Data Reconciliation

1
Data Reconciliation
Illustration Of Random Gross Errors

• abnormality

2
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

3
Data Reconciliation
Measurement Problem Handling
Processing random errors
4
Data Reconciliation

• Data Treatment Typical Strategy
• Establish Plant facilities operating regimes
• Data processing
• Remove random noise
• Detect and correct abnormalities
• Steady state detection
• Identify steady-state duration
• Select data set
• Data reconciliation
• Detect gross errors
• Correct inconsistencies
• Calculate unmeasured parameters

5
Data Reconciliation
METHODOLOGY EMPLOYED
From Plant Facilities
Process data
For simulation and further applications
6
Data Reconciliation
                    
                                                
                

• What is data reconciliation?
• Data reconciliation is the validation of process
  data using knowledge of plant structure and the
  plant measurement system

                  
                                                
           
                  
            
7
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

8
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. US3.5MM 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.

9
Data Reconciliation

• Data Reconciliation Problem of Process Under
  Different Status
• Steady-state data reconciliation
• based on steady-state model
• Using spatial redundancy
• Dynamic data reconciliation
• based on dynamic models
• Using both spatial temporal redundancy

10
Data reconciliation (DR)

• DR Problem Of Process Under Different Status
  (Contd.)
• General expression of conservation law
• input- output generation- consumption-
  accumulation 0
• Steady state case
• no accumulation of any measurement
• Constraints are expressed algebraically
• Dynamic process
• Accumulation cannot be neglected
• Constraints are differential equations

11
Data Reconciliation

• Data Reconciliation of Different Constraints
• Linear data reconciliation
• Only mass balance is considered
• flows are reconciled
• Bilinear data reconciliation
• Component balance imposed as well as energy
  balance
• flows composition measurements are reconciled
• Nonlinear data reconciliation
• Mass/energy/component balances are included
• Flow rate, composition, temperature or pressure
  measurements are reconciled

12
Data Reconciliation
NEXT
13
Pinch Analysis.
14
Pinch Analysis
What is Pinch Analysis?

• 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.

15
Pinch Analysis
FEATURES

• The Pinch analysis is a technique to 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

16
Pinch Analysis

• The Pinch Analysis provides insights about the
  process.

• In Pinch analysis, the design engineering
  controls the design procedure (interactive
  method).

• The pinch Analysis integrates economic parameters

17
Pinch Analysis
The Four phases of pinch analysis in the design
of recovery process
Which involves collecting data for the process
and the utility system
Which establishes figures for the best
performance in various aspects.
Where an initial Heat Exchanger Network is
established by heuristics tools allowing a
minimum target to be reached.
Where an initial design is simplified and
improved economically.
18
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.

19
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 three-generation (use of hot utility to
  produce cold utility and power for things like
  refrigeration).

20
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.

21
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

NEXT
22
Optimization by Mathematical Programming
23
Optimization by Mathematical Programming
introduction

• 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.

24
Optimization by Mathematical Programming
introduction

• A mathematical Model of a system consists of four
  key elements
• Variables ? The variables can take different
  values and their specifications define different
  states of the systems.
• Continuous,
• Integer,
• Mixed set of continuous and integer.
• Parameters ? The parameters are fixed to one or
  multiple specific values, and each fixation
  defines a different model.
• Constraints ? the constraints are fixed
  quantities by the model statement
• Mathematical Relationships ? The mathematical
  model relations can be classified as
• 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.
• Inequalities ? consist of allowable operating
  regimes, specifications on qualities, feasibility
  of heat and mass transfer, performance
  requirements, and bound on availabilities and
  demands.
• Logical conditions ? provide the connection
  between the continuous and integer variables.
• The mathematical relations can be algebraic,
  differential, or a mixed set of both constraints.
  These can be linear or nonlinear.

25
Optimization by Mathematical Programming

• 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 optimization problem features a
number of variables greater than the number of
equality constraints, which implies that there
exist degrees of freedom upon which we optimize.
26
Optimization by Mathematical Programming

• The typical mathematical model structure for an
  optimiztion problem takes the following form

Where x is a vector of n continuous variables, y
is a vector of integer variables, h(x,y) 0 are m
equality constraints, g(x,y) ? 0 are p inequality
constraints, and f(x,y) is the objective function.
27
Optimization by Mathematical Programming

• Classes of Optimization Problems (OP)
• 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 mixed-integer
  nonlinear programming (MINLP) problem.
• Whenever possible, linear programs (LP or MILP)
  are used because they guarantee global solutions.
• MINLP problems features many applications in
  engineering.

28
Optimization by Mathematical Programming

• Applications
• Process Synthesis
• Heat Exchanger Networks
• Distillation Sequencing
• Mass Exchanger Networks
• Reactor-based Systems
• Utility Systems
• Total Process Systems
• Design, Scheduling, and Planning of Process
• Design and Retrofit of Multiproduct Plants
• Design and Scheduling of Multiproduct Plants
• Interaction of Design and Control
• Molecular Product Design
• Facility Location and allocation
• Facility Planning and Scheduling
• Topology of Transport Networks

NEXT
29
Stochastic Search Methods
30
Stochastic Search Methods

• Why stochastic Search Methods
• 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.

