Electrification of Clay Calcination: A First Look

into Dynamic Modeling and Energy Management

for Integration with Sustainable Power Grids
Bruno Laurini Nicola Cantisani
DTU Wind DTU Compute
Technical University of Denmark Technical University of Denmark
Kongens Lyngby, Denmark Kongens Lyngby, Denmark
[email protected] [email protected]

Wilson R. Leal da Silva Yi Zong John Bagterp Jørgensen

Green Innovation DTU Wind DTU Compute
FLSmidth Cement Technical University of Denmark Technical University of Denmark
Copenhagen, Denmark Kongens Lyngby, Denmark Kongens Lyngby, Denmark
[email protected] [email protected] [email protected]

Abstract
This article explores the electrification in clay calcination, proposing a dynamic model and energy management strategy for
the integration of electrified calcination plants into sustainable power grids. A theoretical dynamic modeling of the electrified
calcination process is introduced, aiming at outlining temperature profiles and energy usage - thus exploring the feasibility of
electrification. The model serves as a tool for optimizing parameters, estimating system behavior, and enabling model-based process
control. An innovative energy management model is also presented, ensuring efficient assimilation of electrified calcination plants
into the power grid. It encapsulates demand-supply balancing and optimizes renewable energy usage. In essence, we provide an
insightful pathway to a more sustainable cement production, underlining the value of renewable energy sources and effective
energy management in the context of clay calcination.

Keywords: calcined clay; cement; electrification; energy management; sustainability.

I. I NTRODUCTION
The cement industry is a hard-to-abate sector that contributes to around 8.0% of the world’s CO2 emission [1]. In the
production process of Ordinary Portland Cement, OPC, (i.e. CEM I [2]), the emission of CO2 originates from limestone
calcination and fuel combustion, with a total of 0.80 t CO2 ⁄ t cement, assuming a coal-fired system. The cement industry
has been experimenting with complementary technologies to help reduce its CO2 emissions. The current industrial effort to
deliver a more sustainable cement is towards using more Supplementary Cementitious Materials (SCM) in composite cements,
while reducing emissions from fossil fuels utilisation via electrification. To that end, calcined clay has gained popularity as
SCM due to its worldwide availability, excellent pozzolanic properties, and low cost [3]. The main industrial solutions for
clay calcination rely on the use of fossil fuels. Hence, the replacement of fuel combustion with full electrification can further
reduce the CO2 footprint from composite cements. In particular, the use of electricity in a flash calciners requires a electrical
hot gas generator to heat the exhaust gas from the preheating cyclones to the desired temperature necessary to calcine clays -
usually in the range of 750 - 850◦ C depending on the clay type.
While there exist many positive aspects in electrifying clay calcination process - a comprehensive discussion on technical,
economic, and environmental potentials are provided in [4] and [5] - the integration of such electrified process with the power
grid presents technical challenges. The large increase in electrical demand due to electrification might require investments in
both the plant’s electrical infrastructure and power transmission system [6]. The latter is dependent on the system’s actual
power transfer limits, where power demand fluctuations of such large industrial load are to have a considerable impact on the
power system’s operation. Hence, coordination with system operators and adoption of smart power management strategies are
crucial for guaranteeing stability.
The adoption of demand-side management measures in electrified plants represents an opportunity for innovation, helping
deliver a more sustainable product at a lower operation cost because: a) the use of energy management systems to control
dispatchable units (i.e. flexible loads, storage and on-site generation) can help reduce emissions and operational costs, and b)
EcoClay (Project Agreement No. 64021-7009) is partially funded by the Danish Energy Technology Development and Demonstration Program (EUDP)
under the Danish Energy Agency.
the employment of demand response (DR) strategies by the cement plant can help system operators to handle the intermittence
of Renewable Energy Sources (RES), supporting the integration of such resources into the power grid. In view of that, we
propose a theoretical dynamic model and energy management strategy for the integration of electrified clay calcination plants
into sustainable energy grids.
II. T HEORETICAL MODEL
In this section, we describe the mathematical and physical foundation of the cement plant’s Energy Management System
(EMS) and a dynamic model of the clay calcination process.
A. Energy Management System (EMS)
An EMS aims to optimize the energy generation and usage inside an energy system. It can be applied to systems of different
scales, such as residential and commercial buildings, micro-grids, power plants, and industrial plants. For instance, a residential
building can be equipped with photovoltaic modules, lithium-ion batteries, and a HVAC system. In this case, a Building Energy
Management System (BEMS) could be used to optimize the power dispatch of flexible technologies for given thermal comfort
limits and dynamic electricity contracts (e.g. Time-of-use rates, Real-time pricing) with the utility [7].
The benefit of using smart energy management strategies is rather minimal in small-scale energy systems. Nonetheless, in
an industrial context, they can lead to large costs and CO2 savings. The complexity of EMS increases alongside the scale
of the system due to factors such as: distinct processes taking place simultaneously, large critical loads, power distribution
complexity, and potential participation in power markets.
From a cement plant outlook, we propose an industrial EMS based on an Optimal Power Flow (OPF) algorithm. Most power
distribution networks are based on radial topology, and the power flow within them can be modeled using Branch Flow Model
(BFM), which focuses on describing the physical relationship between branch variables such as Branch current and power
flows. Assuming a radial distribution topology, the directed graph that represents the network can be described by G = (N, E)
. From Ohm’s law, the voltage drop between two buses reads:

