Applied Mathematics for Malting and Brewing Technologists ‘Technological Caleulations, Benchmarks and Correlations for Process Optimization Prof. Dr. sc. techn. Gerolf Annemdller Dr se. techn, Hans-J. Manger glo in gs) BERLIN Published by VLB BerlinDie Deutsche Blolothek lists this publication in the Deutsche Nationalbibliogratie, detailed bibliographic data is available on the Internet at dnd db.ce Contact to the authors: Prof. Dr. so. techn, Gerotf Annemuller Buschiner Str. 34.4 112683 Berlin Germany E-Mall: g.annemueller@t-online.de 4. English Edition 2017 ‘Translated by Christopher Bergtholat ISBN 976-9.921690-83.3, ©VLB Berlin, Seestrate 13, D-13353 Berlin, wrw.vib-berin org All rights reserved by the Versuchs- und’ Lehranstalt fir Brauerei in Berlin (VLE), Seestrasse 13, 19353 Berlin, Germany, ww.vib-berin.org Al rights reserved (Including those of transiation into ether languages), No part of th book may be reproduced in any form ~ by photocopy, scanning or any other means without writen permission from the publishers, Registered names, trademarks, eic. used in this book, even when not specifically ‘marked as such, are not to be considered unprotected by the law. Printing: LB Berlin, PR and Publishing Department Coverphoto: © Gina Sanders, fotolia.com 2 Contents Abbreviations and Symbols Overview of Calculation examples Preface 1. Some hints for rule of three, percentage, and interest calculations and for simple statistics 4.4 Notes on basic math operations 41.2 Rule of three calculations by cirect proportion to base and partial amounts 1.3 Rule of three calculations by Inverse proportion to base and partial amounts 1.4 Percentage calculations 15 Interest calculations 1.8 Mixing calculations and their expanded applications 1.8.1 Requirements and notes for the application of mixing calculations 4.6.2 Calculating with a mixing cross 1.6.3 Mixing calculations with a general equation 1.7 Applicaton of statistical methods for the evaluation of test results {a short overview for beginners) 4.7.1 Preliminary observations 4.7.2 Errortypes 1.7.3 Population 41.7.4 Statistical quality assurance 4.7.5 Sampling 41.7.6 Characterizing the numeric values of a measurement 1.77 Statistical testing methods, statistical reiablity P and probabilty of error a. 1.7.8 Test distributions 1.7.8 Degrees of freedom £ 1.7.10 Confidence interval a% of an average value 1.7.11 Outliers 47-12 Outlier tests 1.7.13 Empinal frequency distributions, 4.7.44 Comparison between two means withthe test 1.7.18 Two-dimensional (Inear, simple) regression and correlation analysis. 4.7.16 Multiple linear correlation and regression analysis Intainer geomet - calculations of areas, volumes and filing capacty in vwalteries and breweries | The calculation of areas for standard shapes 2.1.4 The rectangle and the square 2.4.2 The parallelogram " 12 23 St 58 58 582.19 Tha wapezois 214 Tretange 215 The cree 2:18 The annuus 217 The alpoo 2.2 Example caleulaton using surface area equations 2.3 Galeton fnteralvelumeV from ihe mein vessels that ar of importance for materies ana breweries, 231 The eubod 2.32 The cone and tuncated cone 2.33 The pyremd and tuneate pyramid 23.4 The spe and sporeal dame 23.8 The oynder 24 Same example caculatons for he malery and brewery empoyng volume equations 3. Storage of grains, grain care and grain transport 341 The ceeuation of potential storage losses in feshiy harvested grin 3.1.1 Technological signiicance and standard valise 3112 Balance equations ofthe nate conversion o tore barey 31.3 Galeulation of he loss of substance, oxygen demand Cr ond water formation dung storage of aey 3.14 Catelatng te warming experienced by sored barley 32 The pretminay storage of eshy harvested gain witout preservatives and lino! aeration 33 Aeration of grain cing the prestorage phase wth atmosphere ar and win cooked ay era Phase win tose 3.4 The gran drying 835.The aeration and colng of ran 35.1 Reduied amount ofa 3.52 Pressure loses nthe grain bi 36 imal tansporaton of graine 361 Bek conveyer 38.2€evater 38.3 Screw conveyors 384 Trough can conveyor 38. Tublr crag chan conveyors 3. Preumatic conveying 37 Converting batches of grains to basc motu content 38 Cleaning and sortng barey narge 38.1 Technological am one mallery 2.8 Storage space required fr grains, 4, Malt production 4.1 Calculation of steeping degree 4.2 Caleulation ofthe necessary steeping space 59 59 59 59 61 61 et 83 65 72 2 2 we 74 78 at 87 ar er 89 95 96 or 89 99 100 4.3 Water requirements during steeping 44 Temperature regulation, CO;-remaval and water consumption during steeping 45 Germination box capacity 44.6 Germination air consumption, cooling, humidity ofthe germination air ‘and energy requirements, design of kin ventiation fans and electric Power consumption, pressure drop caleulations in pipe channels of gases and vapors, 477 hxdlagrem in the maltery 47.4 General notes 4.1.2 Thermodynamic laws 4.7.3 The hxcdiagram for moist ait 4.74 Important changes in state 4.8 Assessment of malting processes and malt quality 4.8.1 Overall leat sprout length 4.8.2 Maltng yield, matting losses and malting factor 4.8 Sprouting rootlets malt germs 5. Milling of matt (grist) 5.1 Reference values for malt grist 5.2 Assessment of lauter tun arist 5.3 Welting of grains before miling or grinding 5.4 Grist volume 6. Calculations for brewing water and alkaline cleaners 6.1 Reference values and technological definitions of brewing water 6.2 Useful corversions for water chemistry 6.3 Notes on the analysis of water salts and their calculation 6.4 Calculation ofthe residual alkalinity (RA) of brewing water 68.5 Decartonization of brewing water with ime water 6.5.1 Chemical reactions 6.5.2 Determination of the concentration ofthe lime water 6.5.3 Concentration determination of dissolved CO; content of raw water 6.5.4 Calculation ofthe required amount of lime water forthe reduction ‘of the carbonate hardness in raw water 66.5.5 Required total quantty of ime water 6.6 Estimation ofthe mash pH value as a function ofthe malt quality and the residual alkalinity of brewing water 67 Reduction of the residual alkalinity of mashing water through the ‘addition of Ca ions 6.8 Determination of the cleaning effect of alkaline cleaning solutions 7. Wort production 7.1 Striking and brewhouse yield 7.4.1 Required amount of main strike as a function ofthe desired frst wort concentration 100 102 103 104 404 104 404 108 110 118 118 18 18 121 124 122 123 124 128 128 128 128 130 431 131 131 131 132 132 132 133 134 198 138 138 ‘7.4.2 Caleulation of total mash volume and required mash container volume 130 56 1611 ale ain of Me expe wae qanily forthe sparging 11 ion te want elle Hl worl pe brew 161 carat tes bnewhonea yl i Cabilaion sf lus projected aniount of hot knockout wort 1 Requitedt total evaporation, in elation to kettle fll wort £1181 simaahon of spent grain per brew 11 Rexquitet water quantily forthe production of wort 2 Bejietny the pl value in mash and wort 271 Lechnical definitions: 1 2.2 Guidelines for acidification with lactic acid £1 Calculation of the mash temperature steps (decoetion) 7-4 Mash tun heating 7.4.1 Heat transfer 7.4.2 Calculation of heat quantities and heating surfaces 7.4.3 The design of hest transfer surfaces on brewing veseels 7.4.4 Temperature increase by mixed condensation 7.5 Lautering the wort 7.8.1 Technological summary 7.8.2 Some guidelines forthe lautering process 7.8.3 Demonstrating the influence ofthe grain height and the influence of the material characteristics ofthe lauter wort onthe lautering speed 7.8.4 Influence of lautering technology when discharging the grains on the porosity of the grain cake inthe lauter tun 7.8.8 Calculation ofthe required mashing capacity of a mash fiter 7.56 Requited size of spent grain silos 7.5.7 Extract content of last runnings 7.6 Bolling of wort 7.6.1 Technological goals and important guidelines for wor boiling 7.6.2 The water evaporation during seasoning and the necessary eneray expenditure 7-7 Bitterness dosage and utiization 7.7.1 Orientation values for bitterness utlization (Yx.) In wort and for bitterness losses from the pitching yeast to finished beer resulting from the use of different technological procedures 7.7.2 Calculation ofthe required amount of hope and bitterness 7.7.3 Simplified calculation ofthe yield of bitterness in he brewery and Correction of the a-acid consumption per hectoliter of kettle full wort (cold) 1.7.4 Simplified calculation of the bittemess utiization Yay in relation tothe finished beer 7.8 Calculations to change the grist composition 7.8.1 Caleviation ofthe desired malt color for a malt mix 7.8 2 A simple method forthe conversion of extracts by mal surrogates for extract balancing by means of brewhouse yield 7.9 Extract yield and yield balance 79.1 Standard values for the evaluation of extract yields 158 160 161 181 162 162 165 165 166 49 170 170 170 a7 172 172 7.9.2 Caleulation ofthe classic brewhouse yield Ya 7.9.3 Assessment of extract extraction by the method “Overall Brewhouse Yield” (Your) 7.84 Necessary clarification ofthe addition and recevery of extract by the use of last runnings and trib 7.8.5 Exemple of an extract balance in connection with the corresponding ‘spent grain analysis 7.10 The coelng of the knockout wort to pitching temperature and variants to the utlzation ofthe liquid heat exchange 173 174 175 18 17 7.10.1 Gomparisen of wort cooling variants by means of model calculations 178 7.10.2 Results of a model calculation and conchisinns. 8. Fermentation and maturation of beer 8.1 Calculations for brewery yeast 8.1.1 Physical reference values for yeast cells and their influence on the effective metabolic area ofthe yeast 8.1.2 Thedensity of yeast cells and their sedimentation behavior 8.1.3 The yeast content of different yeast products and their influences on yeast growth 8.4.4 The size of yeast cells and their influence on the clarification behavior 8.1.5 Themutipication kinetics of yeast and thei influence on the interrelation of yeast propagation plants 8.1.6 Calculation ofthe required oxygen and air input for yeast multiplication In beer wort 2.2 Fermentation, degree of fermentation. original gravity, speed of fermentation 8.2.1 Metabolic cycles in the process of fermentation and original graviy of beer 8.22 Fermentation and degree of attenuation 8.2.3 The resulting amount of water from 1000 g of wort 8.24 Volume conversion of wort and beer 8.2 Assessment of a young beer during hosing 8.26 Alcohol and extract calculations according to Tabarié {8.3 The fermentable residual extract at the time of bunging, the maximum possible CO: formation, and the calculation of the required amourt of speise" (feed) for bottle fermentation 8.4 The spsedof fermentation 8.41 Average decrease of the epparent extract in the inital and main fermentation phase every 24 h 8.42 The average fermentation per unit of volume 8.4.3 Technological influence on the average fermentation 8.44 Specfic extract metabolism per yeast cell 8.4.5 Fermentation rate according to Schraderhaim 85 Calculation ofthe bunging pressure 8. Clarification and stabilization of beer 9.1 Objectives and process steps ‘92 Calculating the diatomaceous earth dosage 181 183 183 183 184 185 187 190 197 201 201 202 207 207 208 209 210 213 213 213, 214 216 216 217 219 219 218924 Pre-coating 8.2.2 Filtration time and running dosages 8.2.3 Ditlerential pressure increase an fitrtion time. 23 Fier aid preparation 914 Crossfiow membrane filtration (CMF) 9.5 Preparation using the protein stabilizer slica gel 10. 