14 epfl-valais

©Francois Marechal -IPESE-IGM-STI-EPFL 2014
IPESEIndustrial Process and
Energy Systems Engineering
E-transition : the role of the
E-technology for the Energy
transition
Prof François Marechal
Industrial Process and Energy Systems Engineering
EPFL VALAIS-WALLIS

• E-technology for engineering
– E-technology for modelling
– E-technology for data structuring
– E-technology for decision support
– E-technology for understanding
• E-technology for operating
• E-technology for teaching
E-techonology for the energy transition

Socio - ecologic - economic environment
The energy system
Conversion systems
Waste
Emissions
Raw materials
Resources
Price
Availability
t
t
Products &
Energy services products
products
Demand
Price
t
t

Socio - ecologic - economic environment
The energy transition
Conversion systems
Raw materials
Resources
Price
Availability
t
t
Waste
Emissions
Products &
Energy services products
products
Demand
Price
t
t

• For a given city
Energy services for the energy transition
Energy services
‣Technologies ?
‣Min. Costs and emissions
‣ Heat using existing district heating network (seasonal variation in T and
load): 3357 kWhth/yr/cap
Waste to be treated (existing
facilities for MSW and WWTP):
Resources
‣ Electricity (seasonal variation): 8689 kWhe/yr/cap
‣ Mobility: 11392 pkm/yr/cap
‣ MSW: 1375 kg/yr/cap ‣ Wastewater: 300 m3/yr/cap
‣ Biowaste: 87.5 kg/yr/cap
‣ Woody biomass: 18’900 MWhth/yr
‣ Sun (seasonal variation in T and load): 10‘328 MWhth/yr ‣ Hydro (existing dams): 187‘850 MWhe/yr
‣ Geothermal: 9496 MWhth/yr
?
?
?
?
?
?
?
40’000 inhabitants city in Switzerland (1000m alt.)

• Characterise and localise the technologies to be used to supply the energy
services and the products needed over the lifetime of the system’s element
at the time they are needed with the resources that are available
– Decisions to be taken
• Technology : choice, size and location
• The operating strategy
• The services that are provided
• The resources that are used
– Criteria
• Minimum cost = OPEX + APEX
• Maximum profit = NPV (i,lifetime,cost)
• Minimum environmental impact = LCIA
• Max renewable energy integration
– Constraints
• Mass & Energy Balances
• Context specs
• Bounds
Engineering the energy transition
Yu,l, Su,l
fu,l,t(p) 8p, u, l
fu!s,l,t(p) 8p, u, l, s
fr!u,l,t(p) 8p, u, l, r

E-technology for engineering

The energy systems engineering methodology
Solutions
Energy services
Resources
Context & Constraints
Superstructure
System Boundaries
System performances indicators
•Economic
•Thermodynamic
• Life cycle environmental impact
Results analysis
•Exergy analysis
•Composite curves
•Sensitivity analysis
•Multi-criteria
Technology options
Models
250
300
350
400
450
500
550
600
0 5000 10000 15000 20000 25000 30000
T(K)
Q(kW)
Cold composite curve
Hot composite curve
Heat &  
Mass integration
Decision variables
Solving method
Thermo-economic Pareto
Multi-objective
Optimization
Out=F(In,P)

Multi-scale problem 9
9
Common element is the energy planning
• Energy consumption profile determination
• Energy performance evaluation
• Energy baseline generation
• Identification and evaluation of improvement options
• Implementation and monitoring
Step 1
Step 2
stematic approach
• Methodologies
• Tools
Step 3
Details
Scale
Technologies
Processes
Industrial Clusters
Districts
Regions
E-technology for decision support
* Modelling the interactions
* Identifying the synergies
* Implementing the best interactions 
=> Network or Grid
Materials
Engineering :
from materials development
to system integration

• Model of the interconnectivity (mass, heat, energy)
• Model of the emissions (Equipment, Emissions)
• Model of the cost (size => cost, maintenance)
E-technology for modelling
- Equipment sizing model
- Cost estimation
Heat transfer requirement
Heat transfer
Thermo-chemical conversion
Model
Material streams
Product streams
Electricity
Heat transfer requirement
Water streams
Water streams
Waste streams
Waste streams
Electricity
Unit parameters
Decision Variables
Life cycle emissions
Life cycle of equipment
- Production
- Dismantling
Maintenance
Investment cost
LCA models
LCA models

E-technology for integration
Unit documentation
Description
Purpose
Limitation
Authors
Unit model report
Summary of results
Important data
Errors
Validity
Unit execution handler
Input data
Calculation engines
Existing ﬂowsheeting software
Mimetic models
Output data
Integration/interconnectivity
Costing
LCA => Ecoinvent LCI
Model ontologyModel encapsulation
Common interface of a module for system integration
Flowsheeting tools
•BELSIM-VALI
•gPROMS
•ASPEN plus
•HYSYS
•Matlab
•Simulink
•(CITYSIM)
•MODELICA
•Others possible
•CAPE-OPEN ?
•PROSIM
Unit connectivity
Layers connections
Layer ontology
Data base
Model
Data base
Knowledge
Data base

A variable is the smallest data structure handled by OSMOSE. A variable is defined by

1 Name or tagname that litteraly identifies the variable

2 ID that is tagging the variable

3 Parent indicates the module to which the variable belongs

4 Value that defines one instance of the variable => Matrix

A priori value of the variable will exist for any time in the time sequence, for any period and for any scenario in
which the system is calculated.

