You are on page 1of 55

DESIGN OF THERMAL SYSTEMS

Dr P Venkateswarlu
Professor
Mechanical Engineering

8/31/2017 Department of Mechanical Engineering 1


Modeling
First step in the design of any system is modeling.
Modeling - is the representation of a given system by a
set of equations or by a physical model for
experimentation. Modeling is a cognitive activity.
Normalizing (Nondimensionalizing) - reduces the
number of variables in the equations simplifying the
mathematical or the (physical) experimental modeling
Reynolds No. Re, Prandtl No. Pr are 2 such non-
dimensional variables which combine 7 physical
variables. Very useful in numerical modeling.

8/31/2017 Department Name 2


Modeling
Model: Model is a substitute for a real system. Model
simplifies and streamlines the work in experiment; lab or
computer. Easy to assemble and to make.
Models we made: Paper boats and airplane (glider) are
made and watched their working and dynamics in water
and air.
Mathematical Models: Mathematical models in the form
of equations represent the physical phenomena and the
interconnections in a system. Computer simulation shifts
the tedious computations in solving the equations to the
computer.
8/31/2017 Department Name 3
Mathematical Modeling
Devices and their behaviors can be described by
words, sketches, physical models, computer
programs, or mathematical models.
Mathematical Model is a representation in
mathematical terms the behavior of devices and
objects
Basic idea behind mathematical modeling is the
dimensional consistency and homogeneity.

8/31/2017 Department Name Confidential 4


Modeling
Informal Modeling:

8/31/2017 Department Name 5


Types of Models
Categories:

8/31/2017 Department Name 6


Mathematical Modeling
Conservation principles:

when g =c=0

8/31/2017 Department Name Confidential 7


Mathematical Modeling of Thermal
Systems: Examples
Example 1:Steady State Heat Conduction
i=100 0C, a=30 0C, Ris=0.75 K/W, Rsa= 2.25K/W

----------------------------

8/31/2017 Department Name Confidential 8


Mathematical Modeling of Thermal
Systems: Examples
Example 2:Heating a room(Unsteady)

Example 3: Heating the room when the external temp. is


unsteady

8/31/2017 Department Name Confidential 9


Mathematical Modeling of Thermal
Systems: Examples
Example 4:Building with two rooms

8/31/2017 Department Name Confidential 10


Interpolation & Curve Fitting
Motivation-Finding root of a function. Functions
come from observed data or from physical laws
Interpolation Connect all data points when they are
reliable; linear interpolation. If data has inherent error
or scatter, interpolation has limited purpose. The goal
is to find a and b
Curve fitting is capturing the trend in data by
assigning a single function across the entire range

8/31/2017
Linear Regression (Linear Curve Fitting)
General form: f(x)=ax + b
The one line that provides a
minimum error is the best fit.
Quantifying error: Least squares approach
for minimizing error

8/31/2017
Linear Regression

Putting in matrix form

Solve by Gaussian elimination or matrix inversion,

8/31/2017
Linear Regression
Example 1:

Example 2:

= 4.8515 + 0.3525

8/31/2017 Department Name Confidential 14


Nonlinear Regression
Polynomial Fit: We get the same set of linear
equations which can be solved by Gaussian
elimination or matrix inversion
Exponential Fit: We get nonlinear equations.
Linearize and solve by standard methods. One of the
most common nonlinear models is the exponential
decay or exponential growth model.
= , + , , =1 2
Perform Logarithmic Regression,

8/31/2017 Department Name Confidential 15


Regression
Engineering Examples:
1. Vapor Pressure:
2. Heat Capacity:
3. Thermal Conductivity:
4. Viscosity of Liquid:

5. Heat Transfer Correlation:


6. Sieder-Tate Correlation:

8/31/2017 Department Name Confidential 16


Curve Fitting- Higher Order Polynomials
Nonlinear Regression
Polynomial Curve Fitting:

Least Squares Approach:

n+1 equations:

8/31/2017 Department Name Confidential 17


Curve Fitting- Higher Order Polynomials
Polynomial Curve Fitting:

Least Squares Approach:

n+1 equations:

