Difference between revisions of "Sugar Cane Ethanol Plant"
|Line 745:||Line 745:|
Revision as of 16:06, 15 January 2014
As the alternative energy industry continues to grow in the United States, Global Impact Chemical Corporation (GICC) has taken interest in producing vehicle biofuel from plant matter such as sugarcane. With a capital cost limitation of 1 billion USD, a process plant design and the corresponding economic analysis were devised. Liberia, Guanacaste Province, Costa Rica was chosen as the plant’s location due to inexpensive cost and proximity to rich natural resources. Due to market demand of anhydrous ethanol in the United States and hydrous ethanol in Brazil, both forms of ethanol may be produced with the proposed plant design.
Research on current ethanol manufacturing processes guided the final design of the ethanol plant. The proposed design was divided among three major processes: milling, fermentation, and separations. An electricity cogeneration system and an additional bagasse hydrolysis process assist in keeping the proposed plant self-sufficient and increase ethanol production. Microsoft Visio and Aspen HYSYS were used to design complete process flow diagrams and simulate the fermentation and separations processes. All other calculations were performed in Microsoft Excel. It was estimated that an initial feed of 147 tons per hour of sugarcane, milled and fermented with the bacteria S. Cerevisiae, produced the desired total of 20,000 kg/h of hydrous ethanol and 19,500 kg/h of anhydrous ethanol. Aspen Economic Analyzer aided in estimating the cost of the sized equipment in each of the three major processing steps. The total capital cost of the plant, as designed, is estimated as $465.9 million.
Economic analysis predicts a net present value of $240 million on a twenty-year basis. Furthermore, the estimated rate of return is 19.7% after twenty years, with a pay-back period of 4.10 years, satisfying the desired payback time of three years. A gross margin percentage of 70% after plant startup was calculated, thus the plant’s potential financial performance is significantly better than the average range of 40-50%. According to the analysis, the proposed sugarcane ethanol plant design would be economically viable and would provide GICC with a promising first step towards a biofuel for alternative fuel vehicles.
Fuel ethanol has been an important alternative fuel for decades. With petroleum prices reaching upwards of $100 per barrel on a regular basis, the market for fuel ethanol is poised to grow even further. Currently the largest producer of ethanol in the world is Brazil, where a combination of fuel composition standards, supplier incentives, and low sugarcane prices combine to make the production of sugarcane ethanol a lucrative opportunity. The market in Brazil appears to be nearing saturation, however, as evidenced by a growth rate in production of only 5% from 2009 to 2010 . This pales in comparison to the near 30% growth that the US market has experienced over the last 5 years. It is believed that much of this growth is due to the high gasoline prices in the country and that this may be addressed by an increase in the production of alternative fuels. Our team has decided to pursue the design of a sugarcane ethanol plant to be located in Costa Rica which will serve both the domestic, Costa Rican market, and the US market. Costa Rica offers an area with low sugarcane prices which will allow us to follow the more efficient sugarcane-fed process instead of the cornbased process pursued in the US. Costa Rica also offers free-trade agreements  which allow the ethanol to pass into the US tariff free, providing a significant advantage over Brazilian competitors.
Anhydrous ethanol, contains 99.5 wt% ethanol, while hydrous ethanol contains only 95 wt% ethanol. Anhydrous is required for vehicle use in the United States, as opposed to Brazil which uses hydrous. The key difference is an azeotrope that is required to be crossed in the ethanol-water mixture to distill to anhydrous ethanol.
The actual design of our plant was driven by current manufacturing processes found in literature. The design is split up into four main sections for simplicity which are milling, fermentation, separations, and utilities. Aspen HYSYS was used to model the separations process unit while Microsoft Excel was used in the design of the fermentation and milling units. All Process Flow Diagrams were produced in Microsoft Visio.
This venture seeks to produce anhydrous ethanol for domestic sale in Costa Rica and for export to the United States market. In 2008, the Costa Rican government established a mandate dictating that all gasoline sold domestically must contain 7% ethanol. The Costa Rican government expects to increase the percentage of ethanol mixed with gasoline to 12% in the next 4-5 years. Over the past two decades the U.S. ethanol market has grown dramatically. Between 1990 and 2007, U.S. ethanol consumption increased from 3.4 billion liters per year to 26 billion liters per year. Such a significant increase in demand is attributed to the implementation of the Clean Air Act and the establishment of a renewable fuel standard in the Energy Policy Act of 2005. The renewable fuel standard requires that gasoline sold in the U.S. contain a renewable fuel standard, such as ethanol. The latter mandate required 15 billion liters of renewable fuel in 2006, increasing to 28 billion liters in 2012. The Energy Independence and Security Act of 2007 expanded the renewable fuel standard, reaching an ultimate peak of 136 billion liters in 2022 . In 2009, Costa Rica produced 100 million liters of ethanol, 70 million of which were exported . It is important to note that these values are significantly less than the production values for the years preceding 2009 as a result of the global reduction in demand. It is anticipated that Costa Rica’s annual ethanol export growth mimics the average annual growth rate of the U.S. market of 38 percent. Domestic consumption is anticipated to increase 10% annually. The proposed plant will seek to capture (based on 2015 values, when construction of the plant will be completed) 5% of the domestic market, and 30% of total ethanol exports. In order to achieve this goal, the present ethanol plant is designed to produce 178 million liters of ethanol per year. This information is summarized in the attachment SCEP Design Basis vFP.