1 12, 18, ‘Thermal preservation of beer (pasteurization) 10.1 Aims, definitions and recommended values 102 Flash pasteurization 10.3 Bottle pasteurization in a tunnel pasteurizer 10.4 The D-value and z-value as determinad guideline values for killing special microorganisms |. Energy content of beer and alcohol breakdown in the human body 11.41 Energy equivalence of beer components 11.2 Beer consumption and blood alcohol content Filing 124 Gas aifusion 12.2 Storage capacity of a bottle butfering belt 412.3 Caustic caryover in a bottle cleaning machine (BCM) 12:4 Vapor suction in a bottle cleaning machine (BGM) 12.5 Forkitts 12.6 Acceptance of ing inctallations, guarantees 12.8.1 General information 12.6.2 Results of acceptance and determination of consumption values 426.3 Intemational acceptance and determination of consumption values 12.6 4 important terms for the assessment of filing systems 1265 Time concepis, 12:7 Compliance wath the nominal filing quantity 127.1 Terms 12.7.2 Filing quantity requirements for marking by mass or volume 42.7.3 Calculation notes 12.7.4 Consequences of underfling or overiling 12.8 The space requirement forthe storage of empty and full bottles 12.9 The space ancl room requirement for filing systems Sample calculations for preparation of aleohol-free soft drinks 13.4 Overview and basic requirements, 13.2 Batch calculation for a lemon lemonade 13.3 The stigar-acid ratio 194 Reduced caloilic value of alcohol-free soft inks. 135 The carhonization of alcohol-free soft drinks 19.5.1 CO, solubilly, guideline values and definitions 13.5.2 Calculations to adjust the CO, concentration n water and sodas 219 221 223 225: 227 27 227 209 2a 237 237 242 242 265 247 251 253, 255 255 255 257 250 260 260 261 262 285 285 266 268 289 2 273 275 275 27 4 Product pipelines in the brewery 111 Important aspects for the design of pipelines in the beverage industry 14.4.1 The flow rate 14.4.2 The pressure loss when a pipe or fiting is passed through 14.13 Pressure loss estimation by means of nomogram for liquids. 14.1.3 The Reynolds number 14.4.4 The boundary layer thickness 144.2 The flow rate during product conveyance 14.3 Instructions for the design of pipelines 14.3.1 General information 14.3.2 Thermally induced changes in length 14.33 Bleeding of pipelines, oxygen removal Pumps 15.1 Geodetic height 15.2 Efficiency of the drive motors 18,3 Cavitation 15.4 Power requirement of a centrifugal pump 15.5 Note on pump selection 155.1 Characteristics and ways of influencing therm 185.2 Starting conditions: Compressors 18.1 General information 16.2 Power supply for compressors. 16.3 Notes on the use of compressors 16.3.1 Possibilities for improving efficiency 16.3.2 Notes on compressors in the beverage industry 16.3.3 General information on compressors, Heat exchanger 17-4 Heat transfer 17.2 Heat transter coefficients 417.3 General information on the calculation for heat exchanger 417-4 Thermal dimensioning 47.5 Mean logarithmic temperature difference Indicators for plant planning 18.1 Raw material 18.2 Balance equations respiration and fermentation 18.3 Specific heat capacities 18.4 Specific brewery consumption values 185 Specific characteristics for a maltery 48.5.1 Consumption values 18.5.2 Specific load/eapacity in the maltery 18.5.3 Mating losses 302 302 305 307 310 att 313 317, 317 317 319 ata 320 321 at sat 32 322 324 327 227 227 328 328 328 32818.5.4 Energy consumption values 1855 Electricity 18.5.6 Water demandiwaste water 18.6 Specific consumption values bottle cleaning 18.7 Specific volumes for brewing vessels, characteristic values for brewhouses 18.8 CCV for fermentation, maturation and lageting 18.9 Fitor oysters or beor 18.10 Extract and volume contraction 18.11 Selected values for steam and water 16.12 Characteristics of selected packaging materials 19. Physical-technical units in the brewing and malting industry Index Bibliography and Sources 10 328 330 331 331 331 332 333 335 340 349 369 other measument systems see Chapter 19, aby A Ke OFS DM Abbreviations and symbols Hote: In all calculations the metric system is used. For the conversion of SI units into ‘alcoho! by volume area, surface ‘alcohol content in % rim or viv ‘area ofa sphere barley barley as green malt with bottle filing machine bottle cleaning machine EBC bitterness unit ‘specific heat capacity ‘yindroconical tank cycles per seconds yeast concentration ay diameter diatomaceous earth dosing fiter system ‘dry matter, dry mass, barley dry matter maltdry matter yeast dry matter yeast dry matter inerease removal 1 -onple 4.8 Germination area | onmpte 4.9: Mixture of two air volumes. | -onple 4.10 Mixing oft two moist alr quantities 1 comple 4 41: Humidification 1 spe 4.12: Humidifying and cooling air 1 coniple-4 Lb Dryer, non heated 2 74 74 5 8 76 79 at 83 84 87 88 89 93, 95 96 96 98 99 99 101 401 102 102 103 103 m1 112 113 113 115 15example 4.14: Dryer, heated Example 4.15; Malt drying Example 4.16: Determination of leaf sprouting development degree Example 4.17: Caloulation of the malting yield Example 4.18: Malting losses Example 4.19: Calculation of the total protein in malt rootlets Chapter 5: Milling of malt (grist) Example 5.1; Assessment of a dry grist Example 5.2: Assessment of a conditioned grist Example 5.3: Calculation of the water requirements for conditioning Example 5.4; Water balance in grist Example 5.8: Dimensions of a grist vessel Chapter 6: Calculations for brewing water and alkaline cleaners Example 6.1: Assessment of a prepared brewing water Example 6.2: Lime water addition for reduction of carbonate hardness Example 6.3: Influence of the residual alkalinity on the pH-value of the mash Example 8.4: Calculating the Ca ion addition Example 6.5: Testing and renewing a cleaning solution Chapter 7: Wort production Example 7.1: Calculation of the specific main striking quantity Example 7.2: Required gross volume of a mashing tun Example 7.3: Calculation of the required sparging volume Example 7.4: Calculating the kettle-full wort volume Example 7.5: Calculating the brewhouse yield ‘Example 7.6: Calculating the projected amount of hot knockout wort Example 7.7: Calculating the total evaporation Example 7.8: Calculating the spent grain production Example 7.9: Calculating the required water quantity for the production of wort Example 7.10: pH adjustment by sour wort in the mash Example 7.11: Calculation of the necessary heated mash volume 18 116 116 118 118 119 120 421 122 123 123 123, 124 126 130 132 133 133 138 138, 139 139 140 141 142 142 143 143 144 145 146 Example 7.12: Calculation of heat quantity required for heating Example 7.13: Calculating the required heating area Example 7.14; Calculation of heat transfer coefficient Example 7.18: Mash heating by mixed condensation 4 Example 7.16: Mash heating by mixed condensation 2 Example 7.17: Wort heating by mixed condensation. Example 7.18: Caleulation of the variables in tautering Example 7.19: Estimation of the influence of the change in the ‘grain height on the specific lautering speed Example 7.20: Estimation of the influence of the change in the material characteristics on the specific lautering rate xample 7.21; Distribution of grains depending on the lauter technology in the lauter tun Example 7.22: Calculation of the required number of fiter frames of ‘a mash filter Example 7.23: Calculating the size of a spent grain silo [Example 7.24: Extract content of last runnings -xample 7.25: Effect of the evaporation number on the eneray ‘consumption during wort boiling ‘xample 7.28: Calculation of the bittemess addition and bitterness balance Fcample 7.27: Bitterness dosages of a-acids per brew example 7.28: Calculations average hop boiling time -xample 7.29: Calculation of the simplified bitterness yield | xample 7.30: Calculation of the simplified yield of bitters ample 7.31: Calculation of the desired malt color | xample 7.32: Calculation of the grist contents when using malt surtogates | sample 7.33: Calculation of the classic brewhouse yield | cample 7.34: Calculation with Overall Brewhouse Yield | xample 7.35: Extract corrections by the trub and last runnings ‘management | sample 7.38: Yield balance | sample 7.37: Cooling the wort to the pitching temperature Chapter 8: Fermentation and maturation | ample 8.1: Calculation of the effective mass transfer area | sample 8.2: Density of the yeast 149 181 162 154 155 156 159 160 160 160 161 162 162 164 166 168 168 169 170 170 171 173 174 115 176 179 183 184 185 w‘example 8 3: Liquid yeast pitching per 1 hL of wort Example 8.4: Yeast pitching rate using a dry yeast Example 8.5: Sedimentation rates of diferent yeast cell sizes in beer Example 8.6: Influence of agglomerate size of a fractional yeast on its setting speed and clanfication time Example 8.7: Permissible clearance volume xample 8.8; Calculation of te requited process time for the propagation Example 8.9: Required container volume for a yeast propagation plant Example 8.10: Calculation of the required container volume Example 6.11: Calculation of the total oxygen and air requirement Example 8.12; Calculation of the required oxygen input in the start and end phase of the yeast propagation Example 8.13: Required aeration time in a yeast propagation Example 8.14: Original gravity calculation ofa finished beer Example 8.15: Degree of fermentation Example 8.16: Calculation ofthe alcohol content Ae Example 8.17; Calculation of the alcohol content and the actual residual extract content Example 8.18: Water in the final fermented beer Example 8.19: Wort and beer volume Example 8.20: Calculation of the fermentation cellar degrees Example 8.21: Calculations according to Tabarié Example 8.22: The necessary fermentable residual extract for CO; formation Example €.23: Timing of bunging Example 8.24: Calculation example forthe required of wort fora desired inctease of the CO; content during bottle fermentation Example 8.25: Celeulation ofthe average fermentation Example 8.26: Calculation ofthe impact of individual influencing variables Example 8.27: Caloulation of the required fermentation period Example 8.28: Fermentation rate according to Schréderhoim Example 8.29: Calculation ofthe bunging pressure Chapter 9: Clarification and stabilization Example 9.1; Calculation of the first pre-coating Example 9.2: Calculating the second pre-coaling Example 9.3: Calculating the free room in a filter and possible throughput 18 186 186 188 189 193. v94 194 196 198 199 200 202 205 208 208 207 208 208 209 210 an au 214 215 215, 217 218 219 219 220 221 Example 9.4: Differential pressure inorease, fitration time and fitrate volume Example 9.5: Comparison of the calculations with the recommended ‘values from literature Example 9.6: Calculating the filter aid requirements, Fxample 9.7: Dilution of the beer by the running fiter aid dosage =xample 9.8: Capacity calculations for a crossfiow membrane filter system xample 9.9: DE preparation used in combination with silica gel Shapter 10: Thermal preservation of beer xxample 10.1: Calculation of the required holding temperature xample 10.2: Celeulation ofthe PU at a