b14
b14
b14
b14
b14

Optimization Deals With Power Deregulation

Oct. 9, 2003
Abitibi Consolidated saved more than $1 million the first year it used real-time process and pricing data to control power consumption

Abitibi Consolidated, Montreal, is the world's largest producer of newsprint as well as a top manufacturer of value-added papers, market pulp, and lumber. We have about 16,000 employees around the world, deployed in 27 paper mills, one market pulp mill, 22 sawmills, three remanufacturing centers, and 10 paper recycling centers.

Each year we manufacture approximately six million tons of newsprint, nearly two million tons of value-added papers, more than two million board feet of lumber, and almost a half million tons of market pulp. Our recycling operations consume more than four million tons of recycled fibers.

We faced a huge challenge in 2001 when the Ontario provincial government confirmed a deadline of May 1, 2002, for deregulation of the Ontario power marketplace. We have five paper mills located throughout Ontario, which together historically have purchased more than $100 million in electricity every year. (All figures in this story are in Canadian dollars.) In a regulated power market, this was never a problem. We knew the cost per megawatt/hour and we knew about how much power we would consume each year.

But with deregulation, electricity now is subject to wide swings in both availability and pricing. Heat waves in the summer and intense cold in the winter will elevate power demand. Generating equipment breakdowns and even power availability in neighboring markets affect the supply and cause fluctuations in pricing.

Three of our five Ontario mills have some degree of hydroelectric generating facilities, so they can supplement or replace some of the provincial power as needed. What we don't generate is purchased from the Ontario grid at market price or based on contractual agreement. Poorly managed power purchasing could result in additional costs of hundreds of thousands of dollars.

Responding to this situation required that we put in place a system that could overlay energy market information on our real-time electricity consumption so that we could monitor conditions and make better, more timely business decisions. We first investigated off-the-shelf energy management systems, but the only one that might have been suitable was priced at about $3 million, not counting annual support and maintenance costs. These costs were prohibitive, given the current economic situation in the newsprint market.

Figure 1: An Eye on Consumption

Staff can see every element of the system to monitor current power consumption as well as the status of servers, historians, and the network.

We then examined our operations in Ontario to see if we could build our own application more cost-effectively. After much creative reflection from a team that included personnel from our mills and our head office, we decided the answer was yes, we had the tools in place to create our own system. While it might not offer all the bells and whistles that the packaged systems did, we felt we had all the major elements.

Historians Are Key

One of those elements was our existing Realtime Performance Management (RtPM) solutions from OSI Software (www.osisoft.com). These modules were already used at 18 of our mills, functioning as the nerve centers for all production data and reporting. They were linked to various equipment, such as our DCSs and PLCs, via Windows-based operator interfaces.

Because they were on the critical path for managing production at each mill, each of the RtPM systems could be tapped to provide the raw process data that we needed, including metering information. The primary modules included PI Historian servers, PI Process Book, and PI Data Links.

We added a PI server at one mill that didn't have one and we added a second at our headquarters in downtown Montreal, which served as the control center for the entire application. We redeployed a used PI system as a hot backup, so if anything happened to our primary server we could switch over instantly. We then set up data links between the remote PI servers at each mill and the center in Montreal (Figure 1).

The next step was to deploy OSI's HTML interface to access the new Independent Electricity Market Operator (IMO) web site to obtain market information over the Internet. The IMO is to the electric marketplace what the stock market is to the financial world. This link gives us continuous up-to-the-minute information on the availability and pricing of electricity.

We also set up a link to the Environment Canada web site so we could collect current weather information on the metropolitan Toronto area and at the mill locations, so we can match weather changes to market conditions. We extended the use of PI/HTML by gathering information on the commodities futures markets for natural gas and crude oil.

We began to work on this solution, which now was officially known as the ACI Energy Management System (ACI/EM-SYS), in October 2001. One good thing about dealing with immovable deadlines is they force people to work together to reach a common goal. In this case, we absolutely had to complete our project before May 1, 2002. We were fortunate to have the full support of our executive management, which meant we could request the services of internal resources such as key programmers and specialists from various locations in North America for temporary assignment to our project.

Our IT staff worked closely with both corporate management and the mill energy managers to determine exactly what their needs were. Our design and programming staff from various divisions worked together to build applications specifically to fit those requirements.

Since we were able to gather any production data we needed from any of the mills in real time, we could overlay energy needs on required and actual loads. We also created links between the PI systems and the electric meters at each site, so we knew exactly what the power usage patterns were during any production situation. Finally, we used a small OSI freeware utility called eNotification that allowed us to automatically send e-mails to support groups when a system problem arises or when a major price change occurs. Some mills also integrated their paging system into EM-SYS.

