<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>To address the problems of slow response and insufficient protection sensitivity of existing overcurrent protection techniques, this paper verifies the feasibility of the design of reduced-current overcurrent protection circuits by using the GBDT model. Firstly, the regression tree is used as a circuit learner for iteration, and the CART algorithm is applied to calculate circuit splitting characteristics and optimal splitting points to obtain the gain of splitting. Then the line loss rate is calculated using the differential evolution method to derive the objective function of the minimized line and the optimal value of the circuit. Finally, the mean value is extracted by calculating the optimal line’s statistical characteristics, the regression tree nodes are calculated using the GBDT model, and the root means the square error is used to verify the effectiveness of the circuit protection design proposed in this paper. The experimental results show that the overcurrent protection value of the circuit sample is around 1.35A and the current limit protection time is around 13ms by verifying the current reduction protection function and indication function of the circuit in the full temperature range. This shows that the GBDT model enables the reduced-current overcurrent protection circuit to perform timely and accurate overcurrent detection actions, and the circuit can be reliably protected under different overcurrent conditions.</p>
</abstract>

To address the problems of slow response and insufficient protection sensitivity of existing overcurrent protection techniques, this paper verifies the feasibility of the design of reduced-current overcurrent protection circuits by using the GBDT model. Firstly, the regression tree is used as a circuit learner for iteration, and the CART algorithm is applied to calculate circuit splitting characteristics and optimal splitting points to obtain the gain of splitting. Then the line loss rate is calculated using the differential evolution method to derive the objective function of the minimized line and the optimal value of the circuit. Finally, the mean value is extracted by calculating the optimal line’s statistical characteristics, the regression tree nodes are calculated using the GBDT model, and the root means the square error is used to verify the effectiveness of the circuit protection design proposed in this paper. The experimental results show that the overcurrent protection value of the circuit sample is around 1.35A and the current limit protection time is around 13ms by verifying the current reduction protection function and indication function of the circuit in the full temperature range. This shows that the GBDT model enables the reduced-current overcurrent protection circuit to perform timely and accurate overcurrent detection actions, and the circuit can be reliably protected under different overcurrent conditions.

Design of reduced-current overcurrent protection circuit based on GBDT model

Applied Mathematics and Nonlinear Sciences

This work is licensed under the Creative Commons Attribution 4.0 International License.

{"article-title":"Design of reduced-current overcurrent protection circuit based on GBDT model"}

To address the problems of slow response and insufficient protection sensitivity of existing overcurrent protection techniques, this paper verifies the...

Design of reduced-current overcurrent protection circuit based on GBDT model

Published Online: Jul 19, 2023

Page range: -

Received: Apr 18, 2022

Accepted: Nov 13, 2022

DOI: https://doi.org/10.2478/amns.2023.2.00061

Keywords
GBDT model, Root mean square error, Objective function, Line loss rate, Differential evolution method

© 2023 Zhixian Wen, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Design of reduced-current overcurrent protection circuit based on GBDT model

Published Online: Jul 19, 2023

Page range: -

Received: Apr 18, 2022

Accepted: Nov 13, 2022

DOI: https://doi.org/10.2478/amns.2023.2.00061

KeywordsGBDT model, Root mean square error, Objective function, Line loss rate, Differential evolution method

© 2023 Zhixian Wen, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Keywords
GBDT model, Root mean square error, Objective function, Line loss rate, Differential evolution method