KR102439495B1 - Electricity-safety service system for intelligent control based on an electricity-safety precusory prediction algorithm - Google Patents

Abstract

An electrical safety service system for intelligent control based on an electrical safety precursory prediction algorithm, according to one embodiment of the present invention, comprises: an electrical safety warning device including a first sensor for generating a first sensing value by detecting a current, voltage, or leakage current, a second sensor for generating a second sensing value by detecting temperature or humidity; and an artificial intelligence calculation unit for identifying the type of electrical safety accident by using the first sensing value of the first sensor and the second sensing value of the second sensor, determining whether an electrical safety accident has occurred, by using an artificial neural network for applying a threshold value set according to the identified type of electrical safety accident, and when the electrical safety accident has occurred, generating and transmitting an electrical safety accident warning signal; and an electrical safety warning device including an alarm transmission unit for, when the electrical safety accident warning signal is received from the artificial intelligence calculation unit, transmitting the electrical safety accident warning signal to a target terminal for reception.

Description

Electric Safety Prediction Algorithm-Based Intelligent Control Electric Safety Service System

The present invention relates to an intelligent control electric safety service system based on an electric safety prediction algorithm, and in particular, detects the state of electricity (current, voltage, leakage, etc.) and environment (temperature, humidity, etc.) in an electric safety accident area, and the corresponding electricity It judges the current status of electrical safety accidents by considering the environmental conditions of the safety accident area, and supports electrical safety accident prevention activities by accurately detecting the precursors of electrical safety accidents in practice, and reduces false alarms of electrical safety accidents. It is about an electric safety service system equipped with intelligent technology.

In the case of recent electrical safety accidents and special electrical fires, electrical safety accidents frequently occur in residential areas, commercial districts, communal facilities, and traditional markets due to the aging of electrical equipment, which is emerging as a large-scale casualty and social problem. situation is necessary.

Electric safety warning prediction algorithm-based intelligent control electrical safety service system is a system that can access electrical signals and environmental signals and predict the precursors of electrical safety accidents.

More than 20% of fires in Korea are caused by electrical factors. In particular, in winter and summer, the occurrence rate is increasing in residential areas, business districts, communal facilities, and traditional markets, etc. It is not possible to detect the precursor of a safety accident, and as a result, the initial response and situational judgment cannot be made promptly, resulting in many lives and property damage.

The electrical safety service system collects and manages continuous electrical safety big data to prevent preemptive electrical safety management and false alarm situations, improves the accuracy of AI-based predictive algorithms, and updates related algorithms to on-site alarm devices. By improving the accuracy of the alarm situation and notifying the manager and the relevant disaster safety response agency of the occurrence of an electrical safety accident rather than transmitting whether an electrical safety accident has occurred to a large number of unspecified people, an electrical safety alarm can actually lead to an initial response. A system is needed.

The present invention detects the electrical signal (voltage, leakage current, leakage current) state and environmental signal (temperature, humidity) of the electrical safety accident (electrical leakage, blackout, short circuit) occurrence region, and the environmental condition of the electrical safety accident occurrence region It determines the level of occurrence of electrical safety accidents by considering The purpose is to provide an electrical safety monitoring system equipped with artificial intelligence technology.

Electrical safety omens prediction algorithm-based intelligent control electrical safety service system according to an embodiment of the present invention is a first sensor for generating a first sensed value by sensing current, voltage or leakage current, and a second sensing by detecting temperature or humidity A second sensor for generating a value, and a first sensed value of the first sensor and a second sensed value of the second sensor to distinguish the type of electrical safety accident, and set according to the differentiated type of electrical safety accident An electrical safety alarm device including an artificial intelligence calculation unit that determines whether an electrical safety accident has occurred using an artificial neural network that applies a threshold, and generates and transmits an electrical safety accident warning signal when an electrical safety accident occurs, and the AI calculation unit When the electrical safety accident warning signal is received from the electrical safety alarm device comprising an alarm transmitter for transmitting the electrical safety accident warning signal to the receiving terminal.

