KR101775065B1 - Method, server and system for detecting abnormality of a power plant using solar energy - Google Patents

Abstract

According to the present invention, a method for detecting an abnormality of a power plant using solar energy comprises: a step of acquiring a current meteorological satellite image and current meteorological information observed by a weather station; a step of interpolating the current meteorological information by a distance between the weather station and the power plant to estimate current meteorological information of the power plant; a step of estimating the current amount of solar radiation of the power plant from the current meteorological satellite image, a current altitude of the sun, and a current azimuth of the sun; a step of calculating a probability distribution of a proper power generation amount in accordance with the estimated current meteorological information and the current amount of solar radiation of the power plant; a step of acquiring a current power generation amount of the power plant; a step of calculating a probability of the current power generation amount appearing in the probability distribution; and a step of determining the power plant in an abnormal state if the probability is lower than a prescribed reference value.

Description

METHOD, SERVER AND SYSTEM FOR DETECTING ABNORMALITY OF A POWER PLANT USING SOLAR ENERGY FIELD OF THE INVENTION [0001]

The present invention relates to a method for detecting abnormalities in a power plant using solar energy. And more particularly, to a method for detecting an abnormality in a solar or solar power plant.

There are two main ways to utilize solar energy. One is a circular reflector like the one in the US state of Nevada, or a computer-derived reflector called a heliostat (a daylight reflector) to create steam to get electricity. Another method is to convert the power directly to a solar panel made of semiconductors such as silicon.

Both methods have drawbacks that power generation can vary greatly depending on weather conditions. Therefore, it is difficult to judge whether or not the power plant using solar energy is abnormal only by observing the power generation amount. Therefore, there is a variety of attempts to determine the abnormality of the power plant by observing the amount of generated electricity relative to the weather condition and the irradiation amount by providing a separate meteorological observing apparatus and a radiation amount measuring sensor in the power plant.

Published Patent Publication No. 10-1485051, Announcement 201.21, 2015

A problem to be solved by the present invention is to provide a method, a server and a system for detecting an abnormality of a power plant using solar energy.

The problems to be solved by the present invention are not limited to the above-mentioned problems, and other problems which are not mentioned can be clearly understood by those skilled in the art from the following description.

According to an aspect of the present invention, there is provided a method of detecting an abnormality of a power plant using solar energy, the method comprising: acquiring current weather information and a current weather satellite image observed at a weather station; Estimating current weather information of the power plant by interpolation according to a distance between the weather observing station and the power station, calculating a current solar radiation amount of the power plant from the current weather satellite image, the altitude angle of the present sun and the azimuth of the current sun, Calculating a probability distribution of an appropriate power generation amount according to the current weather information and the current solar radiation amount of the power plant estimated, estimating a current power generation amount of the power plant, calculating a probability that the current power generation amount appears in the probability distribution, And when the probability is lower than a predetermined reference value, And a step of determining to be in the state.

The step of estimating a current solar radiation amount of the power plant may include a step of estimating a current solar radiation amount of the power plant by using the weather satellite image, the altitude angle of the sun at the time of shooting the past weather satellite image, the azimuth angle of the sun, Acquiring a function of taking an altitude angle and an azimuth angle of the sun as input and outputting a predicted solar irradiance as an output and inputting the current weather satellite image, the altitude angle of the current sun and the azimuth of the current sun to the function, And estimating a current solar radiation amount of the solar cell.

The step of calculating the probability distribution of the appropriate generation amount may include the steps of inputting the weather information and the solar radiation amount as the learning data using the actual generation amount according to the past weather information and the solar radiation amount of one or more power plants using solar energy, Obtaining a function having a probability distribution as an output; and inputting the current weather information and the current solar radiation amount of the power plant to the function, and calculating a probability distribution of the optimum power generation amount of the power plant.

