Grate-kiln-cooler has become a major process of producing iron ore pellets in China. Due to the diversity of the raw materials used and the multi-device multi-variable characteristics,this process still encounters wit...Grate-kiln-cooler has become a major process of producing iron ore pellets in China. Due to the diversity of the raw materials used and the multi-device multi-variable characteristics,this process still encounters with control problem. An attempt was proposed to deal with this issue. The three-device-integrated feature of the process was firstly analyzed to obtain control strategy,and then an intelligent control system using a combination of expert system approach and Takagi-Sugeno（ T-S） fuzzy model was developed. Expert system approach was used to diagnose and remedy the abnormal conditions,while T-S fuzzy model was used to stabilize the thermal state. In the construction of T-S fuzzy rules,antecedents were identified by fuzzy c-mean clustering algorithm incorporated with subtractive clustering algorithm,and consequent parameters were identified by recursive least square algorithm. The control system was applied in a Chinese pelletizing plant and the application results demonstrated its effectiveness of stabilizing the thermal states within three devices.展开更多
Mixture of Experts(MoE)regression models are widely studied in statistics and machine learning for modeling heterogeneity in data for regression,clustering and classification.Laplace distribution is one of the most im...Mixture of Experts(MoE)regression models are widely studied in statistics and machine learning for modeling heterogeneity in data for regression,clustering and classification.Laplace distribution is one of the most important statistical tools to analyze thick and tail data.Laplace Mixture of Linear Experts(LMoLE)regression models are based on the Laplace distribution which is more robust.Similar to modelling variance parameter in a homogeneous population,we propose and study a new novel class of models:heteroscedastic Laplace mixture of experts regression models to analyze the heteroscedastic data coming from a heterogeneous population in this paper.The issues of maximum likelihood estimation are addressed.In particular,Minorization-Maximization(MM)algorithm for estimating the regression parameters is developed.Properties of the estimators of the regression coefficients are evaluated through Monte Carlo simulations.Results from the analysis of two real data sets are presented.展开更多
The structure and characteristics of a large multi-parameter environmental simulation cabin are introduced.Due to the diffculties of control methods and the easily damaged characteristics,control systems for the large...The structure and characteristics of a large multi-parameter environmental simulation cabin are introduced.Due to the diffculties of control methods and the easily damaged characteristics,control systems for the large multi-parameter environmental simulation cabin are diffcult to be controlled quickly and accurately with a classical PID algorithm.Considering the dynamic state characteristics of the environmental simulation test chamber,a lumped parameter model of the control system is established to accurately control the multiple parameters of the environmental chamber and a fuzzy control algorithm combined with expert-PID decision is introduced into the temperature,pressure,and rotation speed control systems.Both simulations and experimental results have shown that compared with classical PID control,this fuzzy-expert control method can decrease overshoot as well as enhance the capacity of anti-dynamic disturbance with robustness.It can also resolve the contradiction between rapidity and small overshoot,and is suitable for application in a large multi-parameter environmental simulation cabin control system.展开更多
1 Introduction In early 2025,DeepSeek launched the open-source model DeepSeek Rl,an outstanding technological innovation that marked a significant breakthrough in the field of AI due to its low cost and high efficienc...1 Introduction In early 2025,DeepSeek launched the open-source model DeepSeek Rl,an outstanding technological innovation that marked a significant breakthrough in the field of AI due to its low cost and high efficiency.The technological highlights of this modeclover multiple innovative aspects,such as reinforcement learning,t,hemixture-of-expertss(MoE)model,knowledge distillation,and multi-head latent attention(MLA).DeepSeek excels at Open AI ol at a low cost and opens to the world.This not only injects new vitality into the development of global AI technology but also provides inspiration and far-reaching influence for the re-establishment of China's educational confidence and Chinese-style innovations.展开更多
基金Item Sponsored by National Natural Science Foundation of China(51174253)
文摘Grate-kiln-cooler has become a major process of producing iron ore pellets in China. Due to the diversity of the raw materials used and the multi-device multi-variable characteristics,this process still encounters with control problem. An attempt was proposed to deal with this issue. The three-device-integrated feature of the process was firstly analyzed to obtain control strategy,and then an intelligent control system using a combination of expert system approach and Takagi-Sugeno（ T-S） fuzzy model was developed. Expert system approach was used to diagnose and remedy the abnormal conditions,while T-S fuzzy model was used to stabilize the thermal state. In the construction of T-S fuzzy rules,antecedents were identified by fuzzy c-mean clustering algorithm incorporated with subtractive clustering algorithm,and consequent parameters were identified by recursive least square algorithm. The control system was applied in a Chinese pelletizing plant and the application results demonstrated its effectiveness of stabilizing the thermal states within three devices.
基金the National Natural Science Foundation of China(11861041,11261025).
文摘Mixture of Experts(MoE)regression models are widely studied in statistics and machine learning for modeling heterogeneity in data for regression,clustering and classification.Laplace distribution is one of the most important statistical tools to analyze thick and tail data.Laplace Mixture of Linear Experts(LMoLE)regression models are based on the Laplace distribution which is more robust.Similar to modelling variance parameter in a homogeneous population,we propose and study a new novel class of models:heteroscedastic Laplace mixture of experts regression models to analyze the heteroscedastic data coming from a heterogeneous population in this paper.The issues of maximum likelihood estimation are addressed.In particular,Minorization-Maximization(MM)algorithm for estimating the regression parameters is developed.Properties of the estimators of the regression coefficients are evaluated through Monte Carlo simulations.Results from the analysis of two real data sets are presented.
基金supported by the Aeronautical Science Foundation of China(No.2012ZD51043)‘‘Fanzhou’’ Youth Scientifc Funds(No.20100504)
文摘The structure and characteristics of a large multi-parameter environmental simulation cabin are introduced.Due to the diffculties of control methods and the easily damaged characteristics,control systems for the large multi-parameter environmental simulation cabin are diffcult to be controlled quickly and accurately with a classical PID algorithm.Considering the dynamic state characteristics of the environmental simulation test chamber,a lumped parameter model of the control system is established to accurately control the multiple parameters of the environmental chamber and a fuzzy control algorithm combined with expert-PID decision is introduced into the temperature,pressure,and rotation speed control systems.Both simulations and experimental results have shown that compared with classical PID control,this fuzzy-expert control method can decrease overshoot as well as enhance the capacity of anti-dynamic disturbance with robustness.It can also resolve the contradiction between rapidity and small overshoot,and is suitable for application in a large multi-parameter environmental simulation cabin control system.
基金supported by the Fundamental Research Funds for the Central Scientific Research Institutes of China National Academy Educational Sciences(Grant No.GYB2024009).
文摘1 Introduction In early 2025,DeepSeek launched the open-source model DeepSeek Rl,an outstanding technological innovation that marked a significant breakthrough in the field of AI due to its low cost and high efficiency.The technological highlights of this modeclover multiple innovative aspects,such as reinforcement learning,t,hemixture-of-expertss(MoE)model,knowledge distillation,and multi-head latent attention(MLA).DeepSeek excels at Open AI ol at a low cost and opens to the world.This not only injects new vitality into the development of global AI technology but also provides inspiration and far-reaching influence for the re-establishment of China's educational confidence and Chinese-style innovations.
