Académique Documents
Professionnel Documents
Culture Documents
PIECE
Outline
1.1 Introduction and definition of Process integration. 1.2 Overview of Process Integration tools
NAMP
PIECE
NAMP
PIECE
Linnhoff started the area of pinch (bottleneck identification) at UMIST in the 60s, focusing on the area of Heat Integration UMIST Dept of Process Integration was created in 1984, shortly after the consulting firm Linnhoff-March Inc. was formed
NAMP
PIECE
"Systematic and General Methods for Designing Integrated Production Systems, ranging from Individual Processes to Total Sites, with special emphasis on the Efficient Use of Energy and reducing Environmental Effects"
NAMP
PIECE
Definition of process integration Later, this definition was somewhat broadened and more explicitly stated in the description of its role in the technical sector by this Implementing Agreement: "Process Integration is the common term used for the application of methodologies developed for System-oriented and Integrated approaches to industrial process plant design for both new and retrofit applications. Such methodologies can be mathematical, thermodynamic and economic models, methods and techniques. Examples of these methods include: Artificial Intelligence (AI), Hierarchical Analysis, Pinch Analysis and Mathematical Programming. Process Integration refers to Optimal Design; examples of aspects are: capital investment,energy efficiency, emissions, operability, flexibility, controllability, safety and yields. Process Integration also refers to some aspects of operation and maintenance". Later, based on input from the Swiss National Team, we have found that Sustainable Development should be included in our definition of Process Integration.
rson, International Energy Agency (IEA) Implementing Agreement, A worldwide catalogue on Proc 2001).
Module 8: introduction to process integration 5
NAMP
PIECE
El-Halwagi, M. M., Pollution Prevention through Process Integration: Systema . Academic Press, 1997.
A Chemical Process is an integrated system of interconnected units and streams, and it should be treated as such. Process Integration is a holistic approach to process design, retrofitting, and operation which emphasizes the unity of the process. In light of the strong interaction among process units, streams, and objectives, process integration offers a unique framework for fundamentally understanding the global insights of the process, methodically determining its attainable performance targets, and systematically making decisions leading to the realization of these targets. There are three key components in any comprehensive process integration methodology: synthesis, analysis, and optimization. Module 8: introduction to process integration
NAMP
PIECE
Nick Hallale, Aspentech CEP July 2001 Burning Bright Trends in Process Integration
Process Integration is more than just pinch technology and heat exchanger networks. Today, it has far wider scope and touches every area of process design. Switched-on industries are making more money from their raw materials and capital assets while becoming cleaner and more sustainable
NAMP
PIECE
So What Happened?
In addition to thermodynamics (the foundation of pinch), other techniques are being drawn upon for holistic analysis, in particular:
NAMP
PIECE
Process integration is primarily regarded as process design (both new and retrofits design), but also involve planning and operation. The methods and systems are applied to continuous, semi-batch, and batch process.
Module 8:
NAMP
PIECE
Possible Objectives: Lower capital cost design, for the same design objective Incremental production increase, from the same asset base Marginally-reduced unit production costs Better energy/environmental performance, without compromising competitive position Re d ucing COSTS POLLUTION EN ERGY
Module 8: introduction to process integration
NAMP
PIECE
Modern Process Integration context Among the design activities that these systems and methods address today are: Process Modeling and Simulation, and Validations of the results in order to have information accurate and reliable of the process. Minimize Total Annual Cost by optimal Trade-off between Energy, Equipment and Raw Material Within this trade-off: minimize Energy, improve Raw Material usage and minimize Capital Cost Increase Production Volume by Debottlenecking Reduce Operating Problems by correct (rather than maximum) use of Process Integration Increase Plant Controllability and Flexibility Minimize undesirable Emissions Add to the joint Efforts in the Process Industries and Society for a Sustainable Development.
11
NAMP
PIECE
Pr oce ss D a t a
Pr oce ss k now le d g e
M e t h od ology f or d e t e r m in in g cor r e ct m a r g in a l cost s. Pr oce d u r e s a n d t ools f or En e r g y, W a t e r, a n d raw m a t e r ia l Module 8: se r va t ion a u d it s Con introduction to process integration
12
NAMP
PIECE
In conclusion, process integration has evolved from Heat recovery methodology in the 80s to become what a number of leading industrial companies and research groups in the 20th century regarding the holistic analysis of processes, involving the following elements:
Process data lots of it Systems and tools typically computer-oriented Process engineering principles - in-depth process sector knowledge Targeting - Identification of ideal unit constraints for the overall process
Module 8: introduction to process integration 13
NAMP
PIECE
Outline
1..3 An around--the--world tour of PI practitioners 1 3 An around the world tour of PI practitioners focuses of expertise. focuses of expertise
14
NAMP
PIECE
15
NAMP
PIECE
Op t im iza t ion
St och a st ic Se a r ch M e t h od s
St eady st at e Dynam ic
e g r a t e Pr oce ss D e sig n a n d n t r ol
Module 8: introduction to process integration
Pr oce ss Dat a
16
NAMP
PIECE
Click here
St och a st ic Se a r ch M e t h od s
Click here
St eady st at e Dynam ic
Click here
e g r a t e Pr oce ss D e sig n a n d n t r ol
Module 8: introduction to process integration
Pr oce ss Dat a
17
NAMP
PIECE
Process Simulation
18
NAMP
PIECE
Process Simulation
Process modeling
W h a t is a m od e l? A model is an abstraction of a process operation used to build, change, improve, control, and answer questions about that process
Process modeling is an activity using models to solve problems in the areas of the process design, control, optimization, hazards analysis, operation training, risk assessment, and software engineering for computer aided engineering environments.
