KR-20260063111-A - AI-based preventive diagnostics digital power protection, monitoring and control system

Abstract

The present invention relates to an AI preventive diagnosis digital power protection, monitoring, and control system that prevents arc accidents caused by electric shocks by diagnosing partial discharge in a power distribution board. More specifically, the invention proposes a method for detecting partial discharge in real time by training an AI with the degree of impedance vibration, the degree of shoulder generation, and the degree of ultrasonic generation through an ultrasonic sensor, and by implanting the learned data into hardware to quantify the degree of impedance vibration, the degree of shoulder generation, and the degree of ultrasonic generation.

Inventors

• 이재규
• 김영주

Assignees

• 주식회사 리폼테크

Dates

Publication Date

20260507

Application Date

20241030

Claims (7)

1. AI preventive diagnosis digital power protection, monitoring, and control system characterized by detecting the magnitude of impedance while any one of impedance vibration, shoulder detection, or ultrasound is detected, and determining that partial discharge has occurred if a decrease in magnitude is determined.
2. In paragraph 1, An AI preventive diagnostic digital power protection, monitoring, and control system characterized by training an AI artificial intelligence to determine the degree of impedance vibration, the degree of shoulder occurrence, and the degree of ultrasonic occurrence, which are software operations, during the process of impedance vibration, shoulder detection, ultrasonic detection, and impedance magnitude detection, and by implanting the learned data into hardware to quantify the degree of impedance vibration, the degree of shoulder occurrence, and the degree of ultrasonic occurrence to determine and alarm partial discharge in real time.
3. In paragraph 1, An AI preventive diagnostic digital power protection, monitoring, and control system characterized by determining that the impedance magnitude has decreased due to partial discharge if %Z ≥ %Zset. (Here, %Z(impedance percentage) = ( Zm[memory impedance value]/Z[current impedance value] )*100, %Zset is the setting value of the impedance percentage, and the initial value is 200%).
4. In paragraph 1, impedance vibration detection is ① Calculate %Z ② If %Z ≥ %Zset, increment Nvc_cnt by '1' ③ If Nvc_cnt ≥ Nvc_set within half a cycle, increment Npd_cnt by '1'. ④ An AI preventive diagnosis digital power protection, monitoring, and control system characterized by determining that partial discharge has occurred if Npd_cnt ≥ Npd_set within 50mS.
5. In paragraph 1, Shoulder detection is ① Calculation of upper and lower limits for 'near-zero current magnitude' Izero_curr = ( : Effective memory value from 4 cycles ago) Here, the electrical angle varies depending on the number of samples when calculating Izero_curr (for example, the electrical angle is 15° for 24 samples in one cycle). ② The section where the current magnitude is 'nearly zero' Every accompaniment cycle of each stage An AI preventive diagnosis digital power protection, monitoring, and control system characterized by adding Nzero_cnt (Shoulder Count) by 1 whenever Sample Data satisfying the condition occurs, and determining that Shounder has occurred if the Nzero_cnt value is greater than a certain number of times (Nzero_set).
6. In paragraph 1, ultrasonic detection ① Convert partial discharge output into voltage (4~20mA)x250Ω = 1,000mV ~ 5,000mV (AD input voltage range: 0 ~ 5V) (4~20mA)x500Ω = 2,000mV ~ 10,000mV (AD input voltage range: 0 ~ 10V) ② Convert the transformed voltage value into partial discharge intensity (Nultra_level: 0 ~ 50) (1,000mV~5,000mV)-1,000≥0~4,000≥(0 ~ 4,000)*1.25/100≥0 ~ 50 (2,000mV~10,000mV)-2,000≥0 ~ 8,000≥(0~8,000)*1.25/200≥0 ~ 50 AI preventive diagnostic digital power protection, monitoring, and control system characterized by detection when
7. In paragraph 2, An AI preventive diagnosis digital power protection, monitoring, and control system characterized by applying AI techniques to three types of information—impedance vibration, shoulder detection, and ultrasonic detection—as a partial discharge detection method to detect them using artificial intelligence techniques, and running an AI algorithm using the three types of information—impedance vibration degree, shoulder occurrence degree, and ultrasonic detection degree—as AI input data to infer %Zset, Nvc_set, Npd_set, Nzero_set, and preventive diagnosis operation judgment values.

