without link variable, if so, i.e. i*=i or R
i*=0, make D
i=0, to 4.1 steps.Otherwise, enter the 3.2.4 step.
3.2.4 judgment variable i* has several link variables downwards, if only have link variable, i.e. a R
i*=1, make i*=j, wherein j meets d
i*j=1, walk to 3.2.6.Otherwise the link variable number is greater than 1, walk to 3.2.5.
3.2.5 variable i * has a more than link variable downwards, makes j={j (1), j (2) ... .} meets d
i* (jk)=1.Make i*=j (k*), wherein k* meets C
j (k*)=minC
j.Be about to the link variable of linking number minimum as next variable, walk to 3.2.6.
3.2.6 obtain new link variable i*, the initial variable i distance of distance adds 1.Make SD
i=SD
i+ 1.To 3.2.2.
3.3.1R
i>1, establish j*={j (1), j (2) ..., meet d
i (jk)=1.Make t=1, consider the feasible hazardous location of each j (t), walk to 3.3.2.
3.3.2 make i*=j (t), the possibility of investigating hazardous location under this branch road walks to 3.3.3.
3.3.3 judge whether i* is danger signal, if danger signal, i* is the center, hazardous location met the demands, and makes D
i (t)=i*, walk to 3.3.8.Otherwise, enter the 3.3.4 step.
3.3.4 judge whether that i forms loop to i*, or i* is downwards without link variable, if so, i.e. i*=i or R
i*=0, make D
i (t)=0, walk to 3.3.8.Otherwise, enter the 3.3.5 step.
3.3.5 judgment variable i* has several link variables downwards, if only have link variable, i.e. a R
i*=1, make i*=j, wherein j meets d
i*j=1, walk to 3.3.7.Otherwise the link variable number is greater than 1, walk to 3.3.6.
3.3.6 variable i * has a more than link variable downwards, makes j={j (1), j (2) ..., meet d
i* (jk)=1.Make i*=j (k*), wherein k* meets C
j (k*)=minC
j.Be about to the link variable of linking number minimum as next variable, walk to 3.3.7.
3.3.7 obtain new link variable i*, the initial variable i distance of distance adds 1.Make SD
i (t)=SD
i (t)+ 1.To 3.3.3.
3.3.8 judge whether that all j (t) have calculated, if calculated, i.e. t=R
ito 4.1, otherwise consider next j (t), make t=t+1, walk to 3.3.2.
4.1 whether judgment variable i upwards has link variable, if do not have, i.e. and C
i=0, explanatory variable i does not upwards have the hazardous location connected, and makes U
i=0, SU
i=0, to 5 steps.Otherwise explanatory variable i is at least 1 to the distance of calculating downstream center, immediate hazardous location, makes SU
i=1, to 3.2 steps, calculate the hazardous location that variable i connects downwards.
4.2.1 judgment variable i has several link variables downwards, if only have link variable, i.e. a C
i=1, make i*=j, wherein j meets U
ij=1, walk to 4.2.2.Otherwise link variable is greater than 1, to 4.3.1.
4.2.2 judge whether i* is danger signal, if danger signal, i* is the center, hazardous location met the demands, and makes U
i=i*, to 5 steps.Otherwise, enter the 4.2.3 step.
4.2.3 judge whether that i forms loop to i*, or i* is downwards without link variable, if so, i.e. i*=i or C
i*=0, make U
i=0, to 5 steps.Otherwise, enter the 4.2.4 step.
4.2.4 judgment variable i* has several link variables downwards, if only have link variable, i.e. a C
i*=1, make i*=j, wherein j meets U
i*j=1, walk to 4.2.6.Otherwise the link variable number is greater than 1, walk to 4.2.5.
4.2.5 variable i * has a more than link variable downwards, makes j={j (1), j (2) ..., meet U
i* (jk)=1.Make i*=j (k*), wherein k* meets C
j (k*)=minC
j.Be about to the link variable of linking number minimum as next variable, walk to 4.2.6.
4.2.6 obtain new link variable i*, the initial variable i distance of distance adds 1.Make SU
i=SU
i+ 1.To 4.2.2.
4.3.1C
i>1, establish j*={j (1), j (2) ..., meet U
i (jk)=1.Make t=1, consider the feasible hazardous location of each j (t), walk to 4.3.2.
4.3.2 make i*=j (t), the possibility of investigating hazardous location under this branch road walks to 4.3.3.
