DDGS, WDGS, Modified—Which Makes More Money Today?

Model-predictive control projects can help ethanol producers improve decision making when pricing and producing distillers grains.
By Michael Tay | July 15, 2010
The cost to dry stillage, or any biomass, is generally calculated based on a straightforward linear assumption—the more biomass dried, the higher the cost. We live in a nonlinear world, however, where linear logic doesn't always hold true. When it comes to producing biomass, this is often the case because the energy-related costs required to dry an extra ton of biomass depend on a number of constantly shifting—and not necessarily linear—factors, including wet cake moisture levels, dryer loading and current drying efficiency.

These fluctuating variables present myriad challenges for commodity managers when pricing stillage to sell to a local market. Often the biggest hurdle is trying to identify the ideal amount of wet/local stillage to produce on a given day. This is a difficult decision to make when the manager doesn't have an effective way to determine the market price "tipping point" that should change their decision about how much wet or dry stillage to produce that day.

A dryer/stillage management model project can help commodity managers overcome some of these challenges. Model-based control is based on a mathematical model designed to be optimized based on price. The model incorporates real-time costs of manufacturing, along with scenario optimization, and can tie this information to a browser page available to commodity and plant managers.

Producers can use this model to calculate the ideal business scenario based on varying market pricing for ethanol plant coproducts. To maximize gross revenue, plant managers can use a dryer model (including auxiliary equipment tied to the production of dry or wet distillers grains) to gauge the amounts of dry distillers grain with solubles (DDGS) and wet distillers grain with solubles (WDGS) they should produce at specific prices. Alternately, they can determine what the market price for wet distillers grains needs to be in order to make producing wet grain more attractive than producing dry stillage.

Laying Out the Process
Dry mill ethanol plants that produce DDGS operate using a variety of process configurations with the most common industry layout using two rotary dryers in series. Other layouts may require flash- or steam-tube dryers. As a general rule, plants use from one to four dryers, depending on dryer and plant design and capacity.

In plants using two rotary dryers, producers have the ability to make from zero to 100 percent DDGS. The remaining distillers grains will become wet or modified product, where modified is a combination of centrifuge wet-cake, concentrated stillage syrup and partially dried product. DDGS is comprised of 10 to 15 weight percent water, and WDGS is 45 to 55 weight percent water.

The percentage of grain a plant operator decides to dry is based on orders from the commodity manager, who is selling both WDGS and DDGS. Order volume is determined by both product demand and market prices.

Consider the following example. The key performance indicator calculation sheet (see graphs on page 94) shows the impact on revenue as operators adjust screw speeds for wet-cake, partially dried-cake and syrup on a wet-pad screen in order to increase or decrease the tons per hour of WDGS produced. The calculations show the plant manager his estimated production revenues (in tons per hour multiplied by dollars per ton) of WDGS, DDGS and ethanol. Net gross profit (operating margin) also is shown. This is calculated by subtracting ongoing production costs from corn, natural gas, enzymes, water, electricity and other direct costs of production.

The graphs show that ethanol production has remained fairly stable over this time. However, as the screw conveyor settings are shifted to make varying amounts of wet product, estimated WDGS and DDGS revenues also shift. In the short period trended in the example, the last third of the production is most notable. WDGS production is highest during this period, and as a result, net operating margin drops significantly. This decline indicates that rather than the cost savings intended by producing more wet product, the change in operation actually decreases overall gross margin.

The reasons for negative impact vary, but the ultimate cause is a pricing imbalance. Pricing for WDGS per ton, when compared to the equivalent DDGS that would have been made, minus the costs of drying this additional product, was unbalanced. In this case, the commodity manager priced the WDGS too low, so when the plant produced additional wet stillage, the producer lost money. The producer should have continued to make dried product instead of wet.

Rotary Dryers: Fully Loaded
Avoiding this problem sounds fairly simple—either increase the price of WDGS compared to DDGS, or don't sell the WDGS. But unfortunately, the effectiveness of the producer's dryer performance adds a number of complexities that make the solution far from simple. In our example, the producer used rotary dryers. With this type of dryer, actual effectiveness of drying depends on a number of variables, including moisture content, air temperature, air humidity, air flow rates, feed water content and dryer loading (or bed level).

Looking at a single variable on its own—dryer loading, for example—showcases the complexity of building an accurate model to determine ideal production levels. In dryer loading, load level has a nonlinear impact on drying efficiency, and producers should seek an ideal full load where solid stillage level fills the maximum flight angle with the best air/solids interaction. (See Keey, R. B., "Drying of Loose and Particulate Matter," 1992 Hemisphere Publishing.) At this angle, drying air fully interacts with the maximum surface area of the stillage, which maximizes the effectiveness of the dryer. If the full load is decreased, extra hot air bypasses the stillage, negatively impacting effectiveness. Similarly, increasing the full load increases the "to-be-dried" stillage level, but doesn't add drying surface area, thus limiting effectiveness.

In the example case, there is an ideal load size, but there are two dryers in series, so determining the ideal load becomes slightly more complex. Either one dryer is ideally loaded and the second less than optimally loaded, vice versa, or some sub-optimum loading level is averaged across the two dryers. When a dryer is sub-optimally loaded, the producer spends more money per ton of product dried. As a result, the relative costs to dry or not dry another ton of distillers grains vary any time the loading level is changed.

Each variable in the dryer performance model represents a new and different set of challenges, and dryer loading is only one such example. Despite the varied complexities, it's possible to create a nonlinear dryer performance model that includes key parameters related to dryer performance.

An Added Complexity: RFS2
So far dryer optimization is dependent on two factors—developing a sufficiently accurate model of available drying operating handles and equipment performance, and identifying current market pricing drivers for the business. The U.S. EPA's revised renewable fuel standard (RFS2) and other similar legislative drivers may add complexity by mandating GHG emissions regulation.

RFS2, for example, is anticipated to reward producers with renewable identification number (RIN) credits who dry 50 percent or less of their distillers grains. Given the regulation, the pricing difference between WDGS and DDGS is one decision driver, but there is an added penalty consideration if the plant will be within bounds for RIN credits at 50 percent dried production and out of bounds at 51 percent.

Mathematically, there are two ways for producers to simplify the issue. The simple way, and probably the easiest way from an anticipated reporting perspective, is to constrain or limit the economic optimizer at 50 percent, or at whatever the documented pathway or allowable operating limit is set. In this case, the optimizer would not violate the set limit, even when it may be temporarily opportunistic to make more DDGS.

Another more mathematically elegant solution that is more complex in terms of reporting requirements is to add an additional economic penalty to the optimizer that is equal to the value of the RIN credits. With a penalty in place, the mathematics would allow the producer to decide when it makes economic sense to exceed a 50 percent limit—even though it may include the loss of prospective RIN credits on a fraction of production.

While the elegant solution is interesting from a mathematical and hypothetical perspective, the simpler solution makes more sense in the real world. The same model can be exercised within a variety of scenarios and used to determine when changes to permits and allowable operating conditions are needed.

However, in day-to-day plant operations, decisions regarding how to run on a given day are constrained within the capabilities of the plant's current equipment and operating permits. Nevertheless, significant savings—on the order of 3 to 5 percent increased operating margins or more—may be realized using model-based optimization based on current economics to support operating decisions. EP

Michael Tay is manager sales engineering for Pavilion Technologies, a Rockwell Automation Company. Reach him at mtay@pavtech.com or (512) 438-1482.