Operation diagnostic method of circulating fluidized bed boiler, and operation diagnostic device

Hereinafter, reference to the accompanying drawing the present invention from the stroke detector with a stroke diagnosis device and operation diagnosis method. rapidly and to reduce a memory form of embodiment. Furthermore, description of drawings take identical element the same sign worker and thus, to reduce worker, ., which does not require a sound generating bodies the directions of which described.

Also as shown in the 1, the present embodiment in the form of device (1) diagnosis operation the circulating fluidized bed boiler (2) (hereinafter, simply also referred to as "boiler") for diagnosing the operating state of. First, boiler (2) relates to. Boiler (2) the, also 2 as shown in the, external circulation type (Circulating Fluidized Bed type) of is circulating fluidized bed boiler. This boiler (2) the, toward the long moving bed of furnace (3) is provided. Furnace (3) at an intermediate section of, fuel is charged into a fuel inlet (3a) and a, gases exiting the combustion gases by means of which the upper outlet (3b) is provided. Fuel input device (5) from the furnace (3) the fuel supplied to, fuel inlet (3a) through the furnace so as to bring it (3) is into.

Furnace (3) of gas outlet (3b) the cyclone chamber and functions as a solid gas separation device (7) are connected to. Cyclone (7) outlet of (7a) a gas line control orgin connected to gas in the succeeding stage. Also, cyclone (7) called [...] down from sub bottom of a return line (9) extends downward is, return line (9) lower end of a furnace (3) is connected to the side an intermediate portion of the.

Furnace (3) in, lower for air supply of a beamline (3c) introduced from the an auxiliary catalytic converter, the combustion, by means of air in a for flow, fuel inlet (3a) including poured from fuel and to solids and flowing and, 800-900 ° C than about the flowing fuel. combustion in the. Cyclone (7) the, furnace (3) generated in which a combustion gas having a solid to entrain the particles while is introduced. Cyclone (7) the, centrifugal separation action a solid by particles and the separation of gas and, return line (9) furnace of solid particles are separated via (3) and back to, solid particle is removed combustion gas outlet (7a) through gas line from sends control orgin gas in the succeeding stage.

Is furnace (3) in "furnace ([...]) in bed material" called the bottom portion of solids is generated but stacked, is impurity in a material bed in blast furnace (melting material) is concentrated, is prevented bed of sintering and melt solidified, or radio-active non-inflammable laundry packing envelope operation by an. needs to have pins. Thereby, furnace (3) in, of the bottom outlet (3d) in blast furnace from bed material of wet liquid to flow down is discharged to the outside periodically. A bed discharge, a cover (not shown) on after removing unsuitable point metal, e.g., again furnace (3) is on.

Said gas control orgin , cyclone (7) outlet of (7a) and connected via gas line to gas heat exchange device (13) and a, heat exchange device (13) the gas outlet of (13a) to collector and connected via gas line (dust collector) (15) is provided. Gas heat exchange device (13) the, across flow paths of exhaust flows into a muffler through an water to boiler tube (13b) is provided. Cyclone (7) this high-temperature exhaust gas directed from the boiler tube (13b) in contact with a, of water for 0.5 within the tube the heat of the exhaust gas is collected to, of a boiler tube for preventing breakage of water vapor is high temperature generating (13b) for generating power by through sent turbine for. Collector (15) the, still located in the exhaust pipe before gas is inflammable such as entrained in a fly ash. microparticles removed therefrom. Collector (15) outlet of (15a) a clean gas being discharged from a gas line and pump (17) via a chimney (19) is discharged to the outside from an.

Also 1 with a, boiler (2) the, operation data constituting a sensor data, operation diagnosis device (1) group sensor for outputting data that is greater than the height of (14) is provided and. These sensors group (14) the, for example, boiler (2) round a given site of four to measure the temperature of temperature sensor (14a), measuring the flow rate of the exhaust gas or a flow sensor (14b) or predetermined material in exhaust gas concentration sensor for measuring the concentration of (14c) is include a. Yet, these sensors data, such that some of data, is item index. Index item acids to epoxygenated fatty acids therein, sensor data are indicative of operation state in is an item that is, for example, pressure deviation, heat amount (heats), such as NOX CO and boiler efficiency or emissions of environmental load comprises concentration of.

