CN102890495A - Complete plant diagnosis methods and complete plant diagnosis apparatus - Google Patents
Complete plant diagnosis methods and complete plant diagnosis apparatus Download PDFInfo
- Publication number
- CN102890495A CN102890495A CN2012102458674A CN201210245867A CN102890495A CN 102890495 A CN102890495 A CN 102890495A CN 2012102458674 A CN2012102458674 A CN 2012102458674A CN 201210245867 A CN201210245867 A CN 201210245867A CN 102890495 A CN102890495 A CN 102890495A
- Authority
- CN
- China
- Prior art keywords
- model
- classification
- input
- input variable
- equipments
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 103
- 238000003745 diagnosis Methods 0.000 title abstract description 48
- 230000008569 process Effects 0.000 claims abstract description 90
- 238000012423 maintenance Methods 0.000 claims abstract description 8
- 230000008859 change Effects 0.000 claims description 22
- 238000002405 diagnostic procedure Methods 0.000 claims description 18
- 238000012545 processing Methods 0.000 claims description 18
- 238000010606 normalization Methods 0.000 claims description 7
- 238000012986 modification Methods 0.000 claims description 4
- 230000004048 modification Effects 0.000 claims description 4
- 238000012217 deletion Methods 0.000 abstract description 11
- 230000037430 deletion Effects 0.000 abstract description 11
- 230000005856 abnormality Effects 0.000 abstract description 4
- 230000006870 function Effects 0.000 abstract description 3
- 239000003245 coal Substances 0.000 description 12
- 238000010586 diagram Methods 0.000 description 12
- 230000009471 action Effects 0.000 description 8
- 230000015572 biosynthetic process Effects 0.000 description 7
- 238000012544 monitoring process Methods 0.000 description 7
- 238000010248 power generation Methods 0.000 description 7
- 230000005611 electricity Effects 0.000 description 6
- 230000004087 circulation Effects 0.000 description 5
- 230000000875 corresponding effect Effects 0.000 description 5
- 239000007789 gas Substances 0.000 description 5
- 239000000523 sample Substances 0.000 description 5
- 238000012937 correction Methods 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000013179 statistical model Methods 0.000 description 2
- CVOFKRWYWCSDMA-UHFFFAOYSA-N 2-chloro-n-(2,6-diethylphenyl)-n-(methoxymethyl)acetamide;2,6-dinitro-n,n-dipropyl-4-(trifluoromethyl)aniline Chemical compound CCC1=CC=CC(CC)=C1N(COC)C(=O)CCl.CCCN(CCC)C1=C([N+]([O-])=O)C=C(C(F)(F)F)C=C1[N+]([O-])=O CVOFKRWYWCSDMA-UHFFFAOYSA-N 0.000 description 1
- 102100024405 GPI-linked NAD(P)(+)-arginine ADP-ribosyltransferase 1 Human genes 0.000 description 1
- 101000981252 Homo sapiens GPI-linked NAD(P)(+)-arginine ADP-ribosyltransferase 1 Proteins 0.000 description 1
- 241001272996 Polyphylla fullo Species 0.000 description 1
- 238000010795 Steam Flooding Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000005284 basis set Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 239000012467 final product Substances 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 239000000178 monomer Substances 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000000779 smoke Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000013024 troubleshooting Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Landscapes
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention provides complete plant diagnosis methods and a complete plant diagnosis apparatus. In a case of using a process signal obtained from a complete plant for abnormality or omen diagnosis, if the plant to be diagnosed is replaced or in maintenance, diagnosis results which has been accumulated previously or diagnosis functions which has been constructed in a diagnosis model can not be used, and the diagnosis model is required to be reconstructed. In the complete plant diagnosis method of the invention, a model in which correlations between input variables are modeled is maintained, input data is divided into a plurality of classes based on the correlations between the input variables, and the abnormality of the complete plant is detected based on the classes not belonging to normal classes after the classification, wherein if the complete plant is modified and the model is changed by the addition or deletion of the input variables, a new model is constructed by changing the numbers of the classes of the model.
Description
Technical field
The present invention relates in the abnormity diagnosis of the set of equipments that is consisted of by a plurality of equipment (plant), even newly append in the maintenance after starting operation or also do not make up again during process that deletion is relevant with abnormity diagnosis, can continue diagnostic method and the device of the abnormality diagnostic set of equipments of set of equipments.
Background technology
About take firepower or atomic power device in the generating set of equipments or the industry set of equipments take pharmaceuticals/food/chemical set of equipments as representative of representative, in order to be embodied as the stable utilization of complete equipment, with a plurality of process signals as monitored object.
Particularly, in order to grasp the state of set of equipments, will be arranged at each one for the measuring appliance of gaging pressure, temperature, flow, water level etc., and the process signal value demonstration that obtains will be offered the operations staff.In addition, in nearly all set of equipments, the viewpoint of anomaly-based or trouble shooting or maintenance is kept at the process signal value that obtains in the process computer (process computer) as special purpose computer.
The operations staff exists in the state of set of equipments in the situation about changing, and for whether the value of the process signal confirming to be associated also exists changes, and shows the value of this process signal at monitoring image.In addition, if be necessary, the value in past of this process signal of storing in the process computer is also shown on the monitoring image.Usually, operating control device and the monitoring arrangement of set of equipments become one, and can be implemented as in real time supervision or the control of complete equipment state.
In the prior art, when the set of equipments state changed, the operations staff was presented on the monitoring image value of the process signal related with monitoring arrangement.According to the cycle that is taken into of monitoring arrangement, upgrade online the value of shown process signal.Grasp the running status of set of equipments as the basis take the state of this process values.Operating control device or monitoring arrangement (below, be called " supervision control device ") carry out interlock with alarm device based on the state of process values.For example, when certain force value is higher than setting value, is used in the stand by lamp that the operations staff is notified and lights a lamp.And starting is used for the control from unusual recovering state to normal condition, operates to impel the operations staff.
Recently, investigation before abnormality produces, device or its method that the unusual sign of set of equipments is predicted.In addition, also have after unusual the generation, its reason is carried out servicing unit or its method specific and the analysis reason.
In the system of patent documentation 1, utilize with the corresponding alert level of level and carry out the diagnosis of set of equipments.
In patent documentation 2, disclose and possessed for transmission operation and control information, field data, the network of facility information and device and the method that shows the information terminal device of these information.
Patent documentation 1:JP JP 2005-258649 communique
Patent documentation 2:JP JP 2011-70334 communique
Non-patent literature 1:G.A.Carpenter and S.Grossberg: " ART2 Self-Organization of stable category recognition codes for analog input patterns ", Applied Optics, Vol.26, No.23, (1987)
Set of equipments is made of a plurality of equipment.In addition, the scale of each equipment is also not of uniform size.In diagnosis during set of equipments, according to scale or function, take the process signal that is associated as the basis, uses wherein utilize with statistical processing or neural network more as the such situation of the constructed statistical model such as the study of representative.Statistical model carries out modelling with the correlationship of the process signal inputted.
Patent documentation 1 and patent documentation 2 described methods all are to have utilized the diagnostic method of the set of equipments of self-elevating platform ART (Adaptive Resonance Theory, ART).ART be will input Data classification become sorter a kind of of a plurality of classification.
