برنامه های کاربردی داده کاوی در ارزیابی سیستم های تهویه معدن
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|22253||2012||5 صفحه PDF||سفارش دهید||3542 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Safety Science, Volume 50, Issue 4, April 2012, Pages 918–922
A mine’s ventilation system is an important component of an underground mining system. It provides a sufficient quantity of air to maintain suitable working environment. Therefore, the status of mine ventilation should be tracked and monitored as a timely matter. Based on former findings and in-depth analysis of mine ventilation systems, a proper early warning model is proposed in this paper for such considerations to improve the mine ventilation safety. The model itself is comprised of two sub-models, and two data mining techniques are used to assist in building each sub-model. One is the optimal indexes selection model which applies the Rough Set theory (RS) to assist the selection of best ventilation indexes. The other is the risk evaluation model based on the Support Vector Machine (SVM) to classify the risk ranks for the mine ventilation system. Testing cases have been used to demonstrate the applicability of this integrated model.
A mine’s ventilation system is an important component of an underground mining system. It should provide a sufficient quantity of air to the underground mine workings, to dilute methane and other contaminants, maintain suitable working environment and prevent accidents from happening. Very often, ventilation is a limiting factor for coal mine production (Cheng et al., 2010). In the running stage of mine ventilation, its status could not be kept constant, and may change timely due to production requirements, mining laws and regulation, etc. Generally, the coal mine ventilation is a super complicated system. Lots of influence factors could control or impact the behaviors of system. Thus, stating from the quantitative point, the system shows a fluctuant wave around a certain value. However, if this wave swing is too large to control by the system itself, it may indicate that any potential risks existing. Many methods or models were proposed to evaluate or assess the mine ventilation system. Jalali et al. (2009) proposed their definition of ventilation network evaluation. In this model, the most important factor to influence the network running was considered to be the branch air resistance. Therefore, by using the conventional methods to determine reliability in the transportation or electrical network, the reliability of each branch in the network was defined. Wang (2004) set up an evaluating model for the mine ventilation system reliability based on a BP neural network approach. A number of influence factors associated with the system reliability were considered in his model in order to achieve the most scientific results. Cheng, 2008 and Cheng et al., 2010 introduced an integrated comprehensive method for selecting and evaluating the most suitable mine ventilation system. Using this method, the severe influence caused by anthropogenic factors or single index can be avoided in the final result. Thus, the selection and evaluation procedure of this method would ensure the selected mine ventilation system is the more rational and economical. Hatakeyama et al. (1992) analyzed the dynamic mine ventilation using airflow rate based on anemometer measurements. It could provide a real-time understanding the ventilation conditions, and also was expected to calculate the CH4 or CO gas emission from surrounding strata to minimize the risk of disaster. Mitchell (1996) summarized accident prevention strategies to avoid loss of the property or life during a mine fire event. However, most proposed methods failed to identify the relationship between the historical ventilation records and the potential risk. In this paper, an integrated early warning model is proposed to improve the mine ventilation safety due to such above considerations. The model itself is comprised of two sub-models. One is called the optimal selection model and the other is the risk evaluation model. All of them are based on the data mining technique which is an outstanding tool to mining the connections between the historical data and the induced risks. A test-case demonstration result shows that this integrated model has good applicability and could be applied in practices.
نتیجه گیری انگلیسی
An integrated early warning model for evaluating the mine ventilation system has been proposed. The model is comprised of the optimal indexes selection and the risk evaluation model models. A set of data mining tools have been used to derive solutions that are accurate and scientifically sound. The proposed model could be used as a useful tool not only to track the status of the mine ventilation system timely but also to assess the upgraded one and help mining engineers take proper measurements to avoid any potential accident risks. Testing-case applications show that this integrated model has better accuracy and reliability and could be applied in practices.