JP2772179B2 - Plant operation data management device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a plant operation data management apparatus for managing plant operation data and monitoring the operation state of the plant.

[0002]

2. Description of the Related Art Conventionally, methods for detecting abnormalities in the operation state of a power plant include checking with an alarm limit value,
Some are based on a sensor check function.

The alarm limit value is set on the basis of a value that should not exceed this value due to the design of the plant, and should not exceed because it is controlled so as not to exceed it. . Therefore, under normal operating conditions, the alarm limit value is rarely reached.

[0004] The sensor check function focuses on the output of the plant and determines and checks the value of each input point of the plant when operating at a constant output for a fixed time. is there. Currently, in plants that use the sensor check function,
When 100%, 75%, 50%, and 25% output bands are operated with the same output for more than one hour, sensor checks are performed. The management check value of each point in each output band is manually set one by one, or the past operation state is stored and obtained by statistical calculation.

[0005]

In recent years, in power plants, due to the aging of the power plant and the shortage of electric power, it has been required to establish early detection of abnormalities and establishment of diagnosis technology in order to improve the reliability of plant equipment and to stably supply electric power. Have been.

[0006] Plant abnormalities are roughly classified into two types.

(1) Sudden abnormalities For example, abnormal jellyfishes block the cooling water intake port due to abnormalities, or an earthquake causes the pump to be damaged.

(2) Temporal abnormalities For example, the bearings of the pump are worn out, the shaft vibration is gradually increased, the scale is attached to the piping, and the pressure loss is increased.

[0009] Among them, it is desired to detect the secular abnormality early (2). However, in the conventional method using the alarm limit value, the alarm is notified after the abnormality occurs. When it is difficult to continue the operation due to the large size, it is necessary to immediately stop the pump and eventually the plant. To maintain the equipment, it is sufficient to stop the equipment, but there are many problems in terms of stable power supply. In other words, if one power plant is urgently stopped in a power shortage situation in the summer, some of the customers will experience a power outage.

[0010] If an abnormality is found early, it is reflected in the power generation plan at the central power station, and after taking measures such as starting up a standby thermal power plant, interchanging with other electric power, and limiting supply to large customers, the abnormal plant It is not necessary to stop the power outage, so that the power outage does not occur.
Therefore, it is important how to catch a sign of an abnormality immediately before an alarm and to gain time for performing a necessary treatment.

[0011] From such a viewpoint, there is a method of discovering something deviating therefrom at an early stage on the basis of a normal indicated value of each part of the plant. Sensor check function is part 1
However, this has the following disadvantages: (1) Since the output of the plant is used as a parameter, it is not effective for points that do not have a direct causal relationship with the plant output. For example, in a plant with three pumps, with a plant output of 50%, two of them operate, but it is not known which two. in this case,
It is not possible to perform sensor checks on the flow rate, vibration, and temperature of the three pumps.

(2) Since it is based on a check in a stable state of the plant when operating at the same output for a certain period of time,
It is not suitable for thermal power plants with large output changes.

In the power plant, an operation log is output, and the power generation person looks at the values on this log sheet to determine the validity of the values, grasp the operation status, and estimate the abnormality.

However, the power generation staff who can perform this inspection requires experience and judgment of the plant, and the log point per power generation unit is 100 points or more, and 3 to 8 units are present in the power plant. For this reason, checking logs output daily is a considerable burden for skilled operators.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned conventional problems, and seeks a management limit value suitable for judging a tendency of abnormality based on the operation results of a plant. By checking various quantities of
An object of the present invention is to provide a plant operation data management device capable of early detection of an abnormality.

[0016]

In order to achieve the above object, a plant operation data management device according to the present invention comprises a data storage means for inputting and storing operation data of each point of a plant, a point to be managed, Point setting means for setting at least one reference point corresponding to each management point; and data of the management point set by the point setting means and corresponding reference point data from the data stored by the data storage means. A data search unit that is periodically extracted for a predetermined period, and a management value that statistically processes the data extracted by the data search unit and obtains a management limit function that is a function of a management limit management limit value using the reference point value as a variable. The calculation means and the management restriction function obtained by the management value calculation means are displayed in a graph, Providing a display means for displaying the latest operation data and any past operation data.