The step of acquiring the function may further include the steps of: constructing an artificial neural network that receives the meteorological information and radiation dose as input and outputs a probability distribution of the predicted generation amount; and performs learning by inputting the learning data to the artificial neural network And obtaining at least one parameter that optimizes the function as a result of the learning.

The step of acquiring the function further includes the step of adjusting at least one parameter of the function using the measured power generation amount corresponding to the past weather information and the solar radiation amount of the power plant as learning data, , The current weather information and the present solar radiation amount of the power plant can be input to the function whose parameters are adjusted to calculate the probability distribution of the optimum power generation amount of the power plant.

The step of judging that the power plant is in an abnormal state may further include the step of estimating the cause of the abnormality based on the current weather information of the power plant, the current solar radiation amount, the current power generation amount, and the past power generation history of the power plant .

The step of estimating the cause of the abnormality may further include the steps of: obtaining a past abnormality database of the power plant including a list of one or more abnormality causes and the abnormality causes; And generating a binary classifier for detecting an anomaly cause for each of the one or more anomalous causes based on the learning result, the method comprising the steps of: And estimating a cause of the abnormality by using the binary classifier for detecting an anomaly cause for each of the one or more anomalous causes and the current weather information, the current solar irradiance and the current generated amount of the power plant.

The method includes the steps of acquiring a database of past weather information and an amount of solar radiation, clustering the database to generate clusters of the past weather information and the solar radiation amount, and determining whether the current weather information and the current solar radiation amount of the power plant belong to the clusters And outputting a weather condition statement corresponding to the current weather information of the power plant and the cluster to which the current solar radiation amount belongs.

According to another aspect of the present invention, there is provided a server for detecting an abnormality of a power plant using solar energy, comprising: a current weather information acquiring unit for acquiring current weather information and a current weather satellite image observed at a weather station, And estimating the current weather information of the power plant by interpolating the current weather information according to the distance between the weather station and the power plant, calculating the current weather satellite image, the altitude angle of the current sun and the azimuth of the current sun, Estimating a current solar radiation amount of the power generation plant, calculating a probability distribution of the optimum generation amount according to the current weather information and the current solar radiation amount of the power plant estimated above, calculating a probability that the current generation amount appears in the probability distribution, If it is determined that the power plant is in an abnormal state Processor.

According to another aspect of the present invention, there is provided a system for detecting an abnormality of a power plant using solar energy, the system comprising: a power generation amount measuring device for measuring a current power generation amount of the power plant; Estimating the current weather information of the power plant by interpolating the current weather information with the distance between the weather station and the power plant, calculating the current weather satellite image, the altitude angle of the current sun and the azimuth angle of the current sun Estimating a current solar radiation amount of the power plant, calculating a probability distribution of the optimum generation amount according to the current weather information and the current solar radiation amount of the power plant estimated at the power plant, obtaining the current generation amount of the power plant measured by the power generation amount measurement apparatus, Calculates a probability that the current generation amount appears in the probability distribution, When lower than the predetermined threshold value, and a server in which the power plant is determined to be in the abnormal state.

Other specific details of the invention are included in the detailed description and drawings.

It is difficult to judge whether or not the power plant using solar energy can be operated normally by observing the amount of power generation because the power generation amount of the power plant using solar energy is greatly influenced by the weather condition. In particular, solar power plants are generally located in areas that are difficult to access, making it difficult to quickly diagnose faults.

According to an embodiment of the present invention, it is possible to diagnose the state of a power plant using solar energy without additional equipment for diagnosing the state of the power plant using solar energy. Therefore, there is an effect that the cost for purchasing additional equipment and the cost for maintenance thereof can be reduced.

In addition, it is possible to judge the weather condition by using the meteorological information and the satellite image of the meteorological office which can be easily acquired anywhere, and to detect the abnormality of the power plant.

If an abnormality of a power plant is suspected, it can respond as quickly as possible by presenting the cause as much as possible. Therefore, the maintenance and repair of the power plant can be facilitated, and the power generation amount can be maintained in an appropriate state, thereby improving the profitability of the power plant.