4.3.3 judge whether i* is danger signal, if danger signal, i* is the center, hazardous location met the demands, and makes U
i (t)=i*, walk to 4.3.8.Otherwise, enter the 4.3.4 step.
4.3.4 judge whether that i forms loop to i*, or i* is downwards without link variable, if so, i.e. i*=i or C
i*=0, make U
i (t)=0, walk to 4.3.8.Otherwise, enter the 4.3.5 step.
4.3.5 judgment variable i* has several link variables downwards, if only have link variable, i.e. a C
i*=1, make i*=j, wherein j meets U
i*j=1, walk to 4.3.7.Otherwise the link variable number is greater than 1, walk to 4.3.6.
4.3.6 variable i * has a more than link variable downwards, makes j={j (1), j (2) ..., meet U
i* (jk)=1.Make i*=j (k*), wherein k* meets C
j (k*)=minC
j.Be about to the link variable of linking number minimum as next variable, walk to 4.3.7.
4.3.7 obtain new link variable i*, the initial variable i distance of distance adds 1.Make SU
i (t)=SU
i (t)+ 1.To 4.3.3.
4.3.8 judge whether that all j (t) have calculated, if calculated, i.e. t=C
ito 5, otherwise consider next j (t), make t=t+1, walk to 4.3.2.
5 compare i apart from center, hazardous location, upstream and hazardous location, downstream centre distance.If SD
i<SU
i, Z
i=D
i.If SD
i>SU
i, Z
i=U
i.If SD
i=SU
i, Z
i={ U
i, D
i.So far can determine the hazardous location that variable i is affiliated, to 6 steps.
Whether as calculated 6 check i complete all variablees, if check out to 2.3, under inspection a bit.If all variablees have all calculated, regional partiting step all finishes.
-data acquisition and processing
Immune danger theory diagnostic system according to an embodiment of the invention can directly be connected with production equipment, at step S201 (Fig. 1), by such as PLC (Programmable Logic Controller), DCS (Distributed Control System (DCS)) and LIMS (Laboratory Information Management System) etc. gather required flow process variable data.Wherein known historical data, for system initialization, is set up and is determined dangerous rule and produce original antibody, and real-time online data is for fault detection and diagnosis.All historical datas are introduced into step S202 and carry out normalized by formula (1):
Wherein x is real data,
x
maxwith X
minthe mean value of this variable data, maximal value and minimum value while being respectively nominal situation.
According to one embodiment of present invention, data acquisition and treatment step can be used as implements another preposition step of carrying out before method of the present invention.
-generation antibody library
For the primary fault sample obtained, at step S203, determine its failure mode and time after the normalization of S202 step, for the variable in each hazardous location, get fault and introduce the data of rear set time length, generate first generation antibody, be denoted as Ab_fault=[Ab
1, Ab
2..., Ab
n].The matrix that these antibody all are comprised of the seasonal effect in time series data sample, Ab
itime series for each variable in same hazardous location.Each fault sample will, according to different hazardous locations, generate several different first generation antibody like this.First generation antibody of the same type, through step S204, generates a large amount of second generation antibodies by formula (2) variation and clone, has formed antibody library.
The number that wherein n is known this kind first generation antibody, X
ifor known first generation antibody, X* is the second generation antibody generated through the variation clone.A
ibe the random decimal in 0 to 1, b is the random decimal between-1 to 1.C and d are the random integers between 1 to n.
Calculate again the threshold value of every kind of fault type antibody.
For two antibody (or antigen) of determining, can adopt the Dynamic Time Warping algorithm to calculate the diversity factor of antigen and antibody.At first get DTW for time series (Dynamic Time Warping) the algorithm calculated difference degree of identical variable, also can directly calculate by the straightforward procedure of formula (3):
The diversity factor of calculating all variablees just can obtain the total variances degree vector of antigen and antibody:
η=[η
1,η
2,…,η
n] (4)
Diversity factor in the calculating antibody storehouse between all antibody of the same type, note is η (i, j).The threshold value of calculating respectively each variable of the type antibody is:
-utilize danger signal to carry out fault detect
The real time data obtained by the S201 step, before carrying out fault diagnosis, is introduced into step S205, after carrying out normalization by formula (1), obtains real-time detection data x
new.Enter step S206, utilize danger signal to carry out fault detect, whether broken down by danger signal and dangerous rule monitoring flow process.
-utilize antibody library to carry out fault diagnosis
Once detect confirm to break down at step S206, process and enter step S207, the data of set time length before detecting are constantly generated respectively to different antigen according to the variable of different hazardous locations, be denoted as Ag=[Ag
1, Ag
2..., Ag
n].These antigens are identical with the form of the antibody of identical hazardous location in antibody library, the matrix all be comprised of the time series data sample of hazardous location internal variable, Ag
itime series for each variable.
