Brittle Power- PARTS 1-3 (+Notes) - Natural Capitalism Solutions

Chapter Thirteen: Designing for Resilience 213able to continue to deliver the services of the computing hierarchy, or most ofthem, despite the loss of many subsystems. This design principle, and thebroader philosophy it reflects, have striking parallels in the design of resilientsystems for supply energy.The analogy is not exact. The main reason for ”duplexing” the Tandemcomputing and memory units, for example, is that information is not “fungible”—oneunit is not interchangeable with another—so each transaction mustbe protected from error in each device and operation. In contrast, since oneunit of electricity fed into a grid (for example) is the same as another, suchduplexing of individual power sources would not be required in order to keepthe grid working. Another difference is that a thousand hand calculators simplycannot do the same things as one big computer, whereas a user of oil (forexample) cannot tell whether it comes from the supergiant Ghawar oilfield orfrom a large number of small stripper wells.Despite these differences, the parallel between resilient, distributed dataprocessing systems and resilient, distributed energy systems is illuminating.The design principles that emerge from these two examples and from the earlierdiscussion of biological and engineering resilience can be summarize thus:•A resilient system is made of relatively small modules, dispersed in space,and each having a low cost of failure.•Failed components can be detected and isolated early.•Modules are richly interconnected so that filed nodes or links can bebypassed and heavy dependence on particular nodes or links is avoided.•Links are as short as possible (consistent with the dispersion of the modules)so as to minimize their exposure to hazard.• Numerically or functionally redundant modules can substitute for failedones, and modules isolated by failed links can continue to workautonomously until reconnected.•Components are diverse (to combat common-mode and common-cause failures),but compatible with each other and with varying working conditions.•Components are organized in the hierarchy so that each successive level offunction is little affected by failures or substitutions among components atlower levels.• Buffer storage makes failures occur gradually rather than abruptly: componentsare coupled loosely in time, not tightly.• Components are simple, understandable, maintainable, reproducible,capable of rapid evolution, and socially compatible.The following chapters apply these principles to the problem of designing aninherently resilient energy system.