8/31/2017 Department Name Confidential 18


Gaussian Elimination

8/31/2017 Department Name Confidential 19


Numerical Model of a Thermal System

Discretized FEM model of a steam turbine:


Steam turbine thermal model for transient operation
(Ph D Thesis of Monika Topel, 2014, KTH Royal
Institute of Technology, Stockholm, Sweden)
Key words- steam turbines, transients, start-up, finite
element model, heat transfer

8/31/2017 Department Name Confidential 20


Numerical Model of a Thermal System
Discretized FEM model of a steam turbine:
Objectives:
1. Establish a validated and efficient modeling tool
which can accurately predict the thermal behavior of
steam turbines during transient operation while
being generally applicable in terms of heat transfer
and adaptable in terms of geometry.
2. Implement this tool in parametric studies concerning
improvements of the thermal response of steam
turbines during start-up.

8/31/2017 Department Name Confidential 21


Numerical Model of a Thermal System
Steam turbine transient thermal model:
The modeling scheme uses COMSOL coupled with MATLAB as
software. The integration of COMSOL allows for flexibility
towards multi-physics studies, while keeping the advantages of
MATLAB related to scripting and algorithm automation. The
overall structure of the model consists of three coupled sub-models:
1. A thermodynamic steam expansion model to calculate the
temperatures and mass flows during operation.
2. A gland steam sealing model to calculate the temperatures
and mass flows in the labyrinths seals.
3. A FE heat conduction model to calculate the metal
temperatures within the turbine.

8/31/2017 Department Name Confidential 22


Numerical Model of a Thermal System
Start-up curves

1. Start-up curves set a detailed schedule of the rate for the


turbine to reach nominal speed and full load.
2. Maintain thermal stresses under a given temperature
dependent limit in critically thick components - stator & rotor
8/31/2017 Department Name Confidential 23
Numerical Model of a Thermal System
Steam turbine transient thermal model: MATLAB &
COMSOL

8/31/2017 Department Name Confidential 24


Numerical Model of a Thermal System
Thermodynamic Steam Expansion:
1. Energy conversion from heat to electricity occurs.
2. This energy transformation occurs within a turbine stage, which is
composed of a static blade row, denoted as stator, and a rotating
blade row, denoted as rotor.
3. Essentially, the internal energy of the steam is transformed into
kinetic energy, and the latter in turn is transformed into
mechanical work that rotates the turbine shaft
4. Eulers turbomachine equation in which a change in total
enthalpy 0 is equivalent to a change in tangential flow speed
and/or rotational engine speed (Eulers Equation)

8/31/2017 Department Name Confidential 25


Numerical Model of a Thermal System
Thermodynamic State of Steam:
1. As the steam expands its
enthalpy and pressure decrease.
2. Figure shows the enthalpy-entropy
diagram of steam expanding from
state 1 to state 2.
3. Also represented in this figure,
are the stagnation states denoted
with 0 and the isentropic states denoted with .

8/31/2017 Department Name Confidential 26


Numerical Model of a Thermal System
Gland Steam Sealing:
Sealing glands are located at each end of the shaft, and fulfill two
functions: preventing the leakage of steam and avoiding the ingress
of air through the clearances between the rotating and stationary
parts of the turbine

8/31/2017 Department Name Confidential 27


Numerical Model of a Thermal System
Gland Steam Sealing:
The seal constant Cseal can be expressed as a function of geometric
parameters such as the seal diameter , clearance , number of
teeth , tooth pitch and tooth width , see Equation

8/31/2017 Department Name Confidential 28


Numerical Model of a Thermal System
Gland Steam Sealing:
Steam expansion in a labyrinth seal is described by the
process of throttling, in which the pressure of the steam
is decreased while keeping the same enthalpy level.
The nature of the flow through the seal as the steam
passes under each constriction is described by a Fanno
line, in which the leaked mass flow is dependent on the
inlet conditions of steam, the outlet pressure and the
seal constant, as shown in Equation.