19
NAMP
PIECE
Process Simulation
Syst em Theory
Applicat io n
St at ist ic s
Process m odeling is an underst anding of t he process phenom ena an d t ransform ing t his underst anding int o a m odel.
20
NAMP
PIECE
Process Simulation
Nilsson (1995) presents a generalized model, which, as depicted in the figure below, can be used for different basic problem formulations: Simulation, Identification, estimation and design.
Input I MODEL
Output O
If the model is known, we have two uses for our model: Direct: Input is applied on the model, output is studied (Simulation) Inverse: Output is applied on the model, Input is studied
Module 8: introduction to process integration 21
NAMP
PIECE
Process Simulation
If both Input and Output are Known, we have three formulations (Juha Yaako, 1998):
Identification: We can find the structure and parameters in the model. Estimation: If the internal structure of model is known, we can find the internal states in model. Design: If the structure and internal states of model are known, we can study the parameters in model.
22
NAMP
PIECE
Process Simulation
NAMP
PIECE
Process Simulation The figure below shows a comparison of input and output for a process and its model. Note that always n > m and k > t.
Input
X1, ..., Xn
Input MODEL
X1, ..., Xm
Out put
Y1, ..., Yt
In the process industry we find, two levels of models; Plant models, and models of unit operations such as reactor, columns, pumps, heat exchangers, tanks, etc.
Module 8: introduction to process integration 24
NAMP
PIECE
Process Simulation
Types of models:
Intuitive: the immediate understanding of something without conscious reasoning or study. This are seldom used. Verbal: If an intuitive model can be expressed in words, it becomes a verbal model. First step of model development. Causal: as the name implies, these model are about the causal relations of the processes. Qualitative: These models are a step up in model sophistication from causal models. Quantitative: Mathematical models are an example of quantitative models. These models can be used for (nearly) every application in process engineering. The problem is that these models are not documented or can be too costly to construct when there is not enough knowledge (physical and chemical phenomena are poorly understood). Sometimes the application encountered does not require such model sophistication.
NAMP
PIECE
Process Simulation
Simulation involves performing a series of experiments with a process model. Input Out put Steady State MODEL Snapshot
X1, ..., Xm Y1, ..., Yt
Algebraic equations
Input
X(t)1, ..., X(t)m
MODEL (t)
Output
Dynamic
Movie (time functions) Time is an explicit variable differential equations Certain phenomena require dynamic simulation (e.g. control strategies, real time descition). 26
NAMP
PIECE
Process Simulation
Illustration:
Dynamic simulation of a storage tank
m1
t = time
Simulation unit
m2
m2(t)
0 = m1 m2 + 0 0
m2 m2 t Module 8: introduction to process integration
dM = m1 m2 ( t ) + 0 0 dt
t
27
NAMP
PIECE
Process Simulation
The steady-state simulation does not solve time-dependent equations. The Subroutines simulate the steady-state operation of the process units ( operation subroutines) and estimate the sizes and cost the process units ( cost subroutines). A simulation flowsheet, on the other hand, is a collection of simulation units(e.g., reactor, distillation columns, splitter, mixer, etc.), to represent computer programs (subroutines) to simulate the process units and areas to represent the flow of information among the simulation units represented by arrows.
28
NAMP
PIECE
Process Simulation
To convert from a process flowsheet to a simulation flowsheet, one replaces the process unit with simulation units (Models). For each simulation unit, one assigns a subroutine (or block) to solve its equations. Each of the simulators has a extensive list of subroutines to model and solve the equations for many process units.
The Dynamic simulation enables the process engineer to study the dynamic response of potential process design or the existent Process to typical disturbances and changes in operating conditions, as well as, strategies for the start up and shut down of the potential process design or existing process.
