To investigate the effects of various random factors on the preventive maintenance （PM） decision-making of one type of two-unit series system, an optimal quasi-periodic PM policy is introduced. Assume that PM is perfect for unit 1 and only mechanical service for unit 2 in the model. PM activity is randomly performed according to a dynamic PM plan distributed in each implementation period. A replacement is determined based on the competing results of unplanned and planned replacements. The unplanned replacement is trigged by a catastrophic failure of unit 2, and the planned replacement is executed when the PM number reaches the threshold N. Through modeling and analysis, a solution algorithm for an optimal implementation period and the PM number is given, and optimal process and parametric sensitivity are provided by a numerical example. Results show that the implementation period should be decreased as soon as possible under the condition of meeting the needs of practice, which can increase mean operating time and decrease the long-run cost rate.
An airway pressure and flow data acquisition system is developed to investigate the approach to building the bi-level positive airway pressure BiPAP in a ventilator.A number of experiments under different breathing situations and states are conducted and the experimental data are recorded.According to the data from these experiments the variation characteristics of the pressure and flow are analyzed using Matlab. The data analysis results show that the pressure increases while the flow decreases in the expiratory phase contrarily the pressure decreases while the flow increases in the inspiratory phase during the apnea state both the pressure and the flow remain unchanged. According to the above variation characteristics of breath a feedback-based method for creating bi-level positive airway pressure is proposed. Experiments are implemented to verify the BiPAP model. Results demonstrate that the proposed method works effectively in following respiration and caters well to most polypnea and apnea events.
目的显微细胞成像系统获取的图像序列由于光照、电磁干扰等因素的影响,不可避免地存在一定程度的噪声,消除噪声得到清晰的细胞图像是后续细胞形态特征提取和分析的首要步骤。本文引入细胞图像序列的时域信息来构建时空曲率正则化约束,以实现细胞图像序列的去噪处理。方法首先,利用细胞图像序列的空域和时域相关性,构建基于时空曲率正则化的图像序列去噪模型;然后,通过增广拉格朗日乘子法实现模型的优化求解;最后,分别通过对叠加有不同高斯白噪声水平的纤维母细胞和多能干细胞图像序列进行去噪实验,以验证去噪效果。结果与总变分去噪法、三维阈值剪切去噪法和空间曲率正则化去噪法相比较,基于时空曲率正则化的细胞图像序列去噪方法应用于2组细胞图像序列去噪的视觉效果,及峰值信噪比(peak signal to noise ratio,PSNR)和结构相似度(structural similarity,SSIM)都优于其他3种方法。结论与其他3种去噪方法相比,此方法更加充分利用了细胞图像序列的时域信息,去噪后能有效地维持图像对比度,振铃效应不明显,对高斯噪声具有更好的适应性和稳定性,可应用于细胞形态变化检测的前期处理阶段。
