Boston Dynamics: A Chance to Commercialize

Rising in Robotics
Boston Dynamics is a robotics manufacturing company spun off from MIT in 1992. The company is best known for its human- and dog-like robots that move similarly to their living counterparts. Since its founding, Boston Dynamics has heavily invested into research and development to build robots for a variety of industrial and commercial use cases. Despite this emphasis on research, the company has experienced a tumultuous history, with limited success in commercializing its products. After a 2013 acquisition by Google X and subsequent resale in 2017 to SoftBank, Boston Dynamics is under considerable pressure to begin producing commercially viable products.

Navigating Rough Terrain
Boston Dynamics’ industry peers have traditionally struggled to achieve profitability, with many ventures shutting down after failing to acquire customers. The root of this issue lies in the substantial capital investment needed to build robots, as well as the technical challenges in replicating and automating skilled human tasks. For instance, Boston Dynamics partnered with the U.S. Department of Defense to create four-legged robot BigDog that was subsequently rejected because it was too loud for military applications.

Boston Dynamics has since created an advanced robotic dog known as Spot. Spot can navigate over irregular terrain, across physical barriers, and has its own Software Development Kit (SDK) that allows users to extend functionality and capture telemetry data. In light of pressures to generate revenue, Boston Dynamics should launch a commercialization strategy centered around Spot.

The Winds of Change
Offshore wind farms are particularly well-positioned for a transition to robotic automation. Wind farms currently account for 0.3 per cent of global power generation, with usage expected to increase 15-fold and investment estimated to be $1 trillion by 2040. Governmental shifts toward carbon neutrality policies and sustainable energy sources have made wind farms increasingly popular, and the public reputation Boston Dynamics could gain by servicing this sector could ease its entry into other industries.

Offshore farms benefit from the comparably higher wind speeds available relative to land-based farms and have achieved pricing parity with conventional power sources in Europe. However, generators in offshore wind turbines produce alternating current (AC) electricity, which is susceptible to considerable power loss across the long distances back to shore. Consequently, offshore farms require an electrical substation platform to conduct the complicated conversion process to direct current (DC) electricity that can be brought back to shore with minimal power loss.

Currently, one of the largest challenges facing offshore wind farms is the need for preventative maintenance on converter platforms. Platforms are comprised of several complex piping systems, cooling systems, and wiring systems which require regular inspection. Given that offshore farms are typically 200 to 300 kilometers away from land, using AC current is not a viable option for wind farms. Hence, energy generated by windmills need platforms for successful operations.

From a financial perspective, the high costs of downtime mean that any idle or defective platform parts create substantial opportunity costs for farms. Each additional day of downtime can result in a loss of approximately 20 megawatt-hours of electricity production per turbine. With offshore wind farms facing an annual downtime average of 6.4 days, this would result in lost revenue of approximately $5 million per year. Risks in transmission issues on the platform is a core reason for downtime, which emphasizes a need for preventative maintenance. Additionally, harsh conditions, complex equipment, and the high concentration of flammable, electric materials results in a dangerous environment for offshore workers. With 27.1 annual deaths per 100,000 offshore workers, the mortality rate at offshore wind farms is seven times higher than that of the average U.S. workplace.

Inspection Imperfections
With its reliance on human workers, current preventative platform maintenance efforts are susceptible to several errors. For one, workers lack the ability to replicate a systematic, error-free inspection process. Given that human inspection is observation-based, key risk assessments are both subjective and non-quantifiable. Hence, not only is there variation in exact methodology from worker to worker, but the accuracy of historical data and perceived trends is uncertain. Ultimately, preventing potential downtime is difficult without quantitative measures of risk and highly standardized detection methodology.

Sensors—one of the key pieces of technology currently used in preventative maintenance—also pose various challenges. As they are in fixed positions, sensors are not always optimal for collecting essential data like the thermal imaging of certain pipes or imaging of oil levels. Moreover, sensors can only collect data, but lack the ability to process or analyze trends. In the long-term, not only is it costly to update old sensors, but their fixed locations on paths must be updated regularly to accommodate for changing platforms. Considering these issues, wind farms could be better served with dynamic, adaptable technologies which can process and quantify risks.

New Dog, New Tricks
With its real-time data analysis features and adaptable, mobile body, Spot offers superior preventative monitoring abilities compared to current methodologies. For one, as a nimble robot, Spot can move through sites without difficulty, eliminating the need to reposition sensors. Not only can Spot implement inspections outside a worker’s eight-hour day, but it can also detect problems that are dangerous for humans such as leakages, hot spots, gas leaks, and degradation of equipment. During detection, Spot can quantify any risks, and immediately notify operators, limiting damage and costs.