31
Stochastic Search Methods

• There are probabilistic algorithms, such as
• Simulated annealing (SA)
• Genetic Algorithms (GAs)
• Tabu search
• 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.

32
Stochastic Search Methods

• 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.

33
Stochastic Search Methods

• A population of individuals is maintained within
  search space 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 chromosome (solution) is composed of several
  genes (variables). A fitness score (the best
  objective funtion) is assigned to each solution
  representing the abilities of an individual to
  compete. The individual with the optimal (or
  generally near optimal) fitness score is sought.
  The GA aims to use selective breeding of the
  solutions to produce offspring better than the
  parents by combining information from the
  chromosomes.

34
Stochastic Search Methods

• The general genetic algorithm solution is found
  by

• Start Generate random population of n
  chromosomes (suitable solutions for the problem)
• Fitness Evaluate the fitness f(x) (objective
  function) of each chromosome x in the population.
• New population Create a new population by
  repeating following steps until the new
  populationis complete
• Selection Select two parent chromosomes from a
  population according to their fitness (the better
  fitness, the bigger chance to be selected)
• 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..
• Mutation With a mutation probability mutate new
  offspring at each locus (position in chromosome).
• Accepting Place new offspring in a new
  population 4.
• Replace Use new generated population for a
  further run of algorithm 4.
• Test If the end condition is satisfied, stop,
  and return the best solution in current
  population 5.
• Loop Go to step 2

35
Stochastic Search Methods

• Encoding of a Chromosome
• The chromosome should in some way contain
  information about the solution which it
  represents. The most used way of encoding is a
  binary string. The chromosome then could look
  like this

Each chromosome has one binary string. Each bit
in this string can represent some characteristic
of the solution. Or the whole string can
represent a number Of course, there are many
other ways of encoding. This depends mainly on
the solved problem. For example, one can encode
directly integer or real numbers. Sometimes it is
also useful to encode some permutations.
36
Stochastic Search Methods

• Crossover
• After we have decided what encoding we will use,
  we can make a step to crossover. Crossover
  selects genes from parent chromosomes and creates
  a new offspring. The simplest way how to do this
  is to choose randomly some crossover point and
  everything before this point copy from a first
  parent and then everything after a crossover
  point copy from the second parent.
• Crossover can then look like this ( is the
  crossover point)

There are other ways how to make crossovers, and
we can choose multiple crossover points.
Crossovers can be rather complicated and vary
depending on the encoding of chromosome. Specific
crossovers made for a specific problem can
improve performance of the genetic algorithm.
37
Stochastic Search Methods

• 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 mutation depends on the encoding as well as
the crossover. For example when we are encoding
permutations, mutation could be exchanging two
genes.
38
Stochastic Search Methods

• 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.

39
Stochastic Search Methods
GA Solution Procedure
Start
Initial Population
1st Generation
Get Objective Function Value for Whole
Population (Internal optimization)
Nth Generation
Yes
Optimum?
Stop
No

• Generate New Population
• GA parameters
• GA strategies

(N1)th Generation
40
SA and GA comparation In theory and Practice
NEXT
41
Life Cycle Analysis.
42
Life Cycle Analysis

• What is Life Cycle Analysis?
• Technique for assessing the environmental aspects
  and potential impacts associated with a product
  by
• An inventory of relevant inputs and outputs of a
  system
• Evaluating the potential environmental impacts
  associated with those inputs and outputs
• Interpreting the results of the inventory and
  impact phases in relation to the objectives of
  the study heading
• Evaluation of some aspects of a product system
  through all stages of its life cycle

43
Life Cycle Analysis

• Why LCA is important
• Tool for improvement of environmental performance
• Systematic way of managing an organizations
  environmental affairs
• Way to address immediate and long-term impacts of
  products, services and processes on the
  environment
• Focus on continual improvement of the system

44
Life Cycle Analysis
LCA methodology
LIFE-CYCLE ASSESSMENT
Goal and Scope definition
45
Life Cycle Analysis

• Goal and scope definitions
• goal ? application, use and users
• scope ? borders of the assessment
• functional unit ? scale for comparison
• efficiency
• durability
• performance quality standard
• system boundaries ? process, inputs and outputs
  defined
• data quality ? reflected in the end results
• critical review process ? verification of validity

46
Life Cycle Analysis

• Inventory analysis
• data collection ? qualitative or quantitative,
  most work intensive
• refining system boundaries ? after initial data
  collection
• calculation ? no formal description, software
• validation of data ? assessment of data quality
• relating data to the specific system ? data must
  be ralted to the functional unit
• allocation ? done when not all impacts and
  outputs are within the system boundaries

47
Life Cycle Analysis

• Impact assessment
• category definition ? impact categories defined
• classification ? inventory input and output
  appointed to impact categories
• characterization ? assign relative contribution
• weighting ? when comparison of the impact
  categories is not possible

48
Life Cycle Analysis

• Interpretation/improvement assessment
• identification of significant environmental
  issues ? information structured in order to get a
  clear view on key environmental issues
• evaluation ? completeness analysis, sensitivity
  analysis, consistency analysis
• conclusions and recommendations ? improve
  reporting of the LCA

49
Life Cycle Analysis

• 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

NEXT