Modular Instrumentation and Automated Data Processing for Marine Energy Monitoring

Environmental monitoring of marine energy converters poses a technical challenge both for underwater instrumentation and for its associated operating software. Typical marine energy sites have harsh environments with strong currents or wave action, and the long-term deployments required for continuous monitoring expose instruments to failure from salt-water corrosion and biofouling. Data management for multi-month or multi-year deployments necessitates automated real-time data processing methods to avoid accrual of a data mortgage (acquiring data faster than it can be processed.) Currently available instrumentation is typically ill suited for long term deployment at marine energy sites and there are no software options available that provide this type of automated data processing. The Adaptable Monitoring Package (AMP) addresses these environmental monitoring needs around marine energy converters, with over 1000 days of testing in the water. Key components of the AMP include a stereo-optical camera system, instrument integration hub, passive acoustic monitoring system and biofouling system. Other off the shelf instruments typically include imaging sonars, acoustic doppler current profilers, rotators, wipers, and UVC antifouling lights. In Phase I, the company developed a highly reliable and user-friendly low-cost modular underwater camera system for marine energy device monitoring, achieving a cost reduction of over 50% relative to other commercially available systems. In addition, a software platform called Vision was developed to expand operating system, hardware, and deployment flexibility of the AMP. In Phase II the main objectives were (1) continued hardware development to include the instrument integration hub and lower fabrication costs, (2) software module development for the AMP hub and imaging sonars including automated data processing methods, (3) investigation into the transferability of automated data processing methods between sites and distinct monitoring conditions, and (4) performance assessments and optimizations through tank and field testing. A set of machine learning algorithms were also applied to automate the data processing and image recognition of marine animals. In Phase IIB, there will be continued expansion of the instrumentation suite to enable more monitoring capabilities, further reductions to overall system cost, and improved scalability. Monitoring capabilities will be expanded with the addition of a Passive Acoustic Monitoring hydrophone array and several new components that will further lower hardware system costs by an additional 37%. The software will be updated to improve scalability, decreasing deployment and maintenance costs by >30% while also increasing reliability. Research will continue for the use of machine learning data processing techniques for optical and acoustic datasets monitoring marine energy converters. Finally, a six-month deployment with a local aquarium will show the long-term durability of the system while showcasing the use of machine learning in marine exploration to the public. The ocean covers more than 70% of the earth's surface and our exploration to date has been limited to net trawls and short underwater vehicle deployments. The technology developed in this Phase IIB project enables continuous underwater monitoring with the potential to expand our understanding of marine ecosystems in many unknown ways.