Polygeneration Grid Including Atmospheric Water Generation, Renewable Sources and Storage Technologies: Modelling and Real-Time Management

Since important issues for the human development regard the access to clean water (UN goal N.6 [1]) and clean energy (UN goal N.7 [1]), innovative efforts to integrate these aspects into a smart polygenerative grid have significant potentiality of market penetration. Considering the current water scarcity situation affecting more than two billion people [2] and the future negative forecast due to the climate change process [3], urgent actions are necessary for recycling, better resource use, or the application of technologies based on alternative sources. Considering different alternative solutions, water extraction from air (or Atmospheric Water Generation (AWG)) started to be a competitive approach based on different techniques: radiative cooling, vapour concentration with sorption-desorption techniques, and active cooling [4]. The third technique (the most diffused and commercial one [5]) is based on a reverse cycle, operating with an electricity-powered compressor, including air flow cooling below the dew point [6]. This solution can provide from tens to thousands litres of water per day, depending on the system size and the ambient conditions. However, an important drawback of this water generation regards the energy consumption [7] that, for this reason, links water production to energy management aspects. This is a very important issue in case water needs to be purified due to filtering systems, increasing such consumption [8]. Energy consumption is one of the main concerns about the sustainability of atmospheric water.

A viable way to address the energy issue consists in employing advanced AWGs expressly studied to be integrated with Heating, Ventilation, Air Conditioning (HVAC) systems [8]. The water harvesting cycle, by direct cooling, gives as secondary effects a flux of dry-cooled air and a potential heating flux coming from the condenser. Advanced integrated AWGs are expressly designed in order to transform these secondary effects into useful effects [9]. In such a way, the overall energy efficiency of the system is notably increased, as the machine gives three contributions, water, cooling and heating, with the same energy input [9]. Therefore, atmospheric water becomes a sustainable alternative water source [9]. Nevertheless, in order to be really effective, such a kind of AWGs requires to be duly integrated in the existing HVAC systems.

Due to the extensive application of alternative renewable sources (such as photovoltaic panels (PVs)), important requirement for the decarbonization target [10], the integration with storage solutions is mandatory for both water and energy generation. For this reason, this work includes a tank for clean water and a tank for cooling water. Although more complex devices could be considered (e.g. phase change materials), in this work simple water tanks are considered because attention is focused on the integration and the management approach rather than on innovative technology development. However, the presented approach is general and easily transferable to different types of energy storage technologies. The application of these tanks is essential to store excessive generation and to minimize variable costs due to a proper device management [11].

Due to these integration aspects, energy management is important to pursue optimization objectives, such as cost minimization [12]. Although the application of an Energy Management System (EMS) is an extensively studied topic in smart grids for heat and electricity generation [13-15], the EMS technology is innovative for the application of AWGs. Especially in polygeneration grids equipped with different energy sources, different generation technologies (including the AWG for water), and/or devoted to generation of different useful flows (such as water, cooling energy, etc.), a robust EMS is essential to operate the grid in optimal conditions. This is important not only for the cost point of view, but also for reducing the environmental impact minimizing emissions (especially in case of integration with fossil fuels or the utilization of an electrical grid not based on zero-emission generation). For this reason, this work focuses special attention on the development of a real-time EMS tool to minimize variable costs during operations against standard management approaches.

For the EMS development and the related performance assessment through simulations, it was essential to develop real-time models for the involved components. Due to the necessity to have only global values (e.g., produced, consumed or stored power values) and fast calculation performance, simplified models were used on the basis of 0-D approaches (with input-output balance equations) or interpolation of maps (steady-state part) connected to first order delay tools (transient response). This approach was already presented and successfully validated in different previous works, starting from study [12] (a preliminary activity on AWG integration) to study [21] (an application of 0-D models in real-time calculations).

3.1 Development of component models

The model of the AWA250 AWG system was developed with design data from the manufacturer’s web site [16], using literature data [17] and other data for off-design conditions. So, the steady-state part of the model is based on interpolation of performance maps to calculate the values of produced water, cooling power and electrical power consumed at different operating conditions. Due to the influence of these important parameters, they are functions of ambient temperature and relative humidity. A further model part was implemented for the time-dependent response: two subsequent first order delay blocks were used. Due to this very simplified approach, the time constant setting was carried out on the basis of experimental data. The validation of this model was reported in the study [12] considering the cooling side of a scaled-up heat pump available in in the Innovative Energy Systems laboratory at the University of Genoa [22]. During a component start-up and shutdown phase, a good matching (simulation versus experimental data) was obtained in the study [12]: maximum errors close to ±3% in quasi steady-state conditions and accurate calculation of transient response. Considering that measurement accuracy was in the ±3% range for the power values reported in the study [12], the obtained performance was considered good for the purpose of this analysis.

A similar approach was considered for the heat pump whose cooling side provides the additional cooling energy to the grid (“C. gen. 2” in Figure 1). While the steady-state performance modelling was based on the measured data reported in the study [20], the transient response was based on single first order delay blocks. Following preliminary settings of this time constant, the model was successfully validated against experimental data. For instance, Figure 2 shows the good matching between calculations and experimental measurements (accuracy in the ±3% range) for the cooling power produced by the R600 heat pump during a start-up phase.