The artificial intelligence calculation unit uses Equation 1 below to distinguish the types of electrical safety accidents to generate data for determining power failures during electrical safety accidents.

The artificial intelligence calculation unit generates data for short circuit discrimination by using Equation 2 below to distinguish the types of electrical safety accidents.

The artificial intelligence calculation unit generates data for electric leakage determination by using Equation 3 below to distinguish the types of electrical safety accidents.

The artificial intelligence calculation unit extracts waveform features using the following Equation 4 to which a waveform and a filter image expressed based on the data for determining the power failure, the data for determining the short circuit, and the data for determining the leakage current are applied.

The artificial intelligence calculation unit calculates a specific output value reflecting a weight to the waveform feature extracted by Equation 4 through an artificial neural network to which a predetermined weight is applied, and compares the specific output value reflecting the weight with a preset threshold to alert an electrical safety accident or whether it is normal or not.

The electrical safety platform server includes: an artificial intelligence learning unit that reflects the measured values and output values received from the electrical safety alarm device and changes the weight applied to the extracted waveform features; and a firmware update unit configured to transmit the weight changed by the artificial intelligence learning unit to the artificial intelligence calculation unit through communication so that the artificial intelligence calculation unit can apply the changed weight.

The present invention detects the electrical and environmental conditions of the electrical safety accident occurrence area, and determines the electrical safety accident type (blackout, short circuit, short circuit) and current status in consideration of the environmental conditions of the electrical safety accident occurrence zone, and actually Accurately detect safety precursors to reduce false electrical safety accident alarms, and to improve the reliability and safety of electrical safety monitoring systems through preventive electrical safety management.

1 and 2 are diagrams showing the configuration of an electric safety omens prediction algorithm-based intelligent control electric safety service system according to an embodiment of the present invention.
3 is a flowchart for explaining a driving process of an electric safety omens prediction algorithm-based intelligent control electric safety service system according to an embodiment of the present invention.
4 is a flowchart for explaining a driving process of an electric safety omens prediction algorithm-based intelligent control electric safety service system according to another embodiment of the present invention.

Hereinafter, specific contents for carrying out the invention will be described in detail with reference to the accompanying drawings.

1 and 2 are diagrams showing the configuration of an electric safety omens prediction algorithm-based intelligent control electric safety service system according to an embodiment of the present invention.

1 and 2, the electric safety prediction algorithm-based intelligent control electric safety service system 100 includes an electric safety alarm device 110, an electric safety platform server 120, and a receiving terminal 130. .

The electrical safety alarm device 110 includes a first sensor 111 , a second sensor 112 , an artificial intelligence calculation unit 113 , and a data transmission unit 114 .

The first sensor 111 may be composed of one or more sensor nodes, generates a first sensed value by sensing current, voltage, or leakage current, and uses the generated first sensed value 210 with the artificial intelligence calculator 113 . ) is sent to The first sensor 111 may be an electrical sensor that measures electrical related values such as current, voltage, and leakage current.

The second sensor 112 may include a temperature sensor or a humidity sensor. The second sensor 112 generates a second sensed value 210 by sensing temperature or humidity, and transmits the generated second sensed value to the artificial intelligence calculator 113 . The second sensor 112 of the present invention may be composed of one or more sensor nodes, and may be an environmental sensor capable of measuring temperature, humidity, and the like.

The artificial intelligence calculator 113 may distinguish the type of the electrical safety accident by using the first sensed value of the first sensor 111 and the second sensed value of the second sensor 112 .

The artificial intelligence calculation unit 113 performs the data preprocessing process 220 of Equations 1, 2, and 3 in order to distinguish the types of electrical safety accidents.

Equation 1 is for generating ('pre-processing') data for determining 'power failure' during electrical safety accidents.

The artificial intelligence calculation unit 113 may generate a 'waveform of V _f ', which is data for determining 'power failure', by using Equation (1).