5 MinimumValue and a MaximumValue that defines the bounds for the value of the variables

6 DistributionType and DistributionParameters that defines the uncertainty of the value

7 DefaultValue that defines the value of the variable when it is created. this is also the recommended order of
magnitude of the value recommended for the calculations

8 PhysicalUnit that defines the physical unit of the value of the variable

9 DefaultPhysicalUnits that defines the physical unit with which the calculation have to be made, It means that a
preprocessing phase will have to convert the value from the PhysicalUnits to the DefaultPhysicalUnits as well
as a post-processing phase that is going to convert from the DefaultPhysicalUnits to the PhysicalUnits .

10 Status that identifies if the variable has a fixed value or a value that will be calculated. The status should identify
as well if the value is calculated for each time or if it is constant or if it requires recalculation for each time.
Status is equivalent to ISVIT in OSMOSE.

i x : value is unique for the scenario (i.e. 1 value/scenario)

ii x0 : value is period dependent (i.e. 1 value/period and / period)The value may calculated during
preprocessing phase using the model (i.e. constant that is period dependent).

iii x00 : the value is time dependent (i.e. 1 value/time and period and scenario). 
where x = 
1 : for constant to be fixed by the user or the optimizer
2 : for constant fixed by the user or the optimizer otherwise use the default value
-1 : for values calculated by the model
11 Dependence Matrix a dependence matrix that identifies the dependence of the value with respect to the other
values of the problem is stored. Each of the dependence includes the mention if the dependence is linear, in
which case the value of the dependence is stored in the field as a couple (ID,numerical value), or it can be
stored as a reference to a non linear dependence that could be a surrogate model or the access to the model
that calculates the dependence.
Ontology of a value : E-technology for knowledge structuring
˙mu,l(t), Yi,u,l(t), Xi,u,l(t), ⇧i,u,l(t), KPIk, ˙m⇤
u,l, Y ⇤
i,u,l, X⇤
i,u,l, ⇧⇤
i,u,l

E-technology : Models & knowledge data base

LENI Systems
Integrating heat recovery technologies in the superstructure
Process design
WOOD Natural Gas (SNG) CO2 (pure)

• Explicit
– In flowsheeting software
• Automatic
– Based on the connectivity description
– Restricted matches
– Routing
• Implicit
– Heat cascades
• Combined
– Mass and Energy
Modeling the systems interactions
" p
! le niveau de pression total du flux délivré P p .
En contrepartie, chaque puits est défini par :
! la pureté requise du flux d'entrée Xc
! les limites de composition en impuretés du mélange admis xc , j
max
! un besoin de débit d'hydrogène "mc
! la pression totale requise du mélange à l'entrée Pc .
Le nombre des puits et sources, regroupé au sein d'une même unité, est égal au nombre de niv
de pureté des flux d'hydrogène admis, respectivement rejetés. Dans un premier temps, on peut
chaque unité possède au maximum 1 consommateur et 2 producteurs. Un nombre plus élevé d
éléments par unité correspond à un affinement du modèle, et entraîne une augmentation de la
connexions au sein de la superstructure
On a vu que les unités regroupent, dans une même structure, des puits et des sources qu'on a
consommateurs et producteurs.
Elles sont définies par :
! l'identité des puits et sources qui la compose,
! leurs coordonnées (X,Y) géo-référencées,
! un identifiant, qui permet de les regrouper en sous-systèmes pour l'analyse dans les diagram
Pincement (cf. Chapitre "Représentation graphiques").
Figure 7.1 : Schéma de la superstructure complexe
250
300
350
400
450
500
550
600
0 5000 10000 15000 20000 25000 30000
T(K)
Q(kW)
Hot composite curve

E-technology for decision support
Flowsheeting tools
•BELSIM-VALI
•gPROMS
•ASPEN plus
•HYSYS
•Matlab
•Simulink
•(CITYSIM)
•MODELICA
•Others possible
•CAPE-OPEN ?
•PROSIM
•UNISIM ?
Energy technology data base
•Data/models interfaces
•Simulation
•Process integration interface
•Costing/LCIA performances
•Reporting/documentation
•Certiﬁed dev procedure
Modeling tools integration
MILP/MINLP models
Heat/mass integration
Sub systems analysis
Superstructure
HEN synthesis models
Optimal control models
MILP/ AMPL or GLPK
Multi-period problems
Sizing/costing data base
LCIA database (ECOINVENT)
Process integration
Grid computing
Multi-objective optimisation
Evolutionary - Hybrid
Problem decomposition
Uncertainty
Decision support
GIS data base
Industrial ecology
Urban systems
GUI : Spreadsheets, Matlab
Data Structuring
Technology models data base
Energy conversion
Sharing knowledge

