Search results

Safety management plays an important part in coal mine construction. Due to complex data, the implementation of the construction safety knowledge scattered in standards poses a challenge. This paper aims to develop a knowledge extraction model to automatically and efficiently extract domain knowledge from unstructured texts.

Bidirectional encoder representations from transformers (BERT)-bidirectional long short-term memory (BiLSTM)-conditional random field (CRF) method based on a pre-training language model was applied to carry out knowledge entity recognition in the field of coal mine construction safety in this paper. Firstly, 80 safety standards for coal mine construction were collected, sorted out and marked as a descriptive corpus. Then, the BERT pre-training language model was used to obtain dynamic word vectors. Finally, the BiLSTM-CRF model concluded the entity’s optimal tag sequence.

Accordingly, 11,933 entities and 2,051 relationships in the standard specifications texts of this paper were identified and a language model suitable for coal mine construction safety management was proposed. The experiments showed that F1 values were all above 60% in nine types of entities such as security management. F1 value of this model was more than 60% for entity extraction. The model identified and extracted entities more accurately than conventional methods.

This work completed the domain knowledge query and built a Q&A platform via entities and relationships identified by the standard specifications suitable for coal mines. This paper proposed a systematic framework for texts in coal mine construction safety to improve efficiency and accuracy of domain-specific entity extraction. In addition, the pretraining language model was also introduced into the coal mine construction safety to realize dynamic entity recognition, which provides technical support and theoretical reference for the optimization of safety management platforms.

This paper aims to extend the previous contributions about data-driven control in aircraft control system from academy and practice, respectively, combining iteration and learning strategy. More specifically, after returning output signal to input part, and getting one error signal, three kinds of data are measured to design the unknown controller without any information about the unknown plant. Using the main essence of data-driven control, iterative learning idea is introduced together to yield iterative learning data-driven control strategy. To get the optimal data-driven controller, other factors are considered, for example, adaptation, optimization and learning. After reviewing the aircraft control system in detail, the numerical simulation results have demonstrated the efficiency of the proposed iterative learning data-driven control strategy.

First, considering one closed loop system corresponding to the aircraft control system, data-driven control strategy is used to design the unknown controller without any message about the unknown plant. Second, iterative learning idea is combined with data-driven control to yield iterative learning data-driven control strategy. The optimal data-driven controller is designed by virtue of power spectrum and mathematical optimization. Furthermore, adaptation is tried to combine them together. Third, to achieve the combination with theory and practice, our proposed iterative learning data-driven control is applied into aircraft control system, so that the considered aircraft can fly more promptly.

A novel iterative learning data-driven strategy is proposed to efficiently achieve the combination with theory and practice. First, iterative learning and data-driven control are combined with each other, being dependent of adaptation and optimization. Second, iterative learning data-driven control is proposed to design the flight controller for the aircraft system. Generally, data-driven control is more wide in our living life, so it is important to introduce other fields to improve the performance of data-driven control.

To the best of the authors’ knowledge, this new paper extends the previous contributions about data-driven control by virtue of iterative learning strategy. Specifically, iteration means that the optimal data-driven controller is solved as one recursive form, being related with one gradient descent direction. This novel iterative learning data-driven control has more advanced properties, coming from data driven and adaptive iteration. Furthermore, it is a new subject on applying data-driven control into the aircraft control system.

This new paper aims to extend the authors’ previous contributions about open-loop aircraft flutter test to closed-loop aircraft flutter test by virtue of the proposed direct data–driven strategy. After feeding back the output signal to the input and introducing one feedback controller in the adding feedback loop, two parts, i.e. unknown aircraft flutter model and unknown feedback controller, exist in this closed-loop aircraft flutter system, simultaneously, whose input and output are all corrupted with external noise. Because of the relations between aircraft flutter model parameters and the unknown aircraft model, direct data–driven identification is proposed to identify that aircraft flutter model, then some identification algorithms and their statistical analysis are given through the authors’ own derivations. As the feedback controller can suppress the aircraft flutter or guarantee the flutter response converge to one desired constant value, the direct data–driven control is applied to design that feedback controller only through the observed data sequence directly. Numerical simulation results have demonstrated the efficiency of the proposed direct data–driven strategy. Generally, during our new information age, direct data–driven strategy is widely applied around our living life.

First, consider one more complex closed loop stochastic aircraft flutter model, whose input–output are all corrupted with external noise. Second, for the identification problem of closed-loop aircraft flutter model parameters, new identification algorithm and some considerations are given to the corresponding direct data–driven identification. Third, to design that feedback controller, existing in that closed-loop aircraft flutter model, direct data–driven control is proposed to design the feedback controller, which suppresses the flutter response actively.

A novel direct data–driven strategy is proposed to achieve the dual missions, i.e. identification and control for closed-loop aircraft flutter test. First, direct data–driven identification is applied to identify that unknown aircraft flutter model being related with aircraft flutter model parameters identification. Second, direct data–driven control is proposed to design that feedback controller.

To the best of the authors’ knowledge, this new paper extends the authors’ previous contributions about open-loop aircraft flutter test to closed-loop aircraft flutter test by virtue of the proposed direct data–driven strategy. Consider the identification problem of aircraft flutter model parameters within the presented closed loop environment, direct data–driven identification algorithm is proposed to achieve the identification goal. Direct data–driven control is proposed to design the feedback controller, i.e. only using the observed data to design the feedback controller.