This edition first published 2020
© 2020 John Wiley & Sons Ltd
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.
The right of Sandip Kumar Lahiri to be identified as the author of this work has been asserted in accordance with law.
Registered Offices
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial Office
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats.
Limit of Liability/Disclaimer of Warranty
In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.
Library of Congress Cataloging‐in‐Publication Data
Names: Lahiri, Sandip Kumar, 1970‐ author.
Title: Profit maximization techniques for operating chemical plants / Dr
Sandip Kumar Lahiri.
Description: First edition. | Hoboken, NJ : John Wiley & Sons, Inc., 2020.
| Includes bibliographical references and index.
Identifiers: LCCN 2019058766 (print) | LCCN 2019058767 (ebook) | ISBN
9781119532156 (hardback) | ISBN 9781119532217 (adobe pdf) | ISBN
9781119532170 (epub)
Subjects: LCSH: Chemical engineering--Cost effectiveness. | Engineering
economy. | Profit.
Classification: LCC TP155.2.C67 L34 2020 (print) | LCC TP155.2.C67
(ebook) | DDC 660.068/1--dc23
LC record available at https://lccn.loc.gov/2019058766
LC ebook record available at https://lccn.loc.gov/2019058767
Cover Design: Wiley
Cover Images: © David Burton/Getty Images, © Lightspring/Shutterstock
Dedicated to my Parents, wife Jinia and two lovely children Suchetona and Srijon
Figure 1.1 | Various constraints or limits of chemical processes |
Figure 1.2 | Optimum operating point versus operator comfort zone |
Figure 2.1 | Developing stages of the chemical industry |
Figure 2.2 | Three major ways digital transformation will impact the chemical industry |
Figure 2.3 | Three major impact areas where advance analytic tools will help to increase profit |
Figure 2.4 | Different components of the insights value chain |
Figure 2.5 | Overview of the insights value chain upstream processes (A–B) and downstream activities (D–E) |
Figure 2.6 | Data science is an iterative process that leverages both human domain expertise and advanced AI‐based machine learning techniques |
Figure 3.1 | Different steps in profit maximization project (PMP) implementation |
Figure 4.1 | Different ways to maximize the operating profit of chemical plants |
Figure 4.2 | Schematic diagram of a glycol plant |
Figure 4.3 | Steps to map the whole plant in monetary terms and to gain insights |
Figure 4.4 | Representing the whole plant as a black box |
Figure 4.5 | Mapping the whole plant in monetary terms |
Figure 4.6 | Break‐up of the total cost of production |
Figure 4.7 | Cost of raw material |
Figure 4.8 | Cost of different utilities (USD/h) |
Figure 4.9 | Cost of different chemicals (USD/h) |
Figure 4.10 | Variations of profit margin (USD/h) throughout the year |
Figure 4.11 | Variations of profit margin (USD/MT of product) throughout the year |
Figure 4.12 | Variations of production cost (USD/MT) throughout the year |
Figure 4.13 | Variations of MEG production (MT/h) throughout the year |
Figure 5.1 | Five‐step process of a key parameter identification |
Figure 5.2 | Queries normally asked to perform a process analysis and economic analysis of a whole plant |
Figure 5.3 | Major six categories of limitations in a plant to increase profit |
Figure 5.4 | Some examples of process limitations |
Figure 5.5 | Some examples of equipment limitations |
Figure 5.6 | Examples of instrument limitations |
Figure 5.7 | Guideline questionnaires to initiate the discussion with plant people |
Figure 5.8 | Various causes of catalyst selectivity increase |
Figure 6.1 | Comparison of daily actual profit (sorted) versus best achieved profit in US$/h terms for one year of operation |
Figure 6.2 | Daily opportunity loss in million US$ for one year of operation |
Figure 6.3 | Cumulative opportunity loss in million US$ for one year of operation |
Figure 7.1 | Advantage and disadvantage of the first principle‐based model |
Figure 7.2 | Advantages and disadvantages of data‐driven models |
Figure 7.3 | Advantages and disadvantages of the grey modeling technique |
Figure 7.4 | Advantages and disadvantages of the hybrid modeling technique |
Figure 7.5 | Typical pseudo code of a back‐propagation algorithm |
Figure 7.6 | Architecture of a feed‐forward network with one hidden layer |
Figure 7.7 | Steps followed in data collection and data inspection |
Figure 7.8 | Task performed in the data pre‐processing and data conditioning step |