At this, input to ART with not containing unusual set of equipments process signal, be categorized into a plurality of classification.In the diagnosis when unusual the generation, by process signal is input to ART, when generating the new classification (category) that does not belong to existing classification, owing to produced the set of equipments state that does not have so far, so give a warning.
, generating set of equipments etc. may append new equipment on the way owing to use year number longer, may chase after and establish or remove measuring appliance.In the diagnosis that has utilized ART, because maintenance as described above, the process signal of inputting changes or when increasing and decreasing, can not directly diagnose, and need to again implement the classification classification based on ART.But, when using year number elongated, since huge to this time series data amount for this process signal of accumulating, so be difficult to make up again.
Summary of the invention
The object of the present invention is to provide a kind of diagnostic method and device that solves the set of equipments of above-mentioned problem.
In order to solve aforementioned problems, in the diagnostic method of set of equipments of the present invention, maintenance has been carried out the model after the modelling with the correlationship of input variable, correlationship according to input variable, be a plurality of classification with the Data classification of inputting, and survey the unusual of set of equipments according to the sorted generation that does not belong to the classification of normal classification, wherein, produced appending of input variable or deleted the model of such change about the modification according to set of equipments, the change of numbering by the classification that has utilized this model makes up new model.
In addition, existing model is being appended in the situation of input variable, only make up diagnostic model with the input variable of appending, number to generate the model of new classification numbering based on the classification numbering that obtains from existing model and the classification that obtains from the diagnostic model that appends.
In addition, in the situation of deletion input variable from existing model, to be made as fixed value by the value of the input variable of deleting in the input variable of normalization and input, and utilize the time series data of each input variable in the classification numbering of this model to come the change of execution model.
In addition, in the situation of deletion input variable from existing model, to be made as fixed value by the value of the input variable of deleting in the input variable of also inputting after the normalization, and the time series data of utilizing each input variable in the classification numbering of this model comes the change of execution model, and append and delete processing after processing finishes.
For solving aforementioned problems, in the diagnostic method of set of equipments of the present invention, the correlationship of the input variable that will obtain from monitored object is carried out modelling, the Data classification that is input to model is become a plurality of classification, and survey the unusual of set of equipments according to the sorted generation that does not belong to the classification of normal classification, wherein, the correlationship of the 1st input variable is carried out modelling and keep as feature model, and according to the correlationship of the 1st input variable the Data classification of inputting is become a plurality of classification, and the correlationship of the 2nd input variable of appending is also carried out modelling, the Data classification of inputting is become a plurality of classification, form the unified model based on the 2nd input variable after feature model and the institute's modelling.
In addition, unified model comprises by the determined new classification of combination with the classification of the classification of feature model and the 2nd input variable after institute's modelling.
For solving aforementioned problems, the diagnostic device of set of equipments of the present invention is according to from the control signal of the supervision control device that monitored object is controlled or the measuring-signal that obtains from the process computer of the process signal of input monitored object, survey the unusual of monitored object, the diagnostic device of this set of equipments possesses: feature model section, its correlationship with the signal of input is carried out modelling and is made feature model, the Data classification that will be input in the feature model according to the correlationship of input signal becomes a plurality of classification, and surveys the unusual of set of equipments according to the sorted generation that does not belong to the classification of normal classification; And unified model section, it possesses the unified model that comprises at least the feature model more than 1, the Data classification that will be input to unified model according to the correlationship of input signal becomes a plurality of classification, and survey the unusual of set of equipments according to the sorted generation that does not belong to the classification of normal classification, produced the feature model that appends of input signal about the modification according to control object, only make up diagnostic model with the input signal that appends, number to generate new classification numbering based on the classification numbering that obtains from existing feature model and the classification that obtains from the diagnostic model that appends, and then will generate that new classification is numbered and the model that obtains is stored in unified model section as unified model.
In addition, in the situation of deletion input variable from existing model, to be made as fixed value by the value of the input variable of deleting in the input variable of normalization and input, and utilize the time series data of each input variable in the classification numbering of this model to come the change of execution model.
The invention effect
In the diagnostic method and device of set of equipments of the present invention, even the process signal that should monitor or confirm as the input signal of diagnostic model change occurs or the situation of appending/deleting under, can be constantly with diagnostic result or tendency display reminding to operations staff or maintenance person, the stable utilization of can be conducive to generate electricity set of equipments or industry set of equipments.
Description of drawings
Fig. 1 is be applied to the to generate electricity figure of set of equipments of the diagnostic device with set of equipments of the present invention.
Fig. 2 is that expression is as the system chart of the formation of the thermal power generation complete equipment of diagnosis object.
Fig. 3 is pipe arrangement section in the thermal power generation complete equipment and the enlarged drawing of air heater section.
Fig. 4 is the figure of the data storage form stored in the process computer 300 of expression.
Fig. 5 is the figure of an example of institute's canned data in the representation model information database 450.
Fig. 6 is the figure of an example of institute's canned data in the representation model information database 450.
Fig. 7 is the figure of an example of the information of the diagnostic result stored in the expression diagnostic result database 480.
Fig. 8 is the process flow diagram of contents processing of prime model preparing department 420 of indicating.
Fig. 9 is the process flow diagram of the contents processing of expression unified model preparing department 430.
Figure 10 is the process flow diagram of the contents processing of expression key element diagnosis section 460 and comprehensive diagnos section 470.
Figure 11 is the figure of the shown initial picture of presentation video display device.
Figure 12 is the figure that the shown diagnostic model of presentation video display device is set picture.
Figure 13 is the figure of the tendency chart of the shown process signal of presentation video display device.
Figure 14 is the figure of the shown diagnostic result display setting picture of presentation video display device.
Figure 15 is the figure of the display case of the shown diagnostic result of presentation video display device.
Label declaration
100: the generating set of equipments
200: monitor control device
300: process computer
400: the set of equipments diagnostic device
410: outer input interface
420: feature model preparing department
430: unified model preparing department
450: the model information database
460: key element diagnosis section
470: comprehensive diagnos section
480: the diagnostic result database
490: outside output interface
900: input media
901: keyboard
902: mouse
910: aid
920: outer input interface
930: data send and receive handling part
940: outside output interface
950: image display device
Embodiment
Below, with reference to accompanying drawing, diagnostic method and the device of the set of equipments of preferred forms described.
Embodiment
Fig. 1 is that the diagnostic device of the explanation set of equipments that present embodiment is related is applied to the figure as the example in the generating set of equipments 100 of a certain object.In generating set of equipments 100, be provided with a plurality of measuring appliances for the state of grasping set of equipments.Via industrial siding or general transmission line, will be transferred to by the value of the measured process signal 10 of each measuring appliance and monitor control device 200 and the process computer 300 that is used for store measurement values.
Monitor the control signal 20 that be used for set of equipments operation remained the state of expectation of control device 200 outputs take the value of process signal 10 as the basis.The control signal 20 of output is imported into generating set of equipments 100, and also is input to set of equipments diagnostic device 400.