constant holding time as a function of the hot holding temperature above 60 °C \ ample 10 5: Estimating the PU of a tunnel pasteurizer, variant 1 ‘cample 10.4 Estimating the PU of a tunnel pasteurizer, variant 2 ‘ample 10.5: Calculation of the required holding time for alcohol-free beer, taking into account wid yeast spores ‘example 10.6: Influence of the increase of the het holding temperature on the holding time forthe heterofermentative Lactobacillus strain G | sample 10.7: Required hot holding temperature and holding time for a ‘wort infected with the mesophilic bacterium Clostridium sporogenes Chapter 11: Energy content of beer and alcohol breakdown iple 11.1: Energy content of a full beer | xcimple 11.2: Simple approximation for the estimation of blood alcohol content | xample 11.3: Calculation of the time required for alcohol breakdown in blood | sample 11.4: Conversion of alcohol concentrations from % viv to g/t. | sample 11.5: Approximation calculation of alcoho! intake Chapter 42: Filling | cumple 12.4: How much oxygen diffuses into a PET bottle in 30 days ‘and what amount of CO; is released to the environment fat about 23 °C during the same time 221 222 222 223 228 225 227 228 228 230 238 235 236 237 237 239 239 240 241 242 242 9txample 12.2: What amount of oxygen wil fuse into a PET bottle in 60 days? Example 12.3: What amount of oxygen will diffuse through the sealing ‘of a crown cork into a bottle in 3 months? Example 12.4: How many bottles can be loaded into a buffer with an area As of 3 m?? Example 12.5: Calculation of the caustic concentrations in a bottle ‘leaning machine Example 12.6: Calculating the concentration in the spray zone Example 12.7: Cetermination of the accumulated dirt quantity, the discharged liquid volume, and the caustic concentration Example 12.8: Cow much heat is discharged from an BCM when the machine is set to 80 °C and the fan delivers 3500 m*/h of air at 0°C) Example 12.9: How much Hz must be removed? Example 12.10: Calculate the required forkift mass Example 12.11: Determination of the parameters according Example 12.12: Calculating line efficiency Example 12.13: Calculating the OEE Example 12.14: Check of the filing quantity Example 12.15: Check of the filing quantity Example 12.16: Stacking area requirement Example 12.17: Area requirement for a bottling plant Example 12.18: Required area of a filing plant for reusable glass bottles Chapter 13: Calculations for preparation alcohol-free soft drinks Example 13.1: Calculation of a beverage batch for a citrus lemonade Example 13.2: Calculation of the ratio of the beverage Example 13.3: Comparing the calorie content Example 13.4: Possible CO; concentration in water Example 13.5: Carbonating a citrus lemonade at 20 °C Example 13.6: Carbonating a citrus lemonade at 10 °C. Chapter 14: Product pipelines Example 14.1: Calculation of the flow velocity Example 14.2: Estimation of the pressure loss Example 14.3: Nomogram application Example 14.4: Calculating a Re number 20 245 245 248 248, 249, 250 251 252 254 256 258 258 264 264 265 287 267 268 269 273 274 277 27 278 278 280 282 284 287 Example 14.5: Average flow rate Example 14.8: Boundary ayer thickness Example 147: Flow velocity at the boundary layer surface Example 14.8: Expansion of a pipeline by temperature increase Example 148: Determination of compressive stress in case of resistance to expansion Example 14.10: Removal of gas from a pipatine Example 14.11: Removal of oxygen from a pipeline chapter 15: Pumps Example 15.1: Determination of the pressure difference during conveying Example 15.2: Determination of the drive power of a pump motor Example 15.3: Power requirement of a centrifugal pump xample 15.4; Filter pump hapter 16: Compressors xample 16.1: Design of a kiln fan Chapter 17: Heat exchanger -xample 17.4: Example of a heat recovery \-xample 17.2: Mean logarithmic temperature difference | xample 17.3: Wort cooler 202 293 293 297 297 298 290 302 303 303 308 313 37 319 321 323 328 328 at2 Preface The most wall known texts on applied mathematic for malisters and brewers [1.2 [3] ate over §0 years oli and no longer meet the requirements of the 21st century. A mademized textbook for brewers and masters revisad by R. Simon [¢] was released in 1986 and thus is thity years old. Furthermore in 2003 the Master Brewers Association of Americas released an interesting handbook for basic brewing calculations [s), but it sflzes only Anglo-American measurement units, ‘The purpose of this reference book is to provide an overview of technological ‘aleulations and guidelines from literature, which are supplemented by self-determined ‘ortelations and statistically reliable relationships. These are helpful for trainees, practitioners, and students to optimize process management in beer production, Furthermore, tables and graphs needed for technological calculations are included in » manner to enable rapid solutions without long searches. ‘The information required for assessing the results including reference values found in iierature are presented without much explanation of the technolegleal, biochemical, ‘mierobiological, and technical relationships. For understanding the requisite technclogy ‘udies referenced modern literature see [0], [7]. 8), [1 10} [11 12} {13} [14], 18). ‘n addition to the computational approaches, there are sample calculations with solutions to assist students and skiled workers gain a deeper understanding of the subject matter. From these computational approaches and sample calculations, easy psration-specitic tasks can be derives. Especially for small breweries that do nat have large analytical study capacty, impler technological approximate solutions are proposed. Finding the solution of these tasks require basic knowledge of the handing of a ‘aleulator wth integrated trigonometric, logarithmic, and simple statistical functions. ‘The use of the included equations for creating unversaly ussble eaculaten vcuments with the help of Excel” is strongly recommenced. Power supply for the brewing and malting industry (heat, cold, and electricity), vmpressed alr supply, and CO> recovery are not deait wth in this textbook, because of ‘ie extent of information already available in erature, for example, (16), [17], [18] and ot slin and Frankfurt (Oder), Germany, May 2017 Gerolf Annemailer Hans. Manger 23Chapter 4 1. Some hints for rule of three, percentage, and interest calculations and for simple statistics 411 Notes on basic math operations, Knowledge ofthe four basic arithmetic operations (addition, subtraction, multpicaion and division) and trations ate required of the apprentice brewer. Furthermore, the use of square roots and exponents along with the use of logarithms and trigonometic functions should be mastered with the help of an appropriate calculator. Recommendation: Physical quantities always consist of two parts, the measute and ‘he unit. Example: § iters of water = § L of Water. The number § being the measure and L. the unit. In calculations, i is best to always write down both the measure and the unit. By bringing the unts through the equation, the correct unit ofthe result is “automatically” ‘ound. In addition, seeing the units, may be a reminder of factors that must be used on individual observed values. 1.2 Rule of three calculations by direct proportion to base and partial amounts ‘The direct proportionalty of compared variables are encountered in the following uestions and cthers about for example, The larger, higher, heavier the measure the greater their share, power, scope, ‘olume, or mass. The desired value is tied proportionally to the specified information Example 4.1: Beer pump ‘/\ beer pump delivers at 100 rpm a beer volume of V = 200 L/h of beer. The speed of li pump is steadily increased to 120 rpm with the use of an electronic controller. I the [allow rate increases proportionally to the speed, what is the folow rate of beer? 4100 rpm yields 200 hUh, 120 spmyields how many LIN? Computational approach: 200 ALI 420 rpm =240 ALIN 100 rpm Example 4.2: Storage room for barley ley A has in storage a hL-mass of 68 kg and barley B has a hL-maes of 70 ka. How Innich storage rcom is needed by a maltery to store 120 t barley A and 150 t barley B? Vonversions: 1 hL= 0.1 m° ; 1 t= 1000 kg 25barley A: 6B kg of barley A uses 0.1 m’ of storage room, 120,000 kg of barley A requires how many m” of storage room V7 ‘Computational approach: 01m? -120000k9 | vy = DimE 12000069 975.5 mi? kg Barley B: 70 kg of barley B uses 0.1 m’ of storage room, 150,000 kg of barley A requires how many m of storage room V;? ‘Computational approach: (1m? -150,000 kg 70kg Va 2143. ‘The total amount of required storage space; 76.5 m+ 214.3 m= 390.8 1.3 Rule of three calculations by Inverse proportion to base and partial amounts. Inverse relationships are usually associated with issues that relate to the amount of time spent, the number of workers required, or the extract content after dilution. In Preparing to undertake this computational approach, a logical connection must be noted ‘Whether the requested number must be greater or less than the output reference! Example 1.3: Rule of three calculations Three empioyees need 8 h to clean @ classic fermentation celar department. How much time t do five employees need to lean the same department assuming they Work withthe same intensity? ‘3 employees need & h, 5 employees need how much time (tin hours)? Computational approach: = Semployess-8h Semployess “27 Example 1.4: Original gravity calculation 4000 kg of wort with an original gravity of OG = 12.0 % is diluted with 200 kg of water What original gravity OG does the diluted wort have? 1000 kg of wort has an original gravity of OG = 12 % and an extract content of E= 120kg 41200 kg of diluted wort with an extract content of & oc? 120 kg has what oxiginal gravity 26 Chapter 1 Computational approach: og = 1000k9:12% 10% 1200k9 1.4 Percentage calculations When perceniages are used, a number of 100" are referred to, for example, Tipereants 1 of 100, A per mile Is 1 oF 1000, Example 4.5: Malt calculations From 120 dt of cleaned barley with 2 water content of 14.0 % malt should be produced with a water content of 4 %. How many decitonnes of malt with 4 % water content are produced when the average malting losses (soft and germination losses) based on dry matter content ofthe barley and malt is 15 %2 Conversion of barley to dry matter 100 % of the barley used is 120 at, With 100 9% ~ 14 % = 86 % how much barley dry matter DMs there? Computational approach: 120 at 88% om, = oe 100% 103.241 Caloulation of malt dry matter DMy generated with a malting lass of 18 % from DMs = 103.2 dt 100 % corresponds to DMs = 103.2 dt, 100 % - 15.96 = 85 % is how much malt dry matter Dy? ‘Computational approact Dey = 103.201-85% 100% Converting the generated quantity of malt cry matter to air ied sales malt with a water content of Mc = 4.0 9 100 9%. 