2.png

Figure 2. Heat pump model validation (cooling power)

The model of the thermal storage vessel was based on global power input-output balance to calculate its State of Charge. Moreover, a single first order delay block was included for simulating the time-dependent response. It was modelled considering a system composed of different 500 l chilled water tanks installed in parallel. To be general, since the amount of tanks depends on the operative temperature difference, no detailed design is included in this work and every analysis is reported in the form of inlet-outlet power exchange and stored energy. The validation of this model was performed considering a single 500 l tank internally equipped with a 30 kW electrical heater [23]. In the validation test presented in Figure 3, starting from equilibrium conditions, at time zero the heater set-point was switched to 20 kW. The circuit response that mainly depends on the dimension (thickness of the vessel and the pipes, circuit length, thermal insulation, etc.) was reported in Figure 3. Although the simplified model was not able to calculate details, such as the temperature distribution in the vessel, the model results were able to produce the time-dependent trend with an accuracy level in agreement with the measurements (errors in the ±3% range) and with the accuracy needs of this optimization analysis.

3.png

Figure 3. Thermal storage (chilled water tank) model validation

A further model regards the water storage tank. Also, in this case a very simple tool was implemented. It was based on an integrator block to calculate the mass changes on the basis of the inlet/outlet balance of the fresh water. The model development activity included also the calculation of its state of charge (SoC), as a number in the 0-1 range.

3.2 Energy Management System layout

The core of this activity was the development of the Energy Management System (EMS) starting from the preliminary tool presented in the study [12]. The EMS of the entire polygeneration grid was developed to minimize the related variable costs. Considering that the target of this work was the development of a real-time EMS to manage the grid during operations, no component size optimization was performed. Although this could be another important optimization (as in the study [22]), attention is focused on specific component management. The optimizer core is the cost function in Eq. (1) and the linked functions in Eqs. (2)-(3).

$\begin{array}{r}J_{\text {cost }}=c_W \cdot m_{W_{\text {grid }}}+\left(c_{O \& M}+c_{e l}\right) \cdot P_{e l_{A W A 250}} \\ +c_{e l} \cdot P_{e l_{H P}}+c_{O \& M} \cdot P_{e l_{H P}}  \cdot C O P_{H P}\end{array}$             (1)

$\begin{aligned} & P_{e l_{\text {AWA250 }}}  & =f\left(m_{W_{\text {AWA250 }}}, W E T_{\text {AWA250 }}, R H_{a m b}, T_{a m b}\right)\end{aligned}$             (2)

$C O P_{H P}=f\left(T_{a m b}\right)$             (3)

Water Energy Transformation (WET) is an efficiency parameter related to the atmospheric water generation. It represents the ratio between the needed theoretical energy for condensation and the real machine consumption [6]. Although it is based on the COP concept, for the WET parameter the useful effect is the condensation of atmospheric water.

The objective function in Eq. (1) is minimized by a MATLAB optimizer to have in real-time mode the values of three decision variables: water from the general grid (aqueduct), water produced with the AWA250 $\left(P_{e l_{A W A 250}}\right.$ is dependent on $\left.m_{W_{A W A 250}}\right)$ and the electrical power consumed by the heat pump ($P_{e l_{H P}}$) that is used for the cooling side. The optimization process is based on a constrained non-linear minimization of variable costs [24]. The procedure needs to satisfy both water and cooling demands and the constraints reported in Table 1. Although the application of a MATLAB optimizer implies no specific research activity on the algorithm side, an important innovative aspect of this work regards the application of a real-time EMS on an integrated system including water, cooling energy generation and related storage systems.

Table 1. Optimization process constraints

The management of the storage tanks is based on the cost forecast. As presented in the studies [12, 14], a scheduling is necessary to consider a prediction. For instance, in this analysis, discharging the water tank during low water price periods could be erroneous if the storage system is at low level condition or if the water price is affected by a further decrease. So, in this analysis the following approach is proposed: the tank charging is activated (at its maximum) when the electricity cost is lower than its average value and the AWG is in active phase. In case of the opposite electricity cost situation, the tank is discharged. A similar approach is used for the management of the cooling energy tank. However, in this case the charging of this energy storage system is activable only during operations of one (or both) cooling systems. The utilization of both storage systems is constrained in the 10%-90% range.

The PV panels are included in the EMS considering the related electricity production at zero variable cost. This is included in the EMS as described in the study [12] (a recalculation of the electricity cost through a weighted average with grid electrical energy cost).

This paper regards the development of an innovative EMS for a polygeneration grid integrating advanced AWG technology with a heat pump (cooling side), storage systems and electricity generation from PVs. The main innovation is the application of a management technology usually developed for grids including combined heat and power generation to the water supply issues. The work mainly reports the results summarized in the following points.

• The models of the components included in this polygeneration grid were based on 0-D approaches (interpolation of performance maps plus first order delays for the transient response). The model validation was carried out considering available experimental data, obtaining good reliability in simulations.
• A devoted EMS was developed to manage the grid components with the objective to minimize variable costs and operate in real-time mode.
• The management of the storage systems was based on a scheduling: charging operation when the electricity cost is lower than its average with prime movers (the AWG and/or the heat pump) in active phase; discharging operation in the opposite electricity cost condition.
• The simulation results presented for the base case produced a significant decrease in global variable cost (-20.6%).
• A sensitivity analysis (±20% changes in different parameters) showed significant saving in variable costs for all cases (-14% in the worst case).

The results presented in this paper successfully candidate this polygeneration grids to further analyses in the WISHeR PRIN2022 project. They will be based on laboratory tests in cyber-physical mode (as in study [14]) for assessing the EMS performance from the experimental point of view.