29
NAMP
PIECE
Process Simulation
Steady-State Simulation
Dynamic Simulation
Snapshot of a unit operation or plant Mimic of plant operation Balance at equilibrium condition Equilibrium results for all unit operations Equipment sizes in general not needed Amount of information required: small to medium Time dependent results It doesnt assume equilibrium conditions for all units Equipment sizes needed Amount of information required: medium to large
30
NAMP
PIECE
Process Simulation
Solution Strategies
The Sequent ial Modular St rat egy flowsheet broken int o unit operat ions (m odules) each m odule is calculat ed in sequence problem s wit h recycle loops The Sim ult aneous Modular St rat egy develops a linear m odel for each unit m odules wit h local recycle are solved sim ult aneously flowsheet m odules are solved sequent ially The Sim ult aneous Equat ion-solving St rat egy describe ent ire flow sheet w it h a set of equat ions all equat ions are sort ed and solved t oget her hard t o solve very large equat ions syst em s
31
NAMP
PIECE
Process Simulation
Why steady-state simulation is important: Better understanding of the process Consistent set of typical plant/facility data Objective comparative evaluation of options for Return On Investment (ROI) etc. Identification of bottlenecks, instabilities etc. Perform many experiments cheaply once the model is built Avoid implementing ineffective solutions
32
NAMP
PIECE
Process Simulation
Online system
OPTIMIZATION of plant operations ADVANCEMENT OF PLANT OPERATIONS/ OPERATIONAL SUPPORT / OPTIMIZATION Predictive simulation Optimal conditions EDUCATION, TRAINING CONTROL SYSTEM Operation training simulator DCS control logic Plant diagnosis system PROCESS DESIGN / ANALYSIS Examination of operations Control strategies Advanced control systems Batch scheduling
33
Quasi-online system
Off-line system
NAMP
PIECE
Challenges of simulation
34
NAMP
PIECE
Data Reconciliation
35
NAMP
PIECE
Data Reconciliation Typical Objectives of Data Treatment. Provide reliable information and knowledge of complete data for validation of process simulation and analysis Yield monitoring and accounting Plant facilities management and decision-making Optimization and control Perform instrument maintenance Instrument monitoring Malfunction detection calibration Detect operating problems Process leaks or product loss Estimate unmeasured values Reduce random and gross errors in measurements Detect steady states
Module 8: introduction to process integration 36
NAMP
PIECE
Data Reconciliation
37
NAMP
PIECE
Data Reconciliation
Overview
Plant shutdown Equipment performance Modeling and Simulation Optimization Instrumentation design
Instrument maintenance
38
NAMP
PIECE
Data Reconciliation
39
NAMP
PIECE
Data Reconciliation
Random errors
Features High frequency Unrepeatable: neither magnitude nor sign can be predicted with certitude Sources Power supply fluctuation Signal conversion noise Changes in ambient condition
40
NAMP
PIECE
Data Reconciliation
Process related
Process leaks Solid deposits
41
NAMP
PIECE
Data Reconciliation
bn or ma lity
na
Gr oss err or
Reliable value
t
Module 8: introduction to process integration 42
NAMP
PIECE
Data Reconciliation
Solutions To Problems
Random errors: Data processing Based on successive measurement of each individual variable: Temporal redundancy Traditional filtering techniques Wavelet Transform techniques Inconsistency: Data reconciliation Based on plant structure: Spatial redundancy Subject to conservation laws Unmeasured data Data reconciliation
43
NAMP
PIECE
F
Pro ces sin g ran do m err ors
Module 8: introduction to process integration
t
44
R ec on cil in g Gr os s er ro rs
NAMP
PIECE
Data Reconciliation
1. Establish Plant facilities operating regimes 2. Data processing Remove random noise Detect and correct abnormalities 3. Steady state detection Identify steady-state duration Select data set 4. Data reconciliation Detect gross errors Correct inconsistencies Calculate unmeasured parameters
Module 8: introduction to process integration
45
NAMP
PIECE
Process data Data processing Steady state detection Variables classification Gross error detection Data reconciliation Applications
46
NAMP
PIECE
Data Reconciliation
1 + 1 = 3 !!!
STATISTICAL PRINCIPLES
STATISTICAL PRINCIPLES
NAMP
PIECE
Data Reconciliation
Objectives of Data Reconciliation Optimally adjust measured values within given process constraints mass, heat, component balances Improve consistency of data to calibrate and validate process simulation Estimate unmeasured process values Obtain values not practical to measure directly Substitute calculated values for failed instrument
48
NAMP
PIECE
Data Reconciliation
Possible Benefits:
More accurate and reliable simulation results More reliable data for process analysis and decision making by mill manager Instrument maintenance and loss detection:
e.g. US$3.5MM annually in a refinery by decreasing loss by 0.5% of 100K BPD
Improve measurement layout Decrease number of routine analysis Improve advanced process control Clear picture of plant operating condition Early detections of problems Quality at process level Work Closer to specifications.
Module 8: introduction to process integration 49
NAMP
PIECE
Data Reconciliation
50
NAMP
PIECE
51
NAMP
PIECE
Data Reconciliation
52
NAMP
PIECE
Data Reconciliation
Gross Error Detection Closed Balances Identified Losses Monitored Efficiency Quantified Performance
NAMP
PIECE
Pinch Analysis.
54
NAMP
PIECE
Pinch Analysis
W ha t is Pinch An a lysis?
The prime objective of Pinch Analysis is to achieve financial savings in the process industries by optimizing the ways in which process utilities (particularly energy, mass, water, and hydrogen), are applied for a wide variety of purposes.
The Heat Recovery Pinch(Thermal Pinch Analysis now) was discovered indepently by Hohmann (71), Umeda et al. (78-79) and Linnhoff et al. (78-79).
