Democratizing Electricity Intelligence the AI Way – A cover story by YYC Data

Economic turmoil is enveloping Europe. Putin’s war on Ukraine led the West to impose economic sanctions on Russia, while Russia retaliated by shutting gas to Europe. In Europe, exorbitant energy prices have led to shutdowns in energy-intensive industries, such as aluminum and ammonia. Many power producers in Germany were on the brink as they were forced to buy natural gas at extremely high costs on the spot market.

While Albertans have been spared much of the economic blows wrecked by Putin, the rippling pain can still be felt by much of the province that led to ex-Premier Jason Kenney announcing in March to provide a $50 monthly electricity rebate to eligible Albertans to provide some relief. In the past, we used to read the last single line item total amount on the bill and pay that off every month without paying much attention to the usage, but since the War, many have scrutinized every line of the bill to see the usage and how the total amount is calculated.

The high energy cost effects were not just painful for Alberta households. Alberta is home to many energy-intensive companies, such as timber, steel, oil & gas companies, cement, and other industrial companies, that power our economy, provide many jobs, and contribute taxes for our public service, and they are hurting much more than the rest of Alberta. For some of these companies, energy is one of their highest expenses, and this unplanned cost increase has taken a big hit to their bottom lines. The government rebate is only eligible for businesses that use less than 250,00 kWh per year, and many do not qualify for this relief. Fortunately for these companies, Calgary-based Arcus Power is helping to lower energy expenses for these companies with its AI solutions.

The electricity business is highly complicated because of the nature of real-time supply and demand for both electricity consumers and producers. While our phones and laptops give the impression that storing electrical energy is straightforward, storing electricity on a larger scale is a different challenge. Consider the energy needs of a household refrigerator, which runs continuously, compared to the intermittent charging of a phone. The battery required to power a refrigerator for an extended period would be significantly larger and more expensive than the one in your phone.

Batteries are costly, and the net zero carbon energy transition is driving up the prices of “green metals” such as lithium, cobalt, and manganese, is compounding the energy storage problem further, which economists call “green-inflation.” Therefore, it is difficult for power producers to store excess power generated and sell it to the market when needed.

Further to the supply complication, it is highly capital-intensive to build power plants, whether they are gas, coal, nuclear plants, or wind and solar farms. The permitting and environmental process may take years, and a transparent pricing signal is required to incentivize producers to invest in current production, as well as increased production capacity. Once a power plant is up and running, there are input costs, such as natural gas, coal, or uranium, that need to be factored into the financial analysis to ensure a producer has sustained profitability. A power plant producing electricity is not guaranteed to generate profits. In addition to the cost that needs to be factored in, the demand and the associated market price must be considered when deciding how much electricity to produce. When the market price is high, producers increase production to maximize profits and decrease production when the price is low to save on input costs.

‍The demand side of the electricity business is equally challenging to predict as the supply side. Electricity is unlike an Amazon package. You can’t wait a few days or weeks for it to arrive. It needs to work when we plug our gadgets into the socket at a moment’s notice. Therefore, the first and foremost role of a utility provider is to provide reliable electricity and it is not an easy job to ensure every single transmission line is working and that the grid can handle all our electricity needs. Last summer, with Californians being the biggest purchasers of EVs from generous state credits, California utility companies told EV owners not to plug their cars as the grid was overwhelmed from powering ACs.

Power markets used to be vertically integrated, where generation, transmission and distribution were in a single entity. The absence of competition means the lack of incentive for power companies to reduce costs to be competitive in the market. Today, the power market is deregulated with generation companies separated from transmission and distribution companies.

‍In Alberta, the Alberta Electric System Operator (AESO) is an independent system operator (ISO). This non-profit organization facilitates a competitive electricity market by providing price information to power producers to incentivize power production, as well as power distributors. AESO provides a competitive market by using an electronic auction system to look for the lowest price first and move the price up the bid stack until all consumer demand is met. While weather and various other factors can provide spikes in supply and demand, the day-ahead financial trading market provides market participants with some form of price stability by allowing them to hedge real-time risk with financial contracts traded in financial markets. Having accurate forecasts in the day-ahead price is how Arcus helps industrial companies.

Arcus offers three distinct products, one of which is Nrgstream. This platform serves as one of North America’s largest power market data repositories. The repositories cover all the independent energy operators, like AESO, in North America and include data to help predict electricity prices, such as weather, generation capacity, and fuel storage. The data is available through a front-end portal and web APIs.