After antigen generates, process the failure mode diagnosis that enters step S208, utilize DTW algorithm or formula (6) to calculate the diversity factor of antigen and all antibody of the interior identical hazardous location of antibody library.
When each hazardous location has antibody to calculate diversity factor to be less than corresponding threshold value, think that fault is consistent with the type antibody.Otherwise be judged as a kind of new fault.
The fault diagnosis result of refer step S208, the operator can carry out Artificial Diagnosis (S301), and is input to system, and system enters step S209, accepts the result of input as final diagnostic result.
-antibody library upgrades
At step S203, using antigen as first generation antibody.Enter again step S204, by formula (2), the type antibody is carried out to clonal vaviation, upgrade the antibody in antibody library, and upgraded the threshold value of this antibody by formula (5).
Embodiment:
The method according to this invention has been applied to the fault diagnosis of the reaction realistic model of secondary continuous stir reactor.
The flow process of secondary continuous stir reactor as shown in Figure 3, this flow process be by the hot water in uniform temperature (95 ℃) and cold water (40 ℃) in continuous stir reactor R-1, mix and be cooled to definite temperature (50 ℃) after enter continuous stir reactor R-2, the flow of cold water is fixing (60kg/h) relatively, and the flow of hot water is determined by the fluid level controller of R-1.Obtain determining the chilled water of flow (360kg/h) and temperature (30 ℃) after cooling, and by adding appropriate hot water (95 ℃) to regulate the liquid level in R-2.Two continuous stir reactors have the chilled water of band control to meet the liquid material temperature requirement in still.
Monitored parameters comprises temperature T r1 and the liquid level Lr1 in continuous stir reactor R-1, entrance cold water flow and temperature F101, T101, cold water flow and temperature F102, T102, rate of discharge and temperature F102, T102, the flow of chilled water and temperature F104, T104, the temperature T r2 of continuous stir reactor R-2 and liquid level Lr2, entrance hot water flow and temperature F201, T201, rate of discharge and temperature F102, T102, the flow of chilled water and temperature F104, T103.
The fundamental purpose of this flow process is exactly in order to obtain determining the water of flow and temperature.The control index of whole process main operating parameters and controller is as shown in table 1.
Table 1 nominal situation parameter arranges
Operating parameter | Numerical value | Control target | Numerical value |
The R-1 cold water flow | 60±5kg/h | The R-1 temperature | 50℃ |
Cold water temperature | 40±2℃ | The R-2 temperature | 30℃ |
The hot water temperature | 95±2℃ | The R-1 liquid level | 50% |
Cooling water temperature | 30℃ | The R-2 liquid level | 50% |
The R-2 outlet temperature | 360kg/h |
At first according to step S101, flow process is carried out to HAZOP analysis and SDG analysis.Analyze the dangerous hidden danger of this technological process of discovery and main dangerous matter sources through HAZOP.Analyze the relation obtained between SDG figure confirmation variable through SDG, as shown in Figure 4.
Enter again step S102, confirm danger signal and dangerous rule.According to Dow Chemical index result of calculation and LOPA analysis result, find Tr1, Tr2, the harmful grade of Lr2 is higher, and these three variablees are decided to be to danger signal.Because these three variablees are control variable, according to the HAZOP analysis result, the concrete numerical value of its variable and stability are all very crucial for flow process again.Therefore, these three danger signals are determined to dangerous rule is as follows:
Wherein
for the real time data value of danger signal, V
khistorical data value for danger signal.Subscript i and k mean the time sequence number of variable.V
minand V
maxbe respectively minimum value and the maximal value of danger signal historical data.
According to the SDG of secondary continuous stir reactor figure (seeing Fig. 4), determine three danger signal: Tr1, Tr2, Lr2, then by all variablees according to danger signal step as shown in Figure 2, be divided into three hazardous locations.Take T101 as example, at first make adjacency matrix as shown in table 2.
Table 2 secondary continuous stir reactor fault diagnosis test result
As calculated, the hazardous location 1 centered by Tr1, comprise T101, T102, T103, T104, F101, F102, F103, F104, Tr1; Hazardous location 2 centered by Tr2, comprise T103, T201, T202, T203, F103, F201, F203, Tr2; Hazardous location 3 centered by Lr2, comprise F101, F102, Lr1, F103, F201, F202, Lr1, Lr2.