In process computer 300, accumulate from the value of the process signal 10 that obtains of generating set of equipments 100.The process signal 10 of accumulating is exported to unusual/set of equipments diagnostic device 400 according to its purposes as process signal 30.About storage format, be elaborated with reference to Fig. 4 thereafter.
Set of equipments diagnostic device 400 will be diagnosed required control signal 20 or process signal 30 to be situated between and will be taken into by outer input interface 410.On the other hand, set of equipments diagnostic device 400 is situated between and is connected with aid 910 by outside output interface 490, outer input interface 410, inputs user's operation signal, for example input operator's operation signal and the information of necessity is presented at display device 950.
The outer input interface 410 of set of equipments diagnostic device 400 is according to the instruction from aid 910, and switching model makes up pattern and diagnostic mode.When being the model construction pattern, control signal 20 or process signal 30 that feature model preparing department 420 is inputted, when diagnostic mode, control signal 20 or process signal 30 that key element diagnosis section 460 is inputted.
In the feature model preparing department 420, make and the corresponding ART model of diagnosis object.Having in the situation of a plurality of diagnosis objects, making the ART model of equal number.When making the ART model, the process signal 30 of normal condition is inputed to the ART model.In the ART model, according to the correlationship of the process signal 30 of inputting the input data are classified.Be referred to as classification (category).The resolution of this classification is subjected to about the size institute of alarm parameters.About the suitable establishing method of this alarm parameters, in aforesaid patent documentation 1 or patent documentation 2, propose.In addition, about the detailed action of ART model, it is put down in writing in non-patent literature 1 in detail, so detailed.
The alarm parameters of each ART model or classification number etc. are stored in the model information database 450.In addition, as required, information extraction from model information database 450.About the detailed action in the feature model preparing department 420 or the formation of model information database 450, will be in rear detailed description.The configuration example of model information database 450 such as Fig. 5, shown in Figure 6.
Next, in unified model preparing department 430, making is namely sorted out numbering as the new ART model of input data with the output of the ART model of feature model preparing department 420 interior making.With feature model preparing department 420 in the same manner, with the alarm parameters of ART model or sort out numbering etc., output to model information database 450 and outside output interface 490.In addition, as required, information extraction from model information database 450.About the detailed action in the unified model preparing department 430 or the formation of model information database 450, will be in rear detailed description.
Accepted by outer input interface 410 in the situation of instruction of diagnostic mode, to key element diagnosis section 460 input data.In key element diagnosis section 460, from model information database 450, load the information of diagnostic model.To the diagnostic model that is loaded, input input data are diagnosed.According to the data that are input to diagnostic model, be categorized as the classification of in advance making or make new classification based on the correlationship of new input data.
In the situation of having made new classification, mean to detect the state different from normal condition, be diagnosed as unusual omen.The ground such as output classification numbering that comprise this moment store in the diagnostic result database 480.In addition, also export comprehensive diagnos section 470 to.About the detailed action in the key element diagnosis section 460 or the formation of diagnostic result database 480, will be in rear detailed description.The configuration example of diagnostic result database 480 as shown in Figure 7.
In comprehensive diagnos section 470, from model information database 450, load the information of diagnostic model, will input to diagnostic model from the output of key element diagnosis section 460, diagnose in the same manner.Diagnostic result is stored in outside output interface 490 and the diagnostic result database 480.In addition, about the detailed action in the comprehensive diagnos section 470, will be in rear detailed description.
As the user relevant with generating set of equipments 100, for example the operator can see the various information that generating set of equipments 100 is correlated with by utilizing the input media 900 that is made of keyboard 901 and mouse 902 and the aid 910 that is connected with image display device 950.In addition, can access from the control signal 20 that monitors control device 200, diagnostic result, the model information database 450 from the process signal 30 of process computer 300, set of equipments diagnosis section 400, the information of diagnostic result database 480.
Aid 910 by outer input interface 920, data send and receive handling part 930, outside output interface 940 consists of.
Input signal 91 Jie that generate at input media 900 are taken in the aid 910 by outer input interface 920.In addition, in the aid 910, about from the control signal 20 that monitors control device 200, from the process signal 30 of process computer 300, diagnostic result 40, the model information database 450 from set of equipments diagnostic device 400, the information of diagnostic result database 480, also be taken into by outer input interface 920 in the same manner.Data send and receive in the handling part 930, process input signal 92 according to the information from user's input signal 91, and send to outside output interface 940 as output signal 93.Output signal 94 shows at image display device 950.
In the following description, take the situation that data processing equipment of the present invention is applied to thermal power generation complete equipment as example, describe about the information of preserving in the database and the processing capacity of signal.
Fig. 2 is that expression is as the system chart of the formation of the thermal power generation complete equipment of the object of diagnosis.In this example, the structure of the generating in the thermal power generation complete equipment of coal combustion is described.
In the situation take coal as fuel, provide coal from coal-hole 111 Jie that store coal by 112 pairs of mullers 110 of coal supply device.In muller 110, by the transfer roller of inside, coal grinding is thinned down to the fine coal shape.Jie offers boiler 101 by burner 102 with 1 air of this fine coal and coal transmission usefulness and 2 air of the adjustment usefulness of burning.Fine coal and 1 air are led to 101,2 air of boiler from pipe arrangement 134 and are led to boiler 101 from pipe arrangement 141.In addition, the rear wind (after air) of 2 grades of burning usefulness is put in the boiler 101 by rear air port 103 by pipe arrangement 142 Jie.
After the high-temperature gas that produces by burning of coal flows along the path of boiler 101, by air heater 104.Pump-down process after be situated between by smoke stack emission to atmosphere thereafter.
On the other hand, the feedwater of boiler 101 interior circulations is situated between and is directed to boiler 101 by feed pump 105, is heated by gas in heat exchanger 106, becomes the steam of High Temperature High Pressure.In addition, the number with heat exchanger in the present embodiment is made as 1, but also can dispose a plurality of heat exchangers.
The steam that has passed through the High Temperature High Pressure behind the heat exchanger 106 is situated between and is directed to steam turbine 108 by turbine variable valve 107.The energy that has by steam drives steam turbine 108, generates electricity by generator 109.The electric power that generating obtains offers electric system.
The exhaust of steam turbine 108 is sent to feed pump 105 again after condenser 113 is cooled.In the way, utilize the steam that extracts from turbine, the device that configuration is heated feedwater improves the thermal efficiency.
As above dispose various measuring appliances in the thermal power generation complete equipment of such general formation, be transferred to the supervision control device 200 of Fig. 1 etc. from the obtained information of this measuring appliance as metrical information 10 (process signal 10).For example, in Fig. 2, illustrate flow measuring probe 150, temperature meter 151, pressometer 152, generating output checker 153 and measurement of concetration device 154.
In addition, 150 pairs of flow measuring probes are measured from the flow that feed pump 105 offers the feedwater of boiler 101.In addition, temperature meter 151, pressometer 152 are measured temperature, the pressure of the steam that offers steam turbine 108 respectively.Measure by generating output checker 153 at the generate electricity electric energy that obtains of generator 109.With contained composition (CO, the NO of the gas by boiler 101 just
xDeng) concentration dependent information then can measure by measurement of concetration device 154.In addition, generally speaking, beyond Fig. 2 is illustrated, also dispose numerous measuring appliances at thermal power generation complete equipment, in Fig. 2, omit.