[ Equation 1 ]

Here, 'n' is a sequence number of the measured data, 'T' is a waveform period, 'V' is a 'voltage sensed value' among the first sensing values of the first sensor 111 , and V _rms is the first sensor 111 . represents the magnitude of the RMS voltage with respect to the 'voltage sensed value' among the first sensed values of .

Equation 2 is for generating ('pre-processing') data for 'short-circuit' discrimination among electrical safety accidents.

The artificial intelligence calculation unit 113 may generate a 'waveform of If', which _is data for 'short' discrimination, by using Equation (2).

[ Equation 2 ]

Here, 'n' is a sequence number of the measured data, 'T' is a waveform period, 'V' is a 'current sensing value' among the first sensing values of the first sensor 111 , and I _rms is the first sensor 111 . represents the magnitude of the rms current with respect to the 'current sensed value' among the first sensed values of .

Equation 3 is for generating ('pre-processing') data for determining 'electrical leakage' during electrical safety accidents.

The artificial intelligence calculation unit 113 may generate a 'waveform of Z _f ', which is data for determining 'leakage', by using Equation (3).

[ Equation 3 ]

Here, 'n' is a sequence number of the measured data, 'T' is a waveform period, and 'Z' is a 'leakage current sensing value' among the first sensing values of the first sensor 111 . Z _rms represents an effective leakage current magnitude with respect to a 'leakage current sensed value' among the first sensed values of the first sensor 111 , and 'T _Vrms ' represents a 'voltage sensed value' among the first sensed values of the first sensor 111 . It represents the waveform period of the rms voltage with respect to 'value'.

The artificial intelligence calculation unit 113 generates a waveform (' A waveform feature 230 may be extracted from V _{f ,} I _{f ,} Z _f ′) using Equation 4 .

[ Equation 4 ]

’는 합성곱, ‘F_H’는 필터의 높이, ‘F_W’는 필터의 너비, ‘

’은 정수를 나타낸다. Here, 'G _i,j ' is the i-th row of the matrix-expressed filtered image, the pixels of the j-th column (however, i, j are integers), 'X' is the matrix representation of the image for the waveform, and 'F' is matrix representation of the filter image, '

' is the convolution, 'F _H ' is the height of the filter, 'F _W ' is the width of the filter, '

' represents an integer.

For example, 'X' may be a matrix representation of an image including a waveform generated by the artificial intelligence calculation unit 113 . In applying Equation 4 for 'power outage', a filter of 4×4 and 2nd order is used, and for 'short-circuit' judgment, 2×2, 4×8 filters are used, and 'electric leakage' is judged. For this purpose, 2×2 or 4×8 filters can be used. In this way, the filter can be freely changed according to the situation.

The artificial intelligence calculation unit 113 inputs the waveform features ('G _i,j ') extracted by Equation (4) to the input nodes of the artificial neural network (240), respectively, and applies a predetermined weight according to the input image matrix. A specific output value is calculated by applying a weight to the extracted waveform features ('G _i,j ') through an applicable artificial neural network.

Here, the weight is data derived from the artificial intelligence learning unit 123 accumulated by environmental and seasonal causes, day and night, temperature and humidity, statistical information of past electrical safety accidents, and types of electrical safety accidents (power outage, short circuit, short circuit). In other words, the weight is a value that can be continuously changed and derived through learning of the artificial neural network and Equation (4).

The artificial intelligence calculation unit 113 may determine the electrical safety accident occurrence state based on a specific output value reflecting the weight as shown in Equation 5. When the specific output value reflecting the weight exceeds a preset threshold, the artificial intelligence calculation unit 113 outputs 1 indicating an electrical safety accident alarm, and when the calculated specific output value does not exceed the preset threshold, indicating a normal state outputs 0. The preset threshold may be any one of omen threshold/imminent threshold/occurrence threshold, and the artificial intelligence calculation unit 113 compares the omen threshold/imminent threshold/occurrence threshold with a specific output value that reflects the weight and states (omen, imminent). , occurrence) can be determined.