Description

AI-based preventive diagnostics digital power protection, monitoring and control system The present invention relates to an AI preventive diagnosis digital power protection, monitoring, and control system that prevents arc accidents caused by electric shocks by diagnosing partial discharge in a power distribution board. More specifically, the invention proposes a method for detecting partial discharge in real time by training an AI with the degree of impedance vibration, the degree of shoulder generation, and the degree of ultrasonic generation through an ultrasonic sensor, and by implanting the learned data into hardware to quantify the degree of impedance vibration, the degree of shoulder generation, and the degree of ultrasonic generation. Power receiving and substation facilities are facilities that receive extra-high voltage or high voltage from a power company under a supply contract, step down the voltage required by consumers, and distribute it. The components of these facilities can be broadly classified into three types: receiving facilities that receive extra-high voltage and high voltage power from a substation, substation facilities for stepping down the extra-high voltage and high voltage to the low voltage required by consumers, and distribution facilities for supplying the stepped-down power to consumers. These facilities are housed within a partition structure called a cubicle, and inside the cubicle, switches, extra-high voltage circuit breakers, transformers, low voltage circuit breakers, power fuses, lightning arresters, current transformers, voltage transformers, and various power meters are housed. With the emergence of new industries such as AI and ESS, as well as the diversification and advancement of industries, the increase in power capacity and the digitalization of power protection, monitoring, and control are accelerating, leading to an increase in the installation and use of switchgear and consequently, an increase in accidents involving switchgear. When analyzing the causes of partial discharge accidents in switchgear, they can be broadly classified into four categories: poor contact, insulation degradation due to aging, tracking, and corona. If these four types of accidents progress, they ultimately lead to arc explosion accidents, which are parallel arc accidents. Since it is difficult to restore power in a short period of time after a power outage, economic losses increase significantly. Therefore, it is necessary to prevent power outage damage caused by arc accidents by identifying common phenomena in the causes of switchgear accidents and preventing those phenomena from progressing into arc explosion accidents. The four major causes of switchgear accidents are as follows. 1. Poor contact (mainly series arc) In the high-voltage section, poor contact occurs due to loose screws or loose connections, and as a result, contact resistance increases, generating heat through I₂R . As this heat increases, the air becomes ionized, causing partial discharge (series arc) at the contact point, which progresses into an arc accident. Unlike the high-voltage section, in the low-voltage section, if poor contact occurs, heat is generated through I₂R due to increased contact resistance, and this heat causes the surrounding wire insulation to melt, leading to insulation degradation (or insulation breakdown) of the wires, which progresses into a partial discharge (parallel arc) between the two wires. 2. Insulation degradation due to aging (parallel arc) The process of insulation material deteriorating in power equipment installed inside switchboards can be classified into natural (insulation) deterioration due to aging and insulation deterioration in which the initial physical properties are altered or damaged due to long-term exposure to heat, dust, moisture, overload, and environmental stress. Due to this insulation deterioration, minute partial discharge signals are generated, and repeated partial discharges create a carbonized conductive path, which progresses into an arc accident (parallel arc). 3. Trekking Due to environmental pollution such as dust, moisture, and smoke around the switchgear, the insulation layer between conductors is carbonized by partial discharge (a type of parallel arc) through the wire sheathing in contact with the terminal block, and leakage current gradually increases along this carbonized conductive path, leading to tracking breakdown. 4. Corona Corona discharge is a phenomenon in which sound or light is emitted as the insulation of air in extra-high voltage facilities is partially destroyed. Although the scale varies depending on the structure of the conductor, the magnitude of the applied voltage, and weather conditions, it is an inevitable phenomenon in power facilities that use air as an insulator. In particular, it is known that the corona initiation voltage is lower and the discharge sound is louder when the conductor and the supporting insulation are conta