140 likes | 153 Views
MBA 5002 Analytics For Managers Selecting the Most Cost Effective Additive to Increase Profit. Team: Tan Ker Kuan Kayo Nagazono Eric Koh Wee Koon. Date: 7 Nov 2003. Content. Introduction Building the Model Interpretation of Results Benefits Achieved. Executive Summary.
E N D
MBA 5002 Analytics For ManagersSelecting the Most Cost Effective Additive to Increase Profit Team: Tan Ker Kuan Kayo Nagazono Eric Koh Wee Koon Date: 7 Nov 2003
Content • Introduction • Building the Model • Interpretation of Results • Benefits Achieved
Executive Summary • Multiple Regression Model with Dummy Variable and Interaction Variables was used to determine the most cost effective additive and to effectively convince the management of the soundness of the recommendation. • As a result of implementing the recommendation, the company earns an additional profit of US$ 1.39 Mln a year.
REGENERATOR REACTOR Cat Out Cat In Fluidized Catalytic Cracking (FCC) – Top Profit Earner in the RefineryBut Also Most Complicated Unit to Optimise • Prices of Products • LPG, Propylene, Gasoline, Diesel & Fuel Oil • Base Catalyst and Additives • Selection • Addition rate • Feed Intake & Quality • Density • CCR • Sulphur • Basic Nitrogen • Presence of Metals such Nickel, Vanadium, Sodium etc • Operating Conditions • Reactor Temperature • Reactor Pressure • Feed Intake • Circulating Catalyst to Feed Ratio • Conversion
Case to Conduct Study • Demand for Propylene in Asia Pacific Region is Increasing • Leading to Higher Propylene Price • It becomes more profitable to produce propylene over other products 2. Feed quality, operating conditions and catalyst addition rate was optimized to produce more propylene but to push propylene production to next level, a special additive called ZSM5 is needed. • 3. Several vendors offered their additives. Two were shortlisted for trial – Alpha & Beta
Preliminary results = Use of additive appeared to produce MORE Propylene; Significant Shift in Propylene Yield Pattern Seen • Without Additive • Mean = 4.1 • Standard Deviation = 0.25 • With Additive • Mean = 4.7 • Standard Deviation = 0.30
Building The Model – Multiple regression with dummy variable • However, such conclusion would have drawn criticisms that it could be due to other factors (feed quality, op conditions) and not on the use of additive • The problem questions to solve were • Management needs to be fully convinced that the use of additive can indeed produce more propylene and it is worth the additional cost • There is a need to differentiate and establish which additive (Alpha & Beta) is better • A model should be developed to correlate the amount of additional propylene produced with additive concentration • to address production planning, customer orders, scheduling of ships etc
Building The Model – Multiple regression with dummy variable • Applied what was learned in class into practice - employed multiple regression with dummy variable techniques and interaction variable
Building The Model – Multiple regression with dummy variable • Set type of additives used as dummy variable – No additive (baseline), Alpha & Beta • Use interaction variables such as [Additive Concentration in Base Catalyst] * [Type of Additive Dummy Variable] • Run stepwise regression with STATPRO • Dependent response – Propylene Yield (%) • Variables • Feed Quality • Operating Conditions • Fresh Base Catalyst & Additive Rate • Interaction variables
Interpretation of Results – Model is useful with ~80% variation explained
Interpretation of Results – Beta additive is clearly more effective Propylene Yield (%) = 22.25 + 0.2 [additive concentration][alpha] + 0.293 [additive concentration][beta] + 0.038 [conversion%] – 19.132[density of feed] + 0.442[feed intake] + 0.012[Nickel in Feed] + 0.126[Basic N2 in Feed] + 0.012[Vanadium in Feed] – 0.013[Reactor Temp] Interpretation => Alpha additive can increase propylene yield by 0.2% (20 t/d) for every 1% concentration of alpha added in base catalyst Interpretation => Beta additive can increase propylene yield by 0.293% (29 t/d) for every 1% concentration of beta in base catalyst More EFFECTIVE than Alpha Marginal benefit of 1 additional ton of propylene produced = US$ 200/ton
Benefits Achieved • Beta more effective, however its supplier is aware of its advantage and charges more. • Made selection of optimal additive complicated. • Model can also be used to analyze this problem and propose a recommendation. • Given: • Feed Intake = 10,000 ton/day • Base propylene yield without additive = 4 % = 400 ton/day • Maximum propylene due to unit constraint = 6% = 600 ton/day • Marginal benefit of 1 additional ton of propylene produced = US$ 200/ton • Cost of alpha additive = $10/kg • Cost of beta additive = $20/kg • Base Catalyst addition rate = 10 ton/day
Additive Concentration needed = 10% or Quantity = 1 ton/day Cost of alpha = US$ 10,000/day Increase Profit = US$ 30,000/day Via Alpha Additive Via Beta Additive Additive Concentration needed = 6.9% or Quantity = 0.69 ton/day Cost of beta = US$ 13,800/day Increase Profit = US$ 26,200/day Benefits Achieved Gap Between Base and Max Propylene = 2% * 10,000 = 200 ton/day (Potential Profit = 200*200 = US$ 40,000 day) TO EARN IT Therefore, Alpha additive recommended to and accepted management as most cost effective additive. Additional profit of using alpha over beta = US$ 1.39 Mil a year