LENI Systems
Thermo-economic optimisation
Trade-o s: e⇥ciency and scale vs. investment
E⇥ciency vs. investment and optimal scale-up:
62 63 64 65 66 67 68
900
1000
1100
1200
1300
1400
1500
1600
energy efficiency [%]
specificinvestmentcost[€/kW]
trade-off: efficiency vs.
investment (& complexity)
TECHNOLOGY:
drying: air, T & humidity optimised
gasification: indirectly heated dual fluid. bed (1 bar, 850°C)
methanation: once through fluid. bed,
T, p optimised (p = [1 15] bar)
SNG-upgrade: TSA drying (act. alumina)
3-stage membrane: p, cuts optimised
quality: 96% CH4, 50 bar
heat recovery: steam Rankine cycle
T, p & utilisation levels optimised
input: 20 MW wood at 50% humidity (~4t/h dry)
0 20 40 60 80 100 120 140 160 180
800
1000
1200
1400
1600
1800
2000
2200
2400
input capacity [MW]
specificinvestmentcost[€/kW]
scale-up objective: minimisation of production costs
(incl. investment by depreciation)ε ~ 62%
ε ~ 66%
ε ~ 64%
ε ~ 68%
optimal configurations:
increasing efficiency
discontinuities due to
capacity limitations of
equipment (diameter < 4 m)
1
nb. of
gasifiers: 2 3 4 5 ...
61 / 87
Gassner, Martin, and François Maréchal. Energy & Environmental Science 5, no. 2 (2012): 5768 – 5789.
• Thermo-economic Pareto Front
E-technology for identifying solutions

Some results
Cmparing technologies and processes
Thermo-economic Pareto front
(cost vs e ciency):
LENI Systems
Quelques résultats
Comparaison des technologies
Optimisation de toutes les combinaisions technologiques
(coût et é cacité):
gaz. préssurisé à chau age direct est la meilleure optionThe best solution is the pressurised directly heated gasifier
• Thermo-economic Pareto optimal processes
E-technology : Observing the decision space
Gassner, Martin, and François Maréchal. Energy & Environmental Science 5, no. 2 (2012): 5768 – 5789.
Note : 1.5 years of calculation time ! => parallel computing
Efficiency vs Investment
Synthetic Natural Gas Process

• Graphical representations
– Composite curves
– Sankey diagrams
• PDF reports
• Sensitivity analysis
• Pareto curves => data base of solutions
– Decision variables values
– Uncertainty analysis => most probable optimal solutions
E-technology for understanding the results
250
300
350
400
450
500
550
600
0 5000 10000 15000 20000 25000 30000
T(K)
Q(kW)
Hot composite curve
64 66 68 70 72 74 76 78 80 82 84 86 88
600
700
800
900
1000
1100
1200
1300
SNG efficiency equivalent εchem [%]
SpecificinvestmentcostcGR[USD/kW]
pressurisation
pressurisation, gas turbine integration,
steam drying
& hot gas cleaning
hot gas cleaning & steam dryingwith heat cogeneration
453 K
513 K
5 %
35 %
1073 K
1173 K
573 K
873 K
1 bar
50 bar
573 K
673 K
573 K
673 K
0
1
40 bar
100 bar
623 K
823 K
323 K
523 K
SNG Output
minimal
maximal
min
max
min
max
min
max
Td,in Φw,out Tg Tg,p
pm Tm,in Tm,out yRankine
ps,p Ts,s Ts,u2
0 20 40 60 80 100 120
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
SNG proccess configuration number
Probabilitytobeintop10[−]
Prod. cost
Res. profitability
BM Break even
4.3. COMPARISON WITH CONVENTIONAL LCA 45
-500 %
-400 %
-300 %
-200 %
-100 %
0 %
100 %
200 %
scale-independent,
conventional LCIA
(average lab/pilot tech.)
without cogeneration
Integrated industrial base case scenario (8 MWth)
with cogeneration
(a) Overall impacts
0
5
10
15
20
25
Remaining processes
Rape methyl ester
NOx emissions
Solid waste
Charcoal
Olivine
Infrastructure
Wood chips production
Electricity cons./prod.
Avoided NG extraction
Avoided CO2 emissions
UBP/MJwood
scale-independent,
conventional LCIA
(average lab/pilot tech.)
harmful
beneficial
harmful
beneficial
harmful
beneficial
without cogeneration
Integrated industrial base case scenario (8 MWth)
with cogeneration
(b) Detailed contribution
Figure 4.3: Comparison of LCIA results obtained by conventional LCA and the proposed
methodology for Ecoscarcity06, single score, with detailed process contributions (Gerber
et al. (2011a))
The differences between the conventional LCA and the base case scenario without and with
a Rankine cycle are mainly due to the improved conversion efficiency obtained by applying
process design method, which has a direct influence on the quantity of SNG produced per
unit of biomass. In the base case scenarios, the efficiency is considerably higher because
the energy integration is optimized in the process design. As a direct consequence, the