In the above plant operation data management device, the management value calculating means divides a range between a maximum value and a minimum value that can be taken by the reference point into a mesh having a predetermined interval,
The average, variance, and management limit value of the management points are calculated for each mesh, and the management limit values obtained for the number of meshes are regression-analyzed to obtain a management limit function.

Further, the plant operation data management device displays the operation data used for calculation on a screen displaying a graph of the management restriction function obtained by the management value calculation means, and modifies the coefficient of the management restriction function. Means.

Further, the plant operation data management device includes a comparing means for comparing the latest operation data with a management restriction function set for each management point, and when there is a point exceeding the management restriction, this is used. Output means for notifying.

[0020]

In the plant operation data management device of the present invention, the input data of the plant is fetched at regular intervals by the data storage means and its value is stored, and the relationship between the management target point and its reference point is stored by the point setting means. Based on this, the stored data is extracted for a necessary period by the data search means, statistically processed by the management value calculation means, and the management limit value of the management target point is calculated as a function of the reference point based on the data. By displaying the target date data and the latest data and the function curve of the obtained management limit value on such display means,
It is possible to judge the suitability of the calculated management restriction curve from the difference from the operation results, and to easily catch the tendency of the abnormality by comparing the latest data with the management restriction curve.

In the present invention, seasonal variation is a parameter that cannot be overlooked when statistically processing various plant quantities to obtain a management limit value, and it is necessary to store plant data for several months to one year. In addition, the selection of the reference point can be set to be optimal for the management target point. This is obtained from a defect of the sensor check function, and there is a point where a failure occurs when the reference point is uniformly set to the plant output.

As the management target points, in addition to the raw input of the plant, a value calculated using the raw input and an output point of the operation log can be used. This allows
It can be used to manage the power generation efficiency and plant ratio of the plant, and can also automatically check the operation log log that has been visually checked in the past.

As a method of calculating the management value, there is a method of using a confidence interval used in the statistical processing. However, when this confidence interval is used as the management limit value, the limit value is opened between the maximum and minimum values of the reference point. Sometimes. As for the management of the power plant, the maximum output is the most severe operating state, and it is not desirable that the management limit value in this state becomes loose.

Therefore, according to the present invention, as a management value calculation method, the maximum and minimum of the reference points are divided by a mesh, and the average, variance, and management limit values are obtained from the values of the target points between the maximum and minimum values. It is preferable to use a method of calculating a management restriction function (regression curve) by an operation such as the least square method using the above management restriction value.

Further, it is necessary for the operator to evaluate the management restriction function calculated by the statistical processing and determine whether or not the management can be performed by using the function. Therefore, it is possible to superimpose and display the data used for the statistical calculation and the obtained management restriction curve on the CRT, and to correct the function by the correction means if necessary.

In some cases, a plurality of reference points can be considered for some points of the power plant. For example, for the condenser vacuum degree, two points, the seawater temperature and the generator output, are reference points for management. Even in such a case, in the present invention, a mesh is taken at each reference point,
By performing statistical processing for each block where the mesh intersects, a management limit value can be obtained in the same manner as described above.
In addition, the control limit value of the condenser vacuum degree obtained by this method is
The equation is a curved surface given by seawater temperature and generator output.

Furthermore, since it is not realistic to check the validity of the values of all the points of the plant ranging from several hundreds to several thousands on the screen one by one, the control limit value and the plant value are periodically changed. In comparison, if the system is configured to notify the operator of a point having an abnormality, the abnormality can be more reliably recognized.

[0028]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows a configuration of a plant operation data management device according to an embodiment of the present invention. The data storage means 3 collects and stores operation state data from a plant 1, and
Point setting means 5 for setting points to be managed, that is, plant operation items and reference points
A data search means 7 for periodically extracting data of the management point set by the point setting means 5 and corresponding reference point data from the data stored by the data storage means 3 for a predetermined period; 7. A management value calculating means 9 for statistically processing the data extracted by the control unit 7 to obtain a management limit function which is a function of a management limit value of the management point using the reference point value as a variable. And a display means 11 for displaying the limit function on a graph and displaying the latest operation data and any past operation data on the graph.

As is well known, sensors are set at various places in the power plant 1 in order to manage and monitor its operation state. In the plant operation data management device having the above-described configuration, the data storage unit 3 collects the operation state data of the plant from sensors provided at various parts of the plant 1, records and updates the data in the latest data storage table, and stores the data in the large-capacity storage device. Store sequentially over a long period.