The effects of the present invention are not limited to the above-mentioned effects, and other effects not mentioned can be clearly understood by those skilled in the art from the following description.

1 is a diagram illustrating a system for detecting anomaly in a power plant using solar energy according to an embodiment.
2 is a flowchart illustrating a method for detecting an abnormal state of a power plant using solar energy according to an embodiment.
3 is a view showing an artificial neural network in which a gas-phase satellite image, an altitude angle and an azimuth angle of the sun are input, and a solar radiation amount before the horizontal plane is output.
4 is a diagram showing an artificial neural network in which a weather information vector is input and an average and variance of a beta distribution are output.
5 is a flowchart illustrating a method for estimating an abnormal cause of a power plant using solar energy according to an embodiment.
FIG. 6 is a flowchart illustrating a method of outputting a comment about a current weather condition according to an embodiment.
7 is a configuration diagram showing a server according to an embodiment.

BRIEF DESCRIPTION OF THE DRAWINGS The advantages and features of the present invention and the manner of achieving them will become apparent with reference to the embodiments described in detail below with reference to the accompanying drawings. It should be understood, however, that the invention is not limited to the disclosed embodiments, but may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein, Is provided to fully convey the scope of the present invention to a technician, and the present invention is only defined by the scope of the claims.

The terminology used herein is for the purpose of illustrating embodiments and is not intended to be limiting of the present invention. In the present specification, the singular form includes plural forms unless otherwise specified in the specification. The terms " comprises "and / or" comprising "used in the specification do not exclude the presence or addition of one or more other elements in addition to the stated element. Like reference numerals refer to like elements throughout the specification and "and / or" include each and every combination of one or more of the elements mentioned. Although "first "," second "and the like are used to describe various components, it is needless to say that these components are not limited by these terms. These terms are used only to distinguish one component from another. Therefore, it goes without saying that the first component mentioned below may be the second component within the technical scope of the present invention.

Unless defined otherwise, all terms (including technical and scientific terms) used herein may be used in a sense that is commonly understood by one of ordinary skill in the art to which this invention belongs. In addition, commonly used predefined terms are not ideally or excessively interpreted unless explicitly defined otherwise.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating a system for detecting anomaly in a power plant using solar energy according to an embodiment.

The weather stations 10 and 20 may be weather stations that are managed by the weather station. In addition, the weather stations 10 and 20 may be private weather stations, but are not limited thereto. The meteorological stations 10 and 20 can measure temperature, wind speed, precipitation and humidity.

The meteorological satellite 30 may be a meteorological satellite, but any other meteorological satellite may be used. In one embodiment, the weather satellite 30 can transmit satellite images of five wavelengths (channels) at intervals of about 15 minutes.

The power plant 100 may be a solar power plant or a solar power plant. A power generation amount measuring device 110 for measuring the power generation amount may be installed in the power generation plant 100.

The server 200 can estimate the appropriate amount of power generation of the power plant 100 by using the weather information and the weather satellite image received from the weather stations 10 and 20 and the weather satellite 30. [

The server 200 can estimate the current solar radiation amount of the power plant 100 using the weather satellite image, the current altitude angle and the azimuth angle of the current sun. The server 200 interpolates the weather information received from the weather stations 10 and 20 according to the distance between the weather stations 10 and 20 and the power station 100 to estimate the current weather information of the power station 100 can do.

The server 200 can estimate the appropriate amount of power generation of the power plant 100 by using the estimated weather information and the solar radiation amount of the power plant 100. [ In one embodiment, the server 200 may calculate the probability distribution of the appropriate power generation amount of the power plant 100.

The server 200 can acquire the current power generation amount of the power plant 100 from the power generation amount measurement apparatus 110. [ The server 200 can calculate the probability that the current generation amount appears in the probability distribution of the appropriate generation amount. If the probability is lower than a predetermined reference value, the server 200 may determine that the power plant 100 is in an abnormal state.