And, in Fig. 2 from burner 102 input 1 air and 2 air, describe from the path of rear air port 103 input rear wind.
1 time air is directed to the pipe arrangement 130 from fan 120, branches into pipe arrangement 132 and the pipe arrangement 131 by air heater 104 not by air heater 104 in the way, and again be directed to muller 110 after pipe arrangement 133 collaborates.Air by air heater 104 is by gas-heated.Utilize this 1 air, send the fine coal that generates in the muller 110 to burner 102 via pipe arrangement 134.
2 air and rear wind are directed to the pipe arrangement 140 from fan 121, after air heater 104 heating, branch into the pipe arrangement 142 that pipe arrangement 141 that 2 air use and rear wind are used, and are directed to respectively in burner 102 and the rear air port 103.
Fig. 3 is the pipe arrangement section that passes through of 1 air, 2 air and rear wind and the enlarged drawing of air heater 104.As shown in Figure 3, in pipe arrangement, dispose air door (air damper) 160,161,162,163.By the operation air door, can change the area that the air in the pipe arrangement 133,141,142 passes through, therefore can adjust air mass flow by pipe arrangement by the operation of air door.
Below, the information that the information of the process signal 10 of storing in the declarative procedure computing machine 300, model information database 500 and diagnostic result database are preserved and the calculation function in feature model preparing department 420, unified model preparing department 430, key element diagnosis section 460, comprehensive diagnos section 470.
The information of the process signal 10 of at first, process computer 300 being preserved describes.Fig. 4 is the figure for the example that the information that each process computer 300 is preserved is described.In the generating set of equipments 100 measured information that obtain as shown in Figure 4, measure constantly in the lump preservation by each measuring appliance with each.For example, store in the longitudinal axis project with Fig. 4 and measure constantly, in the transverse axis project, store accordingly classification, the unit of measured value by the PID numbering association that measuring appliance is given inherently.
For example, number about PID, PID150 means the flow measuring probe 150 measured data that obtain at the feedwater flow of Fig. 2, PID151 means the temperature meter 151 measured data that obtain in main steam, PID152 means the pressometer 152 measured data that obtain in main steam, PID153 means that PID154 means the measurement of concetration device 154 measured data that obtain at burning gases in the generating output checker 153 measured data that obtain.In addition, as classification, F means flow, and T means temperature, and P means pressure, and E means generating output, and D means NO contained in the Exhaust Gas
xConcentration.These unit be respectively kg/s, ℃, Mps, MW, ppm.In this wise, the measured value that so PID numbering and classification, unit is showed and the information of time are preserved in the lump.
In addition, in Fig. 4, obtain data and preserve with 1 second cycle, about the sample period of Data Collection, can at random set.In order to be easy to effectively to utilize the data of storing in the process computer 300, can be numbered the basis with the PID to the intrinsic setting of each measured value, process signal is carried out specialization, perhaps can utilize as the key word (key) of process signal when exploring to expectation.
Secondly, the information that model information database 450 is preserved describes.Fig. 5, Fig. 6 are the figure of the pattern of information required in the diagnostic model that is illustrated in based on ART.Such as Fig. 5, shown in Figure 6, prepare to have feature model to use with table TB420 and unified model in the model information database 450 and show TB430.Feature model is the table of the storage information relevant with feature model preparing department 420 handled feature models with table TB420, and unified model is the table of storing the information relevant with unified model preparing department 430 handled unified models with showing TB430.
The maximum numbering of the classification that is equipped with pattern number, model name at the feature model of Fig. 5 in table TB420, generates when inputing to PID numbering, alarm parameters, the modelling of the input variable in the model.In this record example, pattern number is made as E-001, model name is made as main steam model, the input variable relevant with the main steam of the generating set of equipments of Fig. 2 and is made as PID150 (flow measuring probe 150 of feedwater flow), PID151 (temperature meter 151 of main steam), alarm parameters is made as 0.82, and the maximum numbering of the classification that generates during modelling is made as 21.In the example of this such feature model, the main steam model is showed by feedwater flow and main steam temperature.
The maximum numbering of the classification that is equipped with pattern number, model name at the unified model of Fig. 5 in table TB430, generates when inputing to pattern number, alarm parameters, the modelling of the feature model in the model.In this record example, pattern number is made as T-001, model name is made as the main steam circulation model relevant with the main steam of the generating set of equipments of Fig. 2, input model is made as feature model with employed main steam model E-001 among the table TB420 and has added PID152 (pressometer 152 of main steam), alarm parameters is made as 0.89, and the maximum numbering of the classification that generates during modelling is made as 25.In the example of this such unified model, as the main steam circulation model, show as that the main steam model E-001 that feature model is used among the TB420 with table has been added main steam pressure and obtain model.
As known from the above, as unified model, in this example, refer to, feature model has been appended new process variable and the composite model that consists of.Though without diagram, as unified model, also can be with the feature model compound and model that obtains each other in the diagram example.In addition, in this example, as feature model, be the model that simulation main steam monomer obtains, as unified model, be the model of having simulated the main steam circulation.So, unified model can be referred to as performance wider scope, and the model of more upper scope by feature model or process variable are fully utilized, becomes the model that can realize wide Scoped, upperization.
In addition, feature model or unified model can non-patent literature 1 etc. is disclosed to become a plurality of input Data classifications self-elevating platform ART (ART) network of the supervised learning type of a plurality of classification to realize by utilizing, in this case, as long as the next network, then unified model can be orientated feature model as upper network.
In addition, the number of the process variable that uses in feature model (number of times) is made as (being 2 times) N time in the example of Fig. 5, unified model is the model of (N+M) inferior (being 3 times in the example of Fig. 5), is to can be considered the user element model to have made the more model of high order based on this unified model of making.
Based on this, solution problem of the present invention: the change that chases after the diagnostic device inner model when establishing of the appending of the new equipment in the set of equipments, measuring appliance is understood to and feature model can be remained unchanged, and makes unified model and realizes by increasing the process variable that appends.
In addition, the number of alarm parameters along with the adjustment gimmick that is suitable for difference.In this case, utilize alarm parameters to be set with respectively showing in a plurality of situations.Fig. 6 is that alarm parameters is made as the example in the single situation, and Fig. 7 shows the example that alarm parameters is made as each table in a plurality of situations.The difference of Fig. 7 and Fig. 6 is, the pattern that alarm parameters carries according to sorting out numbering.
Fig. 7 is the figure of an example of the information of the diagnostic result stored in the expression diagnostic result database 480.In diagnostic result database 480, prepare to have feature model to use with table TB460 and unified model and show TB470.Feature model is the table of the storage diagnostic result information relevant with the feature model of processing in key element diagnosis section 460 with table TB460, and unified model is the table of the storage diagnostic result information relevant with the unified model of processing in comprehensive diagnos section 470 with table TB470.