Figure 7.9 | Two main univariate approaches to detect outliers |
Figure 7.10 | Guidelines for selection of the relevant input output variables |
Figure 7.11 | Relation between catalyst selectivity and promoter concentration in a commercial ethylene oxide reactor for the latest generation high selectivity catalyst |
Figure 7.12 | Actual selectivity versus ANN model predicted selectivity |
Figure 7.13 | Prediction error percent between actual selectivity and predicted selectivity |
Figure 7.14 | Plot of actual selectivity versus predicted selectivity for testing and training data |
Figure 7.15 | ANN model performance for testing and training data |
Figure 7.16 | Different ANN algorithms developed by different scientists in the last 30 years |
Figure 7.17 | Different activation functions used in an ANN |
Figure 8.1 | Different minimum values of a function depending on different starting points |
Figure 8.2 | Principle features possessed by a genetic algorithm |
Figure 8.3 | Foundation of the genetic algorithm |
Figure 8.4 | Five main phases of a genetic algorithm |
Figure 8.5 | Mechanism of crossover |
Figure 8.6 | Calculations steps performed in DE |
Figure 8.7 | Schematic diagram of DE |
Figure 8.8 | Calculation sequence of a simulated annealing algorithm |
Figure 9.1 | Cause and effect relationship of a steam increase in the distillation column |
Figure 9.2 | KPI‐based process monitoring |
Figure 9.3 | Projection of a three‐dimensional object on a two‐dimensional plane |
Figure 9.4 | Projection of a three‐dimensional object on a two‐dimensional principal component plane |
Figure 9.5 | Projection of data towards a maximum variance plane |
Figure 9.6 | Steps to calculating the principal components |
Figure 9.7 | Normal and abnormal operating zones are clearly different when plotted on the first three principal component planes |
Figure 9.8 | Trends of the first principal component |
Figure 9.9 | Variance explained by the first few principal components |
Figure 9.10 | Front end to detect abnormality in the reciprocating compressor |
Figure 9.11 | Normal and abnormal data projected onto the first two and first three principal component planes |
Figure 10.1 | New business challenges versus improve performance |
Figure 10.2 | Pyramid of a process monitoring system |
Figure 10.3 | Fault diagnosis system |
Figure 10.4 | Characteristics of an automated real–time process monitoring system |
Figure 10.5 | Concerns when building an effective fault diagnosis system |
Figure 10.6 | Different requirements of different stakeholders from fault diagnosis software |
Figure 10.7 | Summary of user perspective and challenges to build an effective fault diagnosis software |
Figure 10.8 | Principal component plot |
Figure 10.9 | Schematic of an ethylene oxide reactor and its associated unit |
Figure 10.10 | EO reactor process parameters along with a schematic |
Figure 10.11 | Various challenges to develop an EO reactor fault diagnosis |
Figure 10.12 | Chloride versus catalyst selectivity plot |
Figure 10.13 | PCA scores plot, T2 plot, and residual plot |
Figure 10.14 | Interface between a data historian and a dedicated PC loaded with PCA and ANN software |
Figure 10.15 | Contribution plots of 15 variables |
Figure 10.16 | Dynamic movement of the reactor status from the normal zone to the overchloride zone |
Figure 10.17 | Steps to build a PCA‐based fault diagnosis system |
Figure 10.18 | Actual versus ANN model predicted selectivity and equivalent ethylene oxide (EOE) |
Figure 10.19 | Integrated robust fault diagnosis system |
Figure 11.1 | Effect of tower loading on the tray efficiency valve versus sieve tray |
Figure 11.2 | Capacity diagram or feasible operating window diagram |
Figure 11.3 | Vapor liquid flow pattern on the tray |
Figure 11.4 | Froth regime versus spray regime operation |
Figure 11.5 | Jet flooding and its impact on entrainment and tray efficiency |
Figure 11.6 | Downcomer choking |
Figure 11.7 | Vapor recycle increases the vapor load |
Figure 11.8 | Downcomer filling |
Figure 11.9 | Effect of weeping on efficiency |
Figure 11.10 | Operational guide for deriving the operating window |
Figure 11.11 | Capacity diagram of the case study |
Figure 12.1 | Operating limits of a distillation column tray |
Figure 12.2 | Various constraints need to be satisfied during a distillation column design |
Figure 12.3 | Various downcomer‐related constraints need to be satisfied during distillation column design |