2 is a flowchart illustrating a method for detecting an abnormal state of a power plant using solar energy according to an embodiment.

Referring to FIG. 2, a method of detecting an abnormal state of a power plant using solar energy is comprised of steps that are processed in a time-series manner in the server 200 shown in FIG. Therefore, even if the method of detecting the abnormal state of the power plant using the solar energy of FIG. 2 is omitted, the description of the server 200 of FIG. 1 is based on the detection of the abnormal state of the power plant using the solar energy of FIG. And the like.

In step S210, the server 200 acquires the current weather information and the current weather satellite image observed at the weather stations 10 and 20.

The present weather information observed at the weather stations 10 and 20 may include, but is not limited to, temperature, humidity, atmospheric pressure, precipitation, solar radiation, daylight hours, cloudiness, wind speed and wind direction.

The weather satellite 30 can transmit satellite images of five wavelengths (channels) at intervals of about 15 minutes. The current weather satellite image obtained from this may include, but is not limited to, an infrared image, a visible image, a steam image, a short-wave infrared image, and an infrared image.

In step S220, the server 200 estimates the current weather information of the power plant 100 by interpolating the current weather information according to the distance between the weather observation stations 10 and 20 and the power generation station 100. [

The server 200 may find one or more meteorological stations 10 and 20 closest to the power station 100 in order to obtain weather information at a place where the powerhouse 100 is located. The server 200 can obtain weather information at a place where the power plant 100 is located by interpolating the weather information obtained from the weather stations 10 and 20.

In one embodiment, the server 200 may perform interpolation using distance inverse distance weighting. The server 200 can obtain the location information of the power plant 100 including the latitude and longitude of the power plant 100. [ The server 200 calculates the number of meteorological observatories having the closest straight line distance d _k to the power plant 100 among the meteorological observatories having equipment capable of measuring the corresponding meteorological variable, Can be selected to obtain the weather variable x _k .

The server 200 may interpolate the vapor phase variable x of the power plant 100 using the acquired K phase variables as shown in the following Equation 1. [

In Equation (1), w _k denotes a weight inversely proportional to the distance.

The server 200 can obtain the necessary weather variables through interpolation. In one embodiment, the server 200 may select a weather station capable of measuring the solar radiation amount, and obtain the solar radiation amount of the power station 100 through interpolation.

In step S230, the server 200 estimates the current solar radiation amount of the power plant 100 from the current weather satellite image, the altitude angle of the current sun, and the azimuth angle of the current sun.

In one embodiment, the solar radiation amount can mean the horizontal solar radiation amount.

The server 200 inputs the weather satellite image, the altitude angle of the sun and the azimuth angle of the sun using the past weather satellite image, the altitude angle of the sun at the time of shooting of the past weather satellite image, the azimuth angle of the sun and the actual solar radiation as learning data And obtain a function of outputting the predicted solar radiation amount.

The server 200 can estimate the current solar radiation amount of the power plant 100 by inputting the current weather satellite image, the altitude angle of the current sun and the azimuth angle of the current sun to the obtained function.

In one embodiment, the server 200 may constitute an artificial neural network that receives the weather satellite image, the altitude angle and the azimuth angle of the sun, and outputs the solar radiation amount before the horizontal plane. Referring to FIG. 3, an artificial neural network 300 is shown in which a weather satellite image, an altitude angle and an azimuth angle of the sun are input, and a solar radiation amount before the horizontal plane is output.

The server 200 can acquire past solar irradiance measurement information of a weather station that can measure the solar irradiance. The server 200 can acquire the past weather satellite images and acquire the azimuth angle, altitude angle, and actual solar radiation of each of the past weather satellite images at the time of shooting. The server 200 can perform the learning of the artificial neural network 300 using the past weather satellite image, the azimuth angle of the sun, the elevation angle, and the actual solar radiation as learning data.

