Monday, April 3, 2017

Resilient Systems and Resilient Design

We hear the word "resilience" a lot these days.  It's a word that's ripe for misuse and vagueness.  
There is a general agreement that resilience is the intrinsic ability of a system to resist disturbances.  Another equivalent definition of resilience is the ability to provide required capability in the face of adversity (sometimes referred to as disaster recovery).  Resilient design is about managing the ubiquitous uncertainties that constrain current design practices as well as finding ways to overcome an imperfect understanding of system requirements that lead to fragile and ineffectual system designs.

The definitions above are fine, but the real question is what is the “scope” of the system – and this is where views vary.  In this blog we focus on electronic products and systems, however, the concepts of resilience and resilient design are more prevalent in the building construction, architecture, and communities design space.

When you talk about resilient systems, several disciplines think they have a handle on the problem.  The control systems folks think that this is their domain, the optimization folks, the PHM (system health management) folks, the reliability engineering folks, sustainability folks, the system engineering folks, machine learning, artificial intelligence, … In my experience, none of these disciplines (with the possible exception of some system engineers) really grasps the total scope of resilience.  Every discipline wants to gather resilience under their umbrella and grant themselves ownership of the problem space based on their narrowed definition of what a system is.

In my opinion, designing resilient hardware and software (which is the focus of most resilient system design activities) is necessary but not sufficient for creating resilient systems.  For a system to be resilient requires:
  • reliable (or self-managing) hardware and software
  • a resilient logistics plan (including supply chain and workforce management)
  • a resilient contract structure
  • resilient legislation (rules, laws, policy)
  • a resilient business model
This represents a somewhat broader scope than what is generally articulated in the engineering literature, however, in practice, neglecting any of these elements potentially creates a legacy system with substantial (and potentially untenable) life-cycle support costs.

Although the discussion in this blog has an electronic product/system focus (our "scope").  Resilient design can certainly be applied to other things, e.g., buildings, communities, furniture, information technology, websites, health care, etc.  In this broader world one could consider adding the following to the bullets above:
  • culturally resilient (does the system transcend cultures and culture changes)
  • environmental resilience (sustainability)
  • resilience to climate change

Thursday, February 16, 2017

Obsolescence Definition

Obsolescence isn’t an uncommon word – a Google search finds more than 5.7 million entries.  However, in the engineering design and product support context it has several different meanings.  First, the dictionary definition of obsolescence is the condition of no longer being used or useful. This definition is not inconsistent with the definitions that follow, but it is also not specific enough to provide much value in a product support context.

Planned Obsolescence – we hear this a lot.  Planned obsolescence refers to products (usually consumer products) that are designed to be rendered obsolete by the introduction of another product that has more functionality, higher performance, smaller size, and/or costs less.  Planned obsolescence is a strategy sometimes followed by companies that design and manufacture consumer electronics.  Planned obsolescence is often confused with “made-to-break” products [1].  Made-to-break products are products that are intentionally designed or manufactured to fail at some point in the future (nominally after the warranty has ended) forcing the customer to purchase a new product.

Sudden or Inventory Obsolescence - Sudden or inventory obsolescence occurs when the product design or system part specifications changes such that existing inventories of components (e.g., spare parts) are no longer required.  This type of obsolescence has been studied in the operations research literature.
DMSMS or Procurement Obsolescence – DMSMS (Diminishing Manufacturing Sources and Material Shortages) obsolescence is the loss or impending loss of original manufacturers of items or suppliers of items or raw materials [2].  This type of obsolescence is caused by the unavailability of technologies or parts that are needed to manufacture or sustain a product.  DMSMS means that due to the length of the system’s manufacturing and support life and possibly unforeseen life extensions to the support of the system, the necessary parts and other resources become unavailable (or at least unavailable from their original manufacturer) before the system’s demand for them is exhausted.  DMSMS obsolescence is the opposite of sudden or inventory obsolescence.

Planned obsolescence is a strategy followed by product manufacturers, DMSMS obsolescence is a situation forced upon system sustainers (these two types of obsolescence are not the same).  DMSMS obsolescence may be the result of the planned obsolescence of products that drive the market for specific types of parts, e.g., if your long field life system depends on the same parts that cell phones depend on, then planned obsolescence strategies for cell phones drive the DMSMS obsolescence of the parts you depend on.  The “poster child” for DMSMS type obsolescence is electronic parts.  For some electronic parts the planned obsolescence of particular products is the primary driver behind the part’s obsolescence, but the discontinuance of parts is also simply the result of falling demand that makes it more profitable to dedicate manufacturing resources to other products, or changes in ownership of product lines or companies.