The total fixed capital cost of the current design is $465.9 million. The ISBL is $267.3 million and the OSBL is $178.2 million, with an engineering and contingency cost of $20.5 million. This cost, and all other price data, were adjusted for Costa Rica’s geographic location and 2011 dollars. The main product revenue from anhydrous ethanol is $117 million annually, with additional revenues of $84 million and $3 million generated from the sale of electricity and food or industrial grade carbon dioxide, respectively. The cost of cost of sugar-cane, our raw material, adds up to $18.4 million annually. Other variable capital costs include waste removal at $6.1 million annually and consumables at $3 million annually. All available capital cost streams can be viewed in the Cost of Production spreadsheet in the attachments. Using the 7 year MACRS depreciation method, a 30% tax rate , and capital available at 12%, the project is estimated to have a simple payback in just over 4 years with a net present value (NPV) of $27.2 million and $240 million at 10 and 20 years, respectively. The 10 year internal rate of return (IRR) is 13.5%, after 20 years IRR is equivalent to 19.7%. While the NPV and IRR financial estimates meet the goals set out by the CFO at the beginning of the project, the simple payback period is one year shy of the CEO’s goal. Two areas of potential error that need to explored future are royalties cost for the Organsolv process used (discussed in Bagasee Hydrolysis section) and the electricity regulatory statutes in Costa Rica. Additionally, the values in this economic analysis may fluctuate slightly as all of the smaller components of the plant such as pipelines undergo final design and sizing. The economic calculations are in the attachments Economic Analysis v4 and ICARUS Project Summary.
Costa Rica was chosen for a host country for several reasons. First, Costa Rica is a signatory of the Central America and Dominican Republic Free Trade Agreement (CAFTA-DR), which facilitates free trade (duty free) between the Unites States and Costa Rica. As per the ethanol provision of the CAFTA-DR, in alignment with the Caribbean Basin Initiative (CBI), which has limited ethanol imports to the United States at 7% of US domestic production, Costa Rica is allotted 117 million liters of ethanol exportation to the US annually. As of 2005 Costa Rica was exporting about 57 million liters annually to the US, while the US market for fuel ethanol consumption has increased by 11% per year from 1995 to 2004. Therefore, Costa Rica can certainly find a market for the addition 57 million liters annually it is allotted to export to the US. Other sugar rich countries such as India and Brazil are not in the free trade agreement, and therefore are subjected to harsh tariffs on exports to the US. CAFTA-DR and CBI also prohibit ethanol with origins other than the signatories (ethanol processed but not synthesized in the signatories) from entering the US, therefore complete production in Costa Rica is necessary . The one unfortunate element of producing in Costa Rica is a lack of government incentives; however, it is possible that they could appear with the increase in fuel blend requirements. With the strongest economy and most stable government in Central America, Costa Rica made an excellent choice for plant location .
Currently, companies are developing ethanol production facilities in Costa Rica. One such company is United Biofuels of America. While these facilities will provide competition to our production plant, the overall trend towards less petroleum dependence within the country will provide business for many facilities. On the raw materials side, there has been a steady increase in sugar cane production in Costa Rica over the last fifty years .
The plant will be located in Liberia, Costa Rica as shown in figure 1. Liberia is the capital of the Guanacaste Province and is the home to a population of over 35,000. Liberia was chosen within Costa Rica due to its proximity to many sugar plantations, Pacific Ocean ports, and the Pan-American Highway. Additionally, the Tempisque River runs adjacent to the town, providing on site fresh water. The proximity to the sugar plantations will reduce the shipping costs associated with procurement of the raw sugar cane. The Pacific Ocean will provide shipping access to foreign nations. The Pan-American Highway will provide necessary infrastructure for materials procurement and product shipping. Additionally, Costa Rica offers potential for shipping from the Atlantic (Caribbean) coast to add lucrative markets such as the European Union.
The design is planning to purchase a site just northwest of the town, at the intersection of the Tempisque River and the Pan-American Highway. This site will provide proximity to Liberia and neighboring towns for our plant staff, without bringing the industrial complex to the cultural and beautiful city center. If adequate public transit does not already exist, GICC will look to partner with city officials to develop a bus system between the city center and the plant. This is just one of many plans GICC has explored to be a good corporate citizen in Costa Rica. The one square kilometer site is estimated to cost $30 million.
Guanacaste Province, is also home to one of Costa Rica’s best technical universities, Invenio, which is located in Canas. Invenio is a premier science and technology university. The Guanacaste Operations team is developing methods to involve the current and future student body through internships on the development, and operation of the plant . In the future, GICC will look to Invenio for a source of engineers and other technical personnel to run the plant.
The production of ethanol through the process of sugarcane fermentation is a three-step process which has been used for many years in such sugar-rich countries, such as Brazil. The entire process may be divided into milling/pretreatment, fermentation, and ethanol purification process units. In the milling section, the raw sugarcane is washed and grinded to form a sucrose product which is then purified using crystallization. The bagasse, which is the leftover product from the sugarcane milling, is separated out and passes through another process in order to convert it into usable feedstocks. The bagasse is a cellulosic material which, in its raw form, cannot be processed by the yeast cells present in the fermentation reactors. To create a usable feedstock from the bagasse, ninety percent of the material undergoes an acid hydrolysis step which converts the material into a glucose solution. The remaining ten percent is sent to steam boilers to be burned as fuel. The steam boilers use the bagasse fuel as well as lignin removed in the milling process and field trash that accompanies the sugar cane from the farmer, to generate enough steam to run the plant and generate plant electricity. Excess electricity generation is sold on the market for additional revenue.