In one embodiment, the server 200 may adjust the neural network connection weights of the artificial neural network 300 using a Stochastic Gradient Descent, but is not limited thereto.

In another embodiment, the server 200 may use other machine learning algorithms based on regression learning with the past weather satellite image, the azimuth angle of the sun, the altitude angle, and the actual solar radiation as learning data.

In step S240, the server 200 calculates the probability distribution of the appropriate generation amount according to the current weather information and the present solar radiation amount of the estimated power plant 100. [

The server 200 obtains a function of inputting weather information and solar radiation as input and outputting a probability distribution of the predicted power generation amount by using the actual power generation amount corresponding to the past weather information and the solar radiation amount of one or more power plants using solar energy as learning data can do.

The server 200 can input the current weather information and the current solar radiation amount of the power plant 100 to the obtained function and calculate the probability distribution of the optimum power generation amount of the power plant 100. [

The server 200 can constitute an artificial neural network that receives the weather information and the solar radiation amount as input and outputs the probability distribution of the predicted generation amount. The server 200 can acquire at least one parameter for performing learning by inputting learning data into the artificial neural network and optimizing the function as a result of learning.

In one embodiment, the server 200 may calculate the beta distribution of the appropriate power generation amount. In this case, the server 200 can calculate the beta distribution of the appropriate power generation amount by using Deep Beta Regression.

The server 200 can generate a weather information vector including a plurality of weather variables and an irradiation dose. The weather information vector may include information about altitude and azimuth of the sun. The server 200 can construct a machine learning model for setting the weather information vector as an input vector and outputting the beta distribution through various nonlinear transformations.

Referring to FIG. 4, an artificial neural network 400 may be constructed in which a weather information vector is input and an average and variance of a beta distribution are output. The artificial neural network 400 may include one input layer, one output layer, and one or more hidden layers.

The input vector can be transformed through each layer. From the input layer to the hidden layer, and from the hidden layer to the hidden layer, as shown in Equation 2 below.

In Equation 2, h _k denotes the vector value of the k th hidden layer and, K is the number of hidden layer. h ₀ denotes the vector value of the input layer, and b _k is a real number. W _k can be a weight in each layer.

σ () may be one of the nonlinear transformations commonly used in deep running. For example, sigma () may be sigmoid, tanh, rectified linear unit, but is not limited thereto.

The final hidden layer to the output layer can be transformed as shown in Equation (3).

In Equation (3), μ denotes the average of the beta distribution, and Φ denotes the variance of the beta distribution. According to Equation (3), mu and each parameter can be forced to have an appropriate range.

The learning data of the artificial neural network 400 may be configured as shown in Equation (4).

In Equation (4), x _n is meteorological information, and y _n is a value obtained by dividing the power generation amount of the power generation plant 100 by the total power generation capacity.

Learning may mean determining the values of the parameters (W _k , b _k , w _μ and w _Φ ) that optimize the objective function. Normal deep running, artificial neural networks, or machine learning methods can be used for learning.

The objective function may consist of negative log likelihood and regularization terms of the data, and the objective function may have to be minimized. The objective function can be constructed as shown in Equation (5).

In Equation 5, [theta] is a value that minimizes the objective function, and [lambda] is a trade-off parameter that controls how much the ordering is performed. R (&thetas;) can generally be a sum of all parameters squared.

Further, p () can be defined by the following equation (6).

According to Equation (6), μ and Φ, which are parameters of the beta distribution, are output to the input vector x _n , and the probability density of the beta distribution generated from the beta distribution is defined as the likelihood for x _n .

In one embodiment, the server 200 may perform zero adjustment according to each of the power plants. Since the server 200 calculates the beta distribution using the power generation amount of the entire power plant as learning data, the characteristics of each power plant can not be reflected. Therefore, if new learning data is generated for each power station and the learning is performed again, the probability distribution reflecting the characteristics of each power station can be calculated.

There may be different parameter settings for each plant depending on location, climate and facilities. Therefore, the server 200 can generate different objective functions for the respective power plants by reflecting characteristics of the respective power plants.