The feature model of Fig. 7 be equipped with in table TB460 pattern number, constantly, sort out numbering, diagnostic result.Storing pattern number in this record example is E-001, is " 2010/01/0100:00:00 " constantly, and the diagnostic result of sorting out numbering 1 is normal.In addition, the unified model of Fig. 7 also consists of with same project with table TB470, and storing pattern number in this record example is T-001, is " 2010/01/0400:10:11 " constantly, and the diagnostic result of sorting out numbering 3 is " unusually ".
When the key element diagnosis section 460 in consisting of the present invention, comprehensive diagnos section 470, applicable patent document 1 and patent documentation 2 described methods.Patent documentation 1 and patent documentation 2 described methods all are to have utilized the diagnostic method of the set of equipments of self-elevating platform ART (Adaptive Resonance Theory, ART).ART is that the Data classification that will input becomes a kind of of sorter in a plurality of classification.
At this, the process signal that does not contain unusual set of equipments is inputed to ART, and be categorized as a plurality of classification.In the diagnosis when unusual the generation, by process signal is inputed among the ART, and generated when not belonging to the existing new classification of sorting out, owing to produced the set of equipments state that does not have so far, thereby give a warning.
In addition, diagnostic result except normally, unusually, generated in addition the situation of the new classification that does not belong to any state in the past.In this case, as the unknown, be situated between and point out this information by aid 910 to the user.As long as this state be normal or unusual judgement determined, just replace diagnostic result diagnostic result database 480 in thereafter.After, in the situation that has produced same classification, export diagnostic result based on this result.
More than, the formation of generating set of equipments has been described as set of equipments that apparatus of the present invention were suitable for.In addition,, be illustrated with the example of the process variable in this situation with table, the unified model concrete example with table about feature model.On the basis of having understood this content, next the contents processing of feature model preparing department 420 described.
Fig. 8 is the process flow diagram that is illustrated in the action of feature model preparing department 420.At first at step S421, judgement is the new model of making or existing model has been applied correction.Being in the new situation of making, enter step S422, not to enter step S425 in the new situation of making.
In addition, the new making of model or revise by the user relevant with generating set of equipments 100 for example the instruction content given according to input media 900 of operator distinguish.Thus, when not giving the relevant indication of change with model from the user, the not processing of execution graph 8.The user gives the indication that model changes according to the change of generating set of equipments 100.Utilize Figure 12 to narrate in the back the concrete gimmick that model changes indication.
Be in the new situation of making in the indication that model changes, in step S422, utilize the process signal in the past that Fig. 4 stores, make new ART model.Among step S423s, the information of the ART model made be stored in model information database 450 thereafter.Make this result, the new ART model of storing at model information database 450 becomes for example main steam model E-001 of Fig. 5, becomes the feature model of determining with feedwater flow and main steam temperature.
In step S424, judge whether all corresponding monitored item purpose model constructions are finished.As all finishing, then finish, then return as unfinished step S421, repeatedly carry out step thereafter, finish all supervision project models until make up.
Indication in the model change is in the situation of existing correction of the model, and in step S425, the input variable of existing model is appended or deleted in judgement.As then entering step S426 for appending, as then entering step S427 for deletion.
Among the step S426 when having appended input variable, only utilize the variable that existing model is appended to make diagnostic model based on ART.At this, suppose that existing model is the main steam model E-001 of Fig. 5, the variable that appends is PID152 (pressometer 152 of main steam).Thus, in step S426, only utilize the variable PID152 append to make diagnostic model based on ART.After finishing, step S423 (information of the ART model made is stored in the model information database 450), enter step S424 (being confirmed whether corresponding with all supervision projects).
In addition, in step S423, existing pattern number E-001 is used in the input model hurdle of table TB430 with the unified model that the pattern number (being made as PID152) that appends stores the model information database 450 of Fig. 5 into.
Among the step S427 when having deleted the input variable of existing model, the input variable that change does not occur will remain unchanged, and input fixed value for the variable after the deletion, in this input 0.5.Wherein, in the present embodiment, suppose that the input variable of ART model all has been carried out normalization.That is, grasp as the value that maximal value is made as " 1 " and minimum value is made as the scope of " 0 ".Thus, " with 0.5 input " means as the value of the not change of intermediate value, processes afterwards.
At this, suppose that existing model is the main steam model E-001 of Fig. 5, deleted variable is PID151 (temperature meter 151 of main steam).In this case, because to PID151 input 0.5, therefore later main steam model E-001 is in fact implemented to process as the model of only determining with the feedwater flow of input variable PID150.
In step S428, accept will deletion signal PID151 be made as 0.5 fixed value, existing classification is transformed to the classification that this model is used.In addition, in the main steam model E-001 of Fig. 5, have 21 classification, so these become the object of conversion.For this conversion, with reference to the feature model of the diagnostic result database 480 of Fig. 7 with table TB460, the moment of the classification numbering that utilization is stored with respect to the hurdle of main steam model E-001, input to the pid information of the variable in this model, take out the time series data of each input variable that process computer 300 preserves.
In the situation of this example, utilization and classification are numbered moment " 2010/01/0100:00:00 " corresponding to " 1 ", are inputed to the pid information (PID150, PID151) of the variable in the main steam model E-001, with reference to Fig. 4, take out the time series data of each input variable (PID150, PID151) that process computer 300 preserves.This time series data is input in the ART model.(21) implement this operation for whole classification.After whole classification is implemented, and hereto in the same manner, enter step S423, step S424.
In addition, same model exist append with the situation of deleting under, carry out first the process flow diagram that appends side at step S425, again be back to step S421 at step S424, again carry out the process flow diagram of deletion at step S425.
Fig. 9 is the process flow diagram that is illustrated in the action of unified model preparing department 430.At first, judge whether newly to have made feature model at step S431.As be new the making, then enter step S436, otherwise (having in the situation of change of feature model) enters step S432.In addition, be in the new situation of making, because the processing that should not process as unified model preparing department 430, so, be transferred to step S431 via step S436 and come other project is judged or end process.
In the situation of the change that feature model is arranged, in step S432, judge the state of input variable.As then entering step S433 for appending, otherwise (in the situation of deletion input variable) enters step S436.In addition, because in the situation of having deleted the input variable deletion, the processing that should not process as unified model preparing department 430, so, be transferred to step S431 via step S436 and come other project is judged or end process.
In step S433, when feature model is made, again input existing model and the new process signal that appends mode input of making.Appending in the case-handling (step S425) at Fig. 8 for example, existing model (main steam model E-001) has been appended PID152, as the process signal related with these, again input to discharge PID150 and main steam temperature PID151 and main steam pressure PID152 with reference to Fig. 4.In addition, in the processing of the step S423 of Fig. 8, with existing pattern number E-001 and the pattern number (being made as PID152) that appends, be stored in the input model hurdle of unified model with table TB430 of model information database 450 of Fig. 5.
In step S434, the classification numbering that structure will be exported from each aforesaid model is as the ART model of input variable.For example the assembled classification of the classification inputted numbering is sorted out for new, and with it as unified model.
Particularly, suppose about existing pattern number E-001, exist 10 to sort out numbering (from A0 to A9), about the pattern number PID152 that appends, exist 5 to sort out numbering (from B0 to B4).In this case, being normally with the combination of B0 (normally) such as A0 (normally), is C0 " normally " with new classification number definition then.When A0 (normally), B1 (unusually), be C1 " unusually " with new classification number definition in addition.Until carry out till will operating final combination, the model that will have the classification numbering of new a group is kept at unified model with among the table TB430 as unified model.
