AI will greatly improve energy efficiency

There has been much talk of the massive amount of energy likely to be used by data centres and artificial intelligence, but could there be another side of the story? Asks Andrew Warren, Chair of the British Energy Efficiency Federation, in an article in the October issue of Energy in Buildings & Industry.

 

Is AI really a drain on global energy?

Worldwide, a  Google search emits about 0.2 grams of carbon dioxide. Every minute spent on TikTok emits about 2.6 grams. Watching one hour of video emits around 36 grams.

But these activities are nothing compared to a relative newcomer to the business world: artificial intelligence (AI). Now  held largely responsible for much recent publicity about  booming consumption and  emissions.

Take Google’s newest search feature: AI-generated answers from its bot Gemini. Any AI-assisted search requires 10 times more power than a traditional search. Converting Google’s entire search engine to an AI chatbot, it might need as much electric power annually as Ireland merely to run Google Search.

This would result in a big increase in carbon emissions. Goldman Sachs is projecting that 47 gigawatts of extra electricity might be required by 2030,  owing to this exponential growth in demand from data centres.

But surely this cannot be the sole outcome that the US Government had in mind when it  earmarked $30 billion in subsidies “to bring cutting-edge AI chip development and manufacturing to American soil”?

Savings potential

In practice, there is a distinct potential upside which should largely  correct this consumption surge.

According to management consultants McKinsey, AI has the potential to deliver energy savings of up to 20% in buildings and 15% in transport systems.

Additionally, AI-driven solutions can help businesses reduce CO2 emissions by up to 10% and cut energy costs by 10-20%- particularly so in the field of industrial energy management.

Manufacturing facilities have a uniquely golden opportunity with the new asset of AI to reduce costs, improve productivity and embrace digital transformation to achieve their top energy-efficient goals. Every operation can use this tool. So where does implementation begin?

Cutting usage helps facilities develop long-term energy efficiency. AI streamlines the lengthy, tedious process of data mining for power metrics across hundreds of thousands of square feet. Necessary technologies include smart meters, data analytics tools, and the Internet of Things to track real-time activity and create performance visualizations.

The AI energy audit quantifies previously nebulous consumption growth assumptions. Stakeholders should then compare their numbers against pessimistic regulatory expectations.

Learning algorithms

Data overload is always a potential hurdle, but curated training and regular data cleaning will maintain integrity and accuracy. AI’s machine learning algorithms become more proficient at determining energetic anomalies over time, the sources pulling the most power, and numeric references the manufacturers’ use to frame their goals.

Once the AI understands a facility’s energy consumption, it can suggest extra areas of opportunity. It may notice the HVAC system is overdue for a filter change and uses more energy than it should for its model and age. AI dashboards will also reveal that lights remain on overnight, even without occupants. After AI discovers unwanted technical behaviour, it will inspire technicians to review faults in the motion-detection system.

Condition monitoring systems are also powerful tools for preserving production quality and preventing downtime. Corporate machine-learning platforms detect faults, predict equipment failures and send reminders when preventive maintenance is due.

These applications may not be compatible with all machinery, particularly legacy systems. However, agencies may undergo gradual retrofits and install middleware to facilitate AI predictive maintenance assistance until a complete rollout can occur.

AI makes maintenance for energy-intensive equipment attentive. Now, it can optimise production workflow habits and employee behaviours to make a more widespread impact. Some of the most obvious energy efficiency optimisation opportunities, lying within these production areas:

Staff may resist change, but management can ease this by offering incentives at training workshops to encourage buy-in. Increasing familiarity with AI tools and learning how important energy efficiency is may be the only difference between stubborn staff and people willing to upskill. Simulation software with AI integrations is useful for modelling process changes before implementing them, to test if they will have the projected energy impact.

AI’s beauty is in its versatility and compatibility with other technologies and realms of the business. For example, if companies want employees to train on the job, AI enhances spatial computing exercises  known as expanded realities (XRs). The data crafts a realistic, virtual setting for operators to test their skills in low-stakes situations with real-time, hands-on interactivity that online training modules lack.

Energy oversight with AI is a constant process. It offer continuous monitoring and feedback loops.  It should never stop because it will discover anomalies, spotting  when employees are deviating from sustainability training. Establishing a perpetual observation system means complacency never develops, avoiding dips in energy efficiency.

For example, AI in data centres is increasing their carbon footprints. However, the industrial and manufacturing industries rely on them to compute vast amounts of information quickly. Leveraging AI in data centres connected to manufacturers could reduce their energy consumption growth by 96% , because of optimizations inspired by round-the-clock monitoring. Continuous surveillance happens within and outside the facility, offering comprehensive energy efficiency gains.

Incorporating AI for energy efficiency gains requires dedication, but the effort yields an optimised, self-sufficient machine. High energy implantation delivers low energy results.

Early AI adopters using it to progress corporate social responsibility and lower building emissions win appreciation among client bases for their initiative. Additionally, they establish thought leadership during this next generation of industrial overhauls powered by the environmental revolution.

The only way to obtain results as swiftly and accurately as the current market demands is with AI. Applying these accessible best practices makes AI central to any sustainability enhancements. And will go  a  very long way to prove those Goldman Sachs gigawatt growth forecasts  to be seriously exaggerated.