The glucose and sucrose solutions are then mixed before being sent to a storage tank prior to the fermentation process unit. This storage tank allows the mill to run continuously while the fermentation unit runs as a staggered, batch process. The sugar solution in the storage tank is then mixed with the growth medium for the fermentation and is sent to an appropriate fermentation vessel which has been preloaded with the required yeast for the process. The fermentation will run for twenty seven hours during which eighty seven percent of the sugar present initially is consumed. At the conclusion of the twenty seven hour period, the batch vessel will contain a 12.5 wt. % solution of ethanol in water which will be sent to another storage tank. This storage tank will feed into a continuously operated hydrocyclone which separates the yeast from the reactor effluent. The purified solution will then be sent to the ethanol purification unit while the yeast is mixed with water and pumped to a propagation vessel to be regenerated for the next fermentation.
The solution entering the ethanol purification unit undergoes a series of distillation steps to separate out the various components of the mixture. Significantly, carbon dioxide is separated from the solution and compressed into cylinders for sale as beverage grade carbon dioxide. These distillations proceed until an azeotrope is formed at 95.63 wt. % ethanol at which point entrainment distillation using benzene is performed to produce anhydrous ethanol with a purity of greater than 99%. Further information about the flow rates are available in the attachments SCEP Design Basis vFP and Cost_of_Production.
The following pages contain a full three page PFD of the process. The first pages is Milling/Pre-treatment of Sugarcane, which inputs raw sugarcane and creates a sucrose and glucose mixture to feed the fermenters. The second page is Fermentation and Separations, which utilizes the sucrose and glucose mixture, fermenting it into an ethanol solution, and then purifying that solution to the final product. The third page is Plant Utilities, which documents the major utility equipment of the facility. Each will be explained further later in the paper, including an equipment list.
Production is scheduled for 300 days of the year, running continuously from February to early December. In December, the plant will shut down for maintenance, cleaning, training, and vacation. In January, the plant can optionally run the dehydration separation steps, bringing in hydrous ethanol as a feedstock and outputting anhydrous product. February will resume full production. This longer than expected shut down is because sugarcane does not grow all year round and can expire. Design Considerations: Milling
Milling and Pre-Treatment
Milling is a very standard process that has been used in the sugarcane industry for many years. With the success that current process equipment has had, the proposed design was modeled very closely to what is in use today . A cane diffuser extracts the desired sucrose needed for ethanol production from the harvested sugarcane. The sucrose mixture is then pretreated and concentrated to the correct concentration needed for efficient fermentation to occur. Based on the amount of ethanol produced, a sugarcane feed of 147 tons per hour is required to ultimately produce an 84 wt.% sucrose solution that will be mixed with the glucose solution that comes from the bagasse hydrolysis process. Table 1 summarizes the equipment chosen for this process.
The total cost of the milling equipment is summarized in Table 2.
A detailed discussion of the milling and pretreatment equipment and material streams is available in the attachment Milling Material Streams.
Key assumptions made for material balances and sizing were found in laboratory papers . in sizing the equipment, the multiple effect evaporators were assumed to be of equal size to make calculations simpler. In reality, each evaporator would be smaller than the previous one.
Although the milling of sugarcane closely followed the standard process used today, the sucrose extraction step and clarification step were researched and investigated for better options. As shown in Table 1, the sucrose extraction step is completed by shredding the sugarcane and thus breaking the cells in which sucrose is located, making it easier to extract. The clarification step is important because it removes the impurities from the cane juice that would otherwise only interfere with the product purity.
Conventionally, roller mills are the equipment of choice used to extract the sucrose from sugarcane. However, sucrose extraction by diffusion has recently emerged in the industry. The sugarcane goes through twelve stages where it is heated co-currently by imbibition water, immediately raising the temperature of the cane to about 75 °C . This heating even opens the cells that were not opened by the previous shredding process. Another advantage of this heating is that sugar destroying bacteria cannot survive at this temperature. However, in a mill tandem, the average temperature is between 30-35 °C, where these bacteria are still active . Extraction by diffusion boasts a higher extraction percentage, lower operating and manufacturing costs, and better mechanical reliability. The ratio of capital costs of a plant using a diffuser compared to one using a mill is about 1:1.5 . One downside of diffusion is that the shredding process is significant because the cane needs to be the right size for efficient diffusion to occur. Therefore, a heavy-duty shredder is needed, which is initially more expensive and requires more power. However, looking in the long run, the appeal of lower capital cost, lower operating and maintenance cost, and high extraction percentages made diffusion the better alternative.
The clarification process is important because it removes the impurities from the cane juice and also neutralizes its acidity. There were many options to consider: defecation (Ca(OH)2), magnesia (MgO2), sulphitation (lime and SO2), carbonation (lime and CO2), and phosphatation (lime and H3PO4) . Each differ in cost, the time needed to precipitate the impurities, and product purity. With these factors in mind, calcium hydroxide, also known as milk of lime, was chosen because its reaction time was the shortest and cost the least. The comparisons are shown in Table 3.
Milling Possible Errors
There were a few sources of possible error in the milling process. When sizing the diffuser, the main equipment piece in this section of the process, dimensions were extracted from existing diffuser models. Using the flow rate of sugarcane needed to produce the desired amount of ethanol, the size of the diffuser was estimated. Also, because the evaporators were assumed to be the same size, ICARUS produced a larger cost than it should be (viewable in ICARUS equipment list attachment). These possible errors would affect the overall cost of the design.