Vi − Vj = zij Iij ∀(i, j) ∈ E, (1)

where (i, j) denotes a network link - connecting bus i and j - and E represents the set of buses. The branch power flow
Si j definition reads:
∗
Sij = Vi Iij ∀(i, j) ∈ E. (2)
Let Np (j) ⊆ N be the bus set containing the parent of j, Nc (j) ⊆ N the bus set of all the children of j, and Nss ⊆ N be
the set of substation buses; then, the power balance in the branch power flow depicted in Fig. 1a can be written as:
X  2
 X
Sij,t − zij |Iij | + Sj,t = Sjk,t ∀j ∈ N \Nss . (3)
i∈Np (j) k∈Nc (j)

Together, (1) to (3) constitute the branch flow equations and can be used for power flow analysis of a given network. After
some manipulation of (1) and (2), we can arrive in a voltage drop formulation that do not contain complex numbers, only
magnitudes, i.e.:
2
+ x2ij lij


Uj = Ui − 2 (Pij rij + Qij xij ) + rij ∀(i, j) ∈ E, (4)
where U and l are squared voltage and current magnitudes. The complex power balance in (3) can be decomposed into its
real and imaginary counterparts by:
X X
(Pij − lij rij ) − Pjk = Pj ∀j ∈ N \Nss (5)
i∈Np (j) k∈Nc (j)

and
X X
(Qij − lij xij ) − Qjk = Qj ∀j ∈ N \Nss (6)
i∈Np (j) k∈Nc (j)

Furthermore, the squared branch current magnitude is related to the voltage drop, and active and reactive power flow through
the branch, according to the apparent power definition, i.e.:

Pij2 + Q2ij
lij = ∀(i, j) ∈ E. (7)
Ui
 BESS PV
WPP

HV grid

MV bus

Department #1 Department #2 Clay calcination

(e.g., Quarry-Crusher) (e.g., Kiln) (Smart load)

(a) Branch power flow (b) Conceptual topology of the plant’s power distribution
Fig. 1: Single line diagrams