In addition, having produced in the situation of newly establishing of measuring appliance etc., may again entirely make about the New model of whole process signals, even for example with reference to the input data in 1 time in the past, and will become the situation of the data of processing huge amount.
About this point, carry out modelling with reference to input data 1 fraction of the year in the past among the present invention, obtain newly that to sort out numbering only be the input variable of appending.The feature model of both having established is paid close attention to sorting out numbering, so the input data volume in 1 time of past is summarized in a plurality of classification numberings.
When finishing, modelling enters step S436.In step S436, judge whether correspondence for whole supervision projects.When not finishing, be back to step S431, carry out repeatedly until whole projects is finished.Finish as finishing then.
Secondly, Figure 10 represents the process flow diagram of the algorithm of key element diagnosis section 460 and comprehensive diagnos section 470.At this, at first at step S461, from the feature model of model information database 450 information that loads each feature model the table TB420.Next in step S462, from the unified model of model information database 450 information that loads each unified model the table TB430.Thus, the data of Fig. 5, Fig. 6 offer key element diagnosis section 460.
In step S463, judgement is the only diagnosis under feature model or comprises diagnosis under the unified model.Only being in the situation of the diagnosis under the feature model, enter step S464, in for the situation that needs the diagnosis under the unified model, enter step S466.
In step S464, implement the diagnosis under each feature model.This is the diagnosis that the Application elements model is used table TB420.In step S465, the diagnostic result of feature model is stored in the feature model table TB460 of diagnostic result database 480.
On the one hand, in step S466, at first implement the diagnosis under each feature model.This is to implement the diagnosis that unified model is used each feature model in the table TB430.For example in the situation that is main steam circulation model T-001, to feature model E-001 with append model PID152 each diagnose.The diagnostic result of feature model is stored in the feature model of diagnostic result database 480 with among the table TB460.
In step S467, with reference to the feature model of diagnostic result database 480 table TB460, will diagnose as input from the classification numbering of each feature model output.
In step S468, diagnostic result is stored in the unified model of diagnostic result database 480 with among the table TB480.
Among step S469s, judged whether to the whole supervision project implementation diagnosis thereafter.When not finishing, return step S461, carry out repeatedly until whole supervision projects is finished.After finishing, finish.
In this process flow diagram, step S467 and step S468 become the action that comprehensive diagnos section 470 carries out, and other step S is implemented by key element diagnosis section 460.
Externally in the output interface 490, each diagnostic result is sent to aid 910 as Output rusults.
Secondly, utilize aid 910 that the method for the information of image display device 950 display control signals 20, process signal 30, diagnostic result 40 or model information database 450 and diagnostic result 480 is described to the user.
Figure 11～Figure 15 is the shown picture example of image display device 950.The user utilizes keyboard 901, mouse 902 to the position that becomes the sky hurdle of these pictures 90, carries out the operation of input parameter value etc.
At first, Figure 11 is the shown initial picture of image display device 950.At picture 90, as initial picture, show that diagnostic model makes button 951 and diagnostic result the Show Button 952, the user selects the button of necessity from these, utilize mouse 902 that cursor 953 is moved, by 902 pictures that show expectation of clicking the mouse.
Figure 12 is feature model shown when having selected that diagnostic model is made button 951 in the initial picture and the setting picture of unified model.Show that in the hurdle, top of picture 90 this picture is the picture that shows that information setting is used.In addition, on the hurdle that each one of picture 90 forms explanation below the demonstration, these hurdles are inputted data or selected to carry out the setting of model by button.
In process signal display field 961, the user will be input in the input field 961 measuring-signal or the operation signal of diagnostic model input, inputs in the lump its scope (upper limit/lower limit).In the illustrated example, as measuring-signal and button has been selected main steam flow, higher limit is made as 300 (kg/s), lower limit is made as 0 and carries out the numerical value input (kg/s).In addition, unit (kg/s) is selecting main steam flow to show together default value as measuring-signal with button.
In addition, when diagnostic model is appended main steam flow, with the time tape input that will show used in the modelling in moment input field 962.In illustrated example, with between 2010/01/01 1 day as beginning, the finish time and set.
On picture 90, by clicking the Show Button 963, as shown in Figure 13, the tendency chart shows at image display device 950.In the example of Figure 13, can show the appearance of the change in time of a plurality of various process variables.By clicking the return push-button 971 of Figure 13, be back to the picture of Figure 12.
Make display field 964 at the feature model of Figure 12 and show the required pattern number of feature model making, model name, input variable etc.The right side that makes display field 964 at feature model shows newly-increased button 965 and revises button 966.
Wherein, when pressing newly-increased button 965, feature model is made display field 964 and is shown blank and become input state, can make new model by user's hand input.
An example of the picture when illustrated feature model making display field 964 has shown that button 966 is revised in selection.Illustrate pattern number E-001, model name is that main steam model and input variable are the example of PID150, PID151.
In addition, New model is made and revised as the basis take existing model is shortcut, in this case, shows the model that becomes object, after having pressed correction button 966, revises and gets final product.Thus, be set with search key input field 967 according to the mode that can retrieve existing model, after having inputted search key, if press index button 968, the information that then becomes the model of object shows in feature model display field 964.
Unified model is made display field 974 and is basically also consisted of in the same manner with feature model making display field 964.In unified model display field 974, show the required pattern number of unified model making, model name, input model etc.When pressing newly-increased button 975, unified model display field 974 becomes input state, can make new model.In situation about revising as the basis take existing model, show the model that becomes object, after having pressed correction button 976, revise.Have search key input field 977 according to the mode that can retrieve existing model, after having inputted search key, if press index button 978, the information that then becomes the model of object will be presented at unified model display field 974.
After the above setting, make button 992 by pressing, make each model.In Figure 12, by clicking return push-button 969, can be back to the picture of Figure 11.
Figure 14 is be used to making diagnostic result be shown in the setting picture 90 of image display device 950.By clicking diagnostic result the Show Button 952 in the initial picture of Figure 11, show the picture of Figure 14.In the hurdle, top of the picture 90 of Figure 14, show that this picture is the picture that the diagnostic result display setting is used.In addition, the hurdle in that each one of picture 90 forms explanation below the demonstration to these hurdles input data or selected by button, carries out the setting of model.
In process signal selectionbar 981, the user will make measuring-signal that the picture 90 of image display device 950 shows or operation signal and its scope (upper limit/lower limit) input in the lump in the input field 981.In illustrated example, show generator output and main steam flow and the picture of its scope (upper limit/lower limit) when inputting in the lump.
The time that will show in addition, inputs to constantly input field 982.After the process signal that will show is determined, make a mark to determine to select by clicking selectionbar.
In feature model selectionbar 983, with the demonstration information setting picture of Figure 12 in the same manner display model numbering, model name, input variable etc.Have search key input field 984 in order to retrieve the model that will show.After having inputted search key, press index button 985 and retrieve.Result for retrieval is presented at feature model selectionbar 983, chooses to determine to select by clicking selectionbar.In illustrated example, the example of display model E-001.