The data stored in the large-capacity storage device is referred to as plant operation history data. This is effective information for accumulating the past operation state data to know the characteristic peculiar to the plant and the state indicated by each device constituting the plant in various operation states. For example, the main devices constituting the plant include a boiler, a turbine, and a generator. The pressure of the main steam generated in the boiler and the rotation speed of the turbine, and the rotation speed of the turbine and the output of the generator are closely related. It is well known that by recording the operating state at the time of starting the plant or at the time of rated output operation of the generator, the characteristic unique to each device constituting the plant can be obtained. In the present invention, each point value is managed based on the operation history data from such a power plant.

The point setting means 3 is provided for receiving an input of setting of an operation item of a plant to be managed and a reference point. Upon receiving an input from an operator, operation history data is recorded. From the large-capacity storage device, the data search means 5 creates a condition for searching for the relevant data together with the reference data closely related to the relevant data. As such a search condition, the search data can be searched for, such as the date of the corresponding data, the value of the data, or the value of other data collected at the same timing as the relevant data. General information is used.

The data search means 5 extracts data from the mass storage device based on the search conditions, and records the extracted data in a search data record table. The extracted data recorded in the retrieval data recording table is a series of data retrieved under specific conditions, and is managed in association with the retrieved reference data at the same time. Therefore, for the data recorded in the search data recording table,
When plotting on a graph with the vertical axis as the corresponding data and the horizontal axis as the reference data, a scatter diagram showing the correlation is obtained.

The management value calculation means 7 obtains a management restriction curve using the reference data as a variable by statistical processing based on the relation between the pertinent data and the reference data from the operation result data extracted by the data search means 5. The display means 11 displays the obtained management restriction curve and any past operation result data on the same screen.

FIG. 2 shows the control limit curves 21 and 23 when the main steam pressure is used as a control point and the generator output is used as a reference point for the control points, from the plant operation data collected from the plant 1. Display means 11 on the same screen together with past operation result data 25
It is an example displayed by. In FIG. 2, the horizontal axis indicates the generator output as a reference point, and the vertical axis indicates the main steam pressure as a control point.

When a display request is made, first, an upper limit curve 21, a lower limit curve 23, and an average curve 27 are displayed as management restriction curves corresponding to the management points selected by the operator. Subsequently, the operation result data selected by the operator (arbitrary past data is used in FIG. 2) is plotted on the same screen. If this plot deviates from the management restriction curve, it can be determined that some abnormality has occurred. In this figure, the mesh width of the reference point is 100MW.
However, this indicates that there is operation performance data that exceeds the upper limit curve 21 with a mesh of 800 MW to 900 MW.

Next, the process for obtaining the management restriction curve will be described in more detail with reference to FIG. In FIG.
The plant operation data management device shown in FIG. 1 includes a means 31 for correcting the management restriction curve and a means for notifying the operator when there is data exceeding the management restriction curve (comparison means 33,
In addition to the output means 35 and the abnormality content recording means 37), the display means 11 has a function of displaying the management restriction curve and the alarm restriction value on the same screen. The management value calculation means 9 includes a distribution means 39 for distributing the data of the management points for each mesh of the corresponding reference points, and a statistical processing means 41 for statistically processing the distributed data.
And the regression analysis means 43 for regression analysis of the statistical processing results.

Now, it is assumed that the operator sets a management point. The registration of the target management point, the reference point for each management point, the maximum value, the minimum value, the mesh width of each reference point, and the alarm limit value of each management point is performed by the point setting means 5, and the distribution condition recording table 51 and the alarm limit value recording table 53, respectively.

FIGS. 4 and 5 show the structures of the distribution condition record table 51 and the alarm limit value record table 53, respectively. In the distribution condition recording table 51, the management point, the reference point, and the maximum value, the minimum value, and the mesh width of the reference point are stored for the number of the management points. The alarm limit value record table 53 stores management points and their alarm limit values.

Meanwhile, plant operation data from the plant 1 is collected by the data storage means 3 and sequentially stored in a history data memory 55 which is a large-capacity storage device.
7 is temporarily stored. When the management point and the reference point are set by the point setting means 5, the corresponding data is extracted from the history data memory 55 by the data search means 7 and stored in the search data recording table 59.

