*5.7. Characterization of the Validated EFM in the Chilled Water Air-Conditioning System*

The characterization framework defines the scope of the validated EFM. The following tables determine or quantify the different parameters and hence the different dimensions as described in Section 4.7. Table 7 describes the functional dimension of the EFM.



<sup>1</sup> MO: Mode of Operation.

In Table 7, the system description, EFM category and operative concept come from the analysis performed in the previous steps. The adjustment factor responds to the control variable in the system, the ramping up and down of cooling generation units. Four different modes of operation (MO) have been defined for the EFM. The division is based on the number of clusters identified in the analysis of the typical load profile and the ability of the EFM to induce an increase (↑) or a decrease (↓) in the electrical consumption. All of these modes of operation are defined as holding because when activated, they only induce one operative state in the system. As the control of the system is performed via a SCADA system the execution level is set on the supervisory level. As expected, the functional dimension responds directly to the physical and operative characteristics of the analyzed industrial system. Table 8 presents the temporal dimension of the identified EFM.


**Table 8.** Temporal Dimension of the identified EFM 1.

<sup>1</sup> MO: Mode of Operation. <sup>2</sup> Calculated for Active Duration, Δ*tActive*, equal to 8 h.

The Active Duration minimum is restricted to avoid compressor short cycling (>5 cycles/h), which might cause the operative failure of the cooling generation units. The maximum Active Duration is limited to one working shift in the facility as an analysis of the typical profiles showed that MOs in the system can change form one shift to the other. The wide range in the Active Duration supports the intended implementation objective as the duration of price volatility can extend over several hours. The Planning Duration is set to zero as no planning is necessary, in both cases, the system can execute its task, provide cooling, without interruption. The Perception Duration depends on the specific market on which electricity is being purchased and hence ranges from 5 min for intra-day handling to 24 h for day-ahead handling. The Decision Duration is considered automatized and hence it is defined by the latency of the components in the EMS. The Shift Duration responds to the ramp-up and down duration of the different cooling generation units. The Activation Duration aggregates the ramping-up of the EFM and it is calculated using Equation (1). The major element deciding the Activation Duration is the Perception duration and hence the identified EFM presents a very high capability to quickly react to price volatility and hence achieved the intended implementation objective. The Deactivation Duration mirrors the Shift Duration and the Regeneration Duration corresponds to the short-cycling avoidance requirement in the cooling generation units. Both are also relatively short, allowing for the EFM to be used to respond to subsequent electricity price variations. The Validity responds to production characteristics of the facility and hence to the operatives clusters determined by the typical profile analysis. The Activation Frequency is calculated using Equation (2).

Table 9 presents the performance dimension of the identified EFM. The given load increase (↑), Δ*Pflex*, values quantify the maximum and average difference between the typical electrical input and the necessary electrical input if the typical cooling output is satisfied by only using the mechanical chillers. The load reduction (↓), Δ*Pflex*, on the other hand, quantify the maximum and average difference between this typical electrical input and the necessary electrical input if the typical cooling output is primarily supplied using the absorption chillers. In reality, the Δ*Pflex* is dynamic and hence a function of the state of operation of the system. The state of operation will depend on the instantaneous cooling demand, in turn, a function of both the outdoor temperature and the level of production in the facility. The given values are hence a static approximation to the dynamic Δ*Pflex*, value.

As previously hinted in the functional dimension, the EFM presents a bidirectional flexibility type without a need for later compensation. The different Δ*Pflex* responds to the different cooling demand of the typical profiles in the facility. A reduction in the cooling demand, MO-3 and MO-4, diminishes the flexible power. Moreover, the flexible power of a load increase is considerably lower than that of a load reduction as in the reference operation, the CHWDXs have operative priority, and hence are already supplying a portion of the cooling demand. The most attractive MO is MO-2, an electrical consumption reduction during production and with an ambient temperature over 10 ◦C. This MO can achieve, on average, a reduction of 88% of the electrical consumption of the system for a period of up to 8 h.


**Table 9.** Performance Dimension of the identified EFM.

<sup>1</sup> Calculated for Active Duration, Δ*tActive*, equal to 8 h.

Regarding the additional characterization parameters, the flexible energy carriers respond to the operative principle of the cooling generation units in the system and the flexible energy is calculated using Equation (3). As the physical and operative characteristics of the chilled water system and, the production characteristics of the facility are considered to calculate the Δ*Pflex* and the Δ*tActive*, they represent the practical EFP.

Finally, in Table 10, the economical dimension of the EFM is presented and the calculation reasoning behind each parameter is described.


**Table 10.** Economic Dimension of the identified EFM.

<sup>1</sup> Calculated for Active Duration, Δ*tActive*, equal to 8 h.

As can be inferred from the descriptions the implementation of the EFM will only represent investment and activation costs. The investment costs relate to additional infrastructure to have a constant supply of HW in the facility, and additional IT-infrastructure to allow the reaction to dynamic electrical prices. The maintenance costs relate, mainly to additional operative hours of the CHWABs which present relatively high maintenance costs, due to their operative principle. Although a relatively high payback period is given, it is clear that MO-3 and MO-4 are prohibitively expensive, based on the average price of electricity in the EU which is approximately 100 €/MWh [41].