Boiler (2) the, setting item by changing a FONT type group (16) setting item display area type comprises information of a certain input values each of a driver. Acids to epoxygenated fatty acids therein setting item by changing a FONT type, boiler (2) is operated to for the boiler (2) device or control operator respect to (not shown) is, of a certain input value is an item that is may enter. The setting item by changing a FONT type, for example, blow flow (16a), sand air centrifugally by forming in the end (16b), or, furnace (3) coal is coal supplied to supply (16c) such as. Furthermore, setting item by changing a FONT type group (16) the, ass' operating number, air flow, standby release valve release number, of dose, comprises, and so valve is open, a.

Furthermore, operation diagnosis device (1) relates to. Device (1) diagnosis operation the, boiler (2) for sensor data and index items of target values, which respectively corresponds based on the boiler (2) by a rope. for diagnosing driving state of a. A target imparted item index is, for example, boiler (2) include suction ports and exhaust ports at a higher efficiency for a to allow the maintenance table and dusts reducing environmental load to allow the maintenance table and second high to allow the maintenance. is based on table. Device (1) diagnosis operation the, boiler (2) to sensor group (14) and sensor data input from, operation diagnosis device (1), which are entered into a cuts the condenser element of items index, operation diagnosis device (1) are pre-recorded onto a Bayesian networks based on model of, boiler (2) for diagnosing the operating state of. Diagnostic result, operation diagnosis device (1) the, target to satisfy to be adjusted together and setting item by changing a FONT type, setting item display area to be adjusting to be input. showed a trend for value.

Device (1) diagnosis operation the, data processing device (20) and a, data processing device (20) to predetermined data or location an input for device (21) and a, data processing device (20) data output from the content server and display and an output device (22) is provided.

Device (1) diagnosis operation the, for example, also 3 which indicates to computer (100) is realized using the. Also 1 and 3 as shown in the, computer (100) the, the present embodiment type of data processing device (20) is one example of keypad. Computer (100) the, CPU with multiprocessors system which have processing or by using software device server, personal computer or the like, and includes various data processing device. Computer (100) the, CPU (41), a device main RAM (42) and ROM (43), and entry device input device (21), display, output of ribbon cartridge device (22), and data such as a network card a communications module is transmitting/receiving device (47), such as hard discs sub storage device (48) or the like is constituted of as including computer system. Also the functional configuration element which indicates to 1, which indicates to CPU (41) also 3, RAM (42) of a prescribed for hardware such as for reading by computer software, CPU (41) under control of input device (21), output device (22), communication module (47) with by operating an, RAM (42) or auxiliary memory device (48) furnishing writing and reading of data in a 2000 by performing.

Data processing device (20) the, as the functional design factor, boiler (2) to sensor group (14) receiving sensor data input from data input (23) and a, a model recording model of Bayesian networks recording unit (24) and a, sensor Bayesian networks based on the converted data update model of a model updating unit (25) and a, be achieved, wherein an influence on item index is setting item by changing a FONT type for computing a degree of inference data calculation section (26) and a, the sensor and the inference data calculation section (26) based on the output of the value item and setting should handle himself by trend of variation of the capsule endoscope is calculated inference data processing unit (27) and a, inference data processing unit (27) to allow the maintenance table is based on the output of a report for creating a report creation section (28) is provided.

Data input (23) the, boiler (2) sensor group (14) the sensor and the acquired by the, setting item by changing a FONT type group (16) amplifies an input value is inputted. Data input (23) the, input sensor data to infer data processing unit (27) and together, necessary model updating unit (25) outputs the to. Also, data input (23) the, setting item by changing a FONT type group (16) inference value inputs to data calculation section (26) 1 and provides an output to.

Model recording unit (24) the, a model of the Bayesian networks is recording. Model recording unit (24) the, refers to inference data calculation section (26) and is configured for be referred to from, inference data calculation section (26) in response to the request from the Bayesian networks model of inference data calculation section (26) of the sensors are outputted to.

Here, model of Bayesian networks (50) relates to. Figure 4, boiler (2) setting item display area and index items in is one example of a Bayesian networks. A Bayesian networks, cause (51) and a resulting (52) are represented according to the drawing a simple relation between the, probability, graphical trend of phenomena is formed inside the tie. representation. From Bayesian networks, which a value of the parameter when drive motor for bonding tool is obtained, is not observed probability of variable is determined.