Along with the power generation/absorption constraints, the aforementioned equations can be cast in a standard optimization
form. Such problem represents an OPF model, where a given objective function is minimized given a set of equalities
and inequalities constraints. For example, a Distribution System Operator (DSO) could run an OPF aiming at power loss
minimization and Conservation of Voltage Reduction (CVR), in a context of high penetration of Distributed Energy Resources
(DERs) [8]. Our proposed EMS for a cement plant is designed in a similar context, with the main difference being that a
DSO cannot control all the loads and generators that are connected to the distribution grid. In the industrial case, the cement
plant owns both the network and the loads/generators, i.e. there are many control actions that can be set to achieve the desired
objective.
Figure 1b shows a conceptual topology of a cement plant’s power distribution, where on-site renewable generation (e.g.
a PV and a Wind power plant), Battery Energy Storage System (BESS) and the main grid are used to feed both flexible
and non-flexible loads. A cement plant is typically divided into functional operating departments according to the different
processes. Usually, separate electrical feeder circuits are used to guarantee electrical power supply for each important department
depending on their size. The loads of each department require different amounts of active and reactive power that are supplied
by the upstream buses.
Moreover, depending on the nature of the process, a load can be flexible or not. In other words, it can modulate its power
absorption within a specific range that depends on the process itself. A clay calcination department - containing an electrified
clay calcination process - can be a source of power flexibility, especially when coupled with energy storage solutions (e.g.
thermal storage). When looking at the plant’s power supply part, the main substation represents the dominant source of supply.
In this area, Photovoltaic, Wind turbines, and Battery Energy Storage Systems are connected to the AC network through power
converters. Such technologies can help increase the plant’s power flexibility and reduce both Scope 2 emissions costs and
costs of buying power from the utility. Note that, while many industrial plants have fixed electricity contracts with the utility
company, there is also an opportunity for the plant to participate in power markets directly, in case it has sufficient power
capacity and smart energy management strategies.
B. Dynamic model
The establishment of a dynamic model of the clay calcination process is the key to assess and simulate the plant’s performance
under changing conditions. Such a model is a proxy to evaluate and optimize controllers for the process. Figure 2 shows the
process diagram of the clay calcination in a loop with gas re-circulation. The process starts by introducing ground clay in the
first cyclone. The clay is pre-heated by two cyclones before entering the calciner. Next, the calcination reaction is completed
in the calciner and the mixture of gas-solid undergoes separation in the last cyclone. Finally, the hot gas is recirculated in a
loop. Our approach towards modelling this dynamic process is presented as follows.
The plant-wide dynamic model consists of connected units. The main units are the calciner and cyclones; and this model
results in the following system of differential algebraic equations (DAEs):
 ẋ(t) = f (x, y, u, d, p), (8a)
0 = g(x, y, u, d, p), (8b)
where x and y are the differential and algebraic variables. u, d and p inputs, disturbances and model parameters, respectively.
The properties and reaction of the materials and substances are evaluated by an independent chemical and thermo-physical
model that is shared by all the units. The chemical model comprise stoichiometry and kinetics of the dehydroxylation reaction
of clay, i.e.:
Al2 O2 · 2SiO2 · 2H2 O(s) →Al2 O2 · 2SiO2 (s)+2H2 O(g). (9)
This reaction is modeled as a third-order reaction as in [9]. The material’s thermo-physical properties are evaluated by building
a thermodynamic library, which can compute the material’s enthalpy H, volume V , and internal energy U as a function of
temperature T , pressure P , and number of models n by means of:

V = V (T, P, n), (10a)

H = H(T, P, n), (10b)
U = H − P V. (10c)
The dynamic behaviour of each unit is modeled by using first principles, i.e. mass and energy balances. Next, the calciner
is modeled as plug flow reactor - as fully described in [10]. The mass and energy balances are expressed as partial differential
equations in length z and time t, i.e.:

∂t c = −∂z N + R, (11a)
∂t ûs = −∂z H̃s + Jˆsg + Q̂amb,s , (11b)
∂t ûg = −∂z H̃g − Jˆsg + Q̂amb,g . (11c)
where c, N, R, û, H̃ are the concentrations, material fluxes, chemical production rates, volumetric internal energy, and
enthalphy flux, respectively. The subscripts s and g indicate solid and gas. Jˆsg and Q̂amb are the solid to gas heat transfer and
heat transfer to the ambient, respectively.
The velocity of the material stream is modeled explicitly as a function of the pressure drop (by means of the Darcy-Weisbach
equation). After spatial discretization of (11), the model is completed by adding the algebraic equations:

V (Ts , P, cs ) + V (Tg , P, cg ) − 1 = 0, (12a)