4.0 % = 98.0 9 mail dry matter corresponds to 87-72 dt, 100 % alr cried malt with 4.0 % of water Is how many decitonnes OM? ‘B72 at ‘Computational approa 87.72dt 100% Diy = 8272-100 % 96% ‘From 420 dt deaned barley (14 % H:0) approximately 91.4 at of sales malt with a waler content of 4 % are generated. ~aledt Note: |i should be kept in mind that the reference value of 100 % can be in either the numerator or denominator depending on the question at hand, ar1.9 Interest calculations Interest calculations function similarly to percentage calculations with an additional time factor included. Interest is the remuneration fora capital loan over a certain period. This time is divided in the following manner fer caleulation: ‘The following equation can be used to calculate the expected amount of interest: pert 100 Resulting amount with interest, ¢.9. in € Lent or borrowed capital e.g. in € Annual percentage rate of interest in percentia Time over which the capitals repaid in years, Equation 1 t ©: t Example 1.6: interest How much interest is obtained at an interest rate of 3.0 9i/a if §,000 € Is borrowed for 1.5 years? Computational approach according to the equation above, 5000 €:3%6-4.5 j= ROOENAGA 100% +a wes 1.6 Mixing calculations and their expanded applications 1.6.1 Requirements and notes for the application of mixing calculations In the fermentation and beverage industries, there are many tasks that can be quickly ‘and simply solved by a mixing calculation. However, there are some preconcitions that ‘must be observed in order to accurately ind the desired resut, 19 The factors involved must have the same base. This base usvally a spect physical property for example volume, mass, temperature, densly, quantiy of heat, specificheat capecity, mass flow, area etc. 2 Temperature dependent variables must always be based on the same reference temperature. ‘©. The factors involved in the “mix” may nat change. When working with mixing calculations in applications that do not require absolute accuracy, often difference inducing factors such as the temperalure dependence of volumes can be omitted. This is perfectly acceptable, but one should be mindful that in such calculations @ fundamental error Is Included which is greater the greater the temperature difference of the volumes involved Example 1.11 shows a model calculation with this error. ‘A simple method to avoid this error source is using temperature independent variables, for example mass, in the calculations. Temperature dependent variables Include volume, density and specific heat capacity, Because many physical parameters are temperature dependent, the comparability of analytical results is ensured by maintaining a fixed reference temperature In the laboratory. This temperature is generally 20 °C in Germany. 28 Chapter 4 ‘The above statement must be considered Ff, for example, the quantities involved react ‘chemically ‘Another example to be mindful of is contraction (Feduction in volume) of iquids during mixing, Water ethanol mixtures show a nonlinear behavior: 1 L water + 1 L ethanol do not necessarily yield 2 L of water-ethanol solution, Because contraction ‘occurs, less than 2 L result from this mivture It is better in this case to work with ‘masses instead: 1 kg water + 1 kg ethanol yield 2 kg of solution, Mixing calculaions can be made easier by employing 2 mixing cross when only 2 ‘components are involved, orn general with an equation (calculation formula). 4.8.2 Calculating with a mixing cross This simple and expedient mixing calculation is limited to calculations with two ‘components and can only solve for one unknown quantity. The parameter of the resulting mix must be between the parameters of the components involved. The dliference between component parameters cannot result a value of zero Generally, mixing crosses are writen (see Figure 1): \ Desited Parameter of he Mxtre (Component 1 with Parameter 1 Porton of Component 1 Component 2 with Paramater 2 Peston of Component 2 Patt Pant? Tela of Components 4 and 2 et {isthe difference between Parameter 2 and the Parameter of the Mixture | an the Parameter othe Moe ana Parameter Figure 1 General mixing cross layout Example 1.7: Water temperature adjustment by mixing There Is 100 kg of water with 2 temperature of 15 °C. This water must be heated through the addition of 30 °C water to 25 °C. How much 30 °C must be added? Solution by mixing cross: 100 ig at 15°C. Spars = 100g Be xkg at 30°C" sOpans= xk 100 kg - 1Oparts x= EAD = 2009 Spars 28Ine solution: 30°C-25°C= 5 pars 25°C. 15°C = 10 parts 200 kg of water with a temperature of 30 °C must be added to obtain @ mixture temperature of 25 °C. 1.6.3 Mixing calculations with a goneral equation The general mixing calculation entation ix Parameter ofthe misture x" sum ofthe amounts volved Au = Parameter 1 (x) ‘Amount 1 (Ar) Paramete! 2 (4) Amount 2 (As) +--+ Parameter i)“ AOU (A) alternatively, simply Hye gy a Aa to Ky A Equation 2 see Sik caunton 28 ‘The parameter x can be a temperature, a specific heat capacity, concentration, density, ‘center of gravity coordinate, a grade or any other parameter. ‘Any physical quality (e.g, mass, volume, heat load, specific heal, area) or other relationship (e.g. numberof students) “The correlation in Equation 2 is universally applicable. Example 1.8: Mixing calculations with several different volumes and temperatures ‘There are the several amounts of water at different temperatures: ‘Amount of Water_| Temperature m= 100 kg 40°C ims = 200g 25°C m= 50Kg aC m= 122g Src When mixed what isthe total amount of water and temperature? Solution: ‘According to Equation 2 and Equation 2a above, x is used forthe individual temperatures and A,for the amount of each subset Yur" Ay = 10°C - 100 kg +25 °C» 200 kg + 41°C - 50 kg +57 °C: 122 kg Xu“ Au = 1000 + 5000 + 2050 + 6954 = 15.004 °C - kg ‘Agy = 100 kg + 200 kg + 50 kg #122 kg xy = 15004 2472 = 31.8°C The total amount of water is 472 ka with 31.8 °C. Chapter 4 Example 1.9: Calculating the center of gravity ‘The center of graviy x forthe area indicated in Figure 2 must be found, Solution: The center of gravity x, is calculated by using Equation 2 or 2a above. The center of ravities ofthe subareas x,t0 x, can be seen in Figure 3. ‘The centers oi gravity ofthe areas are: ‘3 The rectangles inthe intersection of the diagonals, 5 The triangle at one third the height ofthe leg and © The center of the ciel. weese! Ton as Figure 3. Area trom Figure 2, Divided into 4 parts At, A2, A3, Ad The partial areas of At to Ad can be calculated as well. The lack of area in A is represented by a negative value. In the folowing table, the values are shot ‘Area of part Distance to canter Product A, Ai = 15.0 om" ay = 15 mm 225.00 mm: em’ AZ= 125 en" a= 46.7 mm ‘583.75 mm = AS=16.06m" x= 40 mm (640.00 mm = on™ Ad=-49 om" x= 30 mm 447.00 mm ext Total = 38.60m" Total = 1304.75 mm = em at‘The center of gravity distance xs is at 33.72. mm. The height coordinate ys ofthe center of gravity can be calculated in the same manner. Example 1.10: Specific heat capacity Calculation ofthe specific heat capacity of malt ‘The specific heat capacity c of malt at 20 °C should be calculated. The water content of the malt is 4 % and the specific heat capacity of malt dry matter is 1.56 kiikg'K). Solution: Ingenerat coms mys cy + ime © = Specific heat capacity of mait with a 43% moisture content m= Mass of mail with a 4 % moisture content = Specific heat capacity of mait dry matter = 1.55 kii(kg'K) c= Specific heat capacity of water = 4.187 kil{kg-K) m= Matt dry matter = 0.98 kg _m,= Water mass from the 4 % moisture content = 0.04 kgfkg Malt ‘Then follows: © m = 0.98 kg «1.55 kullkorK) + 0.04 kg + 4.187 kha’) +m = 1.488 kg'kil{kg'K) * 0.167 karkllkg:K) = 1.655 kaki kg") osm = 1,855 kg kilkg'k) 1.855 kg KAJIg'K) 1 kg = 1.855 klk: K) The specific heat capacity of malt with 4% moisture content is 1.655 kiM{kg-K) Example 4.11: Water mixture Comparison of water mixing calcuiations with and without consideration of the density: 100 L of water at 4 °C are mixed with water at 84 ‘C so that the resulting mixture reaches 64°C. For calculations inthis example: 100 L water = 100 kg water. Solution: sonny at a°c 30parts2 100g 7 kg at 94°C \ 60 par xkg 100 ka -6Dparts Opes ON v1 100 1 of water at 4 °C have a mass of 100 kg, 200 kg of water at 84 °C must be led to achieve B4 °C inthe resulting mixture, Chapter + ‘The specific heat capaciy fs assumed to be constant in the temperature range of °C fo 100 “C (but thal is not entirely accurate because this parameter depends on temperature, see Chapter 18.3), Ifthe temperature dependence ofthe density of water were taken into account, water ‘at 4 °C would have a density of 0.963 kg/L. and thus a volume of mm__ 20049: 88 for 200g of wate. p 0.963kg 7 Had example 1.11 been calculated simply with Iters instead of kilograms, the result would have been 7,7 L too low! The deviation is 3.85 %. This example illustrates that itis always more accurate to Work with temperature independent measurements for calculation, for example mass, Example 1.12: Grade point average of a class ‘What isthe grade point average of a class when the following scores are earned! mark 2 times mak2 5 times makS 9 times mark 3 times mark § 1 times. Number of students tested: 20 Solution: ‘Average grade + 20 ‘Average grade S18 -249+349-441°5 856 =58/20=28 ‘The average grade from this class is 2.8 Example 1.13: Temperature of milled grain Calculate what the temperature of the milled grain was at mashing in, when 5 t of grit |= mixed (mashed in) with 200 hL of mashing water at a temperature of 55 “C and the ‘emperature of the mash after mashing in is 60 °C (to 0.1 K). At 55 °C water at { bar has a density ofp = 985.71 kgf ‘The specic heat capacity of water 4.19 kul(ka‘K), the malt has a § % water content and a specifc heat capacity of 1.682 ki((kg'K), and the specific heat capacity of the resulting mash is 3.94 ki/(kg'K) (the values were calculated), Solution: 200 hL of water with 55 °C corresponds to 20m? and taking into account the density of rater at 55°C a mass of 19,714 kal Total mash = 5000 kg grist + 19,714 kg mash in water = 26,714 kg pectic heat quantily of the mash: ‘Gum = Specific heat capacity Mus = 3.94 KAMkarK) * 24,714 kg 7373 kK 33Specific heat quantity ofthe water: ‘ay = specific heat capacityy - my = 4.19 klg'K)- 19,714 kg 182,602 kJIK ‘Specific heat quantity of theorist: ‘der specific heat capacityer may = 1.682 kilkg'K) - 6000 kg ‘8410 kuIK ‘Ming caleiation using the specific heat ofthe parts above in kK ‘Bu "Qs = Boe" Gort thy“ Gv ar" don = us" Gas Diy" he ‘= (hats * a= Oy” Qu)! Qe dg. = SOO: STSTBKIK 55°C: 82602 KK _ ag 7 ‘BATOKIIK ‘The arist has a temperature of 38.7 °C at mashing in. 1.7 Application of statistical methods for the evaluation of test results {a short overview for beginners) 4.7.4 Preliminary observations ‘The breadth scientifc works, bath old and new, on the subject of mathematical statistics and their applications and analyses can be intimidating for anyone the when beginning to study the topic (see e.g, [20] and [21)). To help alleviate the anxioties that can arise in the usage and interpretation of collected data, we have included applications for Statistics that have proven useful, These examples and explanations are however incomplete. Essential parts of these applications were omitted due to the scope of this book, for example detailed error analysis of test series, calculations and analysis of statistical results from quality assurance and sensory analysis, the organization and evaluation of ring tests, as well as testing etrar values of international analysis methods, the creation and statistical values in experiment design, etc ‘The following are some useful terms, definitions and methods for statistics. 