E-technology : Large renewable energy integration
Use B M
Use G M
Photo
synthesis
S
CO2 CH4
DH
TS
EL
SOEC M
SOFC
RNE
Biomethane
Gasiﬁcation
Storage
Districtheating
Co-electrolysis
Electrolysis
Renewable Electricity
Sun
Biomass
CO2
H2O2
CH4
Heat
Renewable Electricity
Sun
Wind
Grid Electricity
Fuel appl.
O2
H2O
H2O
P MSun
Capture
atm. CO2
Capture
Flue gases
B : Biomethanisation
C : CO2 Capture
G : Gasiﬁcation
M : methanation
S : CO2 separation
P : Photocatalysis for CO2 reduction
SOFC : Solid Oxide Fuel Cell
SOEC : Solid Oxid electrolysis Cell
ET : Electro thermal storage
CO2 : CO2 Storage
CH4 : Methane Storage
DH : District Heating
CO2
H2
O2
CH4
H2O
Elec
Ren Elec
atm. CO2
CO2
H2O
EPFL Valais-Wallis Demonstrator
Integration of processes in the conversion chain : CO2 valorisation options

•Coordinates attributed to locations
– GIS : Data bases
• Buildings
• Resources
– GIS :Infrastructure
• Roads, district heating network,
electricity grid
– GIS Tools : Routing, Layers, Representations
•Link between GIS data and Models
–Link with unit models parameters
–Access to GIS tools in the workflow
E-technology : Data bases and Geographical Information Systems
Location i, (xi, yi, zi)
Location 1, (x1, y1, z1)
Location 2, (x2, y2, z2)
Geneva
T T
TT -20
0
20
40
60
80
0 50 100 150 200 250 300 350 400
T(C)
Q(kW)
AirWaste Water
Heating Hot water
recovery
Recoverable
Heat requirement
At a given location : combination of energy technologies

E-technology : defining the interactions
Eexp
t,k Eletricity exported to
Egrid
t,k Electricity bought fro
Eaw
t,k Electricity consumed
Eww
t,k Electricity consumed
Eloss
t,k Electricity losses durin
Epump
t Pumping power for th
Etech
t,k,e Electricity produced o
Fm Maintenance factor [-]
Gast Overall gas consumpt
Lmin
e Minimum allowable p
Mbuild
t,k
Mass flow of water fr
building at node k to
Mpipe
t,i,j,p
Mass flow of water flo
j [kg/s]
Mmax
t,p,i,j
Maximum mass flow o
all periods of time [kg
Mtech
t,k
Mass flow from the ne
period t [kW]
MTbuild Mass flow from the ne
Tcons
t,k Temperature at whi
Thot Temperature of the
Tnet Design temperature
v Velocity of the wate
X = 1 if a device exist
Xaw
k = 1 if there is an ai
Xboiler
k = 1 if there is a boi
Xgas
e = 1 if device e need
Xnode
k = 1 if a device e is i
Xtech
k,e = 1 if a device can b
Xww
k = 1 if there is an wa
Y i,j,p = 1 if a connection
Greek letters
Theat Pinch at the heat-exchangers [
Tnet ww Temperature di⇥erence of the
boiler Thermal e⌅ciency of the boile
el
e Electric e⌅ciency of device e [-
Networks superstructure
1 Symbols
Roman letters
An Annuities for a given investment [-]
Ai,j,p Area of the pipe between nodes i an
B Arbitrarily big value
Caw
Investment costs for air/water heat
Cboiler
Annual investment costs for the boi
Cfix
Fix part of the investment costs for
Cgas
Total annual natural gas costs [CHF
cgas
Natural gas costs [CHF/kWh]
Cgrid
Total annual grid costs [CHF/year]
cgrid
Grid costs [CHF/kWh]
Cinv
Investment costs of a given device [C
Cinv
an Annual investment costs of a given
Cpipes
Total annual costs for the piping [C
Cprop
Fix part of the investment costs for
Cref
Investment costs of a device chosen
Q1..T
E1..T
Q1..T
E1..T
Q1..T
E1..T
Q1..T
E1..T
Q1..T
E1..TQ1..T
E1..T
Q1..T
E1..T
Q1..T
E1..T
Cinv
=
ne
e=1
nk
k=1
ae + ae Se,k
Investment
Cgas
=
ne
e=1
nk
k=1
np
t=1
Ht
Qe,k,t
th
e
1 Symbols
Roman letters
An Annuities for a given investment [-]
Ai,j,p Area of the pipe between nodes i and j of network p [m2
]
B Arbitrarily big value
Caw
Investment costs for air/water heat pump(s) [CHF]
Cboiler
Annual investment costs for the boiler(s) [CHF/year]
Cfix
Fix part of the investment costs for a given device [CHF]
Cgas
Total annual natural gas costs [CHF/year]
cgas
Natural gas costs [CHF/kWh]
Cgrid
Total annual grid costs [CHF/year]
cgrid
Grid costs [CHF/kWh]
Cinv
Investment costs of a given device [CHF]
Cinv
an Annual investment costs of a given device [CHF/year]
Cpipes
Total annual costs for the piping [CHF/year]
Cprop
Fix part of the investment costs for a given device [CHF/kW
Cref
Investment costs of a device chosen as reference [CHF]
Cww
Investment costs for water/water heat pump(s) [CHF]
CO2
gas
Total annual CO2 emissions due to the combustion of natura
co gas
CO emissions due to the combustion of natural gas [kg/kW
Roman letters
An Annu
Ai,j,p Area
B Arbi
Caw
Inves
Cboiler
Annu
Cfix
Fix p
Cgas
Tota
cgas
Natu
Cgrid
Tota
cgrid
Grid
Cinv
Inves
Cinv
an Annu
Cpipes
Tota
Cprop
Fix p
Cref
Inves
Cww
Inves
CO2
gas
Tota
co2
gas
CO2
CO2
grid
Tota
co2
grid
CO2
Maintenance
Gas Grid
Electricity
Electricity
Emissions
c Gr
Cinv
In
Cinv
an An
Cpipes
To
Cprop
Fi
Cref
In
Cww
In
CO2
gas
To
co2
gas
CO
CO2
grid
To
co2
grid
CO
COPhp
Co
COPaw
t,k Co
COPww
t,k Co
cp Iso
Disti,j Di
Ht Nu
Econs
t,k El
Eexp
t,e El
Egrid
t,k El
Eaw
t,k El
Eww
t,k El
Eloss
t El
Epump
t Pu
Etech
El
B Arb
Caw
Inve
Cboiler
Ann
Cfix
Fix
Cgas
Tot
cgas
Nat
Cgrid
Tot
cgrid
Grid
Cinv
Inve
Cinv
an Ann
Cpipes
Tot
Cprop
Fix
Cref
Inve
Cww
Inve
CO2
gas
Tot
co2
gas
CO
CO2
grid
Tot
co2
grid
CO
COPhp
Coe
COPaw
t,k Coe
COPww
t,k Coe
cp Isob
Disti,j Dist
H Num
Industry
Power
plant
Waste
treat.
Building plant.
. [5] Francois Marechal, Celine Weber, and Daniel Favrat. Multi-Objective Design and Optimisation of Urban Energy Systems, pages 39–81. Number ISBN: 978-3-527-31694-6.Wiley, 2008.
Water
Fuel
Water
Fuel
Industrial plant.