Bagasse Hydrolysis Results
Bagasse consists of the fibers that come out of the diffuser and in industry, it is generally burned in a cogeneration system that produces electricity to be used in different parts of the ethanol production process. However, recent processes have begun to process the bagasse in a way that it can be added into the fermentation step to increase ethanol production. About 90% of the bagasse produced is used to make ethanol while the remaining 10% is used to generate electricity. This process pretreats the cleaned bagasse with 2 wt.% sulphuric acid  and steam and through reactors, filters, distillation, and evaporation, a 25 wt.% glucose solution is produced. Sterilized sugarcane juice is diluted and added to the glucose solution to before entering the fermenters. The addition of the bagasse sugars increases ethanol production. Table 4 summarizes the equipment chosen for this process.
- All pretreatment and hydrolysis equipment will be made of stainless steel because sulphuric
acid is being used.
The total cost of the bagasse hydrolysis equipment is summarized in Table 5 below.
A detailed discussion of the bagasse hydrolysis equipment and material streams is available in the attachment Milling Material Streams.
Bagasse Hydrolysis Possible Errors
Possible sources of error in the Organosolv process stem from the fact that the process has yet to be produced on an industrial scale. Therefore, all data has been interpolated from laboratory data. This would potentially affect the final glucose yield in the process as well as the overall cost of the design.
The fermentation operation was based on standard batch processes using the well characterized yeast strain, Saccaromyces Cerevisiae. The glucose and sucrose solution from the milling section is sent to two, 100,000 L storage tanks which will provide a buffer between the continuous milling process and the batch fermentation process. As one of our 46 fermentation vessels becomes available for filling, the sugar solution from the storage tanks is pumped out of storage, mixed with the required yeast nutrients, and sent into the fermentation vessel. Yeast culture is then pumped to the fermenter from propagation vessels which are the last stage in our yeast recycle loop. The reaction will proceed within the fermenter for 27 hours which is based on literature values. When the reaction is completed, the solution will pass to one of another two, 100,000 L storage tanks and then to a 41.7 cm diameter, continuous hydrocyclone. This hydrocyclone will separate the yeast from the product mixture which will then be mixed in a 200 L mixing tank to a 50 wt % water slurry for pumping to one of 9 propagation fermenters. This yeast separation step allows for recycle of our used yeast. Meanwhile the remaining product solution is passed to the separations unit for purification. A summary of all required equipment is given in Table 6.
Details pertaining to the design of the equipment listed in table 6 can be found in the attachment entitled Fermentation Calculations.
Mode of Operation
Both continuous and batch processes were considered for our fermentation units. The continuous process was based on immobilized cell bed reactors  though there were also examples of fluidized bed reactors  which also performed the desired process. The batch process, on the other hand, was very simple in that it only involved loading the specified amount of material and reacting these components for a specified period of time. Both modes of operation were compared in deciding our final process. A continuous process such as an Immobilized cell bed offered a higher output rate for a given capital investment while also allowing easy interfacing with the rest of the plant. The batch process, in comparison, offered easy cleaning, flexible unit operation, and easier process control due to the simple reactor design.
When comparing the two modes, our team decided to pursue batch operation for our fermentation units. This decision was largely motivated by the simplicity of the batch reactor and the lack of industrial examples of continuous fermentation. Given the current economic climate, it would be imprudent to pursue a risky venture such as continuous fermentation without significant experience to draw upon so it was decided that batch fermentation was the best route to take. In order to minimize costs in other areas of the plant, it was decided that though the fermentation units will be run as a batch operation, the units will be staggered so as to allow continuous operation in the milling and separation processes. The fermentation vessels will be preceded and followed by large storage tanks which will serve to buffer the small variations in flow rate to or from the vessels. This will allow the overall plant to continuously produce ethanol.
Choice of Bacteria
In deciding what type of organism to use, there were two main options. These two options were Saccharomyces Cerevisiae and Z. Mobilis. Saccharomyces Cerevisiae is a strain of yeast and is the most widely used organism in the beverage and fuel ethanol industry. This organism is widely known and there is a wide range of literature detailing the optimization of its alcohol production. Yeast produces alcohol in concentrations of approximately 10% and can use a wide range of substrates as feed which simplifies the fermentation growth medium. This upper limit for ethanol production is due to the toxicity of ethanol toward the yeast and is the major disadvantage of this organism.
Z. Mobilis is a strain of bacteria which has become a major focus of research in recent years. This recent focus has come about due to a higher ethanol tolerance when compared to the traditional S. Cerevisiae strain of yeast. This higher ethanol tolerance allows more efficient use of batch units as the process can produce more concentrated solutions of ethanol. The major disadvantage of Z. Mobilis is a narrow range of substrates which may be digested for fermentation. This disadvantage is likely a minor issue for sugarcane fermentation, however, due to the simple sugars (glucose and fructose) which are produced from the milling process. In the end, it was decided that S. Cerevisiae was the best organism for the process. Z. Mobilis is largely a “laboratory-scale” strain as there are very few examples of its use in industrial applications. Z. Mobilis is an option that should be under ongoing evaluation in the future, however, due to the promise it shows as an ethanol producer.
The growth medium is a very important part of fermentation because without supplementary nutrition, yeast metabolism and ethanol production is severely hampered. The most important nutrition requirements for fermentation are a Nitrogen source to provide raw material for protein synthesis and a source of Magnesium which allows for rapid yeast growth . The original growth medium to be used in our process is shown in table 6 . This medium, however, was adapted from laboratory practices and was found to be uneconomical on an industrial scale. The laboratory medium was quoted at more than $20 million per year of supplementary nutrients and this was determined to be unacceptable for our commercial process. The most significant contribution to the price was due to the presence of 5 g/L of peptone in the growth medium which was quoted at around five dollars per kilogram or $19.3 million per year.