8/31/2017 Department Name Confidential 29


Numerical Model of a Thermal System
Steam expansion:
1. Steam energy transformation occurs within a turbine
stage, which is composed of a static blade row,
denoted as stator, and a rotating blade row, denoted
as rotor.
2. Essentially, the internal energy of the steam is
transformed into kinetic energy, and the latter in turn
is transformed into mechanical work that rotates the
turbine shaft

8/31/2017 Department Name Confidential 30


Numerical Model of a Thermal System
Steam Expansion Model:
The off-design isotropic efficiency:

Mass flow coefficient ,


is proportional to pressure
ratio across the unit.
Off-design to design ratio:

8/31/2017 Department Name Confidential 31


Numerical Model of a Thermal System
Boundary conditions:

1. The calculated steam properties are used as input for


the BCs of the heat conduction model.
2. The mass flow together with the steam
temperatures, help formulate the BC of the live
steam in the blade passage. Furthermore,
3. The steam conductivity is used in the definition of
the material properties of the turbine.

8/31/2017 Department Name Confidential 32


Numerical Model of a Thermal System
Heat conduction model:
The complex 3D geometries of the turbine units are
modeled as 2D axisymmetric, a simplification made
possible by the inherently cylindrical geometry of axial
turbines.
Note: Not entirely axisymmetric due to the casing
flanges and piping- still a good approximation.

Equivalent thermal conductivity:


8/31/2017 Department Name Confidential 33
Numerical Model of a Thermal System
Geometric model:

Fig: Geometry blocks in heat transfer model superimposed over


an axisymmetric view of the HPT and LPT units
8/31/2017 Department Name Confidential 34
Numerical Model of a Thermal System
Geometric model:
Modular Approach- Three modules
Inlet, blade passage and outlet

Fig: Turbine geometry as defined by the modular approach

8/31/2017 Department Name Confidential 35


Numerical Model of a Thermal System

Boundary conditions and the active range:


1. To solve the heat conduction equations by FE model, BCs
must be defined as shown. The BCs are calculated from
empirical heat transfer correlations obtained from various
sources
2. The definition of the heat transfer BCs is a very important
step towards capturing the correct thermal behavior of a
steam turbine.

8/31/2017 Department Name Confidential 36


Three types of BCs in heat transfer:
1. First type of BC consists of fixing the surface temperature.
2. Second type of BC is a fixed heat flux at a surface. Adiabatic
(zero-flux) surfaces have been defined within the turbine at
locations where heat transfer is negligible.
3. Third type of BC corresponds to convection with a moving
fluid. This condition is the most common in the current
model and is defined on all the surfaces exposed to steam, be
it gland steam or live steam and convection to the ambient air
4. When turbine is idle, neat vacuum conditions exist inside,
only surface to surface radiation from rotor to stator occurs

8/31/2017 Department Name Confidential 37


Numerical Model of a Thermal System
Boundary conditions:

Fig: Different boundary conditions on the LPT and HPT


unit geometries
8/31/2017 Department Name Confidential 38
Numerical Model of a Thermal System
BC-Bearing oil temperature:

1. The bearing surfaces of the shaft are subject to the first type of
BC, in which the outer surface of the shaft is considered as
isothermal and exposed to the lubrication oil temperature.
2. The temperature of the oil is considered increasing dependent
on the rotational speed.
3. Two temperature levels were assumed for the bearing oil,
corresponding to the lower and upper limits reached by the oil
during standby and operation, respectively.

8/31/2017 Department Name Confidential 39


Numerical Model of a Thermal System
Boundary conditions:
1. The gland steam sealing model calculates the thermal properties
of the gland steam by accounting for the mode of operation of
the turbine,
2. Convection to gland steam: The corresponding HTCs were
calculated using the equation below, where is the height of the
expanding chamber and the radial clearance.

3. Convection to Live Steam:

8/31/2017 Department Name Confidential 40


Numerical Model of a Thermal System
Insulation & ambient convection:
1. Conduction through the insulation and convection to ambient
air are combined in their thermal resistances to formulate the
HTC corresponding to this BC.
2. Equation below shows the HTC for the surfaces covered with
insulation, where is the thickness of the insulation layer,
its thermal conductivity and is the external natural
convection heat transfer coefficient.
3. The value of the latter is chosen to be 8 W/m2K in accordance
with the recommended value from for natural convection from
a hot surface inside a building.