• Mid-season typical day demand
E-technology : Integrating renewable sources
Time
ExergyPowerrequirement[kW]
Heat
Electricity
PV production

Fuel-Cell GT
Domestic
Hot water tanks
Heat pump/HVACElectrical gridNatural gas grid
Buildings
Small industries
E-technology : Smart households
Smart Optimal Predictive Control Management Box
Price
T ambient
Comfort
T/Air
People
Light
T storage
Batteries
PV
Solar panels
Heating/Cooling
tanks
Reserve
Cons. proﬁles
Big Data grid
Water grid
Meteo
Smart info (WIFI-GPS-GSM)
T storage
Mass
Waste Water Grid
Cogeneration
CO2 grid

MILP process integration model @location k
Water: mii,
Tin, hin
return : moo,
Tout, hout
Water: mio,
Tin, hin
return : moi,
Tin, hin
Technologies w, @location k period s and time t(s)
Demand @location k, period s and time t(s)
Storage tank
2
Tst,2
HEX h1
(t)
heat demand
Storage tank
1
Tst,1
...
Tst,...
Storage tank
l
Tst,l
...
Tst,...
Storage tank
nl
Tst,nl
HEX h2
(t) HEX h..
(t) HEX hl
(t) HEX hnl-1
(t)
heat demand heat demand heat demand heat demand
heat excess
HEX c1
(t)
heat excess
HEX c2
(t)
heat excess
HEX c..
(t)
heat excess
HEX cl
(t)
heat excess
HEX cnl-1
(t)
Units for heat integration
HL HL HL HL HL HL
HEX: Heat exchangers
HL: Heat losses
Storage system
S. Fazlollahi, G. Becker and F. Maréchal. Multi-Objectives, Mul.-Period Optimization of district energy systems: II-Daily
thermal storage, in Computers & Chemical Engineering, 2013a
Buildings inertia
Heat exchange by heat
cascade model
Electricity balance
Existence : w in location k