Next, the search data record table 59
Is distributed by the distribution unit 39 of the management value calculation unit 9 according to the mesh width of the reference point recorded in the distribution condition recording table 51, and the distribution result recording table 61-i (i) is distributed for each mesh. = 1,
2,..., N). FIG. 6 shows the configuration of the distribution result record table 61-i, in which management point values included in the mesh width of reference points and corresponding reference point values are stored.

Next, the statistical processing means 41 performs statistical processing on the management point values stored in the distribution result recording table 61-i for each mesh, and multiplies the average, variance, and variance by a constant for each mesh. The upper limit obtained by adding the average and the lower limit obtained by multiplying the variance by a constant and subtracted from the average are calculated and stored in the calculation result record table 63-i. FIG. 7 shows the configuration of the calculation result record table 63-i. As shown in this figure, the average, variance, upper limit, lower limit, and corresponding reference point value (representative value of mesh width) of the management point are recorded for each mesh.

The regression analysis means 43 compares the average, upper and lower regression curves of the average, upper and lower limits stored in the calculation result record table 63-i using the least square method or the like. Then, the coefficients of each regression curve are stored in the coefficient recording table 65. By the way, since the number of values of the target points entering the mesh differs for each mesh, the reliability of the management limit value obtained for each mesh differs. In general, the statistical reliability increases as the number of data samples increases, so when calculating the management restriction function, the management restriction value used for the calculation is weighted by the number of data.

FIG. 8 is a diagram showing the configuration of the coefficient recording table 65, in which the coefficients of the average curve, the upper curve, and the lower curve are stored. FIG. 8 shows a case where each curve is represented by a quadratic expression.

Using the coefficients of the curve stored in the coefficient recording table 65, the display means 11 displays the average curve, the upper curve, and the lower curve. Also, the operation result data used for the calculation is displayed on the same screen as needed.

At the same time, the latest data record table 5
7 is compared with the curve in the coefficient record table 65 by the comparing means 33. If the result is out of the management limit, the operator is notified by the output means 35 and the abnormal contents are recorded in the abnormal contents record. Recorded by means 37. The output means 35 for notifying the operator includes an annunciator (buzzer), a CRT display device (for alarm display, a list of abnormal points), and the like. Further, by displaying the latest data together with the average curve, the upper limit curve, and the lower limit curve on the display means 11, it is possible to monitor the plant state.

The correcting means 31 is provided for correcting the contents of the coefficient recording table 65, and thereby makes it possible to correct the curve on the display screen. For example, as shown in FIG.
3 and any of the operation results data 25 and the latest data are displayed on the screen, and the coefficients a, b, and c are corrected, whereby the correction can be arbitrarily performed. Note that, in this screen example, each curve is represented as a quadratic function, but it is also possible to treat each curve as an n-th general formula.

Further, since the purpose of the control limit curve is to detect the abnormality of the plant before reaching the alarm value, the actual limit value registered in the alarm limit value record table 53 is considered. By superimposing it on the display means 11, the operator can evaluate the validity of the management restriction curve. FIG. 10 shows an average curve 27, management restriction curves 21, 23, and operation result data 25.
5 is a screen display example in which an alarm limit value 101 is superimposed and displayed.

In the above embodiment, one reference point having an intimate relationship with an item (management point) to be managed has been described as an example. However, the present invention is not limited to this. On the other hand, a management limit value can be set from a plurality of reference point items. For example, as mentioned above, the output of the generator and the pressure of the main steam are closely related, but the main steam pressure is also closely related to the natural temperature represented by atmospheric temperature and seawater temperature. have. This is due to variations in fuel filling efficiency due to atmospheric temperature,
It is clear from the variation of cooling efficiency with seawater temperature.

That is, by providing a prescribed mesh for each of a plurality of reference points with respect to the control point and obtaining the control restriction curve by the same means as described above, not only the operation mode of the plant but also other elements, for example, natural environment By setting a management limit value that takes into account seasonal fluctuations throughout the year, it is possible to more appropriately manage the data by an easy method.