Compared to fixed sensors, Spot’s removable and replaceable attachments make it both easier and less expensive to upgrade equipment. Whereas changing the layout of platforms would require redesigning the path of sensors, Spot’s dynamism and self-navigation abilities ensures that it can adapt to various environments. As sensor equipment develops and improves, it is also more cost efficient to upgrade sensors on one robotic device compared to replacing fixed sensors dispersed across a platform. Additionally, Spot’s advanced computing power allows it to go beyond the ability of sensors by analyzing trends, generating insights, and ultimately predicting defects.

When inspecting the converter platform, Spot offers comprehensive functions to evaluate the health of equipment parts and pipes. After collecting data from its embedded sensors, the robot uses computer vision to compare real-time data to standard operating ranges. Surpassing the scope of physical checks, the data collected and processed by the SDK allows operators to understand the overall health of their platforms, areas at risk, and the impact of construction or environmental changes. Over time, this data becomes increasingly valuable as Spot’s machine learning capabilities can identify trends in equipment degradation and determine the probability of high-risk situations before physical symptoms appear.

Tilting at Wind Farms
Spot’s technological and navigational capabilities are well-suited for many specific use cases aboard offshore substations with regards to preventative maintenance. Spot’s thermal imaging systems, for example, can develop heatmaps for the oil-based cooling systems on transformer wiring, cable trails, pumps, and other components to detect issues that would otherwise go unnoticed to the human eye. Likewise, 3D imaging technology aboard the unit can dynamically map the environment and floor plan to identify anomalies in structure, malfunctioning equipment, and errors in machine systems. Spot can also perform visual and aural analysis to detect excessive noise or irregular visual cues. The base technology available through these systems are superior to most competitors like ANYBotic’s ANYMal robot. Spot offers 30-times zoom and 360-degree visual cameras for higher fidelity inspections; ANYMal, by comparison, only offers a 10x zoom camera system. Leveraging superior technology and data analytics, Spot’s accuracy and usability will accelerate with time and use.

Finally, compared to competitors like ANYMal, Spot’s main technological advantage lies within its advanced robot arm. This arm allows Spot to complete tasks such as opening doors and grasping objects, which are outside ANYMal’s capabilities. Given the importance for robots to be self-functioning, Spot’s arm substantially increases its tactical usefulness on offshore platforms. By leveraging Spot, wind farms can improve current inspection processes, detect potential equipment defects and ultimately limit the significant annual opportunity cost associated with downtime.

Rise of the Underdog
To commercialize its technology with wind farms, Boston Dynamics should begin by running trials with Spot through a tactical partnership. The goal of the pilot should be to refine pain points in Spot’s interactions with human operators and test the reliability of collected data. These trials will require significant collaboration with wind farms to define the scope of the robot’s tasks and determine optimal hardware attachments. Additionally, to build credibility for its technology and meet safety standards, Boston Dynamics should pursue ATmosphere EXplosible (ATEX) certification from the European Union. This certification ensures that Spot is equipped to deal with potentially explosive equipment found in some wind turbines.

Given the current status of global wind farms, Boston Dynamics should pursue a partnership with Orsted, a company which operates Walney—one of the largest offshore wind farms in the world. With 59 per cent of its revenue derived from offshore wind, Orsted is a suitable partner which could reap material benefits from Boston Dynamics’ technology. In 2019, Orsted generated $1.02 billion in revenue from the sale of power, with employee expenses accounting for 38 per cent of costs. By implementing Spot on farms, Orsted should be able to realize significant cost savings of up to $25 million. Once the initial hurdles of wind farm operations are ironed out and safety certification is obtained for offshore operations, Boston Dynamics can aim to target a broader market with Spot.

Dog Days are Over
As Boston Dynamics aims to shift from functioning as a research and development organization to generating revenues, the company should deploy its robotic dog Spot in wind farms. By capitalizing on nearly 30 years of research expertise, Boston Dynamics’ cutting-edge technology provides wind farms an opportunity to implement superior preventative maintenance methodology which can tackle the multi-million-dollar downtime gap. Looking forward, Boston Dynamics could expand Spot’s use cases to other industries including offshore oil drilling. Within these industries, the risk of human error and the need for real-time analytics means that technologies like Spot would be invaluable. Ultimately, the time has come for Boston Dynamics to embrace Spot’s departure from academic laboratories and towards industry-wide commercialization.