U (Ts , P, cs ) − ûs = 0, (12b)
U (Tg , P, cg ) − ûg = 0. (12c)
The cyclones are modeled as a single cell without spatial discretization. This model consists of mass and energy balances for
solid and gas, formulated as DAEs in the same fashion as the calciner. The material fluxes depend on the cyclone separation
efficiency, which is modeled explicitly as in [11]. The velocities at the inlet and outlet are modeled as a function of the pressure
drop between the input and output streams.
Other minor process units (e.g. electric hot gas generator - EHGG, circulating fan and particle filter) are modeled using static
relationships and incorporated in the model as algebraic equations. In addition, the connection between the units is realised by
lumping the flow resistance in the connecting tube as a pressure node [12]. The flow rate F between 2 units (1 and 2) reads:
s
(P1 − P2 )(P1 + P2 )
F =C , (13)
T1
where C represents the flow resistance.
The inputs u (variables) of the plant-wide model are the clay feed, pressure rise after the circulating fan, and fresh air intake.
The disturbances d are the electrical power to the hot gas generator and the ambient temperature. Notice that the model may
be easily extended to a system of stochastic differential algebraic equations to account for different sources of uncertainty and
model/plant mismatch. Nonetheless, such extension is beyond the scope of this publication.
 Feed CLAY

Cyclone 1

Cyclone 2

Cyclone 3

Preheated solid
Calciner
ACTIVATED CLAY

Electric HGG
ID Fan

HOT GAS

Ceramic Filter

Open loop
Solid
Feed FRESH AIR Gas
Clay bin Mixture
Power

Fig. 2: Clay calcination process diagram.

III. M ODEL INTEGRATION

The integration of the EMS and Dynamic calcination models provides a link between the electricity supply of the cement
plant and the thermodynamics of the clay calcination process. Such a link allows for emulating the physical constraints of both
systems and investigating how the plant can operate close to optimal conditions from costs, emissions, and technical standpoint.
While both models are linked through the power absorbed by the EHGG, they have different temporal/spatial scales, following
a similar modelling approach to the hierarchical algorithm for integrated scheduling and control proposed in [13]. The EMS
is based on a systemic approach, where the focus is on modeling the electrical interaction of different processes through the
power distribution network and scheduling dispatchable units.
The scheduling problem is directly related to daily production plan of the cement plant and to wholesale electricity markets.
Therefore, a model with relatively large time-horizon (e.g. 24 hours) and low frequency (e.g. 1 hour) is used. Conversely, the
dynamic model of the clay calcination plant enables a detailed simulation of the process on the temporal scale of seconds,
capturing transient states and detailed process dynamics.
The dynamic model keeps track of crucial variables such as temperature in the calciner, calcined clay at the outlet, calcination
degree in every step of the process, etc. A local plant controller is implemented to control the process. As mention in Section
II.A, the EMS aims to optimize the energy utilization in the cement plant considering techno-economic and emission-wise
aspects, i.e. the objective function reads:

min ϕ = ϕc + ϕCO2 + ϕu − ϕcc (14a)

s.t. Network constraints, (14b)
Operational/process constraints. (14c)
 Integration framework

Input Input

RES generation
Clay feed
Network topology
EMS CC Dynamic model Pressure drop
Load forecast
Fresh air intake
Market and CO2 signals

Fig. 3: Integration framework of the EMS and dynamic calcination model.