• Energy system = for each VPP (n)
– Investments (expected life time : 20 Years)
• Energy Conversion Units
• Storage
– Operating costs (expected operation time = 25x8760h)
• Management strategy
– Constraints
• Energy services
• Grid capacities
• Grid constraints
Energy system design problem
8u 2 Units; 8n 2 Nodes ) Sizeu,n
8s 2 Storage; 8n 2 Nodes ) Vs,n
8s 2 Storage; 8n 2 Nodes; 8t 2 time ) ˙mu,n(t), Vs,n(t), ˙E(t)
Z T ime
t=1
(
unitsX
u=1
(c+
r (t) ˙mr(u),n(t)) + c+
e (t) ˙E+
(t))dt
8t 2 Time :
NodesX
n=1
˙E(t)  ˙Emax(t)
8n 2 Nodes; 8t 2 Time; 8c 2 Cons. ) ˙Qn,c(t), ˙En,c(t)
NodesX
n=1
UnitsX
u=1
1
⌧y,i
(I(Sizeu,n) + I(Vs,n))
8q 2 quarter(Time) :
NodesX
n=1
Z q+15
t=q
˙En(t) ˙En(t 1) dt  ˙Emax(t)

E-technology : decision support tool
Predeﬁned data
Project data base
OSMOSE LUA
Saved user data
Energy technologies
Problems
Modules
Default values
Thermodynamic
Sizing
Costing Life Cycle Inventory
EcoInvent
Work ﬂow 
Models
Solving methods
in locations
GIS data base
GUI
Results data base
Layer Ontology
Superstructure
Yi,u,l(p(t)) = Fu(Xi,u,l(p(t)), ⇧i,u,l(p(t)))

Design and Control of thermo-electric energy systems
Single family house : 4 pers 160 m2 : with heat pump
E-technology : Optimal design of solar systems
19.11.2015
Fig 4. Pareto front for HP system configuration
I
II
III
Long term storage : 85%
SF =
kWP V
kWbuilding
Self sufﬁciency (SF)
PV production/needs
Self consumption (SC)
PV production used on site
SC
SF
SF and SC as a function of the PV area
Off-site storage
Extra Electricity from PV
Used by the building later
SC =
kWused
kWP V

Single family house : 4 pers 160 m2 : with heat pump
Self sufficiency ?
19.11.2015
Fig 4. Pareto front for HP system configuration
I
II
III
CaseI
Energy system Off-site storage
PV array 88 m2
Battery 4.95 kWh
HW tank 2.43 m3
Heat Pump 3.59 kW
Redox Battery
8.14 MWh
406.9 m3 (20 Wh/l)
4’070’000 € (500 €/kWh)
CaseIII Energy system Off-site storage
PV array 156.9 m2
Battery 8.63 kWh
HW tank 2.39 m3
Heat Pump 3.7 kW
Redox Battery
17.1 MWh
854.6 m3
8’550’000 €
CaseII
Energy system Off-site storage
PV array 109.7 m2
Battery 7 kWh
HW tank 2.46 m3
Heat Pump 3.5 kW
Redox Battery
10.8 MWh
540.2 m3
5’400’000 €
Annual energy balance

Comparison – typical operating periods (winter, summer, mid-season)
E-technology : model predictive control
19.11.2015
II
III
IV
Shaded self-consumption
Blue Import [kW]
Green Export [kW]
Red Generation [kW]
Winter Summer Mid-season
Case III
Case II
Case IV

• Self Sufficiency : + 40 %
• Self Consumption : + 68 %
E-technology : Model predictive control
SF + 41 %
SC + 68 %
NetZero
without MPC

E-technology : technology coordination & market interaction
High electrcity cost during the afternoon
Storage tank = 200 m3
Low cost cost during the afternoon
Storage tank = 200 m3
Heating : 72315 kWh
Electricity : 77897 kWhe
Electricity in : 99596 kWhe
Electricity out : 8710 kWhe
Low price period
Electricity bought : 62894 kWhe
Low price period
Balance : -3138 kWhe
Storage : 22480 kWhe
Engine : 2000 kWe
Heat pump : 2000 kWe
Storage 200 m3
Demand mean heating power = 3000 kW
Storage filling at night
Empty storage tanks before cheap elec price
Fill storage tanks during cheap elec price

E-technology for operation

• From design to operation
E-technology : Optimal management box
Optimization
Problem
Resolution
MANAGEMENT
UNIT
MILP PROBLEM DESCRIPTION
(AMPL )
Optimization Problem
data set -up
Predicted
data
Previous
states
.mods .dats
ush*uvlv*ub*ucg*
System System
parameter
values
System Model
Prediction
Actual
State data
OPTIMIZATION (CPLEX )
States database
(Current state & previous states )
(House structure - Matlab )
Data acquisition
and database
Tariff info
Tdhw
Tsh
Text
Tin
Equations &
constraints
Collazos et al., Computers and Chemical Eng. 2009
Grid connexions
Sensors
Operation set-points

• Predictive Control Algorithm : Moving horizon
– hour 1 : set-point control + 24 h Cyclic : strategy
Predictive Controller

• Electricity cost is changing
Response to the market price variation
Same heat
EEX market price
40% heating cost reduction