For example, the server 200 can adjust at least one parameter of the objective function by using the actual power generation amount corresponding to the past weather information and the solar radiation amount of the power plant 100 as learning data. The server 200 can calculate the probability distribution of the optimum power generation amount of the power plant 100 by inputting the current weather information and the current solar radiation amount of the power plant 100 to the objective function whose parameters are updated.

In step S250, the server 200 acquires the current power generation amount of the power plant 100. [

The server 200 can acquire the power generation amount of the power plant 100 measured by the power generation amount measuring apparatus 110. [ The power generation amount measurement device 110 may be installed separately in the power generation plant 100 or may be the power generation plant 100 itself.

In step S260, the server 200 calculates the probability that the current generation amount will appear in the probability distribution, and if the calculated probability is lower than the predetermined reference value, the server 100 determines that the power generation plant 100 is in an abnormal state.

In one embodiment, given a deep beta regression model with zero point calibration for each plant and input data x _n and y _{obs, which} is the actual power plant operation rate, the server 200 calculates p _a (y _obs ; x _n ) can be defined as Equation (7).

In Equation (7), I (y; [mu], [Phi]) is the cumulative distribution function of the beta distribution.

The server 200 may determine that the power plant is in an abnormal state when p _a is greater than a preset threshold value.

5 is a flowchart illustrating a method for estimating an abnormal cause of a power plant using solar energy according to an embodiment.

When it is determined that the power plant 100 is in an abnormal state, the server 200 determines the cause of the abnormality based on the current weather information of the power plant 100, the current solar radiation amount, the current power generation amount, Can be estimated.

In step S510, the server 200 can perform binary classification learning using the abnormal cause database of the power plant 100. [ The server 200 may obtain a past abnormal case database of the power plant 100 including a list of one or more abnormal causes and abnormal causes.

The server 200 may use the anomaly database to perform binary classification learning to determine whether it corresponds to one or more causes according to weather information, solar radiation amount, and power generation amount. The cause of the abnormality of the power plant 100 may include, but is not limited to, snow, overheating, breakage or communication abnormality.

The server 200 may generate an anomaly detection binary classifier for each of one or more anomalous causes based on the results of the binary classification learning. In one embodiment, the server 200 may generate a support vector machine that determines yes / no as to whether each of the above causes is abnormal. For example, the server 200 receives the current weather information of the power plant 100, the current solar radiation amount, and the current power generation amount as inputs and outputs a support vector (not shown) for outputting the possibility of snow accumulation on the solar panel of the power plant 100 as Yes / You can create a machine.

In step S520, the server 200 can set a rule of an abnormality cause according to each weather information by using a set of rules previously designated by an expert for each cause of abnormality.

In step S530, the server 200 may determine that the power plant 100 is in an abnormal state. The server 200 can obtain current weather information of the power plant 100 when it is determined that the power plant 100 is in an abnormal state.

In step S540, the server 200 can estimate the cause of the abnormality based on the result of the binary classification learning in step S510 and the rule set in step S520.

The server 200 can determine whether the current weather information corresponds to the rule of each cause of abnormality. The server 200 may output a list of the causes having the rule corresponding to the current weather information.

The server 200 can substitute current weather information into a support vector machine corresponding to each of the above-mentioned causes. The server 200 may output a list of the causes for which the output of the support vector machine is YES.

FIG. 6 is a flowchart illustrating a method of outputting a comment about a current weather condition according to an embodiment.

In step S610, the server 200 can acquire the past weather information and the radiation dose database.

In step S620, the server 200 may cluster the database obtained in step S610 to generate clusters of past weather information and solar radiation. In one embodiment, the server 200 may use a K-means clustering algorithm, but is not limited thereto.

The server 200 can construct a database of a weather situation explanation statement showing the characteristics of each community for each community. The weather condition statement can be entered by the person in advance. In one embodiment, the server 200 may determine the characteristics of each cluster and automatically generate a weather condition statement that indicates the characteristics of each cluster.