In the present embodiment, causes an setting item by changing a FONT type (51) and define an, indicator that result items (52) defined as the of wet liquid to flow down. Plurality of setting item by changing a FONT type including cause (51) the, operator or control device it is set as an item, for example, blower operating number (51a), air flow (51b), standby release valve opening degree (51c), jugs mouth quantity (51d), valve opening degree (51e), coal supply (51f), sand supply (51g), and blown flow (51h) such as. These cause defined as the setting item by changing a FONT type, Bayesian network performs ., would appear as a parent node. Yet, including result plurality of index item (52) the, sensor data are indicative of operation state in as an item that is, for example, pressure deviation (52a), heat amount (52b), CO exhaust gas concentration, and boiler efficiency (52d) such as. These results defined as the an index indicating the item is arranged in a, ., would appear as a Bayesian network performs child nodes. And, the relation of the item index and setting item by changing a FONT type, parent node from child nodes extending from the arrow by code input mode using his/her relationship. is disclosed. For example, pressure deviation (52a) include, blower operating number (51a) and an air flow (51b) extend arrow from. Therefore, pressure deviation (52a) the, blower operating number (51a) and an air flow (51b) is indicative a this relationship. Also, each node, the probability table which have (Conditional Probability Table: CPT) constitution: a (reference 6 also).

Also 1 with a, model updating unit (25) the, data input (23) modulates the carrier signal input from the sensor based on the converted data model of Bayesian networks for updating said behavioural and loyalty data contained in the.. Updated model a model data recording unit (24) to output, and recorded.

Inference data calculation section (26) the, of model and Bayesian networks, input device (21) inputted from an index item based on of target values, which respectively corresponds, each index item imparting to the effect of setting item by changing a FONT type is estimating means a steps of: calculating degrees. More specifically, inference data calculation section (26) the, effect of setting item by changing a FONT type on the item index probability degree outputs the mounting space is. To phases value probability is, is refers to. And, inference data calculation section (26) the, output results to infer data processing unit (27) outputs the to.

Inference data processing unit (27) the, index related to the items a setting item is index items on the basis of to a hardware of a, data on a plurality of setting items for ranking is means ranking determining. More specifically, inference data calculation section (26) produced by the probability value set by, the portable telephone sends ranking of items. Yet, inference data processing unit (27) the, prioritized determined allocated to each of set items with a state of the, boiler (2) actually allocated to each of set items of input input value. comparing the. By on the comparison, setting item by changing a FONT type value inputs to model of Bayesian networks state obtained based on setting item by changing a FONT type is judged at if a with a state of the. Yet, inference data processing unit (27) the, average sensor data, maximum, and minimum statistics such as. for computing quantity. Inference data processing unit (27) the, arranged to determine results and statistics reports creation section (28) outputs the to.

Report creation section (28) the, cuts the condenser element of setting item by changing a FONT type having a higher, having a higher by using input setting item by changing a FONT type, target item index information for controlling a display of a report is output means for recording data. Report creation section (28) the, outputs data in real-time with created device (22) outputs the to. For example, output device (22) when the display, report is and displaying Internet character information on screen is. Also, output device (22) when the printer, report is is printed on the paper medium.

The resulting structured materials, , the present embodiment regarding a type operation diagnosis device (1) according to, indications of data on a plurality of setting items of each having multiple may be achieved, wherein an influence degree of each of items, inference data calculation section (26) computes the of wet liquid to flow down. Device (1) diagnosis operation according to, plurality of index item simultaneously target value is imparted in order to satisfy the, setting item by changing a FONT type large in level of influence probability value can be desired product economically since the reaction is. Yet, extracted each of setting item by changing a FONT type tend value to be are taken are obtained. Therefore, this boiler (2) device (1) diagnosis operation of the, unskilled persons without requiring of, boiler (2) operator is capable of checking the driving state of a, plurality of index items simultaneously of target values, which respectively corresponds in order to satisfy the information required for may provide a.

Next, device (1) diagnosis operation driver by using diagnosis method relates to. Figure 5 shows a also, operation diagnosis method main process exhibiting a.. The method diagnosis operation, process a data entry or location sensor data (S1) and, a model read model of Bayesian networks (S2) and read process, for entering of target values, which respectively corresponds item index (S3) and process input target, inference data (S4) and calculating a inference data process, inference for processing data and inference data processing process (S5), (S6) process an output report having.

An interlayer data input process to coat the (S1)

Data input process in (S1), boiler (2) set in the pico-cell information sensor group (14) from the sensor and the, boiler (2) setting item display area device (1) diagnosis operation value is inputs it to a. The (S1) processes data input process, mainly data input (23) by performed at a step. Processes data input process collector includes a plurality of data (S1), 1 ingredient, 1 weeks ingredient, or, such as dog month minute 1. is data in long periods of time. Data, boiler (2) to sensor group (14) and setting item by changing a FONT type data input directly from (23) may is input to, the sensor and the setting item by changing a FONT type the not shown input value recorded on a recording medium, the read from a recording medium, respectively is not used by them.

An interlayer model read process to coat the (S2)

In (S2) process read model, model recording unit (24) Bayesian networks recorded in a model of inference data calculation section (26) a read by. The model the read process (S2), mainly model recording unit (24) by performed at a step. The (S2) process read the model, model of Bayesian networks for updating the probability table which have contained in the determining whether to update judging step (S2a) and, (S2b) and notifies an update process for updating the probability table which have, a reading process model having (S2c).

First, update judging step in (S2a), determines whether a the updating model. Model updating of the when it is determined the (S2a process: YES), the update process of the fluctuation width of the Q (S2b), model updating unit (25), the portable telephone sends update of the model by. Model updating of the determined whether not the (process (S2a): NO), . of the fluctuation width of the Q read process (S2c).

An interlayer (S2b) update process to coat the

The (S2b) update model notifies an update process, mainly model updating unit (25) by performed at a step. First, model updating unit (25) the, data input (23) a plurality of input to the sensor and the, model recording unit (24) having model is recorded in a reads data of a probability table. Here, using a sensor data of model, boiler (2) in the ideal, when comprised of an optical sensing unit requires a is data in.. Next, several value of each sensor data for a predetermined classification into a state, in which the using threshold value of (dispersion) a the dioxide. Furthermore, each classification calculates an of a probability that the situation exists. I.e., data input (23) using sensor data input to, novel probability table data are calculates an. And, recording unit model (24) probability of the probability table which have recorded on data recorded in the machining program, the discharged amount of the gas calculated by introducing a probability data of the probability table which have, probability data for updating a. As such, in a model based [...] bay, easily by introducing model of past, are obtained of the modeled new. Update process (S2b) after executing the, . of the fluctuation width of the Q read process (S2c).

An interlayer (S2c) read process to coat the

Read process in (S2c), model recording unit (24) recorded in a Bayesian networks of model data is inference data calculation section (26) is reading. The read process (S2c), mainly model recording unit (24) and a, inference data calculation section (26) is performed by..

An interlayer (S3) process input target to coat the

In process input target (S3), (S4) process data inference using to accounting for an index item 100 reads of target values, which respectively corresponds. Index items of target values, which respectively corresponds a, boiler (2) or include suction ports and exhaust ports at a higher efficiency a, or, such as reducing dusts environmental load is based on table to allow the maintenance.. These target value, for example, boiler efficiency is "and" and, "low" concentration CO of the exhaust gas such as as, discretized state input device (21) is connected to an input.

An interlayer inference data process (S4) to coat the

Inference data process (S4) the, target an index indicating the input in input process (S3) cuts the condenser element of items, based on model of Bayesian networks, setting data on a plurality of setting item is the degree of be achieved, wherein an influence on item index, index items of target values, which respectively corresponds to satisfy an estimating target value is of setting item by changing a FONT type is process. I.e., index model of Bayesian networks target value is of items is inputted into a, setting corresponding to the use condition item index thereof probability. signal is fed to the input of each item. Inference data process (S4) the, mainly inference data calculation section (26) by performed at a step.

Using a model of Bayesian networks, calculating a probability of setting item by changing a FONT type on the method. A Bayesian networks, based on the conditional probability idea a bay are fabricated by a rope. based on arrangement. First, bay relates to arrangement are fabricated. B idea and a A event occur substantially simultaneously is probability simultaneous of a probability that the.. In this regard, under conditions 404/410/416 A idea B is generated of a probability that the idea, A B under which is conditional probability occurs, after that (1). or presented as a.

[Can 1]

Also, event under conditions 404/410/416 B A is generated of a probability that the idea, B A under which is conditional probability is, after that (2). or presented as a.

[Can 2]

A arrangement are fabricated bay, said type (1) and type (2) after that using (3). is shown.

[Can 3]

Type (3), the laser beam is transmitted through bay are fabricated from arrangement, A B upon generation of the probability of the occurrence of is from ((B |ao A) P), upon generation of B A is the probability of the occurrence of (P (A |ao B)) produce. Generally, and flavors caused by A, B a treated as a result. I.e., when the cause (A), result from the probability of the occurrence of is (B), result (B) upon generation of the probability of the occurrence of the cause (A). calculates a.

Figure 6, boiler (2) setting item display area and index items in is one example of a Bayesian networks. In the present embodiment, and define an (A) causes an setting item by changing a FONT type, index items that result of wet liquid to flow down defined as the (B). Figure 6 shows a also, pixels according to the relation between the water supply flow with a state of the exhaust gas concentration, and water supply temperature and exhaust gas pus relates to a Bayesian networks is shown with codes. Model of Bayesian networks (60) the, cause parent node indicative of flow water supply in (A1) (61) and a, water supply in (A2) cause parent node indicative of a temperature (62) and a, result in exhaust gas concentration (B) includes a root and a node (63) has a. Won in (A1) in parent node (61) the, result in child nodes (B) (63) toward arrows (64) by, child nodes (63) is connected. Yet, cause (A2) in parent node (62) the, result in child nodes (B) (63) toward arrows (65) by, child nodes (63) is connected.

These parent node (61, 62) and child nodes (63) the, the probability table which have (61a, 62a, 63a) constitution: a. Parent node (61) with a probability table number 1 the probability table which have (61a) the, water supply flow to a prescribed threshold using the array block, and by dividing 6 state 2, is indicative of a probability that the each situation exists.. I.e., a random variable, water supply flow rate is "and" 1 on in state, a water supply flow rate is "low" in the range of 0. define an. And, water supply flow is random variable "= 1 and" a2 probability is below a predetermined correspond to when and, when water supply flow rate is "low = 0" is a1 probability is below a predetermined correspond to.

Parent node (62) with a probability table number 1 the probability table which have (62a) the, water supply temperature predetermined threshold using the array block, and by dividing 6 state 2, is indicative of a probability that the each situation exists.. I.e., a random variable, a water supply temperature in the range of 1 on "and", a "low" water supply temperature in the range of 0. define an. And, water supply temperature random variable is "= 1 and" when is a4 probability is below a predetermined correspond to, water supply temperature corresponding to when "low = 0" is a3 size to be.

Child nodes (63) with a probability table number 2 the probability table which have (63a) the, result in child nodes (B) (63) cause of (A1) in parent node (61) and (A2) in parent node (62) is indicative of a probability that the conditional is provided to.. Here, sensor contained in the data index as an item the exemplary exhaust gas concentration. Child nodes (63) with the probability table which have (63a) the, an exhaust gas concentration sensor data is one predetermined threshold using the array block, and by dividing 6 state 2, each sensor data is indicative of a probability that the corresponding to at least two color pixels.. I.e., probability as the parameters, exhaust gas concentrations 1 on "and" in state, exhaust gas concentrations define an a "low" in the range of 0.. For example, water supply flow A1 is "low" and, when A2 is "low" water supply temperature, exhaust of a gas concentration is random variable b11 is one probability of the "low = 0", exhaust of a gas concentration is random variable b21 is one probability of the "= 1 and".

Model of Bayesian networks is (60) from, exhaust of a gas concentration when 1 is (B) random variable, i.e., when "exhaust gas the highly dense", water supply flow rate is constitutively (A1=1) when probability value, water supply temperature caused by probability when (A2=1) constitutively the invention relates to a based aspect arranged by engaging a number.

First, when (B=1) to for higher concentration of exhaust gas, water supply flow rate is constitutively when (A1=1) calculates a probability value (X1). Probability value (X1) the, said type (3) after that the surface of the substrate (4) of hope appeared. such as a. Furthermore, type (4) in, A1 on by merely idea of A1=1, by merely idea of B=1 represented by B of wet liquid to flow down.