All of the types of obsolescence defined above are relevant for real product and system segments.  While it is easiest to think of these as impacting hardware, obsolescence also has a significant impact on software [3], materials, and even the human workforce [4]. 

[1] Slade, G. (2006). Made to break: Technology and obsolescence in America, Harvard University Press.

[2] Sandborn, P. (2008). Trapped on technology's trailing edgeIEEE Spectrum, 45(1), pp. 42-45, April.

[3] Sandborn, P. (2007).  Software obsolescence - Complicating the part and technology obsolescence management problem, IEEE Transactions on Components and Packaging Technologies, 30(4), pp. 886-888, December.

[4] Sandborn, P. and Prabhakar, V.J. (2015).  Forecasting and impact of the loss of the critical skills necessary for supporting legacy systems, IEEE Transactions on Engineering Management, 62(3), pp. 361-371, August.

Monday, January 2, 2017

Sustainment and Sustainment-Dominated Systems

Many people have a preconceived notion that “sustainment” and “sustainability” only refer to environmental sustainability, which is unfortunate.  Sustainment and sustainability are concepts that are much older and broader than just the environmental context that the popular media most often relates them to. 

The origin of the word sustain is the Latin work sustenare, which means “to hold up” or “to support”.  The modern use of the word sustain is to keeping something going or to extend its duration, [1], where the most common synonym for sustain is maintain.  It is not uncommon for sustain and maintain to be interchangeably used, however, maintenance usually refers to activities that are targeted at correcting problems, while sustainment is a more general term referring to the management of system evolution.

The concept of sustainability is connected to nearly every discipline [2], e.g., environmental sustainability, business or corporate sustainability and technology sustainment, however, in this blog we are interested in technology sustainment.  Although sustainability and sustainment are closely related in a semantic sense, environmental sustainability organizations almost never refer to what they are doing as sustainment or sustainment engineering.  However, organizations that maintain systems (sustainment organizations) use both sustainment and sustainability to describe what they do.

Technology sustainment refers to all activities necessary to [2]: a) keep an existing system operational so that it can successfully complete its intended purpose; b) continue to manufacture and install versions of the system that satisfy the original requirements; and c) manufacture and install revised versions of the system that satisfy evolving requirements.

Sustainment Definition

The most widely circulated definition of sustainability is attributed to the Brundtland Report [3], which is often paraphrased as “development that meets the needs of present generations without compromising the ability of future generations to meet their own needs.” This definition was created in the context of environmental sustainability, however, it is useful and applicable for defining all types of sustainability.  For example, for technology sustainment, “present and future generations” in the Brundtland definition can be interpreted as the users and maintainers of a system.  Unfortunately, the definition of sustainability has been customized by many organizations to serve as a means to an end, and in some cases it has been abused to serve special interests and marketing.

A good general definition of sustainment is [4]: “development, production, operation, management, and end-of-life of systems that maximizes the availability of goods and services while minimizing their footprint”.  In this case the terms in the definition mean:
  • “footprint” could represent any kind of impact that is relevant to the system’s stakeholders, e.g., cost (economics), human health, energy required, environmental, and/or other resource consumption (water, materials, labor, expertise, etc.)
  • “availability” represents the fraction of time that a good or service is in the right state, supported by the right resources, and in the right place when the customer requires it
  • “customer” could be an individual, a company, a city, a geographic region, a specific segment of the population, etc.

Note that this definition is consistent with both environmental and technology sustainment concerns.

Sustainment-Dominated Systems

A sustainment-dominated system is defined as a system for which the lifetime footprint significantly exceeds the footprint associated with making it [2].  Where "footprint" has the same definition as above. Defining sustainment-dominated systems provides the opportunity to make a distinction between high-volume, low cost consumer products and more complex, higher-cost systems such as airplanes, infrastructure, and institutions. Non-sustainment-dominated products, which tend to be high-volume products, have relatively little investment in sustainment activities and the total time period associated with the product is short (short manufacturing cycle, short field life).  Alternatively, sustainment-dominated products, which tend to be relatively low-volume expensive systems, have large sustainment costs and long manufacturing and/or field lives (see my "More Than Acquisition Costs - F-35" blog post in December 2016).

[1] Sutton, P. (2004). What is sustainability? Eingana, 27(1), pp. 4-9.
[2] Sandborn, P. and Myers, J. (2008). Designing engineering systems for sustainability, in Handbook of Performability Engineering, K.B. Misra, Editor, pp. 81-103 (Springer, London).
[3] Brundtland Commission (1987). Our Common Future, World Commission on Environment and Development.
[4] Sandborn, P. (2017). Cost Analysis of Electronic Systems, 2nd edition, World Scientific.

