Failure analysis paper in civil engineering

Table of Contents Introduction 2 Common Types of Failure in Electrical Engineering 3 Failure Analysis Method 4 Preventative Methods 4 Forensic Methods 5 References 6

Product failure is inevitable when it comes to product development. But as engineers, our job is to reduce the probability of the product failing while also allowing a product to work efficiently. I am an electrical engineer focusing on the field of microcontrollers and microprocessors, and I have worked with building electrical devices.

One of the most common parts we, as electrical engineers, deal with are resistors. Each of these resistors have a supposed resistance, but depending on the tolerance of the resistor, the said resistance will very. Resistors with less tolerance, meaning that the resistance values will range more closely to the listed resistance, are generally more expensive. Figure 1 illustrates a Monte Carlo analysis of off-the-shelf parts purchased by Dr. Shover and his team [6].

Figure 1: Bimodal distribution of a critical parameter

Here, we see how the user was seeking a parameter of at least 90 and at most 120. When purchasing 10,000 parts, the user received a variety of different parameters which demonstrates that when purchasing a product, its stated value- whether it is resistance, capacitance, voltage, etc.- may vary. So, in order to decrease cost of fabrication, engineers must work with tolerance levels and find the cheapest resistors that will guarantee the functionality of the product. Although engineers do multiple simulations to calculate efficient tolerance levels for resistors, errors still occur in real life. One of the most common types of failure in my field is software “software system failure” [1]. Because software and hardware are directly dependent on each other, if a system (hardware) isn’t built to perform particular tasks, there is no way someone can code (software), or command, the system to perform the task- which is a big reason why software issues occur [1]. There are other technical issues that occur very often like: “slow response from application server, pages don’t download properly, applications are not compatible with browser, etc.” [1]. These issues occur because of the different operating systems that are present, and the compatibility of the software to the operating system. Technical issues are minor setbacks that can be resolved rather easily, but they do take time out of an engineer’s schedule. Other issues are more difficult to resolve, and engineers must do their best to prevent these issues. Because electrical systems are composed of many functioning parts, they can be considered systems of systems (SoS). SoS are more difficult to diagnose because you need to find a failure of a particular system inside of a more complex system [2].

Due to the complexity electrical systems, engineers must do their best in order to minimize product failure. There are 2 ways in which engineers can prevent predicted product failure: using non-destructive or destructive preventative methods. Usually, engineers would prefer non-destructive preventative methods because they allow for larger firms to save money. Many engineers, and engineering firms, use simulation software to design products make it likely that it will function properly in the real world [3]. These software are non-destructive, and allow engineers to predict how a product will operate. The downside to using simulation software to prevent product failure is that they can be time consuming. Some computer architectures are very complex and require many operations. Because of this, simulations can be very tedious and time consuming in performing all the functionalities of a CPU, or other complex systems [4]. Destructive preventative methods are not very common in the field of microprocessors, because they would cause way too much money loss for large corporations. Before simulation technology, engineers would use destructive preventative methods. They would build a product, and simply test it. If it did not work, then they would waste the parts, which costs money, and would lose the time they spent on building the device. Due the popularity and efficiency of simulation software, many people and corporations do not use destructive preventative methods.

Like preventative methods, forensic methods consist of destructive and non-destructive forensic methods. They are used to diagnose a product and look deeper into the functionalities of a system. An example of a forensic method is the “Mobile Device Tool Classification System,” shown in Figure 2[5].

Figure 2: Mobile Device Tool Classification System [5]

As shown by the figure, one of the steps in the Mobile Device Tool Classification System is manual extraction. This is would be considered a destructive forensic method because it requires for engineers to open up the product and see the parts of the electrical device. This procedure can become complicated because “if there is a large amount of data to be captured, a manual extraction can be very time consuming and the data on the device may be inadvertently modified, deleted or overwritten as a result of the examination” [5]. There are also non-destructive preventative methods as part of this forensic procedure. In the logical extraction portion, a device would be connected to another electrical system, usually via wire. Here, a computer would send commands to the device and engineers would see how it responds. Seeing the response of the system can allow engineers to pinpoint a failure, without having to tamper with the device.

[1] S. Dalal and R. Singh Chhillar, “Case Studies of Most Common and Severe Types of Software System Failure,” International Journal of Advanced Research in Computer Science and Software Engineering, vol. 2, no. 8, pp. 341-342, Aug. 2012.

[2] R. Alexander, M. Hall-May and T. Kelly, “Characterization of Systems of Systems Failures,” PROCEEDINGS of the 22nd INTERNATIONAL SYSTEM SAFETY CONFERENCE – 2004, York, England, 2004, pp. 499-508.

[3] S. Eslami, S. Sajadi and A. Kashan, “Selecting a preventive maintenance scheduling method by using simulation and multi criteria decision making,” International Journal of Logistics Systems and Management, vol. 18, no. 2, pp. 250-269, Jan 2014.

[4] B. Schroeder and G. Gibson, “A large-scale study of failures in high-performance computing systems,” Proceedings of the International Conference on Dependable Systems and Networks, Philadelphia, PA, 2006.

[5] R. Ayers, S. Brothers and W. Jansen, “Guidelines on Mobile Device Forensics,” NIST Special Publication, pp. 1-75, May 2014.

[6] M. Shover, “Design for Reliability Techniques- Worst Case Circuits Stress Analysis,” Advanced Energy, pp. 1-9, 2018.