Pinch Analysis does this by making an inventory of all producers and consumers of these utilities and then systematically designing an optimal scheme of utility exchange between these producers and consumers. Energy, Mass, and water re-use are at the heart of Pinch Analysis activities.
With the application of Pinch Analysis, savings can be achieved in both capital investment and operating cost. Emissions can be minimized and throughput maximized.
55
NAMP
PIECE
Pinch Analysis
FEATURES
The Pinch analysis is a t echnique t o design: Recovery Networks (Heat and Mass) Utility Networks (so called Total site Analysis) The basis of Pinch Analysis: The use of thermodynamic principles (first and second law). The use heuristics (insight), about design and economy. The Pinch Analysis makes extensive use of various graphical representations
56
NAMP
PIECE
Pinch Analysis
The Pinch Analysis provides insight s about t he process. In Pinch analysis, t he design engineering cont rols t he design procedure ( int eract ive m et hod ) . The pinch Analysis int egrat e s econom ic param et ers
57
NAMP
PIECE
Pinch Analysis The Four phases of pinch analysis in t he design of recovery process: Process Sim ulat ion Dat a Ext ract ion Target ing
Where an initial design is simplified and improved Which involves collecting data for the process and the
Design
58
NAMP
PIECE
Pinch Analysis
Heat Exchanger Network (HEN) HEN design is the classical domain of Pinch Analysis. By making proper use of temperature driving forces available between process steams, the optimum heat exchanger network can be designed, taking into account constraints of equipment location, materials of construction, safety, control, and operating flexibility. This then sets the hot and cold utility demand profile of the plant. When used correctly, Pinch Analysis yields optimum HEN designs that one would have been unlikely to obtain by experience and intuition alone.
59
NAMP
PIECE
Pinch Analysis
Combined Heat and Power (CHP) CHP is the terminology used to describe plant energy utilities, boilers, steam turbines, gas turbines, heat pumps, etc. Traditionally, these have been referred to as "plant utilities", without distinguishing them from other plant utilities such as cooling water and wastewater treatment. The CHP system supplies the hot utility and power requirements of the process. Pinch Analysis offers a convenient way to guarantee the optimum design, which can include the use of cogeneration or threegeneration (use of hot utility to produce cold utility and power for things like refrigeration).
60
NAMP
PIECE
Pinch Analysis
Possible Benefits:
One of the main advantages of Pinch Analysis over conventional design methods is the ability to set a target energy consumption for an individual process or for an entire production site before to design the processes. The energy target is the minimum theoretical energy demand for the plant or site. Pinch Analysis will therefore quickly identify where energy savings are likely to be found. Reduction of emissions Pinch Analysis enable to the engineer with tool to find the best way to change the process, if the process let it.
61
NAMP
PIECE
Pinch Analysis
In addition, Pinch Analysis allow you to: Update or Development of Process Flow Diagrams Identify the bottleneck in the process Departmental Simulations Full Plant Facilities Simulation Determine Minimal Heating (Steam) and Cooling Requirements Determine Cogeneration and Three-generation Opportunities Determine Projects with Cost Estimates to Achieve Energy Savings Evaluation of New Equipment Configurations for the Most Economical Installation Pinch Replaces the Old Energy Studies with a Live Study that Can Be Easily Updated Using Simulation
62
NAMP
PIECE
63
NAMP
PIECE
A Mathematical Model of a system is a set of mathematical relationships (e.g., equalities, inequalities, logical conditions) which represent an abstraction of the real world system under consideration. A Mathematical Model can be developed using:
Fundamental approaches Accepted theories of sciences are used to derive the equations (e.g., Thermodynamics Laws). Empirical Methods Input-output data are employed in tandem with statistical analysis principles so as to generate empirical or Black box models. Methods Based on analogy Analogy is employed in determining the essential features of the system of interest by studying a similar, well understood system.
Module 8: introduction to process integration 64
NAMP
PIECE
Optimization by Mathematical Programming: introduction A mathematical Model of a system consists of four key elements:
1. Variables The variables can take different values and their specifications define different states of the systems.
1. Continuous, 2. Integer, 3. Mixed set of continuous and integer.
2. Parameters The parameters are fixed to one or multiple specific values, and each fixation defines a different model. 3. Constraints the constraints are fixed quantities by the model statement 4. Mathematical Relationships The mathematical model relations can be classified as:
1. Equalities usually composed of mass balance, energy balance, equilibrium relations, physical property calculations, and engineering design relations which describe the physical phenomena of the system. 2. Inequalities consist of allowable operating regimes, specifications on qualities, feasibility of heat and mass transfer, performance requirements, and bound on availabilities and demands. 3. Logical conditions provide the connection between the continuous and integer variables.
The mathematicalto process integration algebraic, differential, or a mixed set Module 8: introduction relations can be 65
NAMP
PIECE
What is Optimization?
A optimization problem is a mathematical model which in addition to the before mentioned elements contains one or more performance criteria. The performance criteria is denoted as an objective function. It can be minimization of cost, the maximization or profit or yield of a process for instance. If we have multiple performance criteria then the problem is classified as multi-objective optimization problem.