Pwrstream by Arcus is the AI predictive analytics platform that offers energy cost predictions in the coming hours and weeks. Using variables that affect electricity supply and demand, such as line loss, import/export schedule, power supply forecast, and numerous other proprietary data variables, the neural network-trained prediction engine updates every 2 minutes and provides customers with near real-time electricity price and peak demand forecasting (the highest energy grid load and price of the day). Knowing what their energy operating cost is forecasted to be, companies can hedge and lock-in prices when they are forecasted to be high (not just power companies, ask anybody in 2023 with a variable mortgage if hedging with fixed rate is a good idea!), and pay the spot price when they are expected to be low. The vast quantity of data is not the only input to a highly accurate model, the model is also maintained by a team of PhDs with vast domain knowledge in each data input, thereby allowing Arcus to understand the in-depth fundamentals behind each dataset.

‍Sales of Pwrstream have doubled year-over-year in 2022 in the high electricity price environment. Customers see Pwrstream as an energy efficiency investment, with the platform being able to reduce on average a million dollars a month for a 50-megawatt facility, that’s roughly the size of the cement plant in Exshaw you see on the way to Banff. By providing insights into a company’s energy costs and actionable operational recommendations, such as changing interruptible process schedules or using equipment with variable speed drive, a timber mill or a steel mini-mill can reduce peak load reduction by up to 80%, and 96% respectively.

‍Some of Canada’s largest publicly traded midstream pipeline companies that moves oil and natural gas throughout North America are using Pwrstream for pump power optimization that helps with grid stability, and CO2 emission reduction. With natural gas being the dominant input source of electricity, producers are using Pwrstream’s price forecasts to run behind-the-fence natural gas generation to capture power arbitrage (market price less operating costs of running asset)  by strategically deploying their assets. By ramping up production when electricity price is high, and ramping down when electricity price is low, producers can extend the lifespan of their equipments and decrease maintenance operating costs. Being able to look ahead to see when electricity price is low also helps Arcus’ cryptocurrency clients to economically mine coins by also ramping up mining when electricity price is low, and ramping down or shutting off when power cost is high.

With the boom of EVs and the electrification of our economy to get to net zero, we often hear about our current aging grid infrastructure and the massive investment needs to modernize in order to meet the demand in the years to come. There are many government programs, such as Ontario’s Industrial Conservation Initiative (ICI) program or the Global Adjustment (GA) program, that offer preferential electricity rates for large electricity customers to shift their consumption to off-peak hours to take loads off the grid, which is important in extreme weather events to avoid blackouts. With its experience in power analytics and intelligence, Arcus helps these large energy-intensive customers meet the requirements of these savings programs.

The electricity source from fossil fuel is estimated to be ~90%. Despite the massive investments in renewables, the dominant electricity source continues to come from carbon sources. As corporate stakeholders and shareholders become increasingly aware of the importance of protecting our planet, ESG goals are important KPIs for all company levels. To help companies meet their carbon footprint goals, the third product Arcus offers is Crbnstream, a service that forecasts and tracks emissions dynamically. This service lets customers see the carbon intensity of their consumption hours of the day and, if economically viable, shift their consumption to low carbon intensity hours when wind and solar generation is the highest, and coal and natural gas generation is the lowest.

‍

The 2010s’ shale revolution that bestowed the world with low energy prices for eight years from 2014 to 2022 has caused much underinvestment in exploration and development in the oil and gas industry. The Strategic Petroleum Reserve (SPR) release by the Biden administration helped lower gasoline prices at the pump for much of 2022, with the unintended consequence of distorting the price signal for oil and gas producers, causing them to invest less in capital expenditure than if energy prices stayed high. With natural gas still the dominant electricity source, and coupled with material and labour constraints causing skyrocketing renewable energy construction costs, we are likely to see high electricity bills in the years to come. Companies are already struggling with high inflation, high interest rates, and labour shortages. Add on high electricity price and some large power-intensive companies are at their breaking points.

‍The Chinese word crisis (危機) is composed of the characters danger and opportunity, and in today’s difficult economic environment, companies have the opportunity to unlock hidden energy savings with Arcus Power’s innovative AI and Big Data solutions.

Ref & Source: YYCdata.ca cover story on Arcus Power

‍