Case Studies
Mobile Microgrids And The Rugged Edge: Ensuring Power Reliability Everywhere
POSTED 05/16/2023
Background
From California to the East Coast, power grid outages have become all too common. Excessive demand, extreme weather events, cyber attacks, and more have made today’s power grid less than reliable. And resulting power interruptions can lead to everything from mere inconvenience to business losses, and even life-and-death situations. One company specializing in power and fuel solutions for remote and grid-paralleling applications is looking to change that. With its mandate to transform the energy landscape, this company’s small-scale, self-contained microgrid trailers operate independently of major power grids or in tandem with other microgrids to fill grid gaps and keep life running smoothly. Deployed when and where needed, these modular energy resources improve power quality, integrate on-site generation resources, reduce peak demand charges, and provide standby (emergency) power generation.
According to the National Renewable Energy Laboratory (NREL), a national laboratory of the U.S. Department of Energy,“A microgrid is a group of interconnected loads and distributed energy resources that acts as a single controllable entity with respect to the grid. It can connect and disconnect from the grid to operate in grid-connected or island mode. Microgrids can improve customer reliability and resilience to grid disturbances.”
With an array of power solutions, this company offers an advanced energy management and generation platform via an intuitive software portal. Power operators can access power services, insight on emissions reductions, and energy storage for the pressure pumping, remote mining, utility, and distributed generation industries. By incorporating artificial intelligence (AI), the platform manages live emissions tracking, asset carbon intensity, automated back-office management, and environmental, social, and governance (ESG) reporting on a centralized database.
The portal also offers real-time insights into operations, with the ability to view equipment conditions, analyze energy consumption rates, project costs for the rest of the month, approve invoices, and purchase carbon credits to offset emissions.
“Mobile microgrids are where rugged edge computing can really excel – blending performance and reliability in uniquely demanding environments, both in terms of data handling and physical setting,” said the company’s Senior Manager of Network and Communication. “A solution of this magnitude calls for substantial compute power. And its deployment in wide temperature ranges meant we were looking for a robust industrial-grade hardware solution built for next-generation computing anywhere.”
所面临的挑战
To deliver the near-real-time access to critical operational data necessary for reliable microgrid performance, the platform leverages the cloud, machine learning, and AI. Like many AI-centric startups, this power solutions company started out usingconsumer-grade computersto run its algorithms. While the strategy worked early on, technological advances and growth objectives clearly dictated significant hardware improvements. Compute power, ruggedness, component availability, and scalability were top considerations.
在解决计算挑战,这个团队知道their new hardware system would need to incorporate GPU processing. Long used in graphics and gaming, GPUs are particularly well-suited to AI, machine learning, and deep learning applications due to their massive number of cores that can simultaneously perform complex calculations that CPUs find difficult to execute. A GPU essentially serves as a performance accelerator, offloading the CPU’s workload for faster and more efficient processing, a key requirement in edge computing models.
The search for the optimal GPU led the company to the NVIDIA A2, a Tensor Core GPU that provides edge computing inference by incorporating low power and high performance in a small footprint,. But, to take full advantage of NVIDIA AI at the edge, its consumer-grade computers would have to go. Painfully insufficient – lacking the compute power and ruggedness required for this type of mission-critical deployment – existing desktop systems were simply incompatible with the GPU’s extensive list of capabilities. The company needed a significant hardware upgrade to improve its offering and develop a truly portable, scalable, and affordable remote microgrid power solution. The required industrial solution must not only accommodate the NVIDIA GPU but also meet environmental challenges such as wide temperature ranges and shock and vibration situations.
While a pretty straightforward assignment, this could prove difficult, particularly in light of supply chain issues largely brought on by the COVID-19 pandemic – and largely remaining today. In fact, according to a new report from KPMG, “Disruptions to supply chain operations are set to stay in 2023, whether they be related to existing or new geopolitical conflicts, inflationary pressures and the recessionary environment, climate change weather events, or other issues yet to emerge. They can all impact access to goods and how they flow to their final destination, create port holdups, reduce container and ocean freight availability, and surge prices, among other concerns.” The company would need a solution provider who expertly checked the boxes for technology expertise as well as those related to supply chain management and strategy.
The Solution
In investigating its options, the team met with Premio, a leader in rugged edge and embedded computing technology. After learning about the power solutions company’s business, its goals for the new platform, and its intent to leverage an industrial computing solutionwithNVIDIA A2 GPUs, Premio got to work designing the optimal system for the job.
It was clear from the get-go that the platform’s extensive AI capabilities and harsh environment deployments dictated a robust, rugged edge solution. Premio turned to theRCO-6000-CML AI Edge Inference Computerfor its rugged and high-performance socket-type processor design. These units are extremely modular and include PremioEDGEBoost Nodesthat serve as performance-driven building blocks for custom configurations that require hardware acceleration. This approach ensures improved reliability in wider temperatures ( -25C to 70C), wider input voltages (9-48VDC), and tolerance to shock (50G) and vibrations (5GRMS). Tested and validated to ensure reliable performance amid deployments in the harshest environmental settings, Premio then set out to integrate the NVIDIA GPU. Premio’s engineers worked closely with NVIDIA and the customer to combine Premio’s ultra-rugged PC architecture with NVIDIA’s powerful capabilities in running complex algorithms at the edge.
与当前的cpu、gpu由成千上万的cores, making them capable of calculating the linear algebra commonly used in AI models. With pure compute power and environmental issues addressed from an industrial computer standpoint, there was next concern regarding the amount of heat being generated. Premio recommended and designed-in a custom airflow chamber that would alleviate the potential for excessive heat and its related problems due to NVIDIA’s A2 passively cooled design.
The Benefits
By working with Premio, this customer has modernized its offerings in ways it couldn’t have imagined. Integration of the NVIDIA A2 GPU within the Premio RCO-6000-CML AI edge inference computer now enables the platform to run complex algorithms for its edge compute offering. Through low-latency data processing, organizations that leverage mobile microgrids have at their fingertips mission-critical business insights, based on actionable intelligence that allows them to better respond to situational data.
Premio assembled a team of experts in account management, engineering, and supply chain management to work hand in hand with the customer to ensure all questions and concerns were properly addressed and resolved. When the first batch of RCO-6000-CML AI edge inference computers was assembled, the customer’s lead engineer on the project traveled to Premio’s manufacturing facility in Southern California to observe production and take delivery. This simple gesture gave the customer added confidence in the units and their vendor selection. Premio’s sales and fulfillment teams also kept the customer closely apprised of the forecast and delivery schedule, further assuring that defined deadlines would be met.
By leveraging Premio’s engineering and supply chain expertise, this customer has saved significantly on development costs – measured in both time and money. Also, with a consistent configuration, re-engineering and testing new platforms is a thing of the past. Zero defects on all Premio computers have been reported to date, allowing this company the ability to better use its resources for servicing customers and growing its platform and program for mobile microgrids.
“Premio’s partnership in product design and manufacturing has been a game changer for us. Knowing that we are receiving quality products that have met all of our timelines, we can focus our attention elsewhere,” added the customer’s Senior Manager of Network and Communication. “Whether that means growing the program, engaging more closely with our customers, or developing new products or ideas, we can comfortably leave the computing system to the rugged edge hardware experts.”
This customer has found a true partner in Premio – one that understands its challenges and needs, and works together closely as a team to come up with innovative solutions that pioneer new opportunities in the market.