石家莊煤礦風(fēng)機(jī)運(yùn)行狀態(tài)的預(yù)測(cè)研究

 石家莊風(fēng)機(jī)廠 石家莊風(fēng)機(jī) 石家莊市風(fēng)機(jī)廠 石家莊風(fēng)機(jī)維修 石家莊風(fēng)機(jī)銷售

石家莊煤礦風(fēng)機(jī)運(yùn)行狀態(tài)的預(yù)測(cè)研究
煤礦風(fēng)機(jī)是礦井工作人員的呼吸機(jī)，其可靠性直接影響井下生產(chǎn)和工人的生命安全，是重要的通風(fēng)設(shè)備。目前我國(guó)對(duì)礦山設(shè)備的維修很落后，大都采用傳統(tǒng)的定期維修方式，這種維修方式會(huì)造成維修過剩或維修不足，其結(jié)果可能是原本穩(wěn)定的設(shè)備，經(jīng)過維修反而更易出現(xiàn)故障，或設(shè)備“帶病”運(yùn)行造成重大事故。所以一種新的維修方式—按狀態(tài)維修成為設(shè)備維修的發(fā)展方向，這種維修的特點(diǎn)是，沒有具體的維修周期，通過對(duì)設(shè)備運(yùn)行狀態(tài)的實(shí)時(shí)監(jiān)測(cè)，以及歷史數(shù)據(jù)的分析，判斷機(jī)器設(shè)備的運(yùn)轉(zhuǎn)狀態(tài)，并對(duì)未來某段時(shí)間的設(shè)備運(yùn)轉(zhuǎn)狀態(tài)進(jìn)行預(yù)測(cè)，根據(jù)監(jiān)測(cè)數(shù)據(jù)判斷不同的故障類型，制定不同的維修措施。為此，進(jìn)行了煤礦通風(fēng)機(jī)運(yùn)行狀態(tài)預(yù)測(cè)方法的研究。
論文通過對(duì)風(fēng)機(jī)故障機(jī)理的研究提出了基于振動(dòng)信號(hào)的風(fēng)機(jī)運(yùn)行狀態(tài)的預(yù)測(cè)研究，通過對(duì)信號(hào)分析方法以及預(yù)測(cè)方法的歸納分析，同時(shí)考慮到風(fēng)機(jī)振動(dòng)信號(hào)的非平穩(wěn)性，提出了 EMD 與神經(jīng)網(wǎng)絡(luò)相結(jié)合的風(fēng)機(jī)運(yùn)行狀態(tài)預(yù)測(cè)方法。
論文將傳感器技術(shù)與計(jì)算機(jī)技術(shù)相結(jié)合，構(gòu)建了礦井風(fēng)機(jī)振動(dòng)數(shù)據(jù)的實(shí)時(shí)采集系統(tǒng)，完成了系統(tǒng)軟硬件設(shè)計(jì)；將 LabVIEW 與 SQL 數(shù)據(jù)庫技術(shù)相結(jié)合構(gòu)建了礦井風(fēng)機(jī)的數(shù)據(jù)存儲(chǔ)與管理系統(tǒng)，實(shí)現(xiàn)對(duì)實(shí)時(shí)采集數(shù)據(jù)、定周期采集數(shù)據(jù)、故障數(shù)據(jù)、診斷結(jié)果數(shù)據(jù)以及現(xiàn)場(chǎng)技術(shù)人員診斷與維修數(shù)據(jù)的有效管理，為對(duì)風(fēng)機(jī)運(yùn)行狀態(tài)作進(jìn)一步分析提供完整的歷史檔案；將 EMD 與神經(jīng)網(wǎng)絡(luò)相結(jié)合構(gòu)建了基于EMD 的神經(jīng)網(wǎng)絡(luò)風(fēng)機(jī)運(yùn)行狀態(tài)預(yù)測(cè)模型，在 MATLAB 環(huán)境下實(shí)現(xiàn)了風(fēng)機(jī)振動(dòng)信號(hào)的 EMD 分解，完成了直接神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)方法與基于 EMD 的神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)方法的比較研究，結(jié)果表明后者有更好的預(yù)測(cè)準(zhǔn)確性。
Abstract
The coal mine ventilator is the mine workers’ breathing machine. Its spindlereliability influence the mine production and the safety of workers directly. It is animportant ventilative equipment. At present, mining equipment maintenance in ourcountry is developing lag behind. We always use the traditional and regular maintenance.It may cause the excessive or inadequate maintenance. The stable equipment may haveproblems by this maintenance, or cause a major accident by error operating. Thus a newway of maintenance, condition based-maintenance may be the direction of equipmentmaintenance. It doesn’t have specific maintenance cycles. This method can determine theoperation state of the machine and predict a future period of equipment’ running state bymonitoring the operating status of the equipment in real time and analyzing the historydata. According to the monitoring data, we can judge the different fault types and makedifferent maintenance measures. Therefore, this article has researched the predictionmethod of mine ventilator’s running state.
This article has proposed the research of the mine ventilator’s running stateprediction through researching the fault mechanism of the mine ventilator. At the sametime the author has proposed the prediction method of the mine ventilator running stateprediction combing EMD with the neutral network considering the nonstationarity of themine ventilator’ vibration signal.
Firstly, the author built a real-time data acquisition system of the mine ventilator anddesigned the hardware and software of the system by combining the sensor withcomputer technology. Secondly, the author built a data storage and management systemof the mine ventilator by combining the LabVIEW and the SQL database technology. Inthis way, the effective management of the Real-time collected data, fixed cycle data, faultdata, diagnosis result data and field technical personnel diagnosis and maintenance datahas been achieved. They can provide a complete history file for further analysis of themine ventilator running state. Finally, the author built a mine ventilator forecast modelbased on the EMD and neural network by combining the EMD and neural networktechnology. The author used the EMD to decompose mine ventilator signals by theMATLAB completed the comparative study on the method of direct neural networkprediction and the method of EMD-based neural network prediction. The results showthat the latter has better predictive accuracy.Figure 56； Table 22； Reference 60Keywords: condition monitoring, LABVIEW, SQL, Hilbert-Huang analysis, NeuralNetworksChinese books catalog: TH17