[ Equation 5 ]

The artificial intelligence calculation unit 123 determines the state (prediction, imminent, occurrence) for each type of electrical safety accident (blackout, short circuit, short circuit) based on a specific output value reflecting the weight, which is an output node of the artificial neural network. For example, in the output of the artificial neural network, three pieces of safety accident state information for each type of electrical safety accident (power failure, short circuit, and leakage) may be allocated to the output nodes, respectively.

The electrical safety platform server 120 includes a receiving unit 121 , a database unit 122 , an artificial intelligence learning unit 123 , a firmware update unit 124 , and an alarm transmitting unit 125 .

The receiving unit 121 receives the sensing value measured in real time from the electrical safety alarm device 110 and stores it in the database unit 122 .

The database unit 122 may store various types of data.

The artificial intelligence learning unit 123 may calculate new weights or improve the artificial neural network by reflecting various measurement values and output values received from the electrical safety alarm device 110 .

When the artificial intelligence learning unit 123 cannot determine the electrical safety accident state as a precursor, imminent, or occurrence state of an electrical safety accident by a new input node ('actually, an electrical safety accident has occurred, but recognized by the system according to the present invention) '), a new electrical safety accident state is created and added, and the sensing value of the corresponding input node is configured. The artificial intelligence learning unit 123 determines the weight of the artificial neural network to learn the relationship between the new input node and the newly configured output node in the artificial neural network, and the electric safety accident is predicted, imminent, or occurred by the new input node. It can be applied to a new input node of an artificial neural network to improve the case where the state cannot be determined.

The artificial intelligence learning unit 123 does not receive one of the first sensed value of the first sensor 111 or the second sensed value of the second sensor 112 due to other causes such as sensor failure, or does not respond to the sensing condition. If it is not satisfied, a new output node of the artificial neural network may be added to determine the cause of the failure of the sensed value to be received or the occurrence of an abnormality that does not satisfy the sensing condition.

The remaining electrical safety accident imminent status and occurrence status can be applied to the preliminary electrical safety accident imminent status and preliminary electrical safety accident occurrence status in the same way.

When the artificial intelligence calculation unit 123 outputs 1 indicating an electrical safety accident alarm, the artificial intelligence calculation unit 123 generates an electrical safety accident warning signal and transmits it to the alarm transmission unit 125 .

The alarm transmission unit 125 transmits the electrical safety accident warning signal received from the artificial intelligence calculation unit 123 to the reception target terminal 130 through communication.

The firmware update unit 124 transmits the weights updated by the artificial intelligence learning unit 123 or the updated artificial neural network to the artificial intelligence calculation unit 113 of the electrical safety alarm device 110 through communication. This improves the electrical safety alarm accuracy measured by the electrical safety alarm device installed in the field.

3 is a flowchart for explaining a driving process of an electric safety omens prediction algorithm-based intelligent control electric safety service system according to an embodiment of the present invention.

1 and 3, the electric safety omens prediction algorithm-based intelligent control electric safety service system receives a sensing value from the first sensor 111 and the second sensor 112 (S100).

In the electric safety prediction algorithm-based intelligent control electric safety service system, the artificial intelligence calculation unit 113 performs data preprocessing of Equations 1, 2, and 3 in order to distinguish the types of electric safety accidents (S101). ).

Electric safety predictive algorithm-based intelligent control electric safety service system generates waveforms (' V _f, I _f, Z _f '), a waveform feature may be extracted using Equation 4 (S102).

The electrical safety prediction algorithm-based intelligent control electrical safety service system inputs the waveform features ('G _i,j ') extracted by Equation 4 to the input nodes of the artificial neural network, respectively (S103), and to the input image matrix Accordingly, a specific output value in which the weight is reflected in the waveform feature ('G _i,j ') extracted through an artificial neural network to which a predetermined weight can be applied is calculated ( S104 ).

The electric safety omens prediction algorithm-based intelligent control electric safety service system can determine the electric safety accident occurrence state based on a specific output value that reflects the weight as in Equation 5 (S105).