There has been some work in optimizing growth media for industrial applications, however, and Pereira et. Al.  have performed an in depth investigation of this problem. They found that the growth medium presented in Table 8 allowed for optimal yeast nutrition and performance in industrial applications for a minimal price. The savings in this instance are largely a result of substituting corn steep liquor, a complex nitrogen source, for yeast extract and peptone. Corn steep liquor was quoted at five cents per kilogram  and offers a significant economic advantage over typical laboratory supplements. This is the growth medium that was chosen for our fermentation units.
Due to the low volume present in the propagation units, we decided to use a laboratory scale growth medium for our propagation step. The cost of material for this step is insignificant given the scale of our overall process. This growth medium is presented in table 9 .
Yeast Separation Method
Separation Method After the fermentation step, it is desirable to remove the yeast solids from the reactor product before sending the stream to the distillation unit. This prevents the yeast from interfering with any of the distillation calculations but also allows for the possibility of yeast recycle back to the reactor vessel. For the separation, we considered two options. First we considered a disc-stack centrifuge. This type of centrifuge is widely used in the industry and offered an attractive, high efficiency separation. This unit requires the use of centripetal force to separate the yeast solids from the outlet stream and this was the main disadvantage with this option. Spinning forty kilograms of solution per second was an undesirable situation and the high shear stresses inside a disc-stack centrifuge may have caused damage to the yeast which would limit the efficiency of the recycle stream.
The other option for this separation was the use of a hydrocyclone. A hydrocyclone is a cone with the inlet at the top and outlets at the top and bottom for the liquid and solids respectively. The separation is driven by the pressure drop in the system with the mechanism being the vortex created in the interior of the unit. Our specific outlet composition appears to be on the lower end of the range for which hydrocyclones apply according to Towler . It should also be noted that our solids concentration will likely be greater than that quoted in the flow sheets due to excess yeast culture growth in both the propagation and fermentation steps. Due to this, hydrocyclones should do an adequate job of separating the yeast from the ethanol product. Hydrocyclones offer a number of benefits including minimal moving parts (safety and maintenance), low energy requirements compared to centrifugation, and cheap capital costs. For these reasons, a hydrocyclone was chosen for our yeast separation step.
The option of recycling yeast was explored in our design process. The yeast would be separated in the hydrocyclone, then mixed with water into a 50 wt % slurry. This slurry would then be pumped to a fermentation vessel where the yeast would be regenerated to form the culture for the main fermentation tanks. This recycle process was compared to buying new yeast to pitch into the fermentation tanks instead of regenerating the used yeast. We found that fresh yeast may be acquired for approximately $2/kg  and, with only 35 kg required per batch, this was an interesting option for the plant. It was found that pursuing yeast recycle and propagation was likely more economical than the alternative, however, due to relatively equal capital cost requirements but advantageous operating costs. Fresh dried yeast was found to require the use of a tank to soak the yeast in order to activate the culture for fermentation. This increased the capital costs beyond those initially estimated. Buying yeast would also cost at least an additional $730/day extra when comparing the cost of the yeast versus the foregone ethanol production due to sugar consumption in propagation. Due to the advantageous economics, we have decided to install a propagation step in our plant to regenerate used yeast.
Based on the decisions discussed in the previous sections, a set of assumptions was developed for the fermentation section. Information for input prices and key assumptions for the fermentation reaction are given in tables 10 and 11.
Fermentation Possible Errors
There are a number of aspects of the design of our fermentation process unit which deserve further research and verification. The most significant source of possible error is the assumptions for the actual reaction. Most significantly, the reaction time and final ethanol concentration may prove different upon scale up from laboratory scale. The industrial scale yeast growth medium shown in table 8 may also affect the reaction time and final concentration. The cooling requirements for the reactor are also a likely source of error. The reaction profile was assumed to be linear for simplicity and this is known to be incorrect. An actual fermentation reaction will have a slow build up stage, an exponential growth stage, and a slow nutrition limited stage. Therefore, during the exponential growth phase, the cooling requirements will likely be much greater than estimated. All of these sources of error are centered on the actual chemical reaction characteristics. Therefore, it is recommended that a small reactor be used to characterize these key variables prior to the final build.
Research on industrial ethanol-water separations processes yielded an initial design for the production of both hydrous and anhydrous ethanol. However, such industrial separation processes also produce a number of byproducts, such as fusel oil, which ultimately reduce the yield of hydrous ethanol . As the fundamental premise of this project is to produce as much ethanol as possible from a given quantity of sugar cane feedstock, the proposed design differs significantly from traditional ethanol separation processes. The separations for this method were modeled using Aspen Hysys. The model is presented in Figure 2. A full output of the Hysys simulation is available in the attachment Separations Workbook v2.
It was deduced that the most effective means of achieving the desired ethanol product purity was to design a separations process that consists of two parts: 1) the distillation of hydrous ethanol and 2) the dehydration of hydrous ethanol by an azeotropic distillation process . The azeotropic distillation design utilizes benzene as an entrainer . From this method, the desired quality (99.9 mol% ethanol) and quantity (19,500 kg/h) of product is achieved. The recommended design is presented in the Fermentation and Separations PFD and Figure 2. Table 12 summarizes the equipment utilized in the separation process.
From this design, not only are the ethanol production quality and quantity requirements met, but byproduct carbon dioxide (CO2) produced from the fermentation process is successfully separated and purified to 99.9 mol% and produced at a rate of 5200 kg/h. This process introduces an additional $415 per hour cash flow based upon present carbon dioxide prices. This provides not only an additional revenue stream for the plant but also a hedge against predicted new regulations on carbon dioxide emissions.
The total cost of the separations and purification process is summarized in Table 13.