1.7.2 Error types Every analysis of results from technological experiments 5 well as the results from physical, chemical, and microbiological test of raw materials, intermediate samples, and finished products have error and therefore only represent an estimate or approximation of the true results, The difference seen in results from repeat technological tests and ‘multiple measurement of chemical, physical and microbiological tests prove this. With the help of stafisical methods and evaluations, the correctness of measured results can be estimated in diverse tasks and problems, and the results can be properly inlotpreted. The correctness of results in statistics is described by two terms; “accuracy cd “precision”. By using statistical parameters, these terms can be quantified. Error ‘und in analyses can be assigned to one ofthe below categories. Chapter 1 1.72.4 Random error Random err (see Figure 4) will appear: ‘3. Completely irregular, 5 Ave moetly small errors, 12 Alldeterminations are affected by it and 2 The results from mule measurements are scattered around a mean, Figure 4 Distribution of random errors; x * ‘general requirement: more practice! 1.7.2.2 Systematic error Systematic error (see Figure 5) are identified when: 2 The results are always in the same direction, 12 The results of muliple measurements systematically deviate from the tue value, and The analysis results are inaccurate. Possible sourees include: ‘3 The method itself, 3 Human error, from the measurement taker 2 An inaccuracy of the measurement device, {© Improper use of the measurement device, 19. Insufficient degree of purty from the reagents, standards, etc. api ty Figure 5. Possible distribution of systematic errors; They are preciso but not accurate! 1.7.2.3 Gross error Gross error (see Figure 8) occur from: 18 Serious mistakes, 12 Insufficient care from the analyst, 515 Faulty sampling (e.g. losses of bitlemess in foam), 1 Inappropriate storage of the sample (eg oss of aroma compounds), © Selection of inappropriate analytical method or sample preparation (eg. wrong dilution), 19 Inaccuracies in the implementation of the required method, 3 Calculation mistakes inthe evaluation of the analytical results, ete Wis the goal of aity sinayst to keep al eror, random, systematic, and gross, as small as Possible and under constant surveilance. Individual gross errors can be eliminated through the use of outlier tests when using multiple determinations Fine & Te roca cue | 41.7.3 Population The set ofall observed values N, thatthe variable x ean assume, builds a population of values with a mean Hand mean squared deviation or variance o” The population is for example the amount ofall products manufactured in individual packaging (e.g. kegs, bottles) that need to be checked for compliance with certain Quality characteristics and standards. Since the population is too large to be al individually tested (it is uneconomic, involving the destruction of ‘hygienic final Packaging), it is only realistic, from a time and cost standpoint, to measure a representative sample instead. 1.7.4 Statistical quality assurance ‘The quality ofa finished products affected by a plurality of factors that cause variations in quality. The measured actual values of quality characteristics of a certain product are therefore subject to fluctuations and should not exceed the defined safe production + limits (Confidence intervals for random sample mean values) from the set target value. These defined limits form a range that takes Info account both unavoidable production and technological related quality fluctuations, as well as the reproducibility of the applied sampling technique and analytical method. The review and evaluation of a product quality development is carried out with methods of statistical quality assurance Using different statistical models and evaluation tests, Chapter + 4.7.5 Sampling For time and cost reasons itis not possible to test the complete population of a product for compliance to qualty characteristics for example with 3 final product quality determination Even more so, the destruction and loss of product through testing may prohibit the examination of an entire population. Therefore, the testing of quality parameters can only be base samples from the population. The samples must be ‘selected in a manner that they represent the entire population as well as possible. This is better achieved as the scope and randomness of the sampling increases. In principle. the correct collected sample set will be subject fo the random differences trom the Population. Using mathematical methods of statistics, this randomness and the likelihood of their ‘occurrence can be described within the sample, whereby a statement about the quality characteristics ofthe population is made possible. DIN !SO 2859 stipulates the sample ‘size depending on the lot size (Le. population) and recommends acceptance and rejection thresholds based on the number of defective units for the respective sample size. A comprehensive presentation of al eight DIN and DIN ISO standards for different areas of sampling testing are summarized in DIN Standard Collection 225 [22] 41.7.8 Characterizing the numeric values of a measurement ‘The characterization of a sample with a size of n can be carried out by the numerical value mean % und variance s*, They are estimates for the corresponding quantities [ {and oF of the population 176.1The man 5 From a sam, the nial values x, x. (ogeer common efred to as x) tre calculated apd ten te alhmets tan io aloud accorng te be foi equation et hp tte Note: Average caleloons shoud ieude at est 2 india alse (the rests ot A teat ‘3 elvidal anatyet) For hgh cotenly analyaes, Gupte mesauemont Should etic. Equation 3 1.7.6.2 Accuracy and precision Measurements that appear to be precise are not necessary accurate (see Figure 5)! ‘The reproducibility of measurement results is only a statement about the precision of an {analytical method but not its accuracy. ‘The following statistical parameter are used to characterize precision: 2 Variance, 3 Standard deviation, 2 Coefficient of deviation, 2 Repealabily, 2 Comparability. ‘Accurate measurement results are obtained by 15 The use of appropriate measurement of analysis method, 7Proper, ie. without loss or etror, sample preparation and handling, “The use of pure chemicals, ‘The use of appropriate laboratory equipment and measuring devices, ‘The proper application ofthe analytical method, In order to assess the accuracy of @ measurement result, the analyst to technician must ot only work precisely and comply with the aforementioned conditions, but also be able to evaluate the result with regards to the overall technological context to avoid mistakes. 1.76.3 Variance s* The variance of a sample is characterized by the equation found below. Its the root mean square deviation of the measured values from the mean value of the sample. The variance s" in the sample 's generally used to estimate the variance a” of the complete population. Eoin a4 Equation 4 Note: The positive value of the root is called the standard deviation (ie. the meen deviation or scattering). While the variance is required for checking and comparing the frequency of distributions, e.g. when comparing the results of two series of ‘measurements, the standard deviation allows a better interpretation of a single result. A Standard deviation © ‘The standard deviation s is a statistical parameter, with which the distribution of the ‘measured values will be described with relation to the mean. This measure of varabilly characterizes the random error of the method, ‘The standard deviation of a sample is given by the following equation: 21x, Ext Ss iF individual measurement values Mean Number of individual measurements Equation § \Note: The standard deviation is described in amounts wih units that are identical to the Unit of the individual measured values. Ils numerical value after the decimal point should only be given with one point moze that given in the individual values, with the remaining decimals rounded up. Excessive mathematical accuracy’ in the representation of a result by as many decimal places as possible is a sign of ‘mathematical ignorance! 1.7.6.5 Coefficient of variation Ves ‘The coefficient of variation Vax is the standard deviation in percent based on the average value and is calculated using the folawing equation: 38 hapten + Van 2 100% evans Note: The coefficient of variation Vine allows for easier interpretation of random err. A reliable anaiyical method should resut in a ooeficient of variation of V,,..=~'5 % with Identical test material and muitiple determination! When evaluating technological parallel experiments with possible random enors in technological experimental procedure, sampling and chemical analysis of a vakie ‘characteristic, a coefficient of variation of Vine = = 10 % is a good resul 1.7.6.6 Range R ‘The range R joften referred to as variation) refers to the difference between the largest Xou and the smallest value Xan of a seties of measurements and is calculated as follows R= nae Kou Equation 7 Note: This statistical measure is particularly suitable for characterizing the dtribution cof measurements from a small sample set (7 $ 10). 17.8.7 Repeatability and reproducibility ‘According fo DIN ISO 5725 the following definitions apply: ‘The repeatability r is the absolute difference between two single test results obtsined (the same analyst, same apparatus, same laboratory, and over a shor period of time) with the same method on identical test materials and Under the same conditions, below which has a certainty probabilty are the same; unless otherwise stated, the probability is 95 % (1) The reproduobility Ris the the absolute difference between two test results that have been obtained on identical material with the same analytical method but under siferent conditions (diferent analysts, citferent laboratories, at different times) below which the Fesults are exact within a ceriain probabilty. Unless otherwise stated, the probability is 95 % (Ras) ‘The repestabilly and reproducibility are therefore two parameters which describe the precision of @ given test procedure when it is repeated under two different Circumstances. They are, for individual quality parameters of products, in nteraboratory tests determined (see j.a, (231). The results are reported in the standard analysis for collections of individual industries (see for the brewing industy the EBC Analytica and the MEBAK handbooks), ‘These parameters are used to evaluate a lab's own analysis resus. The standard deviation of the own analysis of a quality characteristic s- should be smaller when, properly work than ss (Se < tes). They are used for the monitoring of pemissibie fluctuations in indWvidual agreed quality parameters in supply contracts (e.g. complaint lof the goods when the average value found outside the range contracted: 2 Ra) 39172 Statistical testing methods, statistical reliability P and probability of error « Statistical testing methods are used for objective interpretation of analytical data with the help of test distrbutions (abulated tables given in the appendix of statistical textbooks). When a test method starts from the assumption or hypothesis that the ssamnple or samples of the considered population belong i.e, that the determined mean X and the standard deviation s of a sample of the population mean 11 and its standard deviation c isnot signficanty diferent, and the differences are oniy accidentally. ‘This assumption is called the null hypothesis H,. The other option is referred to as the alternative hypothesis Hs, Le. there is a significant diference between the two. However, it is not saying that this decision is correct in any case, because the result is, based on a sample of size n and there is the possibilty of error. itis estimated at this decision with an error probabilly a, that is generally selected with 05 (25%), a= 0.01 (2 1 %) ora=0.001 (2.0.1 %) From Equation 8 the probability of an outcome P is calculated (also equal to statistical certainty): Pp Pra 1-a respectively, 100- cf) Equation 8 A statistical certainty of P=0.85 (= 95 %) s often referred to as one-star significance (*), 99.9 %) as a three-star significance (*") 1.7.8 Test distributions All test result dstibutions are based on the calculated Gauss or Gaussian distribution, itis a distibution density function whose graphical representation isa bell curve on the y-axis represents the number of frequencies of n characteristic values (x-axis). Mean ‘and variance o* respectively standard deviation ¢ describe this distribution completely. By normalizing the distribution density function with y= 0 and o = 1 is a standardized Gaussian distribution, N (yu; o*) = N (0; 4), the so called normal distribution is reached (see Figure 7). “The curve maximum nue is located at x =, the two turning points can be found at x, + and =H +0. The curve reaches a value of y = 0 when x = + =. However the intercapt values sre already at x =p + 3 « negligible for practical purposes. “The ordinates for drawing a Gaussian curve are summarized in Table 1 below. Tablo 1 Ordinate Values of a Gaussian Curve x] uo UE26 use vy | Yas [TB Yam [TY 180 - Yr thw smaller « is, the narrower and steeper the Bell curve. he deanbutin fnclion of the Gaussian distribution ig called the Gaussian integral or itor intel Il intr Wy the Selected boundaries represent the content of the area _ Chapter 1 lunder the curve plece and gives the probabllty P and the remaining areas the probability of error of a null hypothesis, To test the null hypothesis test distributions are used. These are tables in which the results ofthe distribution functions are tabulated and thus facilitate the analysis, The following table shows the composition of the surface portions depending on the selected limits Tole 2 Toland resin ves ufo thn wrior ier Limits on the x-axis Partoftotalarea__| Residval area a ke pth % % uate bate 8.25 ara pa2e u2o 35.44 456 (ese bao 99.73 027 p-1986 +1066 Ed 3 u=2586 b>2580 3 7 u=3296 u+3290 Eg ot 1 “The last three lines are the known imits of significance tesis, Figure 7 Density function ofthe normal aistribution vith y and ‘on the x-axis and the normalized Irequency on the yraxis ° J (here Yous = 04) Chui Ss he FFor the normal distribution, the ‘distribution has been developed, it is for the ‘oxamination of samples used (flatter due to the larger standard deviation). Other distributions are often to be used are x"-clstibution (ie. the chi-square distribution) and the F- 100, where X ~ and s =o. 41©. talstribution: in sample surveys with relatively low findings, the case for most of pur quay analysis, 9 distribution: is necessary to verify square sizes, has no negative values Corresponds, in certain sections of the normal distribution in the positive x-axis 19 Fedistribution: are based on the 7*-Distibulion here but a two square sizes ae tested together, eg. the comparison of two sample variances or quotients ‘There are here two degrees of freedom required when the larger variance always determines the value f.. The calouated F value must always be >1,0! 41.7.8 Dogroes of freedom f ‘The number f of degrees of freedom (DOF) of a calculated for the sample measure is equal to the sample size (number of independent observations/measurements) minus the number k of from the sample linearly independent estimated parameters that enter Into the calculation of the sample measure: =k Equation 8 When averaging, one may stil be equipped with all the n values of the measurement series, but then a value is the average value determined and one can only through n Values include dates: fena4 Equation 10 When comparing two averages of two independent measurement series or a tWo- imensional regression and correlation analysis therefore result to be observed degrees of freedom from the folowing equation f=n-2 Equation 11 1.7.40 Confidence interval AZ of an average value For chemical analyzes, one can generally build a statement only on a very limited umber of results (parale! determinations). To calculate the confidence inlerval of a thus determined average value in place of the Gaussian distribution the general t- distribution for sampling is applicable. The confidence interval AX of an average value is calculated using the following equation, x sens Equation 12 ‘The results ofthis sample survey can then be stated in the following format: Result: x +x ‘The confidence interval indicates the limits within which with 2 confidence level of P (normally P = 0.98 or 95.0 %) of the true average 4 in comparison to x. There one ses the confidence interval AX also as an error indication forthe sample mean X P = Statistical certainty (normally P = 0.95) Degrees of Freedom, here f= n)— 4 lumber of measurement values per sample (normally: n= 2.4) 42 Chapter 4 fabulated talues e.g. from [26], in par for f= 1...10 (see Table 3) tandard deviation of the measurement Example 1.14: Standard deviation ‘The refractometric ethanol measurement of a full beer yielded the following values in Percent ethanol (% mm): 4.09; 4.12; 4.15 and 4.20. ‘Ine measured values are free from outers (tested according to Chapter 1.7.11. ™), ‘This insure that all values n = 4 can be inclided in the average calculation. First, the averaging of the ethanol content results is as follows: 4.094 6.124 6.15+4.20 4 44 9 mm ethanol Table 3 Excerpt fora the table for t-cstribution (according to (26) with the ‘valves as a function of P and t t P=099 | P= 0869 | Example 1 3.70 | 636.62 2 992 31 - 3 a4 42.92 _| Example 14 4 460 EG : 5 408 a7 6 a7 595__| Example 118 7, 3.50 541 e 3.36 5.04 a 325 478 io] 273 3ai7 459 Then calculation ofthe variance s” using the measured values from the sample results xP (08-414) | 0.0025 (i2=4.14" | 0.0004 (415-414 | 0.0001 (420-414 | 0.0036 = (0.0066 0022 |e calculation of the standard deviation for the sample can then be made with elther of (ne two parts of Equation 5: Eau = PR -0otse- 0.07 % mim tial 8Chapter 4 Grusing the right partof _] Equation 5 with te following e ‘sub-arithmetic calculations: __% 40u 157281 a2 ~[ 16.8744 L 415 17.2205 | f 420 17.6400 | E= 1656 B= 68.5650) “Thus the standard deviation is (0468 = 0.047 % mim ethanol Using the values of from the measurements in the example! nes % = 4.14 % mim ethanol 0.047 % mim ethanol With f= 4~1= 3 and P = 95 %, using Table 3 the t value tr ised to calculate the confidence interval AX 3.18-0.047 3:18: 0.097 9.0747 = 0,075 % nvm ethanol ve ‘Thus result ofthe ethanol measurement when P = 0.95 tA = (4.140.075) % mim ethanol ‘Thus, the true ethanol content of the sample with a confidence level of 95 % is within the limits of 4.06...4.21 % mim ethanol, a 95.0 %) Is Note: ‘2 The computed confidence interval is heavily dependent on the number of parallel ‘measurements. From Table 3, one can see thatthe transition from two to three ‘oF even four simultaneous measurements of the sample has a significant influence on the t-value and thus the calculation ofthe confidence interval and thereby the certainty of the measurement. Increasing the number of parallel determinations past four increases the working effort and only brings @ small improvement. Its recommended that more than four parallel measurements should only be performed in cases requiring high certainty > Ifthe standard deviation of the used analytical method has an estimated value s for the standard deviation from extensive previous tests, you can lise this estimate forthe caleulation ofthe confidence interval 1 ete since normally the number of samples n is significantly greater than the actual measurements forthe unknown sample, there is a higher degree of hhecdom fand thus a smaller value resulting in a smaller confidence interval Hus yiolds more certainty in the final value © Sino the use of an average is required, at least 3 individual values should be tused (ie. the results from three individual analyzes). For highly secure analyzes, duplicates may suffice. 1.7.11 Outliers Outliers ate values that differ so greatly from the other measured values that the fHeviations. ate no longer regarded as accidental and the cause of these variations Cannot be determined. To distinguish between results influenced by random variation ‘and true outlers, an outer test is useful as only values certainly identified as outliers ‘can be elimineted from a measurement series, 1.7.12 Outer tests When working with a repeated measurement, sometimes a reading oan deviate particulary greatly from the other values, To decide whether the value is just a ‘evietion oF major error, that is an outlier, one must test the outlying value of the measurement series with an outlier test. Ifit proves to be a real outlier, this value may be deleted and should be substituted by a newly measured value. There are a number of oul tests, especially for analytical chemistry with options to test even results from minimally repeated results (n 3). A quickly calculated option is the outier test according to Dean and Dixon [24], and another test is according to Janko [25] Furthermore, tne paraliel implementation of the outlier test according to Grubbs and according to Dixon is recommended for determination of repeatability and Feproductbiliy in ring tests of anaiytical methods for inralaboratory measurements by (23). 4.7.42. Outlier test according to Dean and Dixon For the calculation of the fest value Q, measured values are sorted by size and tested by the range R between Xn and Xn aleuated from Equation 7. Equation 13) Q = Test value to be calculated Suspected outlier ‘Adjacent value tothe suspected outlier ‘The calculated value @ is then compared with the table value Q (determined with the Use of P, n) fom Table 4 opposite. The tested value is an outlier ifthe calculated value Qs larger than the citcal value from the table: a> QP, ny! = Statistical certainty of the statement (95 or 99 %) 1 = Number of measured values from the tested replicates‘ave 4 Critica numerical values of @ as a function of P and n, for clinination of gross errors according fo Dean and Dixon [24] 0. P=085 P= 099 3 ost | 0988, 4 0.765 0.889, 5 0.842 0.760, é 0.560 0.898 ul 0.507 0.637 e 0.480 0580 Example 1.18: Outlier test 4 Five water content measurements for a malt 15 Found water contents, in order of size in percent: 3.86; 387; 3.0; 3.91; 3.99 (n= 5) 2 Control: Is the value 3.89 an outlier? 3. From Equation 7 and Equation 13 there isthe following 399-391 e 99-386 @ 9. Therefore: = 0.82 < Qoss 5. The tested value isnot an aber and therefore must be considered in subsequent evaluations together wih the oer 4 values! 