A well defined opt im izat ion problem feat ures a num ber of variables great er t han t he num ber of equalit y const raint s, which im plies t hat t here exist degrees of freedom upon w hich we opt im ize.
Module 8: introduction to process integration 66
NAMP
PIECE
The typical mathematical model structure for an optimiztion problem takes the following form:
min f ( x, y )
x, y
s.t. h( x, y ) = 0 g ( x, y ) 0 x X n y Y integer
Where x is a vect or of n cont inuous variables, y is a vect or of int eger variables, h(x,y)= 0 are m equalit y const raint s, g(x,y) 0 are p inequalit y const raint s, and f(x,y) is t he object ive funct ion.
Module 8: introduction to process integration 67
NAMP
PIECE
If the objective function and constraints are linear without the use of integer variables, then OP becomes a linear programming (LP) problem. If there exist nonlinear terms in the objective function and/or constraints without the use of integer varialbes, the OP becomes a nonlinear programming (NLP) problem. If integer variables are used, they participate linearly and separtly from the continuous variables, and the objective function and constraints are linear, then OP becomes a mixed-integer linear programming (MILP) problem. If integer variables are used, and there exist nonlinear terms in the objective function and/or constraints, then the OP becomes a mixedinteger nonlinear programming (MINLP) problem. Whenever possible, linear programs (LP or MILP) are used because they guarantee global solutions.
68
NAMP
PIECE
Interaction of Design and Control Molecular Product Design Facility Location and allocation Facility Planning and Scheduling Topology of Transport Networks
Module 8: introduction to process integration 69
NAMP
PIECE
70
NAMP
PIECE
All of the model formulations that you have encountered thus far in the Optimization have assumed that the data for the given problem are known accurately. However, for many actual problems, the problem data cannot be known accurately for a variety of reasons. The first reason is due to simple measurement error. The second and more fundamental reason is that some data represent information about the future (e.g., product demand or price for a future time period) and simply cannot be known with certainty.
71
NAMP
PIECE
These are suitable for problems that deal with uncertainty. These computer algorithms or procedure models do not guarantee global optimally but are successful and widely known to come very close to the global optimal solution (if not to the global optimal). GA has the capability of collectively searching for multiple optimal solutions for the same best cost.
Such information could be very useful to a designer, because one configuration could be much easier to build than another.
SA takes one solution and efficiently moves it around in the search space, avoiding local optima.
Module introduction to process integration : 72
NAMP
PIECE
What is GAs?
GAs simulate the survival of the fittest among individuals over consecutive generation for solving a problem. Each individual represents a point in a search space and a possible solution. The individuals in the population are then made to go through a process of evolution. GAs are based on an analogy with the genetic structure and behaviour of chromosomes within a population of individuals using the following foundations:
Individuals in a population compete for resources and mates. Those individuals most successful in each 'competition' will produce more offspring than those individuals that perform poorly. Genes from good individuals propagate throughout the population so that two good parents will sometimes produce offspring that are better than either parent. Thus each successive generation will become more suited to their environment. Module introduction to process integration : 73
NAMP
PIECE
for a GA, each representing a possible solution to a given problem. Each individual is coded as a finite length vector of components, or variables, in terms of some alphabet, usually the binary alphabet {0,1}.
The chrom osom e (solut ion) is com posed of several genes (variables). A f it n e ss scor e (t he best object ive funt ion) is assigned t o each solut ion represent ing t he abilit ies of an individual t o com pet e . The individual wit h t he opt im al (or generally near opt im al) fit ness score is sought . The GA aim s t o use select ive breeding of t he solut ions t o produce offspring bet t er t han t he parent s by com bining inform at ion from t he chrom osom es.
74
NAMP
PIECE
Stochastic Search Methods The general genetic algorithm solution is found by: 1. [ St a rt ] Generat e random populat ion of n chrom osom es
(suitable solutions for the problem) 2.[Fitness] Evaluate the fitness f(x) (objective function) of each chromosome x in the population. 5.[New population] Create a new population by repeating following steps until the new populationis complete
1.
to their fitness (the better fitness, the bigger chance to be selected) 2.[Crossover] With a crossover probability cross over the parents to form a new offspring (children). If no crossover was performed, offspring is an exact copy of parents.. 3.[Mutation] With a mutation probability mutate new offspring at each locus (position in chromosome). 4.[Accepting] Place new offspring in a new population 4.
7.[Replace] Use new generated population for a further run of algorithm 4. 8.[Test] If the end condition is satisfied, stop, and return the best solution in current population 5. 9.[Loop] Go to step 2
75
NAMP
PIECE
Each chrom osom e has one binary st ring. Each bit in t his st ring can represent som e charact erist ic of t he solut ion. Or t he whole st ring can represent a num ber Of course, t here are m any ot her w ays of encoding. This depends m ainly on t he solved problem . For exam ple, one can encode direct ly int eger or real num bers. Som et im es it is also useful t o encode som e perm ut at ions.