• Predictions
– Ambient conditions
– Presence + comfort
– Prices
– Energy services
• Multi-nodes
– Interactions via sub grids
• Electrical (different type)
• Heat/Cold
• Information
• Integration with main grids
– Electrical grid
– Gas grid
Challenges for smart grids systems
Learning algorithm
Images, Courtesy:EngineerLive and Consumer Energy Report
Producers
Consumers

Fuel-Cell GT
Domestic
Hot water tanks
Heat pump/HVACElectrical gridNatural gas grid
Buildings
Small industries
E-technology : smart households operation
Price
T ambient
Comfort
T/Air
People
Light
T storage
Batteries
PV
Solar panels
Heating/Cooling
tanks
Reserve
Cons. proﬁles
Big Data grid
Water grid
Meteo
Smart info (WIFI-GPS-GSM)
T storage
Mass
Waste Water Grid
Cogeneration
CO2 grid

E-technology : Optimal process control
e optimization strategy presented by Weber et al. ([17]) allows to d
ystem and the storage tanks considering the use of a predictive optimal
In addition, methods like the one proposed by Collazos et al. ([11]) can
the predictive optimal management strategy in a control system.
6
Process
hot water
CIP system
Cooling
Glycol
110 °C
50 °C
2 °C
Engine
Gas
Industrial
process
Refrigeration/
Heat Pump
Electricity
Figure 8: Storage tank system conﬁguration

E-technology : car power train management
vehicles, both conventional and hybrid, are assessed on this cycle in terms of fuel co
and emissions. The cycle, illustrated in Figure 6.2, is composed of four repetitions
cyle (UDC) and one extra-urban cycle (EUDC). Basic characteristics of the cycle a
Table 6.2.
0 200 400 600 800 1000 1200
0
20
40
60
80
100
120
140
Time [s]
Vehiclespeed[km/h]
Figure 6.2 New European Driving Cycle.
Table 6.2 NEDC characteristics.
Distance [m] Duration [s] Average speed [km/h] Rep
UDC 1’017 195 18.54
EUDC 6’956 400 62.22
Total cycle 11’023 1’180 32.26
6.3 Simulation results
The engine power and fuel consumption profiles are given in Figure 6.3. During d
phases fuel injection is cut o↵, as is the case for modern engines. Figure 6.4a giv
consumption points relative to the engine consumption point. The NEDC does clear
much di culty for the 2.2 liter engine in terms of load as the consumption points a
in the lower left quadrant. However this zone is not near the region of higher e cie
Table 6.3 compares the simulated emissions result with that of the real vehicl
error of less than 2%, the thermal powertrain model is deemed su ciently precise.
52
Collaboration with PSA
Driving cycle
5 PNEUMATIC HYBRID
Driving cycle
t
V
h
CVT control
Strategy estimator
mode = f(P, T, T0
s, !s)
Vehicle estimator
m ˙V =
P
F
Optimal ratio
2 ( min, max)
min( ˙mfuel)||min( ˙mair)||max( ˙mair)
V ˙V
mode
!w
Tw
Vehicle
m ˙V =
P
F
Vdelay ˙Vdelay
CVT
!s
!w
=
Ps = f(Pw, ⌘)
!w Tw
Energy converter
Alternator
T0
s = Ts + TAT !s > 0
Strategy
mode = f(P, T, T0
s, !s)
Combustion
˙mfuel = f(!s, T0
s)
Pneumatic motor
˙mair = f(P, T, T0
s, !s)
˙hair = f(P, T, T0
s, !s)
Pneumatic pump
˙mair = f(P, T0
s, !s)
˙hair = f(P, T0
s, !s)
Mode filter
!s Ts
!s T0
s
!s T0
s
!s T0
s
!s T0
s
˙hair,pm ˙mair,pm
˙hair,pump
˙mair,pump
˙mfuel
mode
Fuel tank
fEmissions = f(mfuel) [g CO2/km]
˙mfuel
Air tank
Energy balance
Heat transfer model
˙mair ˙hair
Ptank
Ttank
Ptank
Ttank
Figure 5.9 Pneumatic hybrid powertrain model flowchart.
4 THERMAL ELECTRIC HYBRID
DT
C1
G
PSD
C2
ICE
FT
M
⇠PE
BT
SC
Figure 4.3 Parallel thermal electric hybrid powertrain: BT: Battery pack; C1: Primary clutch; C2:
Secondary clutch; D: Di↵erential; FT: Fuel tank; G: Alternator; ICE: Internal combustion engine;
M: Electric motor; PE: Power electronics; PSD: Power split device; SC: Supercapacitor pack; T:
Transmission.
DT
C1
G
PSD
C2
ICE
FT
M
⇠PE
BT
SC
(a) Thermal traction
DT
C1
G
PSD
C2
ICE
FT
M
⇠PE
BT
SC
(b) Electric traction
DT
C1
G
PSD
C2
ICE
FT
M
⇠PE
BT
SC
(c) Hybrid traction
DT
C1
G
PSD
C2
ICE
FT
M
⇠PE
BT
SC
(d) Regenerative braking
Figure 4.4 Parallel hybrid powertrain operating modes.
• ICE -> Drive
• Fuel cell
• Batteries
• El Motor
• Pneumatic storage
• Transmission line
• HVAC
• Energy conversion
=> Size & Weight & Cost
Operation strategy
13 UTILITY INTEGRATION
−25 −20 −15 −10 −5 0 5
250
300
350
400
450
500
550
600
Heat Load [kW]
Temperature[K]
Others
utilities
(a) Operating point: 2000 RPM and 2.2 bars.
−25 −20 −15 −10 −5 0 5
−0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
Heat Load [kW]
CarnotFactor1−Ta/T[−]
Others
utilities
(b) Operating point: 2000 RPM and 2.2 bars.
Figure 13.2 Organic Rankine cycle integration. Text = 0 C
Waste Heat Recovery
* ORC integration
Component set
179
i objective optimization results for hybrid
tric vehicles with different usages
w European Driving Cycle (NEDC)
of a two objectives optimization converged on a Pareto Frontier optimal curve
representing the trade-off between the energy consumption and the cost of the
rmalized driving cycle.
areto curve energy consumption to cost (color bar in thousands of Euros) –
body mass of around 1500 kg is characterizing the D–class vehicles. These
sually used for family transportation or business trips on long distances. The D-
are well equipped and the technology is considered as additional value for the
powertrain technology for low CO2 emissions is an important requirement for
of these vehicles. The official technical characteristics are referenced on the
car makers. The Pareto curve for the NEDC (Figure 8-6) defines a wide range
mproved powertrain efficiencies (25- 45.2%) and fuel emissions (30 to 140 g
contrast, a thermal powertrain D-class vehicle of 1660 kg of mass, with 2.2
ngine yields 17 % of powertrain efficiency and 151 g CO2 / km [147]. To
high powertrain efficiency in the Pareto curves for all usages, the electric half
rain is increased and more electric energy stored in the battery is needed.
ttery capacity increases from 5 to 50 kWh (Figure 8-7 d). This considerably
vehicle mass (Figure 8-7 a). To follow the dynamic requirement on the cycle,
otor power also increases from 20 to 150 kW. The emitted tank-to-wheel CO2
ich are proportional to the diesel consumption, are related to the hybridization
d through the electric motor and the thermal motor size.
HEV P-HEV REX
Pneumatic hybrid powertrain
dation of the thermal powertrain underlying the pneumatic hybrid.
eugeot 308 (Figure 7.1) with a 1.2 liter gasoline engine on the New
Figure 7.1 Peugeot 308.
teristics
parameters are given in Table 7.1.
ble 7.1 Peugeot 308 model characteristics.
ominal mass [kg] 1075
Wheel diameter [m] 0.406
olling friction coe cient cr [-] 0.015
earing friction coe cient µ [-] 0.002
erodynamic coe cient Scx = Ca · A [m2] 0.65
umber of gears [-] 5
ciency [-] 0.94
nal drive1 [-] 4.05
atio 1 [-] 3.02
atio 2 [-] 1.6
atio 3 [-] 1.13
atio 4 [-] 0.86
atio 5 [-] 0.68
isplacement [l] 1.2
umber of cylinders [-] 3
ated power [kW] 60
aximum speed [RPM] 6000
aximum torque [Nm] 120
le speed [RPM] 950
le fuel consumption [l/h] 0.33
eceleration Fuel Cut-O↵ yes
ype Gasoline
ensity [kg/l] 0.745
ower heating value [MJ/kg] 42.7
55
Costs/emissions/efficiency
REX
Elv
Hyb.

