A novel data transformation model for small data-set learning

In most highly competitive manufacturing industries, the sample sizes are usually very small in pilot runs, in order to quickly launch new products. However, it is always difficult for engineers to improve the quality in mass production runs based on the limited data obtained in this way. Past research has demonstrated that adding artificial samples can be an effective approach when learning with small data-sets. However, a prior analysis of the data is needed to deduce the appropriate sample distributions within which the artificial samples are generated. Johnson transformation is one of the well-known models that can be applied to bring data close to a normal distribution with the satisfaction of certain statistical assumptions. The sample size required for such data transformation methods is usually large, and this thus motivates the efforts of the current study to develop a new method which is suitable for small data-sets. Accordingly, this research proposes the small Johnson Data Transformation method to transform small raw data to normal distributions to generate virtual samples. When compared with four other methods, the results obtained with a real small data-set drawn from the Film Transistor Liquid Crystal Display industry in Taiwan demonstrate that the proposed method is able to effectively improve the forecasting ability with small sample sizes.     

基于Bayes法的电主轴极小子样可靠性研究

针对长寿命的磨削电主轴极小子样的可靠性评估问题,提出了Bayes结合虚拟增广样本的分析方法。首先,在Bayes法基本流程的指导下,研究了基于Bayes法的磨削电主轴可靠性评估方法。根据定时截尾试验的原则对电主轴进行可靠性试验,应用Bayes法结合磨削电主轴试验样本的可靠性试验数据,综合虚拟增广样本法对其可靠性进行评估,最终获得电主轴的可靠性评估结果。最后,将基于Bayes法与基于伪寿命分布法的磨削电主轴极小子样可靠性评估结果进行比较,以验证基于Bayes法可靠性评估理论的合理性。     

Using an attribute conversion approach for sample generation to learn small data with highly uncertain features

Accelerating new product development has become an important marketing strategy for manufacturers who are competing globally. However, this may lead to the small data learning issue. Although machine learning algorithms are used to extract knowledge from training samples, algorithms may not output satisfactory predictions when training sizes are small. This paper provides a real case of a TFT-LCD (thin film transistor liquid crystal display) maker when a new strengthened cover glass is developed using chemical processes. With very little prior experience about the processes involved, engineers attempted to improve the yield rates by determining the parameters from a few pilot-run data. However, owing to the fact that the processes were different from those required to make TFT-LCD panels, the highly uncertain characteristics of the processes led to the use of two virtual sample generation (VSG) approaches, bootstrap aggregating (bagging) and the synthetic minority over-sampling technique, from which unsatisfactory results were obtained. Accordingly, this study was used to develop a systematic VSG method based on fuzzy theory to tackle the learning issue. The experimental results show that support vector regressions built with training sets containing the proposed samples present more precise predictions and thus can help engineers infer more correct manufacturing parameters.     

The Safe-SADT method for aiding designers to choose and improve dependable architectures for complex automated systems

Dependability evaluation is crucial to controlling the risks associated with system failure, and for this reason, it is one of the fundamental steps in automated system design. However, the dependability evaluation methods that are currently exploited are not appropriate, given the level of complexity of such industrial systems. The ineffectiveness of the existing methods led us to develop and propose the Safe-SADT (structured analysis and design technique) method. Our method allows the explicit formalization of functional interactions, the identification of the characteristic values affecting the dependability of complex systems, the quantification of the reliability, availability, maintainability, and safety (RAMS) parameters of the system's operational architecture, and the validation of that operational architecture in terms of the dependability objectives and constraints set down in the functional requirement specifications (FRS). The results presented in this paper are limited to RAM quantification.     

通用小型汽油机进排气系统流通特性与排放性能研究

欧美国家对通用小型汽油机排放法规不断加严,目前中国生产的不同批次通用小型汽油机排放性能差异较大已是行业共有的技术难题.油气混合气浓度对通用小型汽油机排放影响很大,利用作者开发的通用小型汽油机进排气系统稳流试验台,对不同汽油机气缸盖进排气道、空滤器等进行稳流试验,对试验数据进行分析,得出进排气系统流通特性的差异和不同批次的生产波动是批量生产小型汽油机混合气浓度改变和排放性能差异的主要原因.研究结果还表明：通用小型汽油机采用压铸气缸盖,进排气道结构受压铸工艺的限制,流通系数偏低,影响发动机性能的提高,进一步提高进排气系统性能是优化通用小型汽油机性能的一个重要途径.     

Using data continualization and expansion to improve small data set learning accuracy for early flexible manufacturing system (FMS) scheduling

Knowledge derived from limited data gathered in the early manufacturing stages is usually too fragile for a flexible manufacturing system (FMS). Unfortunately, production decisions have to be made quickly in a competitive environment. In a previous study, a strategy using continuous data and domain external expansion methods under a known data domain range was proposed to solve the so-called small data set learning problem in FMS. The present paper goes further in seeking a quantitative method to determine the range of domain external expansion under unknown domain bounds. The research considers the data bias phenomenon that often occurs in small data sets and provides a method for its adjustment. Beyond this, the study also compares the learning results among three types of membership functions (Bell, Trapezoid, Triangular) for data fuzzification. The results show that the proposed approach can advance the learning accuracy of a broad range of applications.     

The European regulations for specimen fabrication prior to performance testing of asphalt mixtures

For fundamental performance testing of asphalt mixtures, specimens have to be produced from an asphalt mixture that is compacted in laboratory. For this purpose, different laboratory compaction methods are available today, which are based on the effects of impact, kneading or rolling action and/or vibration. It is well known that the method used for compaction considerably influences the properties of the specimen, and in turn, the specimen properties significantly influence the outcome of fundamental asphalt mix testing. The choice of compaction method is therefore of utmost importance. Hence, there is general consensus among experts that high priority must be given to the harmonisation of laboratory methods for the production of specimens prior to mix testing. Asphalt mix requirements must not be specified independently from the detailed compaction requirements. In Europe, a harmonised quality assurance system for asphalt mix properties was established recently. In order to guarantee its practicability, also a harmonised approach for the fabrication of specimens for fundamental testing becomes indispensable. In this paper, an overview is given on how standardisation of specimen production has progressed in Europe so far, and the need for further harmonisation is pointed out from a European point of view.