The sizing calculations for the separations vessels as well as one of the pumps, one of the compressors, and one of the heat exchangers are available in the attachments BENZ STRIPPER, DEHYDRATOR, DISTILLATION_SIZE, RECTIFIER, Phase_Splitters, P-800, C-2, and E-200.
One of the fundamental assumptions of the separations processes was the composition of the feed from the fermentation units. After consulting a literature source  utilizing a similar fermentation design to the one suggested in this report, and accounting for the ethanol production of the yeast Saccharomyces cerevisiae, the inlet composition as given in Table 14 was determined.
A second assumption was implementing the NRTL fluid package. From the Aspen HYSYS manual, it was deduced that this fluid package would best represent the components in the separation process.
The final separations and purification design implements three rectifying columns, one distillation column, four flash drums and a decanter, as summarized by the Fermentation and Separations PFD and table 12. The flash drums serve a threefold process in the process: 1) to eliminate volatile components from the product stream, 2) purify byproduct carbon dioxide and 3) purify recycle streams. Flash drums are preferred in separation processes as they retain significantly lower capital, operation and maintenance costs, as evident in Table 13.
For the hydrous ethanol purification process, a rectifying and distillation column are implemented in series. The rectifying column removes water and heavy components, while the distillation column purifies the product stream to a composition of 91 wt% ethanol – a value slightly under the ethanol-water azeotrope composition of 95.6 wt% ethanol . The choice to utilize a rectifying column was made for both economic and practical reasons. As demonstrated by the Fermentation and Separations PFD, the distillate vapor collected from the rectifying column is sent directly to the distillation column. Maintaining the feed as a vapor ensures a more effective separation of the ethanol from water and other undesired components. Furthermore, such a design also reduces the energy demand of the distillation column reboiler.
The ethanol dehydration unit consists of two rectifying columns in series. The first column of the dehydration unit is fed with two steams: 1) hydrous ethanol and 2) a recycle stream containing benzene. The addition of benzene, which acts as an entrainer, leads to the formation of a ternary azeotrope that possesses a boiling point significantly different from the boiling point of ethanol. During the azeotropic distillation process, the ternary azeotrope would distills to remove water. Then the binary azeotrope distills to remove any benzene, resulting in 99.9 mol% ethanol as the bottoms product. The distillate water, ethanol and benzene mixture is fed to a decanter, where organic and aqueous phases are separated. The aqueous phase is subsequently fed to a secondary rectifying column where wastewater is collected and benzene is recycled.
Despite its deleterious health effects and environmental effect, benzene was chosen as the entraining agent. Although a number of alternatives act to form a ternary azeotrope with ethanol and water, such as cyclohexane, it was determined that benzene proves to be the most cost effective, as it requires the least amount of material to form the desired ternary azeotrope. It is important to note that 0.33 kg/h of benzene exit the dehydration unit in the anhydrous ethanol stream. Relative to the quantity of ethanol in the stream (19500kg/h), this quantity is minimal. Quantitatively, benzene in the anhydrous product is 0.0015 vol%. This is significantly below the Environmental Protection Agencies (EPA) maximum limit of 0.62 vol% stated in the Mobile Source Air Toxics Rule .
Separations Possible Error
The likely sources of error in the separation process design pertain to the validity of the initial assumptions. Specifically, it is possible that the composition of the fermentation feed stream 1) retains different mole compositions or 2) addition (or fewer) components from those presented in Table 14. Furthermore, it is presumed that the NRTL equation of state accurately describes the vapor-liquid equilibrium (VLE) of all of the components in the separations process. From the simulation output, it is evident that the NRTL fluid package does indeed capture the VLE relationships accurately for simple binary mixtures of ethanol and water. However, for multicomponent distillation systems (such as those in the hydrous purification system) it is likely that some error arises naturally from the inability of NRTL to accurately predict VLE data for such complex mixtures.
The plant utilities are displayed in the third page of the PDF, entitled Plant Utilities. A list of the equipment shown is below in table 15.
Heat Exchanger Network
To economize utilities production, two heat exchanger networks were developed. One is based on the milling and pretreatment area and the other on the separations area. Together, the two networks save the plant the need for an additional 22,500 kW of heating and cooling (45,000 kW total), by using internal heat exchangers. These heat exchanger networks assumed a ΔTmin of 20℃ and can be viewed in the attachment Heat Exchanger Networks. The calculations are with the related sections of the plant in the attachments Milling Material Streams, Separations Workbook v2, and Utilities Streams.
Many sugarcane ethanol plants in the world utilize other materials besides the main sugar cane feedstock for utility stream production . Our plant will separate up to 90% of the waste bagasse material to be processed through hydrolysis and delignification process that will produce glucose liquor. The glucose liquor will be fed to the fermentation process to increase ethanol production. Produced lignin will be added to the remaining 10% of the bagasse material for combustion in the steam boilers as can be seen in the Milling/Pre-treatment PFD and the Plant Utilities PFD. These splits were based on research studies our team found on the utilization of the bagassse not only as a fuel, but also as an additional source of ethanol. Research suggests that these operational splits will allow for complete onsite electricity and steam generation for the production facility in addition to increasing ethanol yields by up to 26% .
The size of the co-generation plant was based on several assumptions about the feed streams and equipment. The plant will make use of the remaining 10% of the bagasse, mixed with the lignin and some of the field trash, excess crop material that is often a waste stream for the farmer, to generate steam. Table 16 contains values from literature and the material balances regarding the boiler fuel .