Easy Navigation Overview

On the IT side, the heart of the ACI/EM-SYS system is the System Watch screen, which monitors all servers, hubs, and links on a 24x7 basis and provides continuous information on how the network is running (Figure 2). This screen provides an overview of all five mills and their consumption, every minute, along with web connections to the IMO, Environment Canada, and other valuable sites. Messages can be posted by EM-SYS users and can be seen throughout the system.

The Control Center screen (Figure 3) allows operators and managers to look at any of the five mills to see their planned demand, their actual load consumption, weather-related factors, and a comparison of actual vs. planned consumption. Staff can also click on the IMO Price Watch to monitor current and forecast prices.

Figure 2: Can Plan Demand

The load planning module screen lets mill operators schedule heavy process equipment based on optimal times for power consumption.

EM-SYS also includes a Load Plan Module (Figure 2) that allows each mill to plan its daily load. This screen covers all the heavy equipment for that mill, such as the thermo-mechanical pulp systems, refiners, paper machines, winders, etc. Users can drill down and see, for example, that the day's production plan in one mill may be to run from 7-7:15 a.m. at 11.6 MWH, or 65% of its capacity, so various strategies can be planned for critical times.

Mills such as Iroquois Falls have their own generation facility, so they can purchase electricity at night and use their own generation during the day. When the price is high they can elect to shut down some equipment or the entire mill if the situation demands, and they can sell their generation to the grid. At times, electricity is the commodity of the day. Having this overview of the entire system is important because these mills with generating capacity are not only concerned with power purchases, but they want to keep their water level behind the dam.

We also monitor aggregate demand for the whole of the Ontario grid to track the demand variation. The market has different rules, regulations, and costs for high and low peaks, both in daytime and at night, so we have to monitor the maximum that we consume during those peak periods to achieve the lowest maximum peak during the daytime. The strategy is to incentivize users not to use peak loads during high-demand periods.

We even factor in the weather. We draw data from Environment Canada's web site and monitor primarily for the Toronto area because it's so heavily populated, therefore demand will vary more widely. We monitor temperature, wind chill, dewpoint, wind speed, relative humidity, atmospheric pressure--even precipitation, which is important for the mills that have generation because they want to know whether the water level in their dams is going to go up or down.

All this information comes together in the pricing arena. We pull pricing information from the IMO site on three price variants (Figure 3). The first is the Real Time Energy Market (RTEM), which basically is the spot market for electricity. This pricing is updated every five minutes.

Figure 3: Keeping Tabs

The IMO Energy Price Watch screen brings together all factors that drive MWH pricing: the Real Time electric Market (RTM), the Hourly Ontario Energy Price (HOEP), and the Pre-Dispatch pricing for the next day.

In addition, we monitor the Hourly Ontario Energy Price (HOEP), which is the hourly average of RTEM pricing. Since pricing is based on the HOEP, it's the most important figure for us. Using RTEM, we calculate a predicted HOEP every five minutes so operators can have a sense, in advance, of what the true HOEP may be at the end of the hour. For us, this five-minute tracking provides a heads-up so we can anticipate what the hourly average will be.

The third price movement we track is what the IMO calls the Pre-dispatch, or the forecast for the next day. This comes to us between noon and one o'clock every day. So far, the IMO's forecast pricing hasn't been very reliable, but we use it as a rough indicator for planning the next day's operations.

The Result: Tightly Controlled Costs

The system has worked very well during its first year of operation. Before the market was deregulated, consumers and corporations alike were paying $43 per megawatt/hour. Since deregulation, the average cost has been $59 per MWH from May 1, 2002, to April 9, 2003. Thanks to our ACI/EM-SYS, we have had an even more granular view, which has helped us fully optimize our energy costs. For example, from May to December of 2002 the average cost was $52 per MWH, but from January to April of 2003 it was $75 per MWH due to weather conditions. We can now view pricing on a basis of five-minute increments or any average up to one year.

The bottom line is that electricity is more expensive now, which makes it critical to manage efficiently when you're consuming millions of megawatt hours. We have to be able to respond to fluctuations on a continuous basis or it can mean huge losses.

As an example, during the first year of operation of the ACI/EM-SYS, the HOEP showed at one point a peak of about $800 for one hour. One mill reacted to it right away and altered production operations. Another looked at it and doubted that pricing was accurate. At the end of the hour the staff called the IMO to verify the data. It was correct, and that hour cost them tens of thousands of dollars.

As Jean-Guy Trudel, ACI vice president-newsprint, puts it: "This new tool enables the five Ontario mills to manage their electricity consumption and save an estimated $1 million per year. In addition, the system can be adapted for the company's other mills in the short term with minimum modifications."

And we accomplished all of this within a total budget of only $77,000, using our outstanding internal staff, instead of $3 million plus an estimated $450,000 a year for support and maintenance.

Guy Roussel, senior analyst-IT planning at Montreal-based Abitibi Consolidated, may be e-mailed at [email protected].