8/31/2017 Department Name Confidential 41


Numerical Model of a Thermal System
Friction & Ventilation Losses:
1. Losses due to friction and ventilation take place in the blades
and discs of the turbine
2. Energy dissipated due to friction is proportional to the steam
density , the square of the disc diameter and the cube of the
circular speed and as shown
3. Friction work is also dependent on an empirical friction factor
fr which is defined by geometric parameters of the disc and the
rotational Reynolds number as shown

8/31/2017 Department Name Confidential 42


Numerical Model of a Thermal System
Ventilation Losses:
1. Similar to friction, ventilation work is also proportional to
density, the cube of the circular speed and a blade aspect
coefficient, as shown This energy dissipation, also known as
windage losses, occurs in the blade passages of the turbine
instead of the discs.
2. When the turbine is partially loaded, a group of blade passages
along the circumference of the disc will be filled with stagnant
steam. Owing to the rotation of the disc, the steam filling these
passages is subject to centrifugal forces and moves
from the roots to the tips of moving blades

8/31/2017 Department Name Confidential 43


Numerical Model of a Thermal System
Ventilation Losses:
1. The appearance of a vortex zone across the blade produces an
increase in steam and blade metal temperatures
2. This effect is a characteristic of the last stage blade (LSB) of a
turbine given the dependency of Equation on blade length
3. This relation is only valid in case of low volumetric flow
through the steam turbine blading.

8/31/2017 Department Name Confidential 44


Numerical Model of a Thermal System
Turbine transients-thermal stress & fatigue:

where is the Young modulus,


is the mean linear expansion coefficient,
is the traverse expansion coefficient,
is the mean integral temperature and
is the initial temperature

8/31/2017 Department Name Confidential 45


Numerical Model of a Thermal System
Thermal Stress & Fatigue
1. Differential thermal expansion

2. Relative thermal expansion axial clearances

8/31/2017 Department Name Confidential 46


Numerical Model of a Thermal System
Solar Steam Turbine- Geometric modeling:
1. A SST-700RH steam turbine, currently employed in
several parabolic trough CSPPs, was used as the
reference turbine in this work.
2. This 50MW turbine includes a reheat system, and is
thus made of HPT and LPT units; both units were
considered in the performed studies.
3. The geometric modeling methods were the modular
approach and the continuous approach based on
previous work performed in KTH

8/31/2017 Department Name Confidential 47


Numerical Model of a Thermal System
Model Validation on Solar Steam Turbine-
Geometric modeling:
1. A SST-700RH steam turbine, currently employed in
several parabolic trough CSPPs, was used as the
reference turbine in this work.
2. This 50MW turbine includes a reheat system, and is
thus made of HPT and LPT units; both units were
considered in the performed studies.
3. The geometric modeling methods were the modular
approach and the continuous approach based on
previous work performed in KTH
8/31/2017 Department Name Confidential 48
Numerical Model of a Thermal System
Geometric modeling-modular vs continuous geometry:
1. It is not simple to access information on the entire shape &
specs making the continuous approach inflexible and
cumbersome to adapt to new geometries.

8/31/2017 Department Name Confidential 49


Numerical Model of a Thermal System
Temperature distribution:
Modular and continuous approaches maintain the same
temperature ranges with the maximum and minimum located at
analogous positions. Comparing each module with the
corresponding area in the continuous geometry it can be seen
that the overall distribution of temperatures within the turbines
is well captured.

8/31/2017 Department Name Confidential 50


Numerical Model of a Thermal System
Measured Data - 96 Hours: Steam Pressure &
Temperature

8/31/2017 Department Name Confidential 51


Numerical Model of a Thermal System
Comparison of simulated data with measured data for
HPT stage:

8/31/2017 Department Name Confidential 52


Numerical Model of a Thermal System
Comparison of simulated data with measured data for
LPT stage:

8/31/2017 Department Name Confidential 53


Numerical Model of a System
Relative errors:

8/31/2017 Department Name Confidential 54


Numerical Model of a System
Relative errors: Large degree of agreement with respect to the
measured data in spite of the different detail levels.

8/31/2017 Department Name Confidential 55