Table 16 Based on steam generation literature for bagasse combustion, the boilers are expect to run at an efficiency of 85% . The cogeneration system will be backed up with #2 Fuel Oil for emergencies, startup, and shutdown. The fuel oil has a lower heating value of 39.9 MJ/L. Current designs include a single high pressure steam (HPS) stream utilized throughout the plant. The HPS will be maintained at 500℃, 20 bar as super saturated steam. our boilers will generate 8.7 million kg of HPS per hour. Based on a 872,000 kg per hour throughput of water and steam, the plant will require 10 field erected boilers. Calculations related to the co-generation steam generation are available in the attachment Utility Streams.
According to EPA studies, the combustion of sugarcane bagasse will not be a significant source of nitrogen oxides and sulfur oxides. It will, however, provide a significant source of volatile organic compounds and particulate matter. The boiler is equipped with wet scrubbers to maintain the boiler exhaust to within environmental regulations. For this matter, in absence of finding environmental regulations from Costa Rica, we have followed the regulations of the US EPA. The scrubbers will also provide relief when using the reserve #2 fuel oil .
Several assumptions were also considered in sizing the electrical generation portion of the plant. After use in operation, the remaining steam will drive electrical turbines (at 98% efficiency) to generate electricity. After the electrical turbines, the steam should contain just enough energy to remain saturated steam vapor. Ultimately, the vapor will be condensed to a liquid in a shell and tube heat exchanger to preheat water entering the boiler. To calculate the energy that can be convert by the steam turbines, it was the energy per hour of the steam leaving the boiler was calculated. Then the base energy (the energy of the water the steam will be after the turbines), and the process steam energy requirements were subtracted from the energy per hour value. Finally, it was assumed that the steam loop would undergo 50% losses of energy to the environment throughout the system loop. The remaining energy was converted at 98% efficiency to electricity . Calculations related to the co-generation electricity generation are available in the attachment Utility Streams.
Additionally, it was estimated that the plant would be capable of selling electricity at a price of $0.08 per kilowatt hour. This is based on an industrial and commercial average electricity price in Costa Rica of $0.12 per kilowatt hour. Since the electricity is highly regulated a conservative estimate was chosen .
The electrical generation portion of the plant is estimated at a rating of 163 MW. A tighter ultimate build on the plant will restrict losses, increasing electrical production. This estimated generation will surpass the expected plant requirement of 19 MW, allowing the plant to sell up to 144 megawatts (MW) of electricity per hour. Based on the electricity price in Costa Rica of 0.08 cents per kilowatt hour, the plant can generate an extra revenue stream of $84 million per year selling electricity . The capital cost of the cogeneration plant is estimated at $143 million installed.
Water Treatment Plant
Treatment Plant In order to provide clean water for plant processes, the plant will utilize a two stage ion exchange water treatment process. The plant will be equipped with 40 units capable of cleaning 52 cubic meters of water per hour. The sizing of these facilities is based on the water recycle ratios and process water needs described below. These treatment units will be used to clean Tempisque River water for use as steam, cooling water, and washing water. The units will also be used to clean wash water after use in the milling process prior to disposal in the Tempisque River. The total cost of the water treatment equipment is estimated at $4.5 million installed. Calculations related to the water treatment plant are available in the attachment Utility Streams.
In-Process Water Recycling
While the design would optimally recycle all of the in-process heating and cooling water, it is realistically infeasible due to fluctuations in production, fowling, and degradation. Therefore, initial set points have set the cooling water at a recycle rate of 90% and the steam boiled water at a recycle rate of 98% (this is due to already accounted for pipe losses). Washing water will be recycled at a rate of 50%. By these ratios the plant will withdraw approximately 675 m3 per hour of water from the Tempisque River. This water will be approximately 18℃ when drawn from the river. After plant startup, studies will evaluate these recycle ratios, optimizing them to save utility costs. Calculations related to the water recycling are available in the attachment Utility Streams.
After any necessary cleaning, water planned for ejection into the Tempisque River will be held in a cooling pond for a period of a week. In accordance with the US Clean Water Act and Costa Rican environmental regulations, these streams will be too warm to be dumped into a river containing any wildlife. To remedy this, the plant will use a cooling pond, modeled off the one used by the Central Utility Plant at Northwestern University. Reject cooling water and steam at the university is emitted into a man-made lagoon allowing it circulate and cool prior to rejoining waters of Lake Michigan. Our plant will use a 150,000 m3 cooling pond, allowing the reject water to cool to ambient air temperature before returning it to the river. This will offset damage that water at the incorrect temperature could do to the river ecosystem . The total cost of the pond is estimated at just under a million dollars installed. Calculations related to the pond are available in the attachment Utility Streams.
The design team has developed the entire plant design thus far with safety as the primary concern. Therefore an effort has been made to reduce the use of carcinogenic and other health hazardous chemicals within the plant. The two exceptions are benzene used for the entrainment distillation and the refrigerant used for the CO2 purification process. The refrigerant was required for cooling to extreme temperatures to ensure CO2 separation from the remainder of the stream. We chose to use the refrigerant R-500, for its properties and common uses . For safety precautions, the refrigerant is in a self contained loop where it cools the process stream in a heat exchanger, then is re-cooled in a condenser. The closed loop system helps to prevent venting the refrigerant, which acts as a greenhouse gas.
Other safety features include adequate eye wash and safety shower locations throughout the operations area for emergency use. Emergency lighting will be supplied on a separate battery and circuitry from plant electricity, and all elevated areas will have appropriate railings and multiple escapes in the event of an emergency. All employees will have full safety training starting their first day of work onsite, and prior to any work they will complete an area specific safety tutorial highlighting the area’s specific hazards and safety operations.