In addition, unified model selectionbar 986 also consists of with feature model selectionbar 983 basically in the same manner.At unified model selectionbar 986 display model numberings, model name, input model etc.Have search key input field 987 in order to retrieve the model that will show.After the search key input, press index button 988 and retrieve.Result for retrieval is presented at feature model selectionbar 986, chooses to determine to select by clicking selectionbar.The example of display model T-001 in illustrated example.
After input more than having carried out or the selection, by clicking the Show Button 989, as shown in Figure 15, the tendency chart is presented at the picture 90 of image display device 950.In the example of Figure 15, by the generator output of process signal selectionbar 981 selections and comparing property of the change demonstration of the time of main steam flow.In addition, feature model E-001 or unified model T-001 and the input variable PID152 that newly appends about selecting according to every classification numbering, carry out time series and show.As shown in Figure 15, the new classification when producing,, the summary part on the tendency picture is emphasized with other color, notifies thus the user.By clicking the return push-button 991 of Figure 15, can return the setting picture of Figure 14.When pressing the return push-button 999 of Figure 14, be back to the initial picture of Figure 11.
Show Figure 15 as the display frame of diagnostic result example, other show the system diagram of the set of equipments that becomes diagnosis object, sort out in the situation about producing new, also can consider this position is significantly changed, when deserving the position by clicking, show that tendency shown in Figure 15 shows such demonstration gimmick.
In addition, show the embodiment that comprises model information database 450 and diagnostic result database 480 at set of equipments diagnostic device 400, also can be respectively: the embodiment that is made as other hardware not being contained in set of equipments diagnostic device 400.
Below, illustrate for generating set of equipments 100, take the result of process diagnosis method of the present invention and device be applied in as the basis set of equipments unusual/effect of omen diagnosis.
With the diagnostic method of set of equipments of the present invention be applied to generate electricity set of equipments unusual/during the omen diagnosis, in a situation that exchanges or keep in repair and diagnostic model is revised owing to equipment in using, can be effectively used to the diagnostic data of accumulating till this, and can diagnose.Its result, again structure that can diagnostic model easy to implement by effective utilization actual achievement so far, can detect unusual or the omen tendency better.
In addition, show owing to the operations staff is also carried out two results' of feature model and unified model tendency, will further become easy based on visual supervision.And come comprehensive a plurality of diagnostic models and be made as stratum's type with unified model, thus, the object that can easily consist of the set of equipments that will generate electricity is made as wider scope and the model diagnosed.
Claims (8)
1. the diagnostic method of a set of equipments, wherein,
Maintenance has been carried out the model after the modelling with the correlationship of input variable,
According to the correlationship of described input variable, be a plurality of classification with the Data classification of inputting, and survey the unusual of set of equipments according to the sorted generation that does not belong to the classification of normal classification,
The diagnostic method of described set of equipments is characterised in that,
Produced appending of input variable or deleted the described model of such change about the modification according to described set of equipments, the change of numbering by the described classification that has utilized this model makes up new model.
2. the diagnostic method of set of equipments according to claim 1 is characterized in that,
Existing model is being appended in the situation of input variable, only make up diagnostic model with the input variable of appending, number to generate the model of new classification numbering based on the classification numbering that obtains from existing model and the classification that obtains from the described diagnostic model that appends.
3. the diagnostic method of set of equipments according to claim 1 is characterized in that,
From existing model, deleting in the situation of input variable, to be made as fixed value by the value of the input variable of deleting in the input variable of normalization and input, and utilize the time series data of each input variable in the classification numbering of this model to come the change of execution model.
4. the diagnostic method of set of equipments according to claim 2 is characterized in that,
From existing model, deleting in the situation of input variable, to be made as fixed value by the value of the input variable of deleting in the input variable of also inputting after the normalization, and the time series data of utilizing each input variable in the classification numbering of this model comes the change of execution model, and deletes processing described appending after processing finishes.
5. the diagnostic method of a set of equipments, wherein,
The correlationship of the input variable that will obtain from monitored object is carried out modelling, and the Data classification that inputs to model is become a plurality of classification, and surveys the sign of the unusual or omen of set of equipments according to the sorted generation that does not belong to the classification of normal classification,
The diagnostic method of described set of equipments is characterised in that,
The correlationship of the 1st input variable is carried out modelling and keep as feature model, and according to the correlationship of described the 1st input variable the Data classification of inputting is become a plurality of classification, and the correlationship of the 2nd input variable of appending is also carried out modelling, the Data classification of inputting is become a plurality of classification, form the unified model based on described the 2nd input variable after described feature model and the modelling.
6. the diagnostic method of set of equipments according to claim 5 is characterized in that,
Described unified model comprises by the determined new classification of combination with the classification of the classification of described feature model and described the 2nd input variable after institute's modelling.
7. the diagnostic device of a set of equipments according to from the control signal of the supervision control device that monitored object is controlled or the measuring-signal that obtains, is surveyed the unusual of described monitored object from the process computer of the process signal of inputting described monitored object,
The diagnostic device of described set of equipments is characterised in that to possess:
Feature model section, its correlationship with the signal of input is carried out modelling and is made feature model, the Data classification that will be input in the feature model according to the correlationship of described input signal becomes a plurality of classification, and surveys the unusual of set of equipments according to the sorted generation that does not belong to the classification of normal classification; And
Unified model section, it possesses the unified model that comprises at least the feature model more than 1, the Data classification that will be input to unified model according to the correlationship of described input signal becomes a plurality of classification, and surveys the unusual of set of equipments according to the sorted generation that does not belong to the classification of normal classification
Produced the described feature model that appends of described input signal about the modification according to described control object, only make up diagnostic model with the input signal that appends, number to generate new classification numbering based on the classification numbering that obtains from existing feature model and the classification that obtains from the described diagnostic model that appends, and then will generate that new classification is numbered and the model that obtains is stored in described unified model section as unified model.