1.7.12.2 Outlier test according to Grubbs To calculate the test stalisic Qo (according to (231) the mean X and the standard 7 and < 13, the following ratios are formed to calculate the test statistic Op: oy=2=%—cauten 17 ard ay =29204 Equation The caiculsted test statistics Qo are compared with the tabulated values of Gnossxy from Table 6. If the calculated test statistic Qo Is greater than the associated table vvalue, the measurement in question is o be treated as an outlier. Table § _Tes! statistics for the outllor tost according fo Dixon n, P=099 a oa70_|_ 0884 4 0.828 0.928 3 710 Oat 6 "0.828 ‘0.740 tor 0.569 0.680 z 0.608 O77 2 564 ora 10 0.530 0.635 if 0.502 (0.805 @ 0479 0.579 arexample 1.17: Outlior tost 3 ‘Again using the measured values from example 1.15 and using Equation 16 to test the largest value as an outier this test statistic is calculated by: 399-3. Q, = 252-391 - 0.62 399-386, ‘Since Qo = 0.82 « Qojooxs) = 0.710, once again ater this test, the suspected value must be considered in future calculations with this data set, Itis not an outer! 4.7.43 Empirical frequency distributions ‘The evaluation of an extensive number of values in a data set (n > 25), from a series of experiments with a single product and a quality characteristic, is faciitated by the Creation of empirical frequency distributions, Graphically this Is an llustration in the form of columns, in which one applies banded classes reading the measurement x on the abscissa and dependent on the frequency the column height fis calculated on the ordinate. With this graphical description you assign the individual values to k classes with a class range dl The number of classes k should according to Doerife! [28] equal the square of the number of measured values. ‘The number of classes ¢ should also be at least five, but should not exceed twenty. Number of classes: cavn Equation 19 Too few classes will not show distribution, while too many blur the borders. The upper Cass limit must always be less than the lower lass limit of the next class. The class ‘icth d can be calculated as a ratio withthe range R of the measured values and the selected number of classes ¢: Class width: au, Equation 20 © Note: By using this graphical representation of your own long-term results of quality characteristics of a product, a first overview of the type of distibulion for the measured value is obtained. With proper testing conditions and constant quality development, a symmetrical distibution with a unique frequency maximum is usually observed. Asymmetric distribution with @ shifled maximum, either to the right or left, points to defects or ‘changes in the testing conditions or the testing procedures (the division of the x-axis must also be checked and possibly corrected!) From the shape of the frequency distribution, qualitative statements about the ‘observed random errors can also be obtained. With the same class divisions, a broad distibution isan indication of @ large random error, while a narrow and sharp distribution is a signal of small random error. Systematic errors of a homogeneous series of measurements are not detectable in {his representation, For joint tials of several analysts or laboratories, inconsistent systematic errors in this frequency distribution give two or more maxima (his can also manifest itsef as a shoulder on the main peak). In most cases, the numerical values {rom a multimodal distribution are not suitable for further evaluation. The source of 48 Chapter 4 these systematic errors should be discovered and resolved. Then a new series of tests ‘can be repeated in good conditions! 1.7.14 Comparison between two means with the t-test From two independent measurements series, the two means are X, and X2 and are determined from n and ne measurements. It is important to test using taistrbution \whether the diferences between the two averages are only due to random error of the testing metnod and ineretore tne average of me same population wan the same mean or if, and X pare significantly different Before performing the procedure, the values must be analyzed using the F- distribution, to determine whether there isa significant diference between the standard deviations 's; and s2 of the two measurement series. When the F-test retums a significant diference, the two series of measurements cannot be compared. The F-test is always completed with the variances, ‘The calculated F-value must always be >1, thats the larger variances is always in the pumeretor: 4. Ftoat of vatiance: Caleulated Fovalue: Fo = $i Equation 21 ‘Comparing the value Faye with the tabulated F-value Fux (P, ff) with f, =n - 1 and =p 1 (see extract of the F-distibution table in Table 7): 9 When Foic> Fi (Pf, f) there isa significant difference between the standard deviations and a comparison of means in not possible. 5 When Faye < Fu (P, ff) the comparison of means are carried out as indicated in Table’? 2, Calculating the standard deviation of the difference between two mean values 84 84 1 wi 2 i Equation 22 3. Calculation of the test value for -listribution fea: tee BO fame Equabon 23 4. Comparison with tabulated t-value tas (See excerpt of tcistrbution values in Table 3): A significant difference between the [wo means is indicated when: tae? tin = be with femeng-2 Equation 24‘example 1.18: Comparison betwoen two means Smal scale brewing trials (cee [20], Aniage 5, Vatiante 61 + 63) wih the malt bil variants x; = 100 % malt and x: = 50 % malt + 50 % un-malted barley (UMB) resulted in the following average values for the apparent final attenuation (Fy } ofthe test works listed in the table below: [ve | ewted ie me pe De aoe mat ease Bar ecb gowmat sorsuMe|4| Tae] tae ee 1. Testing the variation (F-Test, Table 7) joo a= = 928 ‘The comparison of means is permitted! 2, Calculation of ss 1.3064 3. Caloulation of ta: (s24—746) -- 1.3061 ao! Chapter 1 “4, Comparing the calculate value with the tabulated twalue (see Table 3): For f =444-2=6 tose 8448 > tan = be -cese 5.98 ‘Acceptance: Wit a confidence factor of P = 99.9 % diferencesin the averages of eth Variants are highly significant (= **-significance)! 1.7.15 Two-dimensional (Iinoar, simple) regression and correlation analysis ‘Two-dimensional regression and correlation analysis (often refered to as a simple or linear regression and correlation analysis) investigates the dependence or context of two features xand y Correlation analysis determines how large the dependence of the dependent variable y is on the independent variable x by determining the correlation coefficient fy. The Correlation coefficient «, represents the cimensionless variation ranging from 1 (regative correlation, a descending line) and +1 (positive corelaton, an ascending ine) ‘and is calculated using the folowing equation: Se Equation 25, Table 8Crtical correlation coofficiont values rx (according to [26) depending on the degree of freedom and the chosen probabilly of error Table 7 Excerpt from the table for F-distributon values (according to [26) with F-values £ ee aha 195 % must be achieved. If, at higher color concentrations a asymptotic curvature ofthe calibration curve (= reduction of the coefficient of determination) develope, then the curve is suitatle for accurate work only in te linear region In reviewing data from contexts in which an asymptoticaly extending saturation (solution of gases in liquids) or consumption (oxygen consumption of microorganisms at high oxygen panial pressure) © observed, by taking the loganthm of the varabie of measurement series in @ semilogarithmic scale, the data can be transferred to a straight line, allowing easier linear regression and correlation analysis and the verification of their significance using tabulated critical correlation coefficients. The logarithm variable of the measurement series must of course be accounted for in the correlation analy When creating correlation and regression analyzes, only measurement series that are Jn proper logial connection {which requires the competence and expertise of the analyzer), may be compared to each other in ordar to avoid nonsense correlations. It Is important to conduct a proper examination of the provided for the analysis variables for a mesningiul relationship before performing correation and regression analysis, Furthermore, there isa risk of spurious correlations, in which the varlable x and y are both dependent on a third feature 2. Proper selection of variables forthe relationships to be tested alvays requires expertise. Example 1.19: Intorpretation of correlation and regression analysie Below is a simple example (cit. from [27] for a possible interpretation of correlation and Fegression analysis from their own operational tests are displayed. Figure @ shows the relationship between the variable x, the solid volume of yeast after centrifugation in ;poreent (for implementation see [27]}, and the variable y, the yeast concentration of the suspension in 10° celisimL, determined by cell counting by Thoma-chamber. For intorpretatio 3 Caleulaton results ofthe two dimensional linear) corelation and regression analysis © Number of samples n=33 Degrees of freedom 1 =33-2=31 Correlation coefficient (ace. Eq, 25) F =0,98705. Tabilated critical corelation coeticiont fay, osc) = 0.547 Coetcient of determination (B=r°- 100) B = 97,4 % Regression coefficient (acc. Eq. 29) b; = §2.6-10" celisiml for 1% viv Regiession constant (acc. Eq. 90) ay=-204.1-10" cellsimL, 5 Review of results: ~ Singe the caleulated correlation coefficient r= 0.987 >> ras thete exists are significant correlation between the two variables. Since the tabulated critical correlation coefficient determined with a ‘certainty of P = 99.9 % (= error probabilty o.= 0.1 %) and the specified degrees of freedom |s significantly smaller than the calculated correlation ‘coefficient, the relationship is highly significant (= "~-significance)Applied Mathematics ~ The change of y (yeast concentration in 10° celsimL.) is 97.4 % datermined ) by the change of x (solid volume fraction in % vl), ‘The calculated linear regression function ¥ = 52.6x ~ 204.1-10° celisimL, provided with a confidence level of P = 69.9 and a probabilly of error of ‘@= 0.1 %, yields a highly significant correlation between the two variables *-signifeance) and is good for estimating the concentration of yeast cells y by means of a quick determination ofthe yeast solid fraction by centrifugation. A detected increase in the solids volume fraction in the centituge of 1 % viv {corresponds to the application ofthe analtical procedure of inereasing the ‘concentration of yeast cells inthe thick yeast slurry of §2.6+10" calismL. 1» This calibration curve is possible for use with yeast concentration in the range of 250-10" cellsiml. (= >5 % viv solids by volume) to about 4-10" cells/mL. (= about 90 % solids by volume). 