Module 8: introduction to process integration 76
NAMP
PIECE
There are ot her ways how t o m ake crossover s, and we can choose m ult iple crossover point s. Crossover s can be rat her com plicat ed and vary depend ing on t he encoding of chrom osom e. Specific crossover s m ade for a specific problem can im prove perform ance of t he genet ic algorit hm .
77
NAMP
PIECE
Mutation After a crossover is performed, mutation takes place. This is to prevent the falling of all solutions in the population into a local optimum. Mutation changes the new offspring randomly. For binary encoding we can switch a few randomly chosen bits from 1 to 0 or from 0 to 1. Mutation can then be shown as:
The m ut at ion depends on t he encoding as w ell as t he crossover. For exam ple when we are encoding perm ut at ions, m ut at ion could be exchanging t w o genes.
Module 8: introduction to process integration 78
NAMP
PIECE
GAs Characteristics:
A GA makes no assumptions about the function to be optimized (Levine, 1997) and thus can also be used for nonconvex objective functions A GA optimizes the tradeoff between exporting new points in the search space and exploiting the information discovered thus far A GA operates on several solutions simultaneously, gathering information from current search points and using it to direct subsequent searches which makes a GA less susceptible to the problems of local optima and noise A GA only uses the objective function or fitness information, instead of using derivatives or other auxiliary knowledge, as are needed by traditional optimization methods.
Module 8: introduction to process integration 79
NAMP
PIECE
Initial Population
1st Generation
Get Objective Function Value for Whole Population (Internal optimization) Yes
Nth Generation
Optimum ? No
Sto p
(N+1)th Generation
80
NAMP
PIECE
81
NAMP
PIECE
82
NAMP
PIECE
83
NAMP
PIECE
84
NAMP
PIECE
LCA m e t h od olog y:
LIFE-CYCLE ASSESSMENT
Goal and Scope definitio n Inventor y analysis Interpretation
DIRECT APPLICATIONS
Product development and improvement Strategic planification Public policy Marketing Etc.
Impact assessment
OTHER ASPECTS
Technical Economic Market Social etc. 85
NAMP
PIECE
system boundaries ,
Module 8: introduction to process integration 86
NAMP
PIECE
Inventory analysis data collection , refining system boundaries calculation no formal description, software validation of data relating data to the specific system allocation 87 Module 8: introduction to process integration
NAMP
PIECE
NAMP
PIECE
Interpretation/improvement assessment identification of significant environmental issues evaluation , , conclusions and recommendations
89
NAMP
PIECE
Possible Benefits:
Improvements in overall environmental performance and compliance Provides a framework for using pollution prevention practices to meet LCA objectives Increased efficiency and potential cost savings when managing environmental obligations Promotes predictability and consistency in managing environmental obligations More effective measurement of scarce environmental
90
NAMP
PIECE
91
NAMP
PIECE
Many organisat ions t oday are faced wit h t he sam e challenge: TOO MUCH DATA It is t he last it em t hat is of int erest t o us as chem ical engineers
92
NAMP
PIECE
Data-Rich but Knowledge-Poor Far too much data for a human brain Limited to looking at one or two variables at a time:
12
10 8 6 4 2 0 1 2 3 4 5 6 7
Brain
Big Problem: Interesting, useful patterns and relationships not intuitively obvious lie hidden inside enormous, unwieldy databases
Module 8: introduction to process integration 93
NAMP
PIECE
OUTSIDE
IN
Empirical Model
This approach uses the plant process data directly, to establish mathematic correlations. Unlike the theoretical models, empirical models do NOT take the process fundamentals into account. They only use pure mathematical and statistical techniques. Multi-Variable Analysis (MVA) is one such method, because it reveals patterns and correlations independently of any pre-conceived notions. Obviously this approach is very sensitive to Garbage-in, garbage-out which is why validation of the model is so important. Module 8: introduction to process integration 94
NAMP
PIECE
95
NAMP
PIECE
MVA
96
NAMP
PIECE
Colour, shape, firmness, reflectivity, Skin: smoothness, thickness, morphology, Juice: water content, pH, composition, Seeds: colour, weight, size distribution, et cetera
+1
However, always only one latent attribute
-1
Apple or orange?
97
NAMP
PIECE
Statistical Model
. . .. . .. . . . . .