In step S630, the server 200 can determine a cluster corresponding to the current weather information and the solar radiation amount.

In step S640, the server 200 may output a weather condition statement indicating characteristics of the community corresponding to the current weather information and the solar radiation amount.

7 is a configuration diagram showing a server according to an embodiment.

The method for detecting the abnormal state of the power plant using the solar energy described in FIG. 2 can be used to control the server 200 shown in FIG. 7, and the server 200 shown in FIG. Operation can be performed.

Referring to Fig. 7, the server 200 includes a processor 202 and a communication unit 204. Fig. On the other hand, only the components related to the embodiments are shown in the server 200 shown in FIG. Therefore, it will be understood by those skilled in the art that other general-purpose components other than the components shown in FIG. 7 may be further included in the server 200.

The communication unit 204 acquires the current weather information and the current weather satellite image observed at the weather station and acquires the current generation amount of the power station 100.

The processor 202 estimates the current weather information of the power plant 100 by interpolating the current weather information according to the distance between the weather observing station and the power station 100. The processor 202 calculates the current weather satellite image and the current sun altitude angle, The current solar radiation amount of the power plant 100 is estimated.

The processor 202 calculates a probability distribution of the optimum power generation amount according to the current weather information and the current solar radiation amount of the power plant 100, calculates a probability that the current power generation amount will appear in the probability distribution, and when the probability is lower than the predetermined reference value, It is determined that the power plant 100 is in an abnormal state.

The steps of a method or algorithm described in connection with the embodiments of the present invention may be embodied directly in hardware, in software modules executed in hardware, or in a combination of both. The software module may be a random access memory (RAM), a read only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), a flash memory, a hard disk, a removable disk, a CD- May reside in any form of computer readable recording medium known in the art to which the invention pertains.

While the present invention has been described in connection with what is presently considered to be practical exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, You will understand. Therefore, it should be understood that the above-described embodiments are illustrative in all aspects and not restrictive.

10: Weather station
20: Weather station
30: Weather satellite
100: Power station
110: Power generation measuring device
200: Server

Claims (10)