When carrying out preliminary analysis and planning, obtain the detection data of these all variablees by step S201, sampling time interval is 5 seconds.
Bring the normal sample in the historical data after S202 walks normalization into formula (7), help the details about dangerous rule in step S102 to determine.Enter again step S202
In addition, too high cold water flow F102, the V-3 valve turns down, each 2 groups of data of the malfunctioning three kinds of faults of LC-2 controller, carry out normalization at step S202, then near step S203 gets the fault introducing time 10 sampling number certificates, according to the division of three hazardous locations, generate respectively three independently antibody, and carry out variation and clone's generation antibody library of antibody at step S204.
Completed like this initialization of fault diagnosis system, then real time data has been carried out to fault detection and diagnosis after step S205 normalization.This example has adopted altogether 13 groups of real time data samples to verify method.
At first take sample 1 as example, sample 1, for slightly improving the sample produced after cold water flow, after the normalization of S205 step, enters the S206 step and utilizes danger signal to carry out fault detect.The danger signal data that calculate all sampled points all meet dangerous rule, that is to say that this disturbance does not produce actual fault effects for the production of this technological process, so testing result think that this sample is nominal situation.And that the method for diagnosing faults that utilizes other is easy to that this sample is diagnosed as to cold water flow is too high and report to the police.
Take sample 2 as example again, the fault sample that sample 2 produces for significantly improving cold water flow, after the normalization of S205 step, enter the S206 step and utilize danger signal to carry out fault detect, after fault is introduced 15s, the danger signal data have exceeded the restriction range of dangerous rule, think fault has occurred, enter the S207 step by ten sampling number certificates before failure detection time, according to the division of hazardous location, generate 3 independently antigens.Walk by S208 again, utilize the diversity factor of antibody to calculate the tracing trouble kind, find that in 3 antigens and the too high antibody library of cold water flow, 3 kinds of Antibody difference ratios are less than the type fault threshold, and all be greater than its corresponding threshold value with the other types Antibody difference ratio.Therefore diagnostic result is the too high fault of cold water flow, confirms that through step S209 the Artificial Diagnosis result is cold water flow too high, enters the S203 step, by antigen, generates new antibodies.Walk by S204 again, carry out clone and the variation of antibody, upgrade antibody library.
The all fault diagnosis results of this example are as shown in table 3.
Table 3 secondary continuous stir reactor fault diagnosis test result
Sequence number | Fault category | Detect time delay | Diagnostic | Computing time | |
1 | Normal * | - | |
||
2 | 1. cold water flow is too | 15s | Fault | 1 | 0.2458 |
3 | 1. cold water flow is too | 15s | Fault | 1 | 0.1640 |
4 | 1. cold water flow is too | 15s | Fault | 1 | 0.1575s |
5 | 2.V-3 valve turns down | | Fault | 2 | 0.1667 |
6 | 2.V-3 valve turns down | | Fault | 2 | 0.1626s |
7 | 2.V-3 valve turns down | | Fault | 2 | 0.1595 |
8 | 3.LC-2 controller is malfunctioning | | Fault | 3 | 1.1324s |
9 | 3.LC-2 controller is malfunctioning | | Fault | 3 | 0.2985s |
10 | 3.LC-2 controller is malfunctioning | | Fault | 3 | 0.3440s |
11 | 4.TC-2 controller is malfunctioning | 255s | New fault | 0.7584s | |
12 | 4.TC-2 controller is malfunctioning | | Fault | 4 | 0.3105s |
13 | 4.TC-2 controller is malfunctioning | | Fault | 4 | 0.4054s |
* this normal sample, for slightly improving the sample produced after cold water flow, does not destroy dangerous rule.
Should be understood that, the description of in above narration and explanation, the present invention being carried out just illustrates but not is determinate, and do not breaking away under the prerequisite of the present invention limited as appended claims, can carry out various changes, distortion and/or revise above-described embodiment.