FIG. 11 shows a one-dimensional prescribed mesh 111 for distributing the management point value when there is one reference point for the management point. On the other hand, when there are two reference points with respect to the management point, FIG.
As shown in FIG. 2, the prescribed mesh 112 to which the management point values are distributed is a two-dimensional prescribed mesh represented by a two-dimensional array. As described above, when the number of reference points is n with respect to the management point, the distribution of the management point value is performed based on the n-dimensional prescribed mesh. By obtaining the respective limits, it is possible to manage even a plurality of reference points.

FIG. 13 is a screen display example when there are two reference points obtained by the above method. By increasing the number of reference points to two, the management range has a three-dimensional spread, and management more suitable for actual plant operation becomes possible.

As is clear from the above description, in the plant operation data management apparatus of the present invention, for each point to be managed, a reference point closely related to the point can be arbitrarily set. Operation data can be managed in association with reference point data suitable for each operation state, and early detection of abnormalities can be easily performed by comparison with management limit values obtained based on operation result data. Becomes possible.

[0054]

As described above, according to the present invention, the operating state of the plant can be managed as a value corresponding to the environment surrounding the plant and the operating state of the plant, and this management value is obtained from the operating results. By checking with the management limit value, it is possible to realize early detection of an abnormality without imposing an excessive load on the operator.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a plant operation data management device according to an embodiment of the present invention.

FIG. 2 is a diagram showing a screen display example of a display unit.

FIG. 3 is a block diagram showing in detail a configuration of a plant operation data management device according to another embodiment of the present invention.

FIG. 4 is a diagram illustrating a configuration example of a distribution condition recording table.

FIG. 5 is a diagram illustrating a configuration example of an alarm limit value recording table.

FIG. 6 is a diagram illustrating a configuration example of a distribution result recording table.

FIG. 7 is a diagram illustrating a configuration example of a calculation result recording table.

FIG. 8 is a diagram illustrating a configuration example of a coefficient recording table.

FIG. 9 is a diagram illustrating a display example when an average curve, a management restriction curve, and a coefficient thereof are displayed on the same screen.

FIG. 10 is a diagram showing a display example when an average curve, a management restriction curve, and an alarm restriction value are displayed on the same screen.

FIG. 11: Reference point is 1 for one management point
FIG. 7 is a diagram showing a prescribed mesh in the case of two.

FIG. 12: Reference point is 2 for one management point
FIG. 7 is a diagram showing a prescribed mesh when there are two meshes;

FIG. 13: Reference point is 2 for one management point
FIG. 10 is a diagram showing a display example of a management restriction curve when there are two.

[Explanation of symbols]

 3 Data storage means 5 Point setting means 7 Data search means 9 Management value calculation means 11 Display means 31 Correction means 33 Comparison means 35 Output means 37 Abnormality recording means 41 Distribution means 43 Statistical processing means 45 Regression analysis means

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-63-164000 (JP, A) JP-A-1-256000 (JP, A) JP-A-3-216705 (JP, A) JP-A-3-316 233799 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) G08B 23/00 510 G05B 23/02 301 G06F 17/60 G07C 3/00

Claims (4)

(57) [Claims] 1. Data storage means for inputting and storing operation data of each point of a plant, point setting means for setting a point to be managed and at least one reference point corresponding to each management point, Data search means for periodically extracting data of the management point set by the point setting means and corresponding reference point data from the data stored by the data storage means for a predetermined period; and Management data calculating means for statistically processing the obtained data and calculating a management restriction function which is a function of the management restriction value of the management point using the reference point value as a variable; and a graph showing the management restriction function obtained by the management value calculation means. A table that displays and displays the latest operation data and any past operation data on this graph Plant operating data management apparatus characterized by comprising a means. 2. The plant operation data management device according to claim 1, wherein the management value calculation means divides a range between a maximum value and a minimum value that the reference point can take into meshes at predetermined intervals, and A plant operation data management device, which calculates an average, a variance, and a management limit value of management points, and obtains a management limit function by performing regression analysis on management limit values obtained for the number of meshes. 3. The plant operation data management device according to claim 1, wherein the operation data used for the calculation is displayed on a screen displaying a graph of the management restriction function obtained by the management value calculation means, and the management restriction function is provided. A plant operation data management device comprising correction means for changing the coefficient of the plant operation data. 4. The plant operation data management device according to claim 1, wherein the comparison means compares the latest operation data with a management restriction function set for each management point, and a point exceeding the management restriction is provided. An output means for notifying, when there is, a plant operation data management device.