(internal to software)
trends X
trends X
trends X
NAMP
PIECE
OUTLINER
1 component
99
NAMP
PIECE
1 component
NAMP
PIECE
What-if Exercises
Low-cost investigation of options
101
NAMP
PIECE
102
NAMP
PIECE
Control Objectives:
Product specifications variability should be kept to a minimum --> process variability (To Control Product quality). Safety issues(separate equipments), energy costs, environmental concerns have increased complexity and sensitivity of processes
Plants become highly integrated in terms of mass and energy and therefore, process dynamics are often difficult to control. The Control is permanently necessary to do for allowing the process to operate in the best conditions. Module 8: introduction to process integration 103
NAMP
PIECE
CONTROLLABILITY
it is a property of a process that accounts for the ease with which a continuous plant can be held at a specified operating policy, despite external disturbances (resiliency) and uncertainties (flexibility) and regardless of the control system imposed on such a plant. Sources
CONTROL
Changes in Process
NAMP
PIECE
Fundamentals:
Input Variables PROCESS RESILIENCY
Control Loop
Disturbances sen sor
Uncertainties
NAMP
PIECE
Integrate Process Design and Control e.g. Controllability analysis for control structures design
Water, F1 Pulp, F2
C C
F C
C, F
Interac tions
INPUTS (process variables or disturbances) EFFECTS OUTPUTS (Best Selection by Controllability analysis)
106
NAMP
PIECE
Integrate Process Design and Control Why Cont rollabilit y is im port ant :
The process will be more capable to move smoothly around the possible operating edge Stability and better performance of control loops and structures System relatively insensitive to perturbations Efficient management of interacting networks
Module 8: introduction to process integration
Flexibility
107
NAMP
PIECE
The Top level of the process control, Strategic control level is thus concerned with achieving the appropriate values principally of:
108
NAMP
PIECE
109
NAMP
PIECE
The Process Industries are increasingly compelled to operate profitably in very dynamic and global market. The increasing competition in the international area and stringent product requirements mean decreasing profit margins unless plant operations are optimized dynamically to adopt to the changing market conditions and to reduce the operating cost. Hence, the importance of real-time or on-line optimization of an entire plant is rapidly increasing.
110
NAMP
PIECE
What is RTO? Real-time Optimization is a model-based steady-state technology that determines the economically optimal operating policy for a process in the near term The system optimizes a process simulation and not the process directly Performance measured in terms of economic benefit Is an active field of research:
Model accuracy, error transmission, performance evaluation
111
NAMP
PIECE
RTO Schematically Reconciliation And gross Error Detection Updating Process Model (Steady State Dynamic Simulation)
Plant Facility
112
NAMP
PIECE
SETPOINTS (DOFs)
Selected Ouputs
113
NAMP
PIECE
114
NAMP
PIECE
Cost, Process, Cost, Process, Environmental & Environmental & Product Outcomes Product Outcomes
Click here
Process Process Design Design Analysis Analysis And and Synthesis Synthesis
Click Here
Integrated Business Integrated Business & & Click Here Process Model Process Model
Process Process Operation Operation Analysis Analysis and and Optimizati Optimizat
Optimization
Click here
Cost, Process, Cost, Process, Environmental & Environmental & Product Product Data Data
115
NAMP
PIECE
Cost, Process, Environmental & Product Data Integrated Business & Process Model
Reconciled P&E Data
mean all the data are consistent together throughout all the plant facilities Data Reconciliation
Data Reconciliation
Reconciliation
Product Data Market Data
model is built it can be used to validate and reconcile data Plant Facilities
Module 8: introduction to process integration 116
NAMP
PIECE
Integrated Business and Process Model with the classification, recording, allocation, and summarization forAccounting Data of the purpose
Process Data Environmental Data Product Data Market Data
Ac Mo coun de tin l g
Click here
Integrated Business
Co st
Ac Mo coun de tin l g
Co st
and
117
NAMP
PIECE
Distribution
Distribution
Consumer Consumer
W
Supply
Retail Retail
W Collection
Environmental Supply Chain (ESC) holds all the elements a traditional supply chain has but is extended to a semi-closed loop in order to also account for the environmental impact of the supply chain and recycling, re-use and collection of used material (Beamon 1999)
Module 8: introduction to process integration 118
NAMP
PIECE
119
NAMP
PIECE
Process simulation Data Reconciliation MVA using relational database Pinch analysis LCA SC and ESC model analysis Controllability Analysis Optimization ( Deterministic and/or Stochastic)
Process
120
NAMP
PIECE
Process Process Operation Integration Process Design Analysis Analysis and Tools Optimization Loop
and Optimization
Data reconciliation for instrument validation Dynamic simulation Process control strategies MVA (Soft sensor dev.) Real-time optimization Optimizated supply chain Model
NAMP
PIECE
Outline
1.3 An around-the-world tour of PI practitioners 1.3 An around-the-world tour of PI practitioners focuses of expertise focuses of expertise
122
NAMP
PIECE
123
NAMP
PIECE
Courtesy mainly of the www to capture the flavor of the evolution of Process Integration PI is relatively new: Researchers build on their strengths Many of the ground-breaking techniques are coming from universities When techniques become practical, the private sector generally capitalizes and techniques advance more rapidly
124
NAMP
PIECE
Research Area: Recognized as one of the major research groups in the area of Computer Aided Process Design. In Process Integration, the group is recognized for its work in Mathematical Programming, Optimization, Reactor Systems, Separation Systems (especially Distillation), Heat Exchanger Networks, Operability and the synthesis of Operating Procedures.
Current research in Process Integration includes: 1) Insights to Aid and Automate Synthesis (Invention) 2) Structural Optimization of Process Flowsheets 3) Synthesis of Reactor Systems and Separation Systems 4) Synthesis of Heat Exchanger Networks 5) Global Optimization techniques relevant to Process Integration 6) Integrated Design and Scheduling of Batch plants 7) Supply chain dynamics and optimization
Consortium: "Center for Advanced Process Decision-making" with 20 members (2001) including operating companies, engineering & contracting companies, consulting companies and software vendors. The consortium was founded 1986.