8. the diagnostic device of set of equipments according to claim 7 is characterized in that,
From existing model, deleting in the situation of input variable, to be made as fixed value by the value of the input variable of deleting in the input variable of normalization and input, and utilize the time series data of each input variable in the classification numbering of this model to come the change of execution model.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2011-157924 | 2011-07-19 | ||
| JP2011157924A JP5469131B2 (en) | 2011-07-19 | 2011-07-19 | Plant diagnostic method and apparatus. |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102890495A true CN102890495A (en) | 2013-01-23 |
| CN102890495B CN102890495B (en) | 2015-03-18 |
Family
ID=47534023
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210245867.4A Expired - Fee Related CN102890495B (en) | 2011-07-19 | 2012-07-16 | Complete plant diagnosis methods and complete plant diagnosis apparatus |
Country Status (2)
| Country | Link |
|---|---|
| JP (1) | JP5469131B2 (en) |
| CN (1) | CN102890495B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107703912A (en) * | 2017-09-13 | 2018-02-16 | 陕西千山航空电子有限责任公司 | A kind of method for diagnosing faults based on Aerial Electronic Equipment |
| CN107870611A (en) * | 2016-09-28 | 2018-04-03 | 横河电机株式会社 | Workshop analogue means and workshop analogy method |
| CN110291476A (en) * | 2017-02-17 | 2019-09-27 | 三菱日立电力系统株式会社 | The diagnostic device and diagnostic method of complete set of equipments |
| CN112136088A (en) * | 2018-06-28 | 2020-12-25 | 株式会社日立制作所 | Plant diagnostic system and method |
| CN115349111A (en) * | 2020-03-30 | 2022-11-15 | 三菱电机株式会社 | Diagnostic system, diagnostic method, and program |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2016183839A (en) * | 2015-03-26 | 2016-10-20 | 一般財団法人電力中央研究所 | Pulverized coal firing boiler and power generation facility |
| CN105527842B (en) * | 2016-01-25 | 2018-04-24 | 东南大学 | A kind of reheat steam temperature neural Networks Internal Model Control method of compositive economy index |
| JP6919186B2 (en) | 2016-12-14 | 2021-08-18 | オムロン株式会社 | Control system, control program and control method |
| EP3557354B1 (en) | 2016-12-14 | 2023-03-01 | Omron Corporation | Control device, control program, and control method |
| JP6468313B2 (en) * | 2017-06-08 | 2019-02-13 | オムロン株式会社 | Control device, control program, and control method |
| JP6951202B2 (en) * | 2017-11-13 | 2021-10-20 | 株式会社日立製作所 | Data classification device |
| WO2019159883A1 (en) * | 2018-02-13 | 2019-08-22 | 三菱日立パワーシステムズ株式会社 | Model creation method, plant operation support method, model creating device, model, program, and recording medium having program recorded thereon |
| KR20210042928A (en) * | 2018-08-10 | 2021-04-20 | 스미도모쥬기가이고교 가부시키가이샤 | Abnormality monitoring device, abnormality monitoring method and abnormality monitoring program |
| JP2024067066A (en) * | 2022-11-04 | 2024-05-17 | 株式会社日立製作所 | Industrial plant related system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH10207507A (en) * | 1997-01-23 | 1998-08-07 | Hitachi Ltd | Plant simulator |
| JP2005258649A (en) * | 2004-03-10 | 2005-09-22 | Hitachi Ltd | Data sorting method and device |
| CN101361029A (en) * | 2005-11-17 | 2009-02-04 | 霍尼韦尔国际公司 | Apparatus and method for identifying informative data in a process control environment |
| JP2009087082A (en) * | 2007-09-28 | 2009-04-23 | Fuji Electric Systems Co Ltd | Plant operation planning device and its program |
| JP2010237893A (en) * | 2009-03-31 | 2010-10-21 | Hitachi Ltd | Device and method for diagnosing abnormality of plant |
| JP2011070334A (en) * | 2009-09-25 | 2011-04-07 | Hitachi Ltd | Abnormality diagnostic device and abnormality diagnostic method |
-
2011
- 2011-07-19 JP JP2011157924A patent/JP5469131B2/en active Active
-
2012
- 2012-07-16 CN CN201210245867.4A patent/CN102890495B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH10207507A (en) * | 1997-01-23 | 1998-08-07 | Hitachi Ltd | Plant simulator |
| JP2005258649A (en) * | 2004-03-10 | 2005-09-22 | Hitachi Ltd | Data sorting method and device |
| CN101361029A (en) * | 2005-11-17 | 2009-02-04 | 霍尼韦尔国际公司 | Apparatus and method for identifying informative data in a process control environment |
| JP2009087082A (en) * | 2007-09-28 | 2009-04-23 | Fuji Electric Systems Co Ltd | Plant operation planning device and its program |
| JP2010237893A (en) * | 2009-03-31 | 2010-10-21 | Hitachi Ltd | Device and method for diagnosing abnormality of plant |
| JP2011070334A (en) * | 2009-09-25 | 2011-04-07 | Hitachi Ltd | Abnormality diagnostic device and abnormality diagnostic method |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107870611A (en) * | 2016-09-28 | 2018-04-03 | 横河电机株式会社 | Workshop analogue means and workshop analogy method |
| CN110291476A (en) * | 2017-02-17 | 2019-09-27 | 三菱日立电力系统株式会社 | The diagnostic device and diagnostic method of complete set of equipments |
| CN107703912A (en) * | 2017-09-13 | 2018-02-16 | 陕西千山航空电子有限责任公司 | A kind of method for diagnosing faults based on Aerial Electronic Equipment |
| CN112136088A (en) * | 2018-06-28 | 2020-12-25 | 株式会社日立制作所 | Plant diagnostic system and method |
| CN115349111A (en) * | 2020-03-30 | 2022-11-15 | 三菱电机株式会社 | Diagnostic system, diagnostic method, and program |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102890495B (en) | 2015-03-18 |
| JP5469131B2 (en) | 2014-04-09 |
| JP2013025461A (en) | 2013-02-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102890495B (en) | Complete plant diagnosis methods and complete plant diagnosis apparatus | |
| CN109272243B (en) | Carbon asset management method and system | |
| Gerber et al. | Systematic integration of LCA in process systems design: Application to combined fuel and electricity production from lignocellulosic biomass | |
| RU2310226C2 (en) | Method and device for evaluating productivity of steam-gas electric plants | |
| CN104200397B (en) | A kind of nuclear power station DCS system input design file and variable information data base consistency(-tance) method of calibration | |
| CN106302739A (en) | A kind of Industrial Boiler monitoring and analysis aid decision cloud platform system | |
| CN103635910A (en) | Automatic identification of operating parameters for power plants | |
| CN116681187B (en) | Enterprise carbon quota prediction method based on enterprise operation data | |
| CN107291830A (en) | A kind of creation method of equipment knowledge base | |
| JP6742222B2 (en) | Driving support device and program | |
| Sayyaadi et al. | Exergoeconomic optimization of a 1000 MW light water reactor power generation system | |
| CN110134740A (en) | A kind of electric energy efficiency data classification storage method and system | |
| CN103390211A (en) | Thermal generator set overall process energy management and cost analysis system | |
| Lv et al. | A data fusion and data cleaning system for smart grids big data | |
| Wang | Thermo-economic evaluation, optimization and synthesis of large-scale coal-fired power plants | |
| Valero et al. | On-line monitoring of power-plant performance, using exergetic cost techniques | |
| CN105900030A (en) | Method, system, and computer program product for analyzing production and/or process-engineering processes and/or process steps in a plant | |
| CN113722656B (en) | Real-time health evaluation method and system for thermal generator set | |
| CN102830662B (en) | Monitoring system and method of flow industrial pipe network system | |
| CN110162743A (en) | A kind of data administering method based on k neighborhood nonlinear state Eq algorithm | |
| Su et al. | Temporal validation of life cycle greenhouse gas emissions of energy systems in China | |
| Dhanalakshmi et al. | Energy analytics and comparative performance analysis of machine learning classifiers on power boiler dataset | |
| CN207112772U (en) | Boiler hearth temperature on-line prediction system based on core offset minimum binary | |
| Liu et al. | Energy systems engineering: methodologies and applications | |
| Rahman et al. | Environmental impact assessments of compressed air energy storage systems: a review |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20150318 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |