JP3384730B2 - Method and apparatus for monitoring state of turbine generator - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for monitoring the state of a water turbine generator, and more particularly to estimating and predicting a change in bearing temperature of a turbine in advance and determining normality / abnormality of the bearing based on the prediction result. The present invention relates to a method for monitoring the state of a water turbine generator and its device.

[0002]

2. Description of the Related Art As the power generation capacity of a hydroelectric power plant increases, the influence on the power system in the event of an accident increases. For this reason, it is necessary to constantly monitor the operating conditions of power plant facilities and develop technology to prevent the occurrence of serious accidents affecting the power system.

[0003] In a hydroelectric power plant, shaft vibration and bearing temperature of a water turbine generator are regarded as process quantities that need to be monitored, and the bearing temperature has been conventionally monitored. This is to detect an increase in the bearing temperature before burning, because an accident in which the bearing burns may occur if foreign matter or the like enters the bearing.

As a related art related to this, for example,
"Hydroelectric power station condition monitoring system" (Mitsubishi Electric Technical Report, Vol.
63, No. 12, 1980). In this conventional technique, the measured value of the bearing temperature is fetched from the plant, and the absolute value of the bearing temperature, the rate of change of the bearing temperature, the temperature difference of each bearing pad, and the like are monitored using a predetermined limit value, and the bearing temperature is abnormal. / Normal judgment is performed.

[0005]

According to the above-mentioned conventional technique, it is determined that an abnormality has occurred when the measured value of the bearing temperature exceeds the limit value. However, simply by monitoring whether or not the measured value exceeds the limit value, even if the turbine generator is stopped when it exceeds the limit value, it cannot be stopped suddenly due to the large inertia of the turbine generator, and the temperature rises further. There is a risk that a burnout accident will occur due to the progress of the above, and it is assumed that it is difficult to prevent a burnout accident.

An object of the present invention is to provide a method and apparatus for monitoring the condition of a water turbine generator, which enables early detection of abnormalities in the water turbine generator, such as an increase in bearing temperature, at the symptom stage.

[0007]

[Means for Solving the Problems] The above object is to provide a turbine generator.
Assuming the value of the calculation coefficient included in the dynamic characteristic model of
Calculate the value of the state quantity of the generator by the dynamic characteristic model,
The calculated value and the measured value obtained during the previous operation of the turbine generator
The value of the calculation coefficient is determined so that the difference between
The dynamic model of the water turbine generator was updated to the latest model.
The predicted value of the state quantity of the turbine generator is predicted from the latest model of
This predicted value and the value obtained during the operation of the turbine generator
This is achieved by determining the abnormality from the difference between the measured values .

[0008] The above object is also to provide a dynamic model of a turbine generator.
State of the turbine generator assuming the value of the calculation coefficient included in
The value of quantity is calculated by the dynamic characteristic model and the calculated value is
Measured value before input of update command obtained during operation of turbine generator and
The value of the calculation coefficient is determined so that the difference between
The dynamic model of the water turbine generator was updated to the latest model.
Of future state quantities of the turbine generator from the latest model of
Calculate the value, and use this predicted value as the preset value for abnormality determination.
This is achieved by determining an abnormality when exceeding .

The above object is preferably achieved by including a state quantity whose predicted value of the future state quantity cannot be measured.

The above-mentioned object is preferably based on the relationship between the two values of the heat generation amount and cooling amount related to the bearing temperature of the water turbine generator, and the calculation coefficient of the dynamic characteristic model related to this heat generation amount and cooling amount. To calculate the bearing temperature by the dynamic characteristic model, determine the value of the calculation coefficient so as to minimize the difference between the calculation result and the measured value of the bearing temperature, and update it to the latest model. It will be achieved.

The above object is preferably achieved by predicting a future change in at least one of the bearing temperature, the bearing lubricating oil temperature, the shaft temperature, and the bearing gap width from the updated latest model.

The above object can also be achieved by detecting an abnormality using the value of the determined operation coefficient.

[0013]

According to the present invention, it is possible to detect an abnormal change in the bearing temperature or the like from the prediction result before the measured value such as the bearing temperature reaches the limit value. Furthermore, using the estimated values of the calculation coefficients related to the heat generation capacity and the cooling capacity, abnormalities occurring in the bearings of the turbine generator etc. are detected early before they have a large effect on temperature changes. It becomes possible.

[0015]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block configuration diagram of a state monitoring device for a water turbine generator according to an embodiment of the present invention. In FIG. 1, 1 is a turbine generator, 2 is a data capturing device for capturing measurement data from the turbine generator, and 3 is a data storage unit for storing the captured measurement data. Reference numeral 4 denotes a calculation coefficient estimation unit, which uses the data stored in the data storage unit 3 and the dynamic characteristic model 6 that simulates the dynamic characteristic of the water turbine generator 1 to generate heat and remove heat included in the dynamic characteristic model 6. Estimate the value of the calculation coefficient. The calculation coefficient will be described with reference to FIG. The value of the estimated calculation coefficient is stored in the calculation coefficient storage unit 7
Prediction of the bearing temperature is carried out by the prediction processing unit 8 using this value. This prediction result is stored in the prediction result storage unit 13.

Reference numeral 5 denotes a processing control unit, which is an arithmetic coefficient estimation unit 4
And the display processing on the prediction processing unit 8 and the display device 10. Further, the input device 9 is an input device for a user such as an operator to perform input related to processing and display. 11
Is an abnormality determination unit that determines an abnormality of the water turbine generator using the calculation coefficient value stored in the calculation coefficient storage unit 7, the prediction result from the prediction processing unit 8 and the abnormality determination data.

FIG. 2 is a diagram for explaining the outline of the turbine generator. As shown in FIG. 2, in the turbine generator, the rotor 20 and the turbine runner 21 of the generator are fixed to one shaft 22, and these rotate together. Upper bearing 23, lower bearing 24, thrust bearing 25, and turbine bearing 26 are provided for supporting the shaft 22 and for smooth rotation.

The water turbine bearing 26 is provided with an oil tank 28 for removing heat generated in the bearing gap 27 between the bearing 26 and the shaft 22. Further, the cooling pipe 29 is provided in the oil, and the heat transferred to the oil in the oil tank 28 from the bearing gap 27 is removed by the cooling water.

Although the heat removal mechanism using the oil tank is shown by taking the water wheel bearing 26 as an example, the other bearings 23, 24 and 25 are also provided with the same heat removal mechanism.

FIG. 3 is a diagram summarizing the main factors related to the bearing temperature in the water turbine generator shown in FIG. As shown in the figure, the bearing temperature is determined by the relationship between the heat generated in the bearing gap, that is, the heat generation amount, and the heat removed from the bearing, that is, the heat removal amount. The main factors that determine the amount of heat generation are the bearing gap width, the rotational speed of the turbine and the oil temperature. On the other hand, as factors related to the amount of heat removed, the heat passage rate, oil temperature, which is a coefficient related to heat conduction and heat transfer in a cooling pipe,
Examples include cooling water temperature and shaft temperature.

FIG. 4 is a dynamic characteristic model in which the relationship between these factors is analyzed and associated with each other. The heat generation amount determined depending on the oil viscosity determined by the bearing gap width, the rotation speed, and the temperature of the lubricating oil in the oil tank. , The cooling water, the amount of heat removed by the surrounding air, etc. are expressed. In this dynamic characteristic model, the relationship between the measured value and the state inside the model that cannot be measured is shown. Further, in this dynamic characteristic model, the heat transmission rate of the cooling pipe, the initial value of the axial temperature, and the reference value of the gap width, which cannot be directly determined from the measured values, are treated as calculation coefficients related to heat generation and heat removal.

The turbine generator is repeatedly started and stopped more frequently than nuclear power generation or thermal power generation in order to respond to changes in power demand. Therefore, the values of the above-described calculation coefficients are estimated and determined based on the measured values obtained during the operation of the water turbine generator. That is, the model is updated to the latest model based on the measurement data obtained during operation. Then, using this latest model, future changes in the bearing temperature and oil temperature of the turbine generator are predicted. Regarding the calculation coefficient, it is possible to arbitrarily select an amount that is related to the heat generation amount or the heat removal amount and cannot be directly determined from the measured value. For example, in the device of the present embodiment, it is also possible to treat the heat transmission rate on the bearing pad surface, which is treated as a constant, as a calculation coefficient. In this embodiment, three state quantities that have a great influence on the heat removal amount and the heat generation amount are selected as the calculation coefficients.

FIG. 5 is a diagram for explaining the outline of the processing in the state monitoring apparatus according to the above embodiment. In this state monitoring device, the data capturing device 2 captures each measured value of the bearing temperature, the rotation speed, the oil temperature, the cooling water temperature, and the temperature of the surrounding environment of the turbine generator. After this,
Using the dynamic characteristic model and each measurement value, the calculation coefficient estimation unit 4 estimates the value of each calculation coefficient described above (gap width reference value, cooling pipe heat passage rate and axial temperature initial value). Once the values of these calculation coefficients are determined, the dynamic characteristic model is confirmed,
By giving the rotation speed of the turbine generator as the boundary condition, it is possible to predict the state change in the future operation based on the latest dynamic characteristic model. As the prediction result, a result related to the state quantity included in the dynamic characteristic model of FIG. 4 is obtained. That is, changes in the bearing temperature, oil temperature, shaft temperature, bearing gap width, etc. can be predicted.

FIG. 6 is a flow chart showing the processing carried out by the above-mentioned arithmetic coefficient estimating section. The arithmetic coefficient estimator 4 (FIG. 1) periodically determines whether or not a processing start command has been input (step 101). This processing start command is input as a demand from the user or automatically by the processing control unit 5. The automatic process start command is generated, for example, at a preset cycle.

When the processing start command is input, the measurement data from time t0 to time ts is fetched from the data storage unit 3 (step 102). Here, the time ts is the time when the processing start command is input, and the time t0 is the time before the time ts by the predetermined data fetching time. Then, the value of the calculation coefficient of the dynamic characteristic model is assumed (step 10
3), changes in bearing temperature, oil temperature, etc. from time t0 to time ts are calculated by the dynamic characteristic model (step 10).
Four).

Calculation processing using this dynamic characteristic model (step
The processing procedure of 104) is shown in FIG. First, the time t is set to t0 (step 201). Next, the initial value of the dynamic characteristic model is set using the measured value and the value of the assumed calculation coefficient (step 202). The dynamic characteristic model in Fig. 4 is the oil temperature T
It is expressed by three first-order time differential equations for o, shaft temperature TA, and bearing temperature TJ. In setting the initial values, the values of the oil temperature To, the bearing temperature TA, and the shaft temperature TJ and their time differential values are set.

The values of the oil temperature To, the bearing temperature TA and the shaft temperature TJ after Δt are calculated using the set initial values (step 203). In this calculation, the value of the state quantity at time t + δt is calculated by adding the value obtained by multiplying the time derivative of each state quantity at time t by the time step width δt to the value at time t. This method is often used to calculate the time change of the state quantity using the equation of time differential, and is usually called the Euler method.

Next, using the value of each state quantity at the time (t + δt) calculated in step 203, the time (t + δ)
Calculate the time derivative of each state quantity at (t) (step 2
04). Next, it is determined whether or not the time t is equal to or longer than the calculation end time tE (step 205), and the process is ended if the condition is satisfied. On the other hand, if the time tE is not yet exceeded, the time t is set to t + δt (step 20
6), the processing from step 203 is repeated and executed. Here, the calculation end time tE is set to the time ts when the processing instruction is input in the calculation coefficient estimation.

When the above-described calculation procedure of step 104 is completed, the process proceeds to step 105 of FIG. 6, and the difference J is calculated by the following equation 1 using the calculated value and the measured value.

[0030]

[Equation 1]

As shown in FIG. 8, the difference J is an amount showing the difference between the state quantity calculated by the dynamic characteristic model and the measured value of the plant. That is, Equation 1 is an amount indicating the difference between the calculated value and the measured value for the oil temperature and the bearing temperature. Here, both oil temperature and bearing temperature were used to calculate the objective function. However, it is also possible to use either one of them, for example, only the bearing temperature, in the calculation of the objective function. That is, the objective function can be calculated by using the state quantity that can obtain both the measured value and the calculated value.

As shown in FIG. 6, the calculation coefficient estimation unit obtains the value of the calculation coefficient that minimizes the difference J (step 10).
6). In the state monitoring device according to the present embodiment, the steepest descent method, which is well known as a non-linear optimization method, is used to obtain the calculation coefficient value. That is, the difference J between the calculated value and the measured value is used as the objective function, and the value of the operation coefficient that minimizes this value is determined by the steepest descent method.

FIG. 9 conceptually shows this method. FIG. 9 shows a state of optimization in a two-dimensional space, that is, when there are two calculation coefficients. In the optimization, P0 is given as the initial search point. This point corresponds to the value of the calculation coefficient assumed in step 103 of FIG. The gradient ∇J at this point P0 is calculated, and a one-dimensional search is performed in the direction of −∇J, that is, in the direction in which the objective function value decreases most, and on the line along this direction, the objective function value J is Find the minimum point.
Then the gradient at this point is calculated and a one-dimensional search is performed in this direction. The same process is repeated until the value of the objective function J is no longer improved, and the optimum point is obtained. The coordinate value of this optimum point corresponds to the variable value that minimizes the objective function.

FIG. 10 is a flow chart showing the processing procedure of this optimization processing. First, the value of the assumed calculation coefficient is set as an initial value (step 301). Then, the objective function value J and its gradient ∇J at the point in the search space corresponding to the value of the calculation coefficient, that is, the initial search point are calculated (step 302). A one-dimensional search is performed in the direction in which the objective function value decreases most, that is, in the direction of −∇J, and the point at which the objective function value J is minimized is obtained (step 303). Based on this result, it is determined whether the objective function value J has become smaller than the value ε given in advance as the ending condition (step 30).
Four). If it becomes smaller, the value of the operation coefficient is determined and the process is terminated (step 306). On the other hand, when J is ε or more, the value of the operation coefficient is corrected corresponding to the coordinates of the point obtained by the one-dimensional search (step 305) and the processing from step 302 is repeated.

In this way, when the value of the arithmetic coefficient is determined in step 106 of FIG. 6, the value of the arithmetic coefficient is stored in the arithmetic coefficient storage unit 7 in the next step 107, and the estimation of the arithmetic coefficient is finished. Notify the processing control unit 5.

Next, the processing in the prediction processing unit 8 for predicting future changes in the state quantity using the estimation result of the calculation coefficient will be described. FIG. 11 is a flowchart showing the processing in the prediction processing unit. The prediction processing unit periodically determines whether or not a processing start command is input from the processing control unit 5 (step 401). This processing start command is output from the processing control unit 5, for example, at the time when the processing coefficient estimation unit 4 notifies the processing control unit 5 of the end notification of the processing for estimating the calculation coefficient.

When the processing start command is input, first, the value of the calculation coefficient is read from the calculation coefficient storage unit 7 (step 4
02). Next, the value of the measurement data at time t0 is fetched from the data storage unit 3 (step 403), and the initial value of the dynamic characteristic model is set using this value and the value of the calculation coefficient (step 404). Next, the rotational speed change and the prediction end time tE are fetched from the processing control unit (step 40).
5), change in state quantity until time tE is calculated (step 40)
6). This calculation is performed by the method shown in FIG.
After the calculation is completed, the prediction result about the state quantity change is
Output to the display device 10. Further, the processing control unit 5 is notified that the prediction calculation is completed (step 408).

FIG. 12 is a flow chart showing a processing procedure in the processing control section 5. In the processing control unit, when a prediction request is input (Step 501), or when the prediction processing is automatically executed (Step 502) or when the plant state is reached (Step 503), the calculation coefficient estimation unit A processing start command is output to (step 504). Here, for automatic execution of the prediction process,
The execution time and the plant status to be executed are given in advance. On the other hand, the calculation coefficient estimation end signal indicates that the calculation coefficient estimation unit
When input from 4 (step 601), the prediction processing unit 8
The prediction start instruction is output to (step 602).

In the processing control section 5, as other processing,
A processing command is transmitted to the abnormality determination unit 11 based on the operation statuses of the calculation coefficient estimation unit 4 and the prediction processing unit 8. The process relating to this abnormality determination will be described later.

FIG. 13 is a diagram showing a state of the operation relating to the state prediction. FIG. 13 is a graph comparing the plant measured value and the model calculated value for the bearing temperature according to the operation of the condition monitoring apparatus according to the present embodiment. In the figure, the time
At ts , the processing coefficient estimation unit 4 receives a processing start command. When this processing start command is input, the calculation coefficient estimation unit fetches the measurement data from time t0 to ts immediately before the input of the processing start command, out of the measurement data that are sequentially measured and stored in the storage device. Then, assuming the initial value of the calculation coefficient of the dynamic characteristic model, the model is initialized at time t0, and then the changes in the bearing temperature and the oil temperature from time t0 to ts are calculated by the dynamic characteristic model. This result is shown in
It is a graph of 1001 of FIG.

The calculation coefficients relating to the heat generation amount and the heat removal amount are estimated so that the difference between the result 1001 and the plant measured value 1002 is minimized in the time from time t0 to time ts. The graph 1003 in the figure shows the result of calculation using the dynamic characteristic model using the calculation coefficient. In this way, the difference between the measured value and the model calculated value from time t0 to ts becomes small. Further, by extending the calculation in the dynamic characteristic model until time tE, the prediction result of the bearing temperature can be obtained as shown in the figure.

FIG. 14 is a diagram showing an example of a display screen of the status monitoring apparatus according to this embodiment. In this figure, 1101
Is a window for the operator to input the future operation policy, that is, the stop time or the time change of the rotation speed of the water turbine generator.In the present embodiment, this window is used for simulating the water turbine generator by the latest model. Using 1101, the operator inputs the operation policy of the water turbine generator after the present time, that is, the rotation speed change policy. Here, when the stop time is input to stop after 100 seconds in this illustrated example, at a constant rotation speed until the stop time, after that, based on the change curve of the rotation speed prepared in advance, the dynamic characteristics Determine the input value of the rotation speed to the model. The change situation of the rotation speed given to the dynamic characteristic model is shown in the window 1102.
Is displayed in. In this window 1102, the measured value of the rotation speed up to the present time is also displayed.

As a result of the prediction using the dynamic characteristic model, the prediction results of the oil temperature and the bearing temperature can be obtained. In the example of FIG. 14, the prediction result of the oil temperature and the measured value up to the present time are combined and displayed in the window.
Displayed in 1103. The bearing temperature is also displayed like the window 1104. In this window, the set value of the protection relay is displayed in addition to the predicted temperature value.If the bearing temperature exceeds the set value of the protection relay, a warning is given to the operator as shown in the figure. Is displayed. This allows the operator to know in the future that the bearing temperature may exceed the set value of the protection relay and how many minutes there is until that time. Here, the abnormality determination unit 11 determines whether the predicted value of the bearing temperature exceeds the set value of the protection relay or the like. The abnormality determination unit 11 determines that the prediction result has been obtained based on the information from the process control unit 5, and uses the prediction result and the set value for abnormality determination prepared in the abnormality determination data 12, Determine the temperature abnormality. The result is
It is displayed on the display device 10 via the processing control unit 5 as shown in FIG.

In addition to the temperature change, the condition monitoring apparatus according to the present embodiment can also obtain a prediction result of the bearing gap width, which is a main factor that determines the heat generation amount in the bearing gap. The result is shown as the window 1105 in FIG. 14, and the operator can know the change in the gap width. As a result, the operator can judge whether the change in the bearing gap width that cannot be measured is normal or abnormal.

An example of the abnormality determination processing in the abnormality determination unit 11 has been shown above, but another abnormality determination processing executed in the abnormality determination unit will be described next. FIG. 15 shows one of the abnormality determination methods, for example, when the turbine generator is started. In this method, of the calculation coefficient values obtained based on the data from the previous plant operation, the gap width reference value and the cooling pipe heat transfer rate are used to predict the state quantities such as the bearing temperature at the start of the plant. . For the initial value of shaft temperature,
The measured value of the bearing temperature is used instead because it is generally different from the time when the calculation coefficient value was estimated. This is based on the assumption that the difference between the shaft temperature and the bearing temperature is small when the turbine generator is stopped. The prediction result and the measurement value from the plant are sequentially compared, and when the difference ΔT between the prediction result and the measurement value exceeds a predetermined determination reference value, it is determined to be abnormal.

In this method, since the calculation coefficient value estimated by using the measured data in the previous plant operation is used,
It is not necessary to collect measurement data for obtaining the calculation coefficient value at the time of starting this time, and the prediction calculation can be started within a short time after starting. Therefore, it is possible to deal with the occurrence of an abnormality immediately after startup. Further, by using the calculation coefficient value based on the previous operation data, it is possible to detect the difference from the plant state at the previous operation. That is,
In the method of periodically repeating the above-described prediction using the calculation coefficient value based on the measurement data from time t0 to ts, the value of the calculation coefficient is determined by reflecting the plant state at that time. That is, when an abnormality occurs, a value representing the plant characteristic at the time of abnormality is set as the calculation coefficient, and the model changes in accordance with the abnormal state.

This method is effective in accurately predicting future conditions such as bearing temperature. However, from the viewpoint of detecting the deviation from the normal state, the method shown in FIG. 15 is more effective. In other words, assuming that the plant state during the previous operation is normal, the prediction result based on this is considered as a normal change in bearing temperature, and when a change different from this normal change appears, it is possible to detect the occurrence of an abnormality. it can.

In this method, as the value of the operation coefficient based on the previous operation data, the above-mentioned operation coefficient value obtained using the data from time t0 to ts is the final value in the operating state, that is, immediately before the stop operation. A calculation coefficient value can be used. It is also possible to select a part of the operation data obtained from the previous start-up to immediately before the stop, and calculate and use the calculation coefficient value based on it. For example, it is possible to use the calculation coefficient value estimated based on the measurement data in the entire operation time from the time of starting to the time immediately before stopping. It is also possible to use a calculation coefficient value based on the data at the time when the bearing temperature rises slowly and the temperature is almost saturated. By selecting an appropriate method from these among them according to the characteristics of the turbine generator, it is possible to carry out the abnormality determination with high accuracy.

Here, for example, in the case of using the calculation coefficient value obtained based on the measurement data in the entire operation time from the start-up to immediately before the stop, the calculation coefficient estimation process is performed to request the stop of the turbine generator. This is performed by outputting a processing request from the processing control unit 5 to the arithmetic coefficient estimation unit 4 when a signal is input. Further, the prediction process is performed by outputting the process request from the process control unit 5 to the prediction processing unit 8 at the time when the start request signal for the water turbine generator is input.

Up to now, the state prediction and abnormality determination at the time of starting the water turbine generator have been described. The same process as this can be performed in the same way as the process of stopping the turbine generator. FIG. 16 shows a method for determining an abnormality in the process of stopping the turbine generator. As shown in FIG. 15, similarly to FIG. 15, the prediction using the operation coefficient value estimated from the last operation data, that is, the measurement data in the stop process of the last operation is performed by using the stop request signal of the water turbine generator as a trigger. However, by successively monitoring the difference between the prediction result and the measured value, it is possible to detect the abnormality of the turbine generator. In this case as well, of the calculation coefficient values obtained based on the data from the previous plant operation, the gap width reference value and cooling pipe heat transfer rate are used to predict the state quantities such as bearing temperature at plant startup. To do. Since the initial value of the shaft temperature is generally different from the time when the calculation coefficient value is estimated, the measured value of the bearing temperature is used instead.

Here, the case where the calculation coefficient value based on the operation data in the last stop process is used is shown. Of the measured data from the start to the stop in the present operation,
It is also possible to obtain a calculation coefficient value based on data of an appropriate time and use it to predict the stopping process to detect an abnormality. It is also possible to detect an abnormality by obtaining the calculation coefficient value based on the measurement data obtained immediately before during the stop process of the present operation and performing the subsequent prediction of the present stop process. .

Next, another abnormality detecting method will be described. In the condition monitoring device according to the present embodiment, the gap width reference value is obtained as the calculation coefficient related to the bearing gap width. This state quantity is an amount corresponding to the gap width when there is no change in the bearing gap width due to temperature change, that is, when the plant is cold stopped. Using this estimate,
Abnormality detection as shown in FIG. 17 can be performed. That is, one or a plurality of suitable bearings are selected from the plurality of bearings installed around the shaft, and the gap width reference value for this is stored. By monitoring the change, it is possible to detect the change state of the bearing gap width, that is, the inclination of the shaft, the deviation of the bearing, the deformation, and the like.

As described above, according to this embodiment,
It is possible to obtain the prediction result about the future state of the state quantity such as the bearing temperature related to the bearing. Using this result,
It is possible to detect an abnormality early before the state quantity of the plant reaches a limit value or the like. Further, by comparing the estimated prediction result in the model based on the normal operating state in the past with the sequentially measured data, the deviation from the normal state can be detected at an early stage. Further, it is possible to detect the deformation of the bearing, the deformation of the shaft, etc. based on the change history of the calculation coefficient such as the reference value of the bearing gap width.

In the determination of the calculation coefficient in the above-mentioned embodiment, the optimization method is implemented so as to reduce the difference between the result calculated using the calculation coefficient and the measurement data, and the determination can be made by a simple process. Is. In addition to the bearing temperature and the oil temperature, it is possible to display to the operator, etc., changes in the bearing temperature, oil temperature, and unmeasured shaft temperature, bearing gap width, cooling pipe heat transfer rate, and the like. . As a result, effective information can be provided to the operator, and it becomes possible to assist the operator in determining the operation policy.

In the above-described embodiment, the model is updated to the latest model based on the measured value, and the example of performing the simulation of the water turbine generator using this latest model has been described. However, the model is included in the updated latest model. It is used to perform offline simulations by changing the values of parameters such as heat transfer rate, predict changes in bearing temperature with changes in parameters, obtain knowledge about plant characteristics, and study the causes of changes in plant characteristics. It is also possible.

[0056]

According to the present invention, it is possible to accurately detect an abnormality related to a bearing of a water turbine generator or the like at a symptom stage before the abnormality is actualized and present it to an operator. In addition, it is possible to display the predicted results of the bearing temperature, the oil temperature, etc., and the change status of the shaft temperature, the bearing gap width, the heat transfer coefficient of the cooling pipe, etc., which are not measured, to the operator or the like. Based on such information, operators can recognize abnormal events at an early stage, easily and quickly formulate a plant operation policy when an abnormality occurs, and take corrective actions to maintain plant safety and soundness. It is possible to execute surely and quickly. As a result, the economical efficiency and safety of the hydroelectric power plant can be greatly improved.

[Brief description of drawings]

FIG. 1 is a block configuration diagram of a state monitoring device according to an embodiment of the present invention.

FIG. 2 is a schematic diagram showing the structure of a turbine generator.

FIG. 3 is a schematic diagram summarizing factors related to bearing temperature.

4 is a block diagram showing a dynamic characteristic model used in the state monitoring device shown in FIG. 1. FIG.

5 is a schematic diagram showing an outline of processing in the state monitoring device shown in FIG.

6 is a flowchart showing a process in a calculation coefficient estimation unit shown in FIG.

FIG. 7 is a flowchart showing a process of calculation related to a dynamic characteristic model.

FIG. 8 is a schematic diagram showing a method of calculating an objective function.

FIG. 9 is a schematic diagram showing an optimization method.

FIG. 10 is a flowchart showing an optimization process.

FIG. 11 is a flowchart showing processing in a prediction processing unit shown in FIG.

12 is a flowchart showing a process in a process control unit shown in FIG.

FIG. 13 is a schematic diagram showing an operation example of the state monitoring device shown in FIG. 1.

FIG. 14 is a schematic diagram showing an example of a screen display of the state monitoring device shown in FIG.

15 is a schematic diagram showing a method of determining an abnormality in bearing temperature of the condition monitoring device shown in FIG.

16 is a schematic diagram showing a method for determining an abnormality in bearing temperature of the condition monitoring device shown in FIG.

FIG. 17 is a schematic diagram showing a method for detecting an abnormality in a bearing gap.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Water turbine generator, 2 ... Data acquisition device, 3 ... Data storage part, 4 ... Calculation coefficient estimation part, 5 ... Processing control part, 6 ... Dynamic characteristic model, 7 ... Calculation coefficient storage part, 8 ... Prediction processing part, 9 ...
Input device, 10 ... Display device, 11 ... Abnormality determination unit, 12 ...
Abnormality determination data, 13 ... Prediction result storage unit 101 to 1
08 ... Process steps, 201-206 ... Process steps, 301-306 ... Process steps, 401-408
… Processing steps, 501 to 504… Processing steps,
601, 602 ... Process steps, 1001 ... Model calculated value, 1002 ... Plant measured value, 1003 ... Model calculated value, 1101-1105 ... Display window.

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kazuo Moriguchi 3-1-1, Saiwaicho, Hitachi, Ibaraki Hitachi Ltd. Hitachi factory (72) Inventor Tominobu Uchida 4-6, Kanda Surugadai, Chiyoda-ku, Tokyo Hitachi, Ltd. (72) Inventor Jun Yamamoto 3-3-22 Nakanoshima, Kita-ku, Osaka City, Osaka Prefecture Kansai Electric Power Co., Inc. (72) Shunji Komon 3-32 Nakanoshima, Kita-ku, Osaka City, Osaka Prefecture No. Kansai Electric Power Co., Inc. (72) Inventor Hiroto Nakagawa 3-22 Nakanoshima 3-chome, Kita-ku, Osaka-shi, Osaka Seisai Honda Co., Ltd. (72) Seiji Honda 3-chome Nakanoshima, Kita-ku, Osaka-shi, Osaka No.22 in Kansai Electric Power Co., Inc. (56) References JP-A-9-111320 (JP, A) JP-A-64-26912 (JP, A) Actual development Shou 63-92215 (JP, U) (58) Field (Int.Cl. ^7, DB name) H02P 9/00 G05B 13/04 G05B 23/02

Claims (12)

(57) [Claims] 1. A method of monitoring the state of a turbine generator, wherein the value of the state quantity of the turbine generator is calculated by the dynamic model assuming the value of a calculation coefficient included in the dynamic model of the turbine generator. , Updating the dynamic model of the turbine generator to the latest model by determining the value of the calculation coefficient so that the difference between the calculated value and the measurement value obtained at the previous operation of the turbine generator is minimized, A turbine generator characterized in that a predicted value of the state quantity of the turbine generator is obtained in advance from the latest model, and an abnormality is determined from a difference between the predicted value and a measurement value obtained while the turbine generator is operating. How to monitor machine status. 2. A method of monitoring the state of a turbine generator, wherein the value of the state quantity of the turbine generator is calculated on the basis of the value of a calculation coefficient included in the dynamic model of the turbine generator. , The value of the calculation coefficient is determined so that the difference between the calculated value and the measured value before the input of the update command obtained during the operation of the turbine generator is minimized, and the dynamic characteristic model of the turbine generator is updated to the latest model. To obtain a predicted value of the future state quantity of the water turbine generator from the latest model, and to determine an abnormality when the predicted value exceeds a preset set value for abnormality determination. Generator status monitoring method. 3. The method for monitoring the state of a water turbine generator according to claim 2, wherein the predicted value of the future state quantity includes a state quantity that cannot be measured. Updates to the claim 4, wherein the latest model, based on the two amounts of the relationship between the calorific value and the amount of cooling of the bearing temperature of the water turbine generator, the dynamic characteristic model of the emitting amount of heat and the amount of cooling And calculating the bearing temperature by the dynamic characteristic model, and determining the value of the calculation coefficient so that the difference between the calculation result and the measured value of the bearing temperature is minimized. The method for monitoring the state of a water turbine generator according to any one of claims 1 to 3. The latest model wherein the said updating, bearing temperature, either the lubricating oil temperature of the bearing, shaft temperature, claim 1 characterized in that predicting at least one future change in bearing gap width 4 The method for monitoring the condition of the turbine generator described in. 6. A condition monitoring method of water turbine generator according to any one of claims 1 to 4 which is characterized in that the abnormality is detected using the value of the determined operation coefficient. 7. A state monitoring device for monitoring the state of a hydraulic turbine generator, wherein the value of the state quantity of the hydraulic turbine generator is determined by the dynamic characteristic model assuming the value of a calculation coefficient included in the dynamic characteristic model of the hydraulic turbine generator. Calculate and determine the value of the calculation coefficient so that the difference between the calculated value and the measured value obtained during the previous operation of the turbine generator is minimized and update the dynamic model of the turbine generator to the latest model Means, a means for obtaining a predicted value of the state quantity of the turbine generator from the latest model in advance, and a means for determining an abnormality from a difference between the predicted value and a measured value obtained while the turbine generator is in operation. And a condition monitoring device for a turbine generator. 8. A state monitoring device for monitoring the state of a hydraulic turbine generator, wherein the value of the state quantity of the hydraulic turbine generator is calculated by the dynamic characteristic model assuming the value of a calculation coefficient included in the dynamic characteristic model of the hydraulic turbine generator. Calculate the dynamic characteristic model of the water turbine generator by determining the value of the calculation coefficient so that the difference between the calculated value and the measured value before the input of the update command obtained during operation of the water turbine generator is minimized. Means for updating to the latest model, means for obtaining a predicted value of the future state quantity of the water turbine generator from the latest model, and determining that the abnormality is abnormal when the predicted value exceeds a predetermined set value for abnormality determination And a state monitoring device for a water turbine generator. 9. The state monitoring device for a water turbine generator according to claim 8, wherein the means for obtaining a predicted value of the future state quantity obtains a predicted value including a state quantity that cannot be measured. 10. The dynamic means relating to the heat generation amount and the cooling amount is based on the relationship between the heat generation amount relating to the bearing temperature of the water turbine generator and the cooling amount, and the dynamic characteristic relating to the heat generation amount. The calculation coefficient of the model is adjusted, the bearing temperature is calculated by the dynamic characteristic model, and the value of the calculation coefficient is determined so that the difference between the calculation result and the measured value of the bearing temperature is minimized. The condition monitoring device for a water turbine generator according to any one of claims 7 to 9. 11. The means for obtaining the predicted value is characterized by predicting at least one future change in bearing temperature, bearing lubricating oil temperature, shaft temperature, and bearing gap width from the updated latest model. A condition monitoring device for a water turbine generator according to any one of claims 7 to 10. 12. The method of claim 11, wherein the abnormality determination means, the state monitor hydraulic turbine generator according to any one of 10 claims 7 characterized in that to detect an abnormality by using the value of the determined operation coefficient.