Figure 12.4 | Various process constraints need to be satisfied during distillation column design |
Figure 13.1 | Some chemical engineering applications of genetic programming |
Figure 13.2 | Five major preparatory steps for the basic version of genetic programming that the human user is required to specify |
Figure 13.3 | Flow chart of genetic programming |
Figure 13.4 | A typical individual that returns 5(x + 7) |
Figure 13.5 | Two‐offspring crossover genetic operation |
Figure 13.6 | Example of sub‐tree mutation |
Figure 13.7 | Initial population of four randomly created individuals of generation 0 |
Figure 13.8 | Fitness of the evolved functions from generation 0 |
Figure 13.9 | Population of generation 1 (after one reproduction, one mutation, and one two‐offspring crossover operations) |
Figure 14.1 | Different ways to increase plant throughput |
Figure 14.2 | Schematic diagram of strategy 2 of the maximum capacity test run |
Figure 14.3 | Schematic diagram of strategy 3 of the maximum capacity test run |
Figure 14.4 | Schematic diagram of strategy 4 of the maximum capacity test run |
Figure 15.1 | Different low grade heat recovery options |
Figure 16.1 | Flow scheme of a simple cracking furnace using an advance process controller |
Figure 16.2 | Hierarchy of the plant‐wide control framework |
Figure 16.3 | Features of potential plants for APC implementation |
Figure 16.4 | Capital investment versus benefits for different levels of controls |
Figure 16.5 | Typical benefits of APC |
Figure 16.6 | APC stabilization effect can increase plant capacity closer to its maximum limit |
Figure 16.7 | Reduced variability allows operation closer to constraints by shifting the set point |
Figure 16.8 | Operating zone limited by multiple constraints |
Figure 16.9 | Typical intangible benefits of APC |
Figure 16.10 | Typical payback period of APC |
Figure 16.11 | Typical benefits of APC implementation in CPI |
Figure 16.12 | Advance control implementations by one of the major APC vendors |
Figure 16.13 | Spread of APC application across the whole spectrum of the chemical process industries |
Figure 16.14 | Different steps in the APC implementation project |
Figure 16.15 | Steps in the functional design stage |
Table 2.1 | Comparisons between smart and conventional chemical industries |
Table 4.1 | Representing the whole plant as a black box with consumption and cost data |
Table 4.2 | Summary of profit margin and cost intensity |
Table 6.1 | Table to calculate production cost, cost intensity, profit, and profit/MT of product |
Table 6.2 | Table to relate production cost, cost intensity, with key parameters |
Table 6.3 | Plant reliability assessment |
Table 6.4 | Typical performance of control loops in industry |
Table 7.1 | Input and output variables for the ANN model |
Table 8.1 | Initial Population of x1 and x2 and Their Fitness |
Table 8.2 | Mutation and Crossover |
Table 8.3 | New Generation Populations |
Table 8.4 | Optimum Value of Input Variables Corresponding to the Maximum Value of Selectivity |
Table 9.1 | Performance Parameter for Major Process Equipment |
Table 10.1 | Input Parameters of a PCA‐based EO Reactor Model |
Table 10.2 | Input and output parameters of an ANN‐based EO reactor model |
Table 10.3 | Prediction performance of an ANN model |
Table 10.4 | Comparison of PCA and ANN input data |
Table 11.1 | Conditions of the Most Constrained Tray |
Table 11.2 | Tower and Plate Dimensions |
Table 12.1 | Simulation results |
Table 12.2 | Simulation results for different feed tray locations |
Table 12.3 | Optimization variables with their upper and lower limits |
Table 12.4 | Different constraints and their limits |
Table 12.5 | Optimal column geometry using improved PSACO methods |
Table 12.6 | Value of constraints corresponding to the optimum solution |
Table 13.1 | Examples of primitives used in GP functions and terminal sets |
Table 13.2 | Best model generated by the GP algorithm and corresponding RMS error |
Table 13.3 | Best model generated by the GP algorithm and the corresponding RMS error |
Table 15.1 | List of Process Coolers (Water Cooler and Fin Fan Air Cooler) along with Their Duty and Money Lost |
Table 15.2 | Calculation of Money Loss |
Table 15.3 | Table to Estimate the Money Lost from an Entire Plant Due to the Drain |
Table 15.4 | Table to Estimate the Money Lost from an Entire Plant Due to Vent and Flaring |
Table 16.1 | Typical benefits of APC implementation in refinery |