When the electric safety warning prediction algorithm-based intelligent control electric safety service system outputs 1 indicating an electric safety accident alarm, it generates an electric safety accident alarm signal and transmits it to the alarm transmitter (S106, S107).

The electric safety warning prediction algorithm-based intelligent control electric safety service system transmits the received electric safety accident warning signal to the receiving terminal through communication (S108).

4 is a flowchart for explaining a driving process of an electric safety omens prediction algorithm-based intelligent control electric safety service system according to another embodiment of the present invention.

1 and 4, the electric safety prediction algorithm-based intelligent control electric safety service system receives a sensing value from the first sensor 111 and the second sensor 112 (S200).

In the electric safety omens prediction algorithm-based intelligent control electric safety service system, the artificial intelligence calculation unit 113 performs the data preprocessing process of Equations 1, 2, and 3 in order to distinguish the types of electric safety accidents (S201) ).

Electric safety predictive algorithm-based intelligent control electric safety service system generates waveforms (' From V _f, I _f, Z _f '), a waveform feature may be extracted using Equation 4 (S202).

The electric safety omen prediction algorithm-based intelligent control electric safety service system inputs the waveform features ('G _i,j ') extracted by Equation 4 to the input nodes of the artificial neural network, respectively, and is preset according to the input image matrix. The weights can be changed by repeatedly learning the weights that can be applied to the artificial neural network to which the weights can be applied ( S203 ).

The electrical safety omen prediction algorithm-based intelligent control electrical safety service system can calculate the output value of the output node of the artificial neural network (S204).

The electric safety omens prediction algorithm-based intelligent control electric safety service system may determine whether the predictive power of electric safety accidents is improved (S205).

When it is determined that the electrical safety accident predictive ability is improved, the electrical safety omens prediction algorithm-based intelligent control electrical safety service system may generate an updated artificial neural network (electrical safety accident prediction model) (S206).

Electrical safety omens prediction algorithm-based intelligent control electrical safety service system transmits the updated artificial neural network to the artificial intelligence calculation unit through the firmware updater unit, so that the artificial intelligence calculation unit can use the updated artificial neural network (S207).

The described embodiments may be configured by selectively combining all or part of each of the embodiments so that various modifications can be made.

In addition, it should be noted that the examples are for illustrative purposes only, and not for their limitation. In addition, those skilled in the art will understand that various embodiments are possible within the scope of the technical idea of the present invention.

100: Electric safety prediction algorithm-based intelligent control electric safety service system
110: electrical safety alarm device
120: electrical safety platform server
130: reception target terminal
111: first sensor
112: second sensor
113: artificial intelligence calculation unit
114: data transmission unit

Claims (7)

A first sensor for generating a first sensed value by sensing current, voltage, or leakage current, a second sensor for generating a second sensed value by detecting temperature or humidity, and the first sensed value of the first sensor and the The type of electrical safety accident is distinguished using the second sensing value of the second sensor, and it is determined whether an electrical safety accident has occurred using an artificial neural network that applies a threshold set according to the type of the identified electrical safety accident, Electrical safety alarm device including an artificial intelligence calculation unit that generates and transmits an electrical safety accident warning signal when an accident occurs; and
When the electrical safety accident warning signal is received from the artificial intelligence calculation unit, an electrical safety alarm device including an alarm transmitter for transmitting the electrical safety accident warning signal to a reception target terminal;
The artificial intelligence calculation unit,
In order to distinguish the types of electrical safety accidents, using Equation 1 below to generate data for determining power failure during electrical safety accidents,

[ Equation 1 ]