Sunday, December 18, 2016

2nd Edition of Cost Modeling Book Published

This book provides an introduction to cost modeling for electronic systems that is suitable for professionals involved with electronic systems development, management and sustainment, and advanced undergraduate and graduate students in electrical, mechanical and industrial engineering.

This book melds elements of traditional engineering economics with manufacturing process and life-cycle cost management concepts to form a practical foundation for predicting the cost of electronic products and systems. Various manufacturing cost analysis methods are addressed including: process-flow, parametric, cost of ownership, and activity based costing. The effects of learning curves, data uncertainty, test and rework processes, and defects are considered. Aspects of system sustainment and life-cycle cost modeling including: reliability (warranty, burn-in), maintenance (sparing and availability), and obsolescence are treated. Finally, the total cost of ownership of systems, return on investment, cost-benefit analysis, and real options analysis are addressed.

The book includes a large number of  quantitative examples solved in the text, over 230 problems (detailed solutions to over 90% of the problems are available to practitioners, researchers and instructors using the book in classes), and over 300 references.

Table of Contents (571 pages)

Chapter 1 Introduction

Part I Manufacturing Cost Modeling
Chapter 2 Process-Flow Analysis
Chapter 3 Yield
Chapter 4 Equipment/Facilities Cost of Ownership (COO)
Chapter 5 Activity-Based Costing (ABC)
Chapter 6 Parametric Cost Modeling
Chapter 7 Test Economics
Chapter 8 Diagnosis and Rework
Chapter 9 Uncertainty Modeling - Monte Carlo Analysis
Chapter 10 Learning Curves

Part II Life-Cycle Cost Modeling
Chapter 11 Reliability
Chapter 12 Sparing
Chapter 13 Warranty Cost Analysis
Chapter 14 Burn-In Cost Modeling
Chapter 15 Availability
Chapter 16 The Cost Ramifications of Obsolescence
Chapter 17 Return on Investment (ROI)
Chapter 18 The Cost of Service
Chapter 19 Software Development and Support Costs
Chapter 20 Total Cost of Ownership Examples
Chapter 21 Cost, Benefit and Risk Tradeoffs
Chapter 22 Real Options Analysis

Appendix A Notation
Appendix B Weighted Average Cost of Capital (WACC)
Appendix C Discrete-Event Simulation (DES)

Index (over 900 entries)

Wednesday, December 14, 2016

More Than Acquisition Cost – F-35

Aircraft are “sustainment-dominated” systems.  These are systems for which the lifetime footprint significantly exceeds the footprint associated with making it [1].  In the case of aircraft, the footprint we are talking about includes cost.   Lockheed Martin’s official response to President-elect Trump’s recent tweet about out-of-control costs for the F-35 included the following statement [2]:

“The cost doesn't just include the acquisition price. Lockheed Martin and its industry partners are also investing in reducing the sustainment costs of the aircraft recognizing that much of the cost of owning and operating an aircraft is after it's delivered. We're investing hundreds of millions of dollars to reduce the cost of sustaining the airplane over its 30-40 year lifespan. We understand the importance of affordability and that's what the F-35 has been about.”

It is not uncommon for 70% of more of the life-cycle cost of a sustainment-dominated system (e.g., commercial and military aircraft, ships, power plants, and other high-cost, long-life items), to be incurred after the design, development, and procurement of the system.  These life-cycle costs can include: operation, maintenance, upgrade, spare parts, testing, training, documentation, unplanned life extensions, obsolescence management, and many more things that contribute to the logistics footprint of a complex system.  As an example, consider obsolescence management [3].  The majority of the electronic systems in the aircraft are not constructed from “custom” parts, but rather from the same parts that are in consumer products (phones, computers, etc.).  Most of these parts have a procurement life of a few years at best, but an airplane has to be supported for 30+ years.  Sourcing these parts after they are discontinued (obsoleted) by their original manufacturer can be expensive and risky.  The problem is that aircraft are safety-critical systems that are highly qualified and certified, replacing obsolete parts with newer versions of parts may be a very expensive proposition (may require re-qualification of critical subsystems or even the entire aircraft); alternatively using aftermarket suppliers exposes systems to the risk of counterfeit parts [4]. Obsolescence is only one example of how high procurement cost systems can become even more (much more) expensive to sustain.