[Can 4]

Here, the P (B) molecules, formula (5). or presented as a.

[5 can]

Also, the P (B |ao A1) denominator, formula (6). is shown.

[Can 6]

Said type (4), (5), (6) by, when for higher concentration of exhaust gas, water supply flow provide a high degree of probability value are obtained is (X1).

Next, when to (B=1) for higher concentration of exhaust gas, water supply temperature when (A2=1) constitutively calculates a probability value (X2). Probability value the (X2), said type (3) after that the surface of the substrate (7), such as of hope appeared.. Just, type (7) in, by merely idea of A2 on A2=1, exhibits to B by merely idea of B=1.

[Can 7]

Here, the P (B) molecules said type (5) is calculated by. Also, the P (B |ao A2) denominator, after that (8) or presented as a..

[8 can]

Said type (5), (7), (8) by, when for higher concentration of exhaust gas, water supply temperature is high probability value (X2) is are obtained.

From calculated or more, for higher concentration of exhaust gas when the water flow rate is for higher (B=1) (A1=1) probability of value (X1) and a, for higher concentration of exhaust gas when the water for higher temperature (B=1) (A2=1) probability of value is is calculated on the basis of a (X2). These probability value (X2) by (X1) and probability value, such as temperature water supply flow rate and water supply each setting item is, exhaust gas concentration the degree of be achieved, wherein an influence on items the earth, index items of target values, which respectively corresponds setting item by changing a FONT type of target values, which respectively corresponds to satisfy enabled is to estimate a. For example, probability value the probability value is (X1) (X2) if the reacquisition time is greater than, an exhaust gas concentration index item is the degrees of be achieved, wherein an influence on, water supply than room temperature, . envisioned in of large water supply the flow rate.

As described said, boiler model of Bayesian networks (2) operation of the when opened diagnosis, cause defines item setting (A), the sensor data defined by the result (B). And, setting item by changing a FONT type predetermined condition is when a sensor obtained in a step-of items contained in the data index probability (P (B |ao A)) if of data from the data segment past a, index item is a predetermined target amount when the probability of setting item display area (P (A |ao B)) can be to infer.

Just, a value is taken random variable (described in examples of said, 0 or 1) the, by constraints for of calculated, is handling mounting space is discrete. Therefore, a sensor data take values continuous when the for storing/searching recorded, using predetermined threshold, and, during, low such as discretized performs a process, it is necessary that.

An interlayer inference data processing process (S5) to coat the

As seen to 5 also, inference data processing process (S5) the, inference data calculation section (26) the data in, of report to the window search button includes a window data processing, the portable telephone sends a. Deduced data processing process (S5) the, ranking process (S5a) and, having (S5b) process compared. Ranking process in (S5a), be achieved, wherein an influence on item index is setting item by changing a FONT type (probability value) based on degree of, and decides the data on a plurality of setting items for ranking. Ranking process the (S5a), mainly inference data processing unit (27) by performed at a step. For example, ideal boiler (2) as of target values, which respectively corresponds driving state of a, second reference one, or as small high boiler efficiency, emissions of environmental load. the concentration. These cuts the condenser element from model of Bayesian networks, each indicator corresponding to the use condition item and setting item by changing a FONT type, item for setting the same may take a state where are obtained by the probability. Setting item by changing a FONT type largest value probability is the over number 1, probability in the size through which degrees influence the magnitude of the ranking unit.. Probability of setting item by changing a FONT type item index large value and to determine an order in an order, from large probability setting item by changing a FONT type surface on improved, boiler (2) driving state of a. closer to an ideal state. Also, in (S5b) process compared, boiler (2) setting item display area is state and value, inference data process (S4) calculating an index indicating the item of target values, which respectively corresponds in satisfy a. comparing the setting item by changing a FONT type. By on the comparison, boiler (2) setting item display area is state value, inference data process (S4) calculating an index indicating the item of target values, which respectively corresponds in the setting item by changing a FONT type is judged at if a with a state of the. Furthermore, in (S5) process data processing inference, for each data sensor, and calculates a statistical amount-be evaluated.. The amounts of statistics, average, maximum, such as minimum.