Claims (4)
1. the method for diagnosing faults of the process industry based on immune danger theory is characterized in that comprising:
The risk zontation step, for dividing hazardous location;
Antibody generates step, and the historical data in the primary fault sample in each described hazardous location generates antibody;
Whether failure detection steps, break down for monitoring flow process;
SDG and Safety Evaluation Analysis step, for carrying out corresponding analysis to the original process flow process;
Danger signal and dangerous regular determining step, for being determined danger signal and dangerous rule by described SDG and HAZOP analysis result;
Antigen generates step, in the situation that described failure detection steps is determined with fault, exists, and by real time data in each hazardous location of schedule time length before the detection constantly of described failure detection steps, generates respectively a plurality of antigen;
The failure mode diagnosis algorithm, utilize described antigen and described antibody, judges the kind of fault;
Data acquisition and treatment step, gather required flow process variable data, and described flow process variable data comprises historical data and described real time data, and described historical data comprises original normal sample and described primary fault sample,
Described danger signal and dangerous regular determining step comprise:
Thereby the hazard level of the judgment variable as a result of analyzing with Dow Chemical index and LOPA is determined danger signal,
Determine the danger rule of described danger signal with described HAZOP analysis result, and
Divide danger signal with described SDG and HAZOP analysis result,
Wherein said risk zontation step further comprises: centered by described danger signal, according to the SDG analysis result, divide described hazardous location.
2. according to the method for diagnosing faults of the process industry based on immune danger theory of claim 1, it is characterized in that
Described SDG and Safety Evaluation Analysis step comprise:
At least one analysis of carrying out that SDG analyzes and analyzing, select analysis based on the safety checklist method and the analysis based on fault tree analysis from HAZOP,
Calculate the Dow Chemical index, and
Carry out the LOPA analysis.
3. the fault diagnosis system of the process industry based on immune danger theory is characterized in that comprising:
The risk zontation device, for dividing hazardous location;
The antibody generating apparatus, the historical data in the primary fault sample in each described hazardous location generates antibody;
Whether failure detector, break down for monitoring flow process;
The Safety Evaluation Analysis device, for carrying out SDG and reaching Safety Evaluation Analysis to the original process flow process;
Danger signal and dangerous rule are determined device, for determined danger signal and dangerous rule by described SDG and HAZOP analysis result;
The antigen generating apparatus, exist in the situation that described failure detector is determined with fault, by real time data in each hazardous location of schedule time length before the detection constantly of described failure detector, generates respectively a plurality of antigen;
The failure mode diagnostic device, utilize described antigen and described antibody, judges the kind of fault;
Data acquisition and treating apparatus, gather required flow process variable data, and described flow process variable data comprises historical data and described real time data, and described historical data comprises original normal sample and described primary fault sample,
Described danger signal and dangerous rule determine that device comprises:
Thereby the hazard level of the judgment variable as a result of analyzing with Dow Chemical index and LOPA is determined the part of danger signal,
Determine the part of the danger rule of described danger signal with described HAZOP analysis result, and
Divide the part of danger signal with described SDG and HAZOP analysis result,
Wherein said risk zontation device further comprises: centered by described danger signal, according to the SDG analysis result, divide described hazardous location.
4. according to the fault diagnosis system of the process industry based on immune danger theory of claim 3, it is characterized in that
Described Safety Evaluation Analysis device comprises:
For carrying out that SDG analyzes and from the part of at least one analysis that HAZOP analyzes, analysis based on the safety checklist method and the analysis based on fault tree analysis are selected,
For calculating the part of Dow Chemical index, and
Carry out the part of LOPA analysis.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201010299142.4A CN102436252B (en) | 2010-09-29 | 2010-09-29 | Process industry fault diagnosis method and system based on immune hazard theory |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201010299142.4A CN102436252B (en) | 2010-09-29 | 2010-09-29 | Process industry fault diagnosis method and system based on immune hazard theory |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102436252A CN102436252A (en) | 2012-05-02 |
CN102436252B true CN102436252B (en) | 2014-01-01 |
Family
ID=45984354