[1] Sandborn, P. and Myers, J. (2008). Designing engineering systems for sustainability, in Handbook of Performability Engineering, K.B. Misra, Editor, pp. 81-103 (Springer, London).
[3] Sandborn, P. (2008). Trapped on technology’s trailing edge. IEEE Spectrum, 45(1), pp. 42-45.
[4] Pecht, M. and Tiku, S. (2006). Electronic manufacturing and consumers confront a rising tide of counterfeit electronics. IEEE Spectrum 43(5), pp. 37-46.

Sunday, December 4, 2016

Engineering Economics vs. Cost Modeling

Engineering economics treats the analysis of the economic effects of engineering decisions and is often identified with capital allocation problems.  Engineering economics provides a rigorous methodology for comparing investment or disinvestment alternatives that include the time value of money, equivalence, present and future value, rate of return, depreciation, break-even analysis, cash flow, inflation, taxes, and so forth. Cost modeling is a different beast.

While traditional engineering economics is focused on the financial aspects of cost, cost modeling deals with modeling the processes and activities associated with the manufacturing and support of products and systems, i.e., determining the actual costs that engineering economics uses within its cash flow oriented decision making processes.

Cost modeling is one of the most common business activities performed in an organization.  But what is cost modeling, or maybe more importantly, what is it not?  The goal of cost modeling is to enable the estimation of product or system life-cycle costs.  Cost analyses generally take one of two forms:

       Ex post facto (after the event) – Cost is often computed after expenditures have been made.  Accounting represents the use of cost as an objective measure for recording and assessing the financial performance of an organization and deals with what either has been done or what is currently being done within an organization, not what will be done in the future.  The accountant’s cost is a financial snapshot of the organization at one particular moment in time.
       A priori (prior to) – These cost estimations are made before manufacturing, operation and support activities take place.

Cost modeling is an a priori analysis.  It is the imposition of structure, incorporation of knowledge, and inclusion of technology in order to map the description of a product (geometry, materials, design rules, and architecture), conditions for its manufacture (processes, resources, etc.), and conditions for its use (usage environment, lifetime expectation, training and support requirements) into a forecast of the required monetary expenditures.  Note, this definition does not specify from whom the monetary resources will be required--that is, they may be required from the manufacturer, the customer, or a combination of both.

Sunday, November 27, 2016

Absolute vs. Relative Accuracy in Cost Modeling

Cost modeling, like all other modeling activities, is fraught with weaknesses.  A well-known quote from George Box, “Essentially, all models are wrong, but some are useful,” [1] is appropriate for describing cost modeling.  First, cost modeling is a “garbage in, garbage out” activity – if the input data is inaccurate, the values predicted by the model will be inaccurate. That said, cost modeling is generally combined with various uncertainty analysis techniques that allow inputs to be expressed as ranges and distributions rather than point values.  Obtaining absolute accuracy from cost models depends on having some sort of real-world data to use for calibration.  To this end, the essence of cost modeling is summed up by the following observation from Norm Augustine [2]:

“Much cost estimation seems to use an approach descended from the technique widely used to weigh hogs in Texas.  It is alleged that in this process, after catching the hog and tying it to one end of a teeter-totter arrangement, everyone searches for a stone which, when placed on the other end of the apparatus, exactly balances the weight of the hog.  When such a stone is eventually found, everyone gathers around and tries to guess the weight of the stone.  Such is the science of cost estimating.”

Nonetheless, when absolute accuracy is impossible, relatively accurate costs models can often be very useful.  Relatively accurate cost models produce cost predictions that have limited (or unknown) absolute accuracy, but the differences between model predictions can be extremely accurate if the cost of the effects omitted from the model are a “wash” between the cases considered--that is, when errors are systematic and identical in magnitude between the cases considered.  While an absolute prediction of cost is necessary to support the quoting or bidding process, an accurate relative cost can be successfully used to support making a business case for selecting one alternative over another.

As an example for a relatively accurate cost model (with little or no absolute accuracy), consider a model for selecting the best way to maintain a system.  In this case the question to be answered is not whether or not to maintain a system, nor how much to budget to maintain a system, but rather which of several alternative approaches one should use.  Presumably once an approach is chosen, more accurate modeling can be performed for budgeting purposes.  The alternatives might be preventative, predictive, and corrective maintenance, and combinations thereof.

[1] Box, G. E. P. and Draper, N. R. (1987). Empirical Model-Building and Response Surfaces (Wiley, Hoboken, NJ).
[2] Augustine, N. R. (1997). Augustine’s Laws, 6th Edition (AIAA, Reston, VA).