E-technology for education

E-technology for education : www.energyscope.ch

E-technologies : Enabling the energy transition
Smart metering
• Electricity consumptions
• Heating/cooling
• Renewable energy
Smart command
• WIFI set points
• Tracking
• Distance management
Smart Control
• Adaptative/predictive
• Optimal
• Fitted
• Renewable energy integration
• Demand side Management
Price signals
Predictions
Power/Energy reserve
Smartoptimalcontrolandmanagementbox
Equipment
• Heat pump
• Cogeneration
• PV
• Batteries
• Storage tanks
• Thermal solar
• HVAC
• Refrigeration
• Biofuels
• Distribution
• …
End-Users
• Power plants
• Industry
• Buildings
• District systems
• Storage
Building stocks/micro grids
Retroﬁt
Refurbishment
Sizing
Energy conversion
Services
Buildings
Building complexes
Small and medium industries
Districts
Renewable energy
Smart engineers Smart Systems
Eﬃciency
System design

Thank you for your attention
Prof François Marechal
http://ipese.epfl.ch
EPFL Valais-Wallis

14 epfl-valais

Recommandé

Recommandé

Contenu connexe

En vedette

En vedette (18)

Similaire à 14 epfl-valais

Similaire à 14 epfl-valais (20)

Plus de TechnoArk

Plus de TechnoArk (20)

Dernier

Dernier (20)

14 epfl-valais