2018
DOI: 10.1177/0020294018789202
|View full text |Cite
|
Sign up to set email alerts
|

Reinforcement learning–based fault-tolerant control with application to flux cored wire system

Abstract: Background: Processes and systems are always subjected to faults or malfunctions due to age or unexpected events, which would degrade the operation performance and even lead to operation failure. Therefore, it is motivated to develop fault-tolerant control strategy so that the system can operate with tolerated performance degradation. Methods: In this paper, a reinforcement learning -based fault-tolerant control method is proposed without need of the system model and the information of faults. Results and Conc… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
13
0

Year Published

2019
2019
2024
2024

Publication Types

Select...
5
2

Relationship

3
4

Authors

Journals

citations
Cited by 11 publications
(13 citation statements)
references
References 35 publications
0
13
0
Order By: Relevance
“…Remark 2: The optimal control ( ) can be obtained only using the state information and the immediate cost because there are only ( ), ( + 1) and in formulas (22) and (23).…”
Section: A Reinforcement Learningmentioning
confidence: 99%
See 1 more Smart Citation
“…Remark 2: The optimal control ( ) can be obtained only using the state information and the immediate cost because there are only ( ), ( + 1) and in formulas (22) and (23).…”
Section: A Reinforcement Learningmentioning
confidence: 99%
“…Reference [18] presented a brief survey on the advances that have occurred in the area of deep learning. From engineering application aspect, the RL/DRL showed an excellent performance after a good training in UAV [19], air-conditioning refrigeration [20], smart power control [21], fault tolerant control [22,23] and so forth [24,25].…”
Section: Introductionmentioning
confidence: 99%
“…c2 = c2 + ∆ c2 (21) For an action network, a1 is denoted as the connected weight between the input layer and hidden layer, a2 as the connected weigh between the hidden layer and output layer, sigmoid function as the activation function with the same form of formula (12). One will update the connected weights a1 and a2 of the action network according to the formulas (22) and 23:…”
Section: Figure 1 Schematic Block-diagram Of the Caftcmentioning
confidence: 99%
“…Recently, RL based FTC have paid attention and some results were reported such as FTC tracking control for MIMO discretetime systems [19], FTC design for a class of nonlinear MIMO discrete-time systems [20] and RL-based FTC for linear discrete-time dynamic systems [21]. It is noticed that learning of RL is a very slow process which makes RL be limited to a small FTC system [22] although RL has succeeded in game, computer vision [23] and robot [24]. Several methods accelerating on RL which aim to a huge system have proposed such as trajectory trace [18], trust region policy optimization(TRPO) [25], actor-critic with experience replay (ACER) [26], proximal policy optimization algorithm(PPO) [27], FPGA [28] and so forth.…”
Section: Introductionmentioning
confidence: 99%
“…In this way, the RL can achieve optimal action based on the current states and environment. Along with incredible success in games [8], the RL has attracted great interest in various industries, such as robot [9,10], fault detection [11], and fault-tolerant control [12,13].…”
Section: Background and Motivationmentioning
confidence: 99%