An interlayer output process (S6) to coat the

Output process in (S6), inference data processing process (S5) to given by the cuts the condenser element of setting item by changing a FONT type having a higher, having a higher by using input setting item by changing a FONT type, index target item outputs the information for controlling a. The output process (S6) the, mainly report creation section (28) by performed at a step. Output process created by the data in (S6), output device (22) is output or rinter a line display. Exhibits to 7 having a one example of report. Also as shown in the 7, report (70) the, boiler (2) driving state of a indications of tables displaying objects (71) and a, required to improved operation state for displaying information table (72) constitution: a.

Report (70) of tables, via (71) the, boiler (2) information indicating the operating state of a second window displays IDS. Is table (71) the, index email widow, a web page or displaying objects (71a) and, sensor group (14) an index indicating the acquired by the statistical amount of data item email widow, a web page or for displaying (71b) and, a discretized data of items index indicating the results of this a email widow, a web page or (71c) is connected to the semiconductor layer.. The amounts of statistics, for example, average, maximum, and set that a minimum value in a. Yet, the results discretized, for example, boiler heat, such as hole transport material, "low", "during", state and dioxide such as "and", less than "low" the 90.2, 93.4 less than 90.2 or more "during" the, "and" such as dioxide the 93.4 of at least the used in to a threshold and, the 13.5% and "low", "during" is the 74.5%, "and" such as the 12% that of hope appeared. probability distribution that of each status.

Report (70) of tables, via (72) the, controlled to target item index reproduction status which is information used for boiler (2) of ideal operation state required for the proximity state of information is set. Is table (72) the, ranking process (S5a) to setting having a higher given by the email widow, a web page or displaying objects (72a) and, a preferred setting item by changing a FONT type email widow, a web page or transistor and method for manufacturing the same (72b) and, the ADC input setting item by changing a FONT type having a higher setting item by changing a FONT type email widow, a web page or, a status representing the state of (72c) and, in order to improve representation operation setting item by changing a FONT type email widow, a web page or any diagnostic representing the result (72d) is connected to the semiconductor layer.. For example, blow flow indicative of a heat, is preferably a large flow blow, the current state is during (to, during, bovine), for ratio is 0%, blow flow arc shape to input adjusting the where it is desirable it is found that the.

The resulting structured materials, , the present embodiment regarding a type operation diagnosis according to method, indications of data on a plurality of setting items of each having multiple may be achieved, wherein an influence degree of each of items, using Bayesian networks calculates a. According to method is, plurality of index item simultaneously target value is imparted in order to satisfy the, influence in level of large. is capable of extracting setting item by changing a FONT type. Yet, extracted each of setting item by changing a FONT type tend value to be are taken are obtained. Therefore, this boiler (2) operation of the diagnosis method, unskilled persons without requiring of, plurality of indicator of items at the same time target value is that which is necessary to produce the setting by using the mask pattern..

Also, the present embodiment in regarding a type operation diagnosis method, value is setting item display area, boiler (2) preparing high-purity chiral alcohol an index indicating the acquired by the sensor measuring the ADC sensor in order to acquire data of items data input process (S1) and, based on the converted data and sensor input, selection, parent node (61, 62) with probability table number 1 (61a, 62a) and a, child nodes (63) with number 2 the probability table which have (63a) (S2b) allowing for updating of a updating process is may respectively have functional further. Case an, Bayesian network and each node (61, 62, 63) is including the probability table which have (61a, 62a, 63a) since the is updated, the probability table which have (61a, 62a, 63a) with probability data position 2000. Therefore, setting item by changing a FONT type large in level of influence for extracting can be much accuracy.

Just, said device (1) diagnosis the driver form a surfactant embodiment and operation diagnosis method is one example of is indicated. Device (1) diagnosis operation is provided to the present invention the method and operation diagnosis, said a surfactant embodiment form and not limited to, the claimed in one aspect in a range in which the output without changing the, a surfactant said device (1) diagnosis operation regarding a type embodiment diagnosis and deform a method, and others, and is applied to the is may.

Index item is arranged in a, said described with items that may be utilized in sensor data other.. Yet, the setting item by changing a FONT type, said described item other with items that is may setting item by changing a FONT type.

Industrial applicability

Circulating fluidized bed for boiler according to diagnosis method and operation diagnosis device, unskilled persons without requiring of, plurality of indicator of items at the same time target value is the setting to by using the mask pattern..