On the design side, all pressure vessels were design to the ASME BPV code and are equipped with pressure relief devices. All valves and rotating equipment were designed with fail positions to prevent dangerous situations such as pressure build up in power or control loss 26 situations. Since our facility will be operating with extensive rolling and crushing machinery in the milling section, those machines will be designed with guards and two-handed required operating controls to keep appendages away from the pinch points. Additionally, throughout the indoor processes of the plant, mostly the milling and bagasse hydrolysis steps, the air handling system will be designed to adequately remove dust and other particulate matter from the air to prevent dust explosions. All electrical equipment will be rated for anti-spark areas and electrical rooms will be pressure positive to prevent dust from travelling into them. Later in the process when there is an extensive use of hot and cold streams, the design includes pipe insulation, even if it is not needed to maintain process temperature, but to protect operators from accidental scalds and burns.
Finally, Personal Protective Equipment (PPE) will be a must. All operators will be outfitted with hard hats, chemical resistant work gloves, chemical goggles, fire and chemical retardant jump suits and steel toed boots. Those in the milling area will be outfitted with ear protection as well, as the machines will be loud. Lastly, a smoke free site and a cell phone ban in vehicles will help prevent industrial accidents.
Several measures have already been incorporated to increase the reliability of the plant. First is a through storage system necessary to take the batch-wise fermentation and have it run continuously with the remainder of the plant. The storage system offers two days of production storage before and after the fermentation, and prior to the entrainment final dehydration step. This will allow for the separations processes to continue running for two days in the event of a fermentation issue. Additionally, in the fermenter sizing, the calculations allotted for each fermenter to be taken out of service and cleaned for 4 hours at the end of each fermentation run. This will increase product quality. Lastly, a large holding area was sized for incoming sugarcane with the capacity to store up to a week supply at a time. The design specifications of the storage vessels can be viewed in the attachment Storage Sizes.
Another reliability design feature is redundant pumps. Not only was each pump purchased in double or triple, most of the pumps throughout the plant share a make, type, size, and rating with several other pumps. This way, if one is broken and its reserve is broken, it may be able to use a “brother” pump’s reserve.
Lastly, as the project cost was being completed, seeing that we were on target for payback, we provided liberal corrosion allowances and materials selection on material to improve the lifetime of the equipment.
Figure 3 is an in-depth process controls diagram for the refrigeration loop, used in the CO2 processing equipment. Stream FD2-V is a vapor stream, containing CO2, water, ethanol, and other byproducts. For the final CO2 bottling stream, pure CO2 is required, with impurities only in the few parts-per-million. The stream is at 25℃ and 1.4 bar, with a flowrate of about 5300 kg per hour. To achieve the desired purity, contents of FD2-V will be chilled to -78℃, then sent through a flash separator, isolating the CO2. We are looking at the chilling process, FD2-VC (leaving the diagram), takes the stream to the flash separator.
To chill the stream, we are using Refrigerant-500, which is in the Refrigerant and Refrig Recycle streams at 100 wt% concentration. The R-500 is fed to the heat exchanger at -90℃, cools the FD2-V, and leaves the heat exchanger at 10℃. A compressor unit re-cools the refrigerant to -90℃. Due to fluctuations in previous steps, it is possible that FD2-V will not always be at the same flowrate and temperature. It is key to know the values of the incoming stream to ensure enough cooling without too much cooling. If FD2-VC is too warm, the separator will not work as well, and the CO2 will be filled with impurities. If FD2-VC is too cold, then CO2 will sublime, no separation will occur, and no CO2 will be captured.
The first set of control systems controls the flowrate of the refrigerant based on the inlet flowrate and temperature of FD2-V. A flow meter (FE1) and a temperature element (TE1) on FD2-V are read by a flow transmitter (FT1) and a temperature transmitter (TT1), respectively, sending a signal to a flow indicating controller (FIC1) and a computer, which using the information, calculates the correct flow rate needed of refrigerant for the heat exchanger. The computer and the FIC send a signal to a flow controlling automated screw valve (FV1) which controls the flow of the refrigerant. The screw valve is fail closed, so that in a situation where some power is lost, but the pump still has power, the valve will stop refrigerant flow, causing the pressure relief (explained below) on the pump to circulate fluid rather than send it through the condenser.
The refrigerant cycles in a loop, which is driven by a centrifugal pump. If the valve is sufficiently closed, or some other back up occurs, increasing the pressure in the pipe, a relief valve (V-5) will be triggered. This relief valve will signal an alarm, and circulate the refrigerant around the pump, not causing excess pressure on the valve or the compressor. The alarm will notify personnel to rectify the situation.
An additional instrument is on the refrigerant leaving the condenser. Since the computer calculating the flow of the refrigerant is assuming the refrigerant is within certain temperature constraints, there is a temperature element (TE3), a temperature transmitter (TT3) and a temperature indicating controller (TIC3) in the refrigerant line. This sensor is equipped with both high and low alarms to indicate the operators of a failed compressor.
If the refrigerant is returning colder than expected, the work of the compressor can be adjusted. This is controlled by temperature element (TE2), a temperature transmitter (TT2) and a temperature indicating controller (TIC2) which adjusts a variable speed drive on the drive shaft of the compressor. A cold refrigerant recycle will slow the drive as less compression is needed. A warm recycle will increase drive speed for more compression. Finally there is a check valve (V-2) on FD2-V ensuring it does not back feed the previous flash separator. There are also two check valves in the refrigeration loop; one to prevent back feed into the compressor (V-3) and one to prevent the compressor from back feeding the pump (V-4).