'n' is the sequence number of the measured data, 'T' is the waveform period, 'V' is the voltage sensed value among the first sensed values of the first sensor, and V _rms is the voltage sensed value of the first sensed values of the first sensor. RMS voltage for
Electric safety predictive algorithm-based intelligent control electric safety service system.
delete A first sensor for generating a first sensed value by sensing current, voltage, or leakage current, a second sensor for generating a second sensed value by detecting temperature or humidity, and the first sensed value of the first sensor and the The type of electrical safety accident is distinguished using the second sensing value of the second sensor, and it is determined whether an electrical safety accident has occurred using an artificial neural network that applies a threshold set according to the type of the identified electrical safety accident, Electrical safety alarm device including an artificial intelligence calculation unit that generates and transmits an electrical safety accident warning signal when an accident occurs; and
When the electrical safety accident warning signal is received from the artificial intelligence calculation unit, an electrical safety alarm device including an alarm transmitter for transmitting the electrical safety accident warning signal to a reception target terminal;
The artificial intelligence calculation unit,
Generating data for short circuit discrimination using Equation 2 below to distinguish the types of electrical safety accidents,
[ Equation 2 ]

'n' is the sequence number of the measured data, 'T' is the waveform period, 'V' is the current sensed value among the first sensed values of the first sensor, and I _rms is the current sensed value from the first sensed values of the first sensor. The magnitude of the rms current for
Electric safety predictive algorithm-based intelligent control electric safety service system.
A first sensor for generating a first sensed value by sensing current, voltage, or leakage current, a second sensor for generating a second sensed value by detecting temperature or humidity, and the first sensed value of the first sensor and the The type of electrical safety accident is distinguished using the second sensing value of the second sensor, and it is determined whether an electrical safety accident has occurred using an artificial neural network that applies a threshold set according to the type of the identified electrical safety accident, Electrical safety alarm device including an artificial intelligence calculation unit that generates and transmits an electrical safety accident warning signal when an accident occurs; and
When the electrical safety accident warning signal is received from the artificial intelligence calculation unit, an electrical safety alarm device including an alarm transmitter for transmitting the electrical safety accident warning signal to a reception target terminal;
The artificial intelligence calculation unit,
In order to distinguish the type of electrical safety accident, using Equation 3 below to generate the data for electric leakage determination,
[ Equation 3 ]

'n' is the sequence number of the measured data, 'T' is the waveform period, 'Z' is the leakage current sensing value among the first sensing values of the first sensor, and Z _rms is the leakage current sensing among the first sensing values of the first sensor RMS leakage current magnitude with respect to the value, 'T _Vrms ' is the waveform period of the RMS voltage with respect to the voltage sensed value among the first sensed values of the first sensor
Electric safety predictive algorithm-based intelligent control electric safety service system.
5. The method of any one of claims 1, 3 and 4,
The artificial intelligence calculation unit,
Extracting the waveform features using Equation 4 below to which the waveform and filter image expressed based on the data for power failure determination, the data for short circuit determination, or the data for the leakage determination are applied,

[ Equation 4 ]

'G _i,j ' is the matrix representation of the i-th row and j-th column pixel of the filtered image (however, i, j are integers), 'X' is the matrix representation of the image for the waveform, and 'F' is the filter image The matrix representation of '

' is the convolution, 'F _H ' is the height of the filter, 'F _W ' is the width of the filter, '

' is an integer
Electric safety predictive algorithm-based intelligent control electric safety service system.
6. The method of claim 5,
The artificial intelligence calculation unit,
A specific output value that reflects the weight on the waveform feature extracted by Equation 4 is calculated through an artificial neural network that applies a predetermined weight, and the specific output value that reflects the weight is compared with a preset threshold to determine whether an electrical safety accident alarm or normal An intelligent control electric safety service system based on electric safety prediction algorithm that judges.
7. The method of claim 6,
An artificial intelligence learning unit that changes a weight applied to the extracted waveform feature by reflecting the measured value and output value received from the electrical safety alarm device, and the weight changed by the artificial intelligence learning unit is transmitted to the artificial intelligence calculation unit through communication Thus, the electric safety omens prediction algorithm-based intelligent control electric safety service system further comprising an electric safety platform server comprising a firmware update unit that allows the artificial intelligence calculation unit to apply the changed weight.

Images