Owing to the increasing complexity of the power system, the difficulty of fault diagnosis in the power grid is also increasing. In response to the issue of continuously decreasing fault diagnosis accuracy, a power grid fault diagnosis model based on an improved deep Q-network model is raised. This model enhances its information integration capability by constructing a fault parameter recognition model and introducing alarm information for text processing. By identifying power grid fault parameters and processing alarm information, the efficiency and accuracy of fault diagnosis can be improved. The experimental results show that the model shows significant performance improvement in multiple state dimensions, and is significantly better than the traditional algorithm in single fault diagnosis and multi-fault diagnosis scenarios. The results show that the proposed method has significant advantages in the accuracy of fault prediction, processing efficiency and anti-noise ability, which verifies the validity and practicability of applying this model in complex power systems. It also emphasises the importance of combining reinforcement learning with unstructured data to further promote the development of smart grid technology.

Inderscience is a global company, a dynamic leading independent journal publisher disseminates the latest research across the broad fields of science, engineering and technology; management, public and business administration; environment, ecological economics and sustainable development; computing, ICT and internet/web services, and related areas.

Construction workers are prone to construction fatigue in high-intensity working environments, and failure to receive effective rest may result in casualties and property damage. The study uses deep learning algorithms to construct an intelligent fatigue prediction model aimed at accurately assessing the fatigue status of construction workers. The study takes smartphones to collect basic data and inputs it into an improved random forest algorithm for fatigue feature recognition. Then, an intelligent construction fatigue recognition model is established based on the improved random forest algorithm. The research model had an accuracy rate of 94.7% in recognising different human movements, and an accuracy rate of 91% in predicting construction fatigue. The designed method accurately predicts the complete exhaustion, fatigue, concentration and excitement states of workers, and its predictive ability is superior to other prediction models. The research model can effectively assist construction managers in accurately detecting workers' fatigue status and taking timely intervention measures to reduce safety accidents.

This study focuses on estimating conditional stress strength reliability of a system using ranked set sampling, when stress and strength variables follow independent exponential distributions. Two estimation methods are used, namely Maximum Likelihood Estimation (MLE) and bootstrap estimation. The asymptotic confidence interval is constructed based on a maximum likelihood estimator and the Boot-P confidence interval is constructed. A simulation study is carried out to determine the Mean Square Error (MSE) and length of the confidence interval. This study uses MSE and the length of the confidence interval to compare the estimator based on ranked set sampling to that based on simple random sampling in the context of exponential distribution.

Robot navigation in complex and dynamic environments remains a challenging problem, requiring methods that can efficiently adapt to unforeseen obstacles and goal-oriented tasks. This paper presents a novel approach that combines the biologically-inspired Attractor Dynamics Approach with the Deep Deterministic Policy Gradient (DDPG) algorithm to enable a mobile robot, specifically the e-puck robot, to navigate through cluttered spaces while avoiding collisions with moving obstacles effectively. The Attractor Dynamics Approach utilises attractors as goals and repulsive forces to avoid obstacles, offering robust and goal-oriented navigation even with very low-level sensory information. In parallel, the DDPG-based reinforcement learning component fine-tunes the robot's motion controls based on range sensor readings, ensuring precise and adaptive obstacle avoidance. The integration of these two techniques empowers the robot to autonomously explore its environment, dynamically adjust its trajectory and reach predefined targets successfully and safely.

Inderscience is a global company, a dynamic leading independent journal publisher disseminates the latest research across the broad fields of science, engineering and technology; management, public and business administration; environment, ecological economics and sustainable development; computing, ICT and internet/web services, and related areas.

Denitration System (DS) is an important part of the production of thermal power plants, and any failure that occurs during its operation will seriously affect the productivity and safety of thermal power plants. Therefore, this paper proposes a Fuzzy Fault Tree (FFT) analysis method for DS of thermal power plants. Firstly, the theory of FFT is described, then the process flow of DS is introduced, the main faults in the operation of the system are analysed and the FFT is established for the main faults. The ambiguity of the basic event probability is solved by fuzzy set theory and expert judgment. After qualitative and quantitative analysis, the weak parts of the system and the degree of importance of the basic events are evaluated, so that it's of great significance to formulate protective measures in advance for the weak links of the system.

Through the analysis of numerous case samples, it is easy to find that the numerous reasons for the frequent occurrence of coal mine accidents are the human factors, and human-caused accidents have become the main cause of coal mine safety. The causes of human-caused accidents are very complex as well as the characteristic mechanisms of accidents are difficult to extract and carry out accurate data measurement, so the traditional engineering safety measurement methods have great limitations. They cannot accurately evaluate the causes of human-caused accidents in coal mines in the problem of high-dimensional pattern recognition. Based on this, the main work of the study focused on selecting the algorithm with good performance of simulated evolution-ant colony clustering algorithm as the technical support for the evaluation work, combining with the current mine production practice, four categories of factors from individual worker factors, environmental factors, life event factors and biological rhythm factors were clustered and analysed, scientific classification patterns as well as mathematical models were established to quantitatively analyse. The analysis of human misbehaviour is analysed and predicted. A behaviour correction model is established based on the human causes derived from the analysis so as to achieve the effect of accident prevention.