1.7.16 Multiple linear correlation and regression analysie Hf there is more than one influencing factor in the correlation and regression approach then one speaks of a multiple inear regression and correlation. ‘These regression approaches have had the folowing appearance forthe a ‘2 Dual linear regression: Y= a9 + Dixy + boxe Equation 31 9. Tiilinear regression: F = ay + bx + bees + Baty ‘Since the computational complexity is now significantly greater than in a simple correlation and regression analysis, their implementation is only useful with ‘computerized calculation programs and cannot be further elaborated herein Equation 32 etc. Note: ‘The factors used in an analysis from x; to x, must be independent from each ether. You ‘may not correlate the same variables with each other to avoid misjudgments. In clearly significant internal correlation of two factors of size from the calculation must be eliminated. It is therefore in uncertain factors useful before a linear correlation {0 create a matrix between al factors (x, «xs x1~ Xs; x2 = Xs; ett) fo ensure exclusion of internal correlations, ‘The influence of the different influence variables that determine the target is calculated using a partial correlation while keeping the other influence variables constant, allowing the corretation coefficient to be treated as i it were a simple inear correlation for significance testing (*- to **-signiicance is required!) and the result is reported as @ Partial coefficient of determination Boas, ‘Non-significant factors (especially important when more than two influence variables ‘are present in the computational approach) should be eliminated from the conelation and regression analysis successively and the calculation analysis repeated with the ‘reduced computational approach. 54 Chapter 4 | sco Jes2sor-20005 °° R= 0.9743 2° 8 | ms | 0 | 5 \| 1} | ee 1 | al? | 0» » * @ 9 @ w © o% w 7 Sedimentin vv ——_! igure € Determining the volumetric sold content and the associated yeast cell counts Of thin and think pasty yeast samples (1. [27)) Ihe significance of the regression function then increases with the total coefficient of letemination of muttiple corrstaton anaiysis, The determined mutiple coefficient of Jetermination Bm: can be tested for significance using the formula of Forster and Ron [28] and the F-cisinbution. The test variable forthe F-distibution Fy 8 caleuated using w following equation using the determined muliple coefficient of datermination Bru rai —m—1) ™ (~Brn) lumber ef value pairs lumber of explanatory variables (x-values) Equation 33, ‘he caleulated value Fins Is compared with the associated tabulated values Fx from Ihe F-distbution. The associated tabulated value Fig = Fi.ow results from the F-table vs the following conditions for values (see Table 7) =m f= n-m-1 und «= Probabilly of error 5% When Fin > Fue the designated multple functions dependent of the usable o. value with + certainly of P= 95 % or P = 99 % Is significant. 55Example 1.20: Partial coofficient of determination A simple example (from [23}) will be presented for possible interpretation: With fn = 108 small-scale brewing trials with malt bils of 40 to 100 % malt (17 diferent malts) and the equivalent of © to 60 % unmalted barley (malt comparison value: 125 kg of unmalted barley = 100 kg of mait) were completed with a standara infusion mash process using a bacterial protease- and f-glucanase contain enzyme mixture (only for brews brewed with adjuncts) and the influenoe on the free a-amino ‘itogen of the wort (FAN) was examined. The following multiple regiession function ith ther correlations ware found from the tal 1.67 + 0.218 x; + 0.100 x2 0.086%; Bun = 88.4 % Byar X1 = 7.6 % "* (r= 0.881) %4= 45.4 % "(5 7, = 0.674) X= 15.2% * (5 r= 0,390) 7 = FAN content of the wort in mg/100 mL (calculated to 0% %1= Mail n the matt bill as a percentage ofthe total grain load {adjusted with the mait comparison factor) %2= Soluble nitrogen content of the malt from a congress wort mg/100 mL (adjusted to OG = 12% P) %5= Dosed bacterial protease quantiy in tyrosine units TUIkg of milled grain (absolute), (TU at pH 56; 50°C; casein) 12% P) Examination of the mutiple coefcient of determination with F-clstribution (see Table 7) ‘and using Equation 33: ‘Number of explanatory variable m=3 and therefore f, = 3. Number of value pairs n= 408 and thus f,= 108-3-1= 104 From Table 7 resuits for c= 1 % Fux Foot a=i9) "387 ‘The calculation ofthe test variable is @s follows: 0.884(108-3-1) Ft bapa)” BAe Result: Since Fis = 264.2 >> Fis = 2.97, the result of Brus = 88.4 % with P confirmed, ‘The examination of the patil coeficient of determination forthe three influencing Variables x), x2, and x; is performed using the tabulated critical correlation coetficonts (see 19% certainly is Table 6) and the calculated degrees of freedom 08-2 = 106 From Table &, the following critical correlation coefficients are calculated for the degrees of ‘freedom and the selected probability of ervor of P= 98.8 9: 314 06 na 56 Chapter 4 Since the calculated partial coefficients of determination for the three influence variables are caleulated as correlation coeticients and ry, tats > for = 0.314, te three Parla resuits are confirmed to be signfcant with a certanty of P = 69.9% (2 =, significance), Review: {actors are constant or even eliminated ine Pattialcoefcients of determination show thatthe main facto infuencing soluble ‘nitrogen was the malt portion of the malt bil and the second important factor wae the malt quailty, LA change in the addition ofthe bacterial protease info the mash only accounted for a Ho change in the FAN content, This means thatthe protease dosages has only a limited corectve effect Al (Slatoments are also highly significant because of the number of repettions performed. syApplied Mathematics 2. Container geometry - calculations of areas, volumes and capacity in malteries and breweries 2.4 The calculation of areas for standard shapes 2.4. The rectangle and the square For a rectangle, the edge lengths are labeled ax a and b. itis a square wei w= b. The area A of a rectangle is obtained as such: Azad Equation 34 5 The edge length may be expressed in any unit of length, - The area of a square is calculated as a follows: Aza Equation 35 Figure 9 Rectangle 2.4.2 The parallelogram | heel he parateogram {9= Baseline of the parallelogram ‘The area of a parallelogram is calculated 3 as follows: Azan Equation 36 Figure 10 Paraitelogram 2.1.8 The trapezoid G = Longer parallel side ‘Shorter parallel side Height ofthe trapezoid Center li +6)-05 Aetts h Equation 37 Figure 14 Trapezoid Chapter 2 2.4.4 The triangle Baseline ofthe triangle = Height ofthe triangle D “Tho area of a riangle is calculated as | follows: pe ana Equation 38 Figure 12 Tangle 24.6 The elrcle 2m = Radius of the circle rr 42S te cs: | nia 8 Seeman on Cande2n-r Equation 38 Tha sofa citi clad fai: pee EE extene ret Figure 13 Circe 2.4.6 The annulus Radius of the outer (larger) circle Radius of the inner (smaller) ciclo Diameter ofthe outer (larger) circle = Diameter of the inner (smaller) cicle: d= 2 r ‘The annulus i calculated as follows Ane (-}no( af") Ente a4} R D 3 Figure 14 Annulus 24.7 The ellipse D = Longer diameter ofthe elise J d = Shorter diameter ofthe ellpse Pe = Approximate perimeter ofthe ellipse Ded Equation 42 Figure 15 Etipso 59mppreu masonnaucs ‘The area of an elise is calculated as follows aa Db Euston ss 2.2 Example calculations using surface area equations. Example 2.1:Grain transport For tnls example, the characteristics for cereals found in Chapter 18 and in Table 13 and Table 14 are used. 10m 040m Figure 16 Grain bulk surface on a flat conveyor belt — (On a flat conveyor belt (1.00 m width) malting barley (maximum slope angle c= 30°) are transported so that 10 om remain free on both sides of the conveyor belt The bulk density of malting bariey is assumed to be p = 700 kgim’. The conveyor belt has a belt speed of v= 3 mis. How many kilograms of malting barley can be transported per hour? Fr calculation, the following equations can be used: fn =A v" 93600 s/h Equation 44 Mass flow in kg/h Buk cross-sectional area in m? Belt speed in mis = Bulk density in kg/m” Calculation ofthe cross-sectional area A as a nearly isosceles triangle with g ‘The height h can then be calculated using the angle and the tangent function: h tan 30° = ose h=0.5-0.8m-tan30°=0.5-0.8m-05774=023m ‘The bulk cross-sectional area A is thus: Aw 2:0_0.8m-023m ‘Therefore, there is an hourly flow rate of Ti = 0.082 mi + 3 mis + 700 kg/m’ - 3600 e/h = 695,520 kg/h ~ 695.5 Ub Example 2.2: Lauter tun bottom stress Calculating the potential lauter tun bottom stress for @lauter tun with an annular lauter ‘surface (see the Pegasus System from Steinecken). A lauter tun with an annular lauter surface has an inner dlameter of d, = 2m and an ‘external diameter of d, = 4.50 m, 60 Chapter2 {tis used with conditioned dry grist (100 % malt) and an average lauter tun load of m= 280 kg gristim"lauter area is calculated. How large is the malt grist oad per brew? ‘According to the equation forthe annular area, it follows that the effective lauter surface ores Ais Ae \ (= Za) 9.14 = (600-1) -3.14=1278n8 ‘The malt grist load Mg is then calculated as: Mn = 12.75 m 280 kg/m? = 3569.55 kg/brew ~ 3570 kolbrew 2.3 Calculation of internal volume V from the main vessels that are of Importance for malteries and breweries Depending on the dimensions that are used (e.g. In mm, cm, m) volume V can be indicated with a variely of units such as in mm, cm, dm, mt, of km. Inthe following ‘drawings, the cimensions needed for calculations is designated in each equation 23.4 The cuboid Figure 17 Horizontal cuboid (let) and vertical cubotd (right) For the calculation of the volume of either horizontal or vertical cubods, the following ‘equation is used: Vo =a-beh Equation 45 4, = Euge lengths of the base hh = Height ofthe cuboia 2.3.2 The cone and truncated cone Using the dimensions from the form shown in Figure 18, the folowing equations ean be used fo calculate the volume ofa cone Vey anda truncated cone or fustum Vie: zee, Equation 46 Veo 3 61led Mathematics Vio = Sn RP +r Ree?) Equation 47 When the radi rand R (or their corresponding diameters) do not differ significantly from teach other, the approximate volume of the frustum can also be calculated using the following equation: Vico Fn? +R?) Equation 48, Ia cone-shaped container is only partially filed to a determined level h, the associated outer radius 1, can be calculated using the intercept theorem using the following ‘equation, and then can be used with Vr.-equation to calculate the fil vlume: HERR Equation 49 k—R Figure 18 Volume of a cone Vx, and a trumoated cone or frustum Vico 2.3.3 The pyramid and truncated pyramid The following equations are used forthe calculation of the volume of a pyramid Ve and the truncated pyramid Vrp (for dimensions see Figure 19): 1 nat. Equation 50 yaph at Sn ie he SAF) ceuatnt By + Bp ba + fy By a -Ba) Equation 52 62 i Figure 19 Volume of a pyramid (lef) Ve and volume of a truncated pyramid Ve (right) .3.4 The sphere and spherical dome In order to calculate the volume of @ sphere Vs and a spherical dome Veo the folowing ‘oquations can be used (see also Figure 20) Vex Seer Equation 53 Vso. a bat +t?) Equation 64 Radus of the sphere Radus ofthe circle of intersection of the spherical dome, light of the spherical dome When the intersection of the spherical dome is not at the middle of the sphere, the sphere radius r will aways be greater than the radius of the Intersectional erce:r> a! ‘The surface area of a sphere Asis calculated using the following equation dene? Equation 55 Asp=x'd+ Equation 58a