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201010299142.4A Active CN102436252B (en) | 2010-09-29 | 2010-09-29 | Process industry fault diagnosis method and system based on immune hazard theory |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102436252B (en) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102760208B (en) * | 2012-07-03 | 2015-07-29 | 清华大学 | Based on the Dynamic artificial immune method for diagnosing faults of simulating vaccine |
CN103745005A (en) * | 2014-01-24 | 2014-04-23 | 清华大学 | Dynamic artificial immunity fault diagnosis method based on vaccine transplantation |
CN104238545B (en) * | 2014-07-10 | 2017-02-01 | 中国石油大学(北京) | Fault diagnosis and pre-warning system in oil refining production process and establishment method thereof |
CN105787555B (en) * | 2016-02-25 | 2018-06-29 | 湖北第二师范学院 | Abnormal learning behavior based on artificial immunity danger theory finds method |
CN110148072B (en) * | 2018-02-12 | 2023-05-02 | 庄龙飞 | Sport course scoring method and system |
CN112859725B (en) * | 2021-01-04 | 2022-06-24 | 中国核电工程有限公司 | Monitoring system for status characteristic signals of plant equipment |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1866267A (en) * | 2006-06-14 | 2006-11-22 | 龚涛 | Artificial immune system based on normal model |
CN101458751A (en) * | 2009-01-06 | 2009-06-17 | 华中科技大学 | Storage abnormal detecting method based on artificial immunity |
CN101561880A (en) * | 2009-05-11 | 2009-10-21 | 华北电力大学(保定) | Pattern recognition method based on immune antibody network |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8015128B2 (en) * | 2006-10-01 | 2011-09-06 | International Business Machines Corporation | Biometric security using neuroplastic fidelity |
-
2010
- 2010-09-29 CN CN201010299142.4A patent/CN102436252B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1866267A (en) * | 2006-06-14 | 2006-11-22 | 龚涛 | Artificial immune system based on normal model |
CN101458751A (en) * | 2009-01-06 | 2009-06-17 | 华中科技大学 | Storage abnormal detecting method based on artificial immunity |
CN101561880A (en) * | 2009-05-11 | 2009-10-21 | 华北电力大学(保定) | Pattern recognition method based on immune antibody network |
Non-Patent Citations (4)
Title |
---|
化工过程混合故障诊断系统的应用研究;戴一阳等;《2009中国过程系统工程年会暨中国mes年会论文集》;20090930;第129页第1段-第132页第1段 * |
基于SDG-HAZOP的系统风险定量分析研究;孟令慧;《中国优秀硕士学位论文全文数据库-信息科技辑》;20081115(第11期);正文第2页第1段-第61页第1段 * |
孟令慧.基于SDG-HAZOP的系统风险定量分析研究.《中国优秀硕士学位论文全文数据库-信息科技辑》.2008,(第11期),正文第2页第1段-第61页第1段. |
戴一阳等.化工过程混合故障诊断系统的应用研究.《2009中国过程系统工程年会暨中国mes年会论文集》.2009,第129页-第132页. |
Also Published As
Publication number | Publication date |
---|---|
CN102436252A (en) | 2012-05-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102436252B (en) | Process industry fault diagnosis method and system based on immune hazard theory | |
Shu et al. | Abnormal situation management: Challenges and opportunities in the big data era | |
Bhatia et al. | Dynamic risk-based inspection methodology | |
CN102682560B (en) | Device for assessing level of fire interlock alarming in ship cabin | |
CN104573850A (en) | Method for evaluating state of thermal power plant equipment | |
CN102110187B (en) | Method and system for diagnosing mixed failure based on PCA and artificial immune system | |
CN104792468B (en) | A kind of oil refining apparatus gas detecting and alarming instrument optimization placement method | |
BahooToroody et al. | A condition monitoring based signal filtering approach for dynamic time dependent safety assessment of natural gas distribution process | |
WO2019211288A1 (en) | A method and system for discovering and visualizing potential operational problems of processes running in equipment and systems in an installation | |
Kumari et al. | An integrated risk prediction model for corrosion-induced pipeline incidents using artificial neural network and Bayesian analysis | |
CN103942735A (en) | Method for evaluating relay protection states | |
Raoni et al. | Employing process simulation for hazardous process deviation identification and analysis | |
CN115099360A (en) | Intelligent sewage treatment equipment fault assessment method and system | |
Meng et al. | Alarm management optimization in chemical installations based on adapted HAZOP reports | |
Chen et al. | Discovery of potential risks for the gas transmission station using monitoring data and the OOBN method | |
Wang et al. | Improved methods of online monitoring and prediction in condensate and feed water system of nuclear power plant | |
Wang et al. | An integrated methodology for assessing accident probability of natural gas distribution station with data uncertainty | |
Deng et al. | Misoperation monitoring and early warning during startup and shutdown of petrochemical units | |
CN104460337B (en) | A kind of control system analysis of common cause failure method based on the amendment β factors | |
CN110348148A (en) | A kind of Key experiments process recognition methods of Kernel-based methods FMEA | |
CN107784148A (en) | A kind of gathering line crash rate appraisal procedure and device | |
Milazzo et al. | The influence of risk prevention measures on the frequency of failure of piping | |
Satur et al. | Cost effective alternative in meeting the required Safety Integrity Level (SIL) | |
CN103337000A (en) | Safety monitoring and prewarning method for oil-gas gathering and transferring system | |
Wang et al. | Leakage risk quantitative calculation model and its application for anaerobic reactor |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |