Examples of successful applications

 

Reasonable and successful applications of statistics in practice.

 

Every day we are making an effort to meet the expectations of our clients. The following examples provide evidence therefore:

 

1) Mobility

a)  Failure behavior of a vehicle component

 

The statistical analysis of few failures occurred at a vehicle component during customer usage showed a significant deviation from the reliability target. By the use of a prediction model, a pending series damage has been identified in an early phase. The clear result motivated the responsible engineers to fastly react on the problem. After the improved component has been validated, it has been released for series production and  prevented  high warranty costs.

 

b) Investigation of wheel forces at a rail car

 

The design of wheels of rail cars requires the knowledge of the corresponding wheel forces. In particular, the effects of t he car design, the routes and the track condition have tob e considered. To quantify and compare these effects, simple but effective regression models have been applied. During the project work, the engineers learned the required statistics know-how to carry out such analyses autonomously in future applications.

 

c) Less effort – more power

 

The performance of shaft-hub joints in steering systems has been optimized by using statistical design of experiments. A short report has been published (in German) in Automobil Produktion Nr. 1/2004, S. 39, (2004).

 

 

2) Life Sciences

d) Investigation of effects on the learning aptitude of children


To assess the effects of regular musical practice of children on their learning aptitude, longitudinal and profile studies have been planned. The success in learning  has been measured by standardized reading tests and calculation checks. The longitudinal studies have been planned to investigate the time effect oft the musical activities while the profile studies were for the comparison of different groups of children (separated by behavioural and socio-economic characteristics etc.). The data have been collected either by survey/observation or by measurement of brain activity.

 

The statistical methods have been applied to estimate required group sizes and to analyse the effects of interest.

 

JeKi, Verbundprojekt AMseL, 2010.


e) Packaging of a pharmaceutical


The package material of a pharmaceutical is encountered to different stresses and hast o fulfill very strict quality requirements. To support the application of statistical methods during the development of the packaging, the responsible engineers have been trained specifically.

 

f) Kontrollkarte statt Poker. (Application of statistical process control, in German) Process Nr. 11, S. 26-27, (1/2004).

 

g) Gelungene Mischung. (Application of experimental design in plastic industry, in German) Pharma and Food Nr. 7, S. 36-38, (2/2004).

 

h) Statistisch optimierter Gourmet-Mix. (Biometric application of experimental design, in German). Chemische Rundschau Nr. 21, S. 24, (2002).

 

 

3) Energy

i) Prediction of wind speed


When selecting the location for a new wind farm, beside the ecological factors the expected wind speed plays an important role. To estimate the expected energy yield, the statistical analysis of historical meteorological data is required. To compare different locations,  corresponding wind signals have been compared. This provided the basis to select the most effective place for the new wind farm.

 

j) Battery of electric vehicle


To check the suitability for daily use of a mobile energy storage, it has to be tested under various requirements. Not only the variety of environmental conditions (temperature, …) but also the variability oft he users‘ dis/charging behaviour are important input parameters fort he generation of an adequate test plan. In addition, the prototypes of the storage are expensive, the lab infrastructure complex and a certain test duration is required to see potential degradation phenomena.

 

In this situation, a statistical experimental design provided the basis to reach with a minimum number of tests a maximum of information about the relation between the investigated factors and the ageing behaviour of the storage. After execution of the tests, the statistical analysis of the data showed clearly the dominating factors which are responsible for the degradation.