125
NAMP
PIECE
Consortium: "Process Systems Engineering" with 17 members (2003) including operating, engineering & contracting companies, software vendors. introduction to process integration Module 8: 126
NAMP
PIECE
Current research is organized in three major areas: 1) Efficient Use of Raw Materials (including Water) 2) Energy Efficiency 3) Emissions Reduction 4) Eefficient use of capital.
Consortium: "Process Integration Research Consortium" with 27 members (2003) including operating companies, engineering & contracting companies, consulting companies and software vendors. The consortium was founded in 1984 by six multinational companies.
127
NAMP
PIECE
Industry: Close co-operation with some of the major pulp and paper industry Module 8: introduction to process integration groups, including training courses, consulting, etc. 128
NAMP
PIECE
Consortium: "Process Integration Research Consortium" with 13 members (2003) including operating companies, engineering & contracting companies, consulting companies and software vendors in pulp and paper industry.
129
NAMP
PIECE
130
NAMP
PIECE
1) Global allocation of Mass and Energy 2) Synthesis of Waste Allocation and Species Interception Networks 3) Physical and Reactive Mass Pinch Analysis 4) Synthesis of Heat-Induced Networks 5) Design of Membrane-Hybrid Systems 6) Design of Environmentally acceptable Reactions 7) Integration of Reaction and Separation Systems 8) Flexibility and Scheduling Systems 9) Simultaneous Design and Control 10) Global Optimization via Interval Analysis
131
NAMP
PIECE
Research areas are: 1) Synthesis of Processes; Modeling, Simulation, Control and Optimization of Processes; New Processes and Materials. 2) Recovery systems of Heat; Renewable sources of Energy; Thermodynamic Optimization. 3) Contaminated Atmosphere rehabilitation; Treatment of Effluents; Environmental Processes.
132
NAMP
PIECE
Clients: they have founded your own consultancy enterprise the name Wits Enterprise.
133
NAMP
PIECE
Linnhoff March Ltd., Northwich, Cheshire, UK Web: http://www.linnhoffmarch.com/ List of Services in the area of Process Integration: Linnhoff March is the pioneering company of Pinch Technology and has built a reputation for being the "Pinch Company", encompassing: Project execution and consulting Software development and support Training assistance PI Technologies: Pinch Technology (Analysis and HEN DesignTotal Site Analysis) Water Pinch for Wastewater minimization Combined Thermal and Hydraulic Analysis of Distillation Columns PI Software: Extensively proven state-of-the-art industry including iency and profitability in the hydrocarbon processingsoftware worldwide. KBC analyses SuperTarget, PinchExpress, WaterTarget and Steam97. and offers Implementation Services to assist clients in realising measurable financial im Typical Projects: 1200 assignments over 18 years - or over 50 studies per year in PI, making them the unquestionable world leader (27th February 2002)Was acquired last year by KBC process technology
134
NAMP
PIECE
We are the premier consulting engineering specialists dedicated to the pulp and paper industry. Prom. energy and water reduction to planning new power islands. American Process can provide solutions through practical experience, process integration, troubleshooting, and project implementation.
Founded in 1994, with offices in Atlanta, GA, Athens, Greece, and ClujNapoca, Romania, American Process is the premier specialist firm dedicated to reducing energy, water, and other operating costs for the pulp and paper industry.
Energy Targeting Using Pinch Analysis, PARIS (Decision-Making Tool for Optimizing Pulp and Paper Mill Operations)
Production Analysis for Rate and Inventories Strategies.
integration
135
NAMP
PIECE
Around the World tour of PI practitioners focuses of experience Process Systems Enterprise Ltd., london, UK. Web: http://www.psenterprise.com
Process Systems Enterprise Limited (PSE) is a provider of advanced model-based technology and services to the process industries. These technologies address pressing needs in fastgrowing engineering and automation market segments of the chemicals, petrochemicals, oil & gas, pulp & paper, power, fine chemicals, food, pharmaceuticals and biotech industries. gPROMS, for general PROcess Modelling System
Steady-state and dynamic process simulation, optimization (MINLP) and parameter estimation software, packaged for different users.
Model Enterprise - Supply chain modeling and execution environment. Module 8: introduction to process integration Model Care - Business model 136
NAMP
PIECE
137
NAMP
PIECE
138
NAMP
PIECE
139
NAMP
PIECE
http://www.combination.com Advanced Process Combinatorics (APC) Aspen Technology Inc. (AspenTech)http://www.aspentech.com and http://www.hyprotech.com http://www.ipa-scotland.org.uk/members/nel.htm National Engineering Laboratory (NEL) http://www.quantisci.co.uk/ QuantiSci Limited
...
...
140
NAMP
PIECE
End of Tier 1
At the moment we are assuming that you have done all the reading, this is the end of Tier 1. We do not have doubt that much of this information seems fuzzy, but we are only trying to set all the pieces in the Process Integration scope.
141
NAMP
PIECE
QUIZ
142