In a method for detecting abnormalities in a power plant using solar energy,
Obtaining current weather information and a current weather satellite image observed at a weather station;
Estimating current weather information of the power plant by interpolating the current weather information according to the distance between the weather station and the power station;
Estimating a current solar radiation amount of the power plant from the current weather satellite image, the altitude angle of the current sun and the azimuth angle of the current sun;
Calculating a probability distribution of an appropriate power generation amount according to the current weather information and the current solar radiation amount of the power plant estimated;
Obtaining a current generation amount of the power plant;
Calculating a probability that the current generation amount appears in the probability distribution; And
Determining that the power plant is in an abnormal state when the probability is lower than a predetermined reference value; Lt; / RTI >
Wherein estimating a current solar radiation amount of the power plant comprises:
Constructing an artificial neural network having an input of an aerial satellite image, an altitude angle and an azimuth angle of the sun, and an output of solar radiation of the power plant;
Learning the artificial neural network using the past weather satellite image, the altitude angle of the sun corresponding to the time of shooting the past weather satellite image, the azimuth angle of the sun and the actual solar radiation as learning data; And
Estimating a current solar radiation amount of the power plant by inputting the current weather satellite image, the altitude angle of the current sun and the azimuth angle of the current sun to the learned artificial neural network; / RTI >
delete The method according to claim 1, wherein the step of calculating the probability distribution of the appropriate generation amount comprises:
Acquiring a function of inputting weather information and solar radiation as input and outputting a probability distribution of the predicted power generation amount using the actual power generation amount corresponding to the past weather information and the solar radiation of one or more power plants using solar energy as learning data; And
Inputting the current weather information and the current solar radiation amount of the power plant to the function, and calculating a probability distribution of the optimum power generation amount of the power plant; / RTI >
4. The method of claim 3, wherein obtaining the function further comprises:
Constructing an artificial neural network having the input of the weather information and the solar radiation amount and outputting the probability distribution of the predicted generation amount as an output;
Performing learning by inputting the learning data to the artificial neural network; And
Obtaining at least one parameter that optimizes the function as a result of the learning; / RTI >
4. The method of claim 3, wherein obtaining the function further comprises:
Adjusting at least one parameter of the function using actual measured power generation amount corresponding to the past weather information and the solar radiation amount of the power plant as learning data; Further comprising:
Wherein the calculating the probability distribution comprises:
And inputting the current weather information and the current solar radiation amount of the power plant into the function whose parameters are adjusted so as to calculate a probability distribution of the optimum power generation amount of the power plant.
The method of claim 1, wherein determining that the power plant is in an abnormal state comprises:
Estimating the cause of the abnormality based on the current weather information of the power plant, the current solar radiation amount, the current power generation amount, and the past power generation history of the power plant; ≪ / RTI >
7. The method of claim 6, wherein estimating the cause of the abnormality includes:
Obtaining a past abnormal case database of the plant including a list of one or more abnormal causes and the abnormal causes;
Performing binary classification learning by using the abnormal case database to judge whether each of the one or more abnormal causes corresponds to weather information, solar radiation amount, and power generation amount;
Generating an anomaly detection binary classifier for each of the one or more anomalous causes based on the learning result; And
Estimating the cause of the abnormality using the current weather information, the current solar radiation amount, and the current generation amount of the power plant; / RTI >
The method according to claim 1,
Obtaining a database of past weather information and solar radiation;
Clustering the database to generate clusters of past weather information and solar radiation; And
Determining whether the current weather information and the current solar radiation amount of the power plant belong to one of the communities and outputting a weather condition statement corresponding to the current weather information and the current solar radiation amount of the community; ≪ / RTI >
1. A server for detecting abnormalities in a power plant using solar energy,
A communication unit for acquiring the current weather information and the current weather satellite image observed at the weather station and acquiring the current generation amount of the power station; And
Estimating the current weather information of the power plant by interpolating the current weather information according to the distance between the weather station and the power plant, inputting the weather satellite image, the altitude angle and azimuth angle of the sun, And the artificial neural network is learned by using the past weather satellite image, the altitude angle of the sun corresponding to the time of shooting the past weather satellite image, the azimuth angle of the sun and the actual solar radiation as learning data, Estimating a current solar radiation amount of the power plant by inputting the current weather satellite image, the altitude angle of the current sun and the azimuth angle of the current sun to the artificial neural network, and calculating a probability of a proper generation amount according to the current weather information and the present solar radiation amount of the estimated power plant Calculates a probability that the current generation amount appears in the probability distribution, If the probability is below a predetermined threshold value, the processor that the power plant is determined to be in the abnormal state; .
1. A system for detecting abnormalities in a power plant utilizing solar energy,
A power generation amount measuring device for measuring a current power generation amount of the power plant; And
A current weather information acquisition unit for acquiring current weather information and a current weather satellite image observed at a weather station and estimating current weather information of the power station by interpolating the current weather information according to a distance between the weather station and the power station, An altitude angle and an azimuth angle of the sun, and an output of solar radiation of the power plant as an output, wherein the altitude angle and the azimuth angle of the sun and the actual azimuth angle of the sun Estimating a current solar radiation amount of the power plant by inputting the current weather satellite image, an altitude angle of the current sun and an azimuth angle of the current sun to the learned artificial neural network using the solar radiation amount as learning data, Obtain the probability distribution of the appropriate generation amount according to the current weather information of the power plant and the present solar radiation amount estimated The current power generation amount of the power plant measured by the power generation amount measuring device is obtained and the probability that the current power generation amount appears in the probability distribution is calculated and when the probability is lower than a predetermined reference value, A server to judge; .

Images