Engineering:Power Analytics

Power Analytics is a San Diego, California -based software developer specializing in the field of model-validated analytics for complex electrical power infrastructure, an emerging branch of analytics focusing on the CAD design and predictive analytics surrounding complex electric power infrastructure. Analogous to business analytics – massive mathematical models and algorithms used to predict financial behavior and potential risks – in power infrastructure model-validated analytics use the data in the original CAD model of an electrical power system to assess the impending likelihood of a short circuit, arc flash, or other power engineering problems.

Model-validated analytics can also optimize power consumption energy conservation in mission-critical facilities by ensuring – like the engine management system in an automobile engine – that every component is operating within its as-designed operating specifications. This component-by-component calibration results in what is effectively a 24/7, system-wide tune-up, that occurs every millisecond. Thus they play a vital role in help organizations minimize their carbon footprint and work toward a carbon-neutral footprint.

The trademark for "Power Analytics" was awarded to Power Analytics Corporation in 2006.

Where model-validated analytics are used

Model-validated analytics are used in mission-critical facilities (those requiring 99.9% uptime or better) marked by very dense electrical power usage, and where operational reliability is paramount. Facilities who are candidates for employing power analytics include transaction-intensive data centers, government and military installations, offshore oil platforms, air traffic control hubs, hospitals, casinos, and electric power generation and power transmission networks. In such facilities, electrical power problems can have very costly consequences, costing millions of dollars (U.S.) per minutes.

How they work

Model-validated analytics employ and integrate three primary sources of power-related data:

• A power systems CAD model, containing symbols of electrical components (power cables, circuit breakers, uninterruptible power supply, electrical generators, etc.) These components are connected to one another to form a virtual “power network,” in which a power system designer can easily test, trouble-shoot, and re-test the behavior of the system until no discernible design flaws remain. This can be called “as-designed” data.
• A data acquisition system (or SCADA system, or Building Management System) that aggregates live operational data about the electrical power infrastructure in real-time. This can be called “as-is” data.
• A diagnostics engine that continually (at millisecond speeds) compares the “as-is” values coming from the data acquisition system to the “as-designed” values encoded in the CAD model.

Once the system is deployed, the three software platforms interoperate with one another. If there are no deviations between the “as-is” and “as-designed,” specifications, no actions are taken. However, the instant any “as-is” value begin to drift from the corresponding value specified in the CAD model, however slightly, the diagnostics engine immediately begins to isolate the deviation, assess its implications, devise a remedy, and alert the facility operator to the problem and its resolution. Because noticeable electrical power problems are frequently the result of minute anomalies that went undetected – and thus, worsened over time – model-validated analytics are recognized as a very effective technology for preempting serious power faults before they can pose a threat to the operational effectiveness of a facility.