The terms in the objective function are operational cost including electricity purchase (ϕc ), CO2 emissions (ϕCO2 ), voltage
deviation (ϕu ), and calcined clay production (ϕcc ).
Solving the optimization problem (14) will provide scheduling decisions for the chosen time horizon. These may include
power set-points for generation units and storage, or the operational status of voltage controlling devices. The use of different
objective terms characterizes a multi-objective optimization and allows for the EMS to make scheduling decisions based on
a combination of different objectives. Weights may need to be used in the objective function to tune the performance of the
optimizer. Amongst the decision variables, PeEHGG represents the power set-point sent to the clay calcination plant controller.
Note that the EMS sets the active power consumed by the EHGG based on a higher-level model: as such, it does not capture
the dynamics of the clay calcination process. The actual power used by the EHGG in real-time will actually depend on the
process dynamics and the plant controller. Fig. 3 shows the proposed model integration framework.
IV. C ONCLUSION
This paper presents a theoretical methodology aimed at optimizing energy management in cement plants, with a focus on an
electrified clay calcination process. We propose an EMS model alongside a dynamic model for the clay calcination. Also, we
discuss an integration framework to link the power management of the plant with the electricity requirements of the electrified
calcination process. This framework is designed to reduce operational costs associated with electricity procurement, while
mitigating indirect carbon dioxide emissions from the power grid. By addressing technical and production-related constraints,
our approach aims to achieve these objectives while ensuring the fulfillment of production targets and technical requirements.
Future work should be carried out to extend the EMS and dynamic models including a market bidding strategy that encompasses
participation in Day-ahead and balancing electricity markets and advanced control strategies based on model predictive control.
V. ACKNOWLEDGMENTS
The EcoClay project (Project Agreement No. 64021-7009) - which is partially funded by the Danish Energy Technology
Development and Demonstration Program (EUDP) under the Danish Energy Agency.
R EFERENCES
[1] A. H. P. Monteiro, S. Miller, “Towards sustainable concrete,” Nature Materials, vol. 16, no. 207, pp. 698–699, 2017.
[2] EN 197-1 (2011) Cement Part 1 Composition, Specifications and Conformity Criteria for Common Cements. Brussels, Belgium: European Committee
for Standardization (CEN), 2011.
[3] R. Jaskulski, D. Jóźwiak-Niedźwiedzka, and Y. Yakymechko, “Calcined clay as supplementary cementitious material,” Materials, vol. 13, no. 21, 2020.
[4] M. Nakhaei, L. Hu, W. R. Leal da Silva, B. Laurini, Y. Zong, and C. Træholt, “Electrification of calcined clay systems in the cement industry – technical,
economic and environmental potentials,” in The 16th International Congress on the Chemistry of Cement 2023 (ICCC2023), pp. 1–4, 2023.
[5] B. Laurini, Y. Zong, C. Træholt, S. Parra, and M. Shanmugasundaram, “Techno-economic analysis on the integration of an electrified clay calcination
process into a cement plant,” in 15th International Conference on Applied Energy - Doha, Qatar, 2024.
[6] M. C. Wei, M. and S. de la Rue du Can, “Electrification of industry: Potential, challenges and outlook,” Current Sustainable / Renewable Energy Rep,
vol. 6, no. 2019, pp. 140–148, 2019.
[7] M. Manic, D. Wijayasekara, K. Amarasinghe, and J. J. Rodriguez-Andina, “Building energy management systems: The age of intelligent and adaptive
buildings,” IEEE Industrial Electronics Magazine, vol. 10, no. 1, pp. 25–39, 2016.
[8] M. Farivar, R. Neal, C. Clarke, and S. Low, “Optimal inverter var control in distribution systems with high pv penetration,” in 2012 IEEE Power and
Energy Society General Meeting, pp. 1–7, 2012.
[9] P. Ptáček, D. Kubátová, J. Havlica, J. Brandštetr, F. Šoukal, and T. Opravil, “Isothermal kinetic analysis of the thermal decomposition of kaolinite: The
thermogravimetric study,” Thermochimica Acta, vol. 501, no. 1, pp. 24–29, 2010.
[10] N. Cantisani, J. L. Svensen, O. F. Hansen, and J. B. Jørgensen, “Dynamic modeling and simulation of a flash clay calciner,” in 12th IFAC International
Symposium on Advanced Control of Chemical Processes (ADCHEM 2024), July 14-17, 2024, Toronto, Canada, 2024.
[11] M. Trefz and E. Muschelknautz, “Extended cyclone theory for gas flows with high solids concentrations,” Chemical Engineering & Technology, vol. 16,
no. 3, pp. 153–160, 1993.
[12] J. E. Hansen, Plant wide dynamic simulation and control of chemical processes. PhD thesis, Technical University of Denmark, 1998.
[13] L. E. Sokoler, P. J. Dinesen, and J. B. Jørgensen, “A hierarchical algorithm for integrated scheduling and control with applications to power systems,”
IEEE Transactions on Control Systems Technology, vol. 25, no. 2, pp. 590–599, 2017.