Profit Maximization Techniques for Operating Chemical Plants, I by Lehiri

Profit Maximization Techniques for Operating Chemical Plants

Sandip Kumar Lahiri

National Institute Of Technology, Durgapur, India

 

 

 

 

 

Wiley Logo

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

Dedicated to my Parents, wife Jinia and two lovely children Suchetona and Srijon

Figure List

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 List

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