Engineering design tools intended for education must pave pathways to explicitly connect science and engineering education practices [3,4]. Based on our own research on engineering design at the secondary level [5][6][7], a cognitive barrier for students to connect science and engineering arises from the fact that science concepts are often less evident than other design variables such as price and shape, especially when students are overwhelmed by the complexity of design. Furthermore, the ability to apply a science concept to improve a design rests not only on the understanding of the concept per se, but also on the understanding of its interplay with all the other variables that may affect the overall function of the whole system. The latter understanding, which is related to systems thinking, is a prerequisite to core engineering practices such as making tradeoffs that satisfy multiple criteria under multiple constraints. In a CAD environment, we can visualize the underlying science concepts on top of design artifacts to render their effects in an engineering context. Such scientific visualizations can create learning opportunities for students to see science concepts at work individually or concertedly in their own designs, thereby fostering the integration of different concepts and the linkage between science and engineering [6].
Apart from "creating the demand" for science [8] in an engineering project, another way to inspire students to learn and apply science in a design process is to continuously provide formative feedback [9] to illuminate how science concepts play out in their own designs. Specific feedback can help students learn the effect of a science concept and then make an actionable design choice. In typical engineering projects, students are challenged to design and construct an artifact that performs desirable functions under constraints. In a conventional design project, to determine whether a change of form can result in a specific function, students have to build and test a physical prototype or rely on the opinions of an instructor. This creates a delay in getting feedback at the most critical stage of the learning process, slowing down the iterative cycle of design and cutting short the exploration in the solution space. When students' time for a design project is limited, a long delay in the feedback loop can be detrimental to learning. Even if there is enough instructional time, not all teachers are expert at evaluating complex student designs and not all students have access to resources for meaningful prototyping and testing. All these problems can be addressed by supporting engineering design with an intelligent CAD system empowered by physics-based numerical simulations [5] and artificial intelligence [10] capable of analyzing students' design artifacts and then generate visual feedback to them on an ongoing basis.
To the best of our knowledge, there are currently very few developmentally appropriate CAD software available to K-12 students that can achieve the above instructional goals-most CAD software used in industry not only appear to be science "black boxes" to students, but also require a complicated procedure of tool chaining involving pre-processors, solvers, and post-processors, making them more challenging to use in schools. To fill the gap, the Institute for Future Intelligence has developed the Web-based Aladdin CAD tool that opens the "black box" of science with rich interactive visualizations and integrates design and analysis seamlessly to eliminate tool chaining. As the numerical simulation capabilities of Aladdin are key to achieving the aforementioned instructiveness of the tool, it may be appropriate for us to begin the introduction of Aladdin with its accuracy. In the following section, we compare its predictions of solar irradiance in different locations with available sensor data.
Aladdin is currently equipped with computational engines that support two types of physics simulation: solar energy simulation and building energy simulation. Since this paper focuses on concentrated solar power (CSP), we only describe solar energy simulations. Solar energy simulations in Aladdin are powered by a revised solar radiation model based on ray optics [11] that also considers the effects of sunlight attenuation in the atmosphere, diffuse sky radiation due to scattering by air molecules, reflection by the ground, and sky clearness from the weather data of the current location (as of August, 2022, Aladdin includes the weather data from a total of 762 locations worldwide). The credibility of the simulation engine is important to engineering education as the accuracy of CAD simulations is responsible for the authenticity of the engineering projects that it supports. To some extent, the authenticity helps justify replacing traditional prototype-based design activities with CAD-based design activities-while CAD-based activities remove some hands-on steps in prototype-based activities, accurate CAD simulations may boost the fidelity and quality of the design work as a compensation. For example, it is nearly impossible to calculate the annual outputs of solar panels without using a predictive simulation tool.
Fig. 1 provides some benchmark results of the average daily solar irradiance throughout a year for four geographically diverse U.S. locations. In each case, a virtual sensor is placed on a horizontal surface and a south-facing vertical surface, respectively, in Aladdin. The time series data of solar irradiance logged by the virtual sensors are generated based on simulating the movement of the sun across the sky from sunrise to sunset on the 22 nd of each month. Aladdin calculates the intensity of solar radiation that strikes a sensor any time during a day, the daily average of which is then plotted against the month of a year with a built-in graph shown in Fig. 1. As the graphs illustrate, the calculated results agree reasonably well with the sensor data collected by the National Renewable Energy Laboratory (NREL) from 1961 to 1990 [12]. This validity was also confirmed independently by recent studies [13,14] that used it to model both photovoltaic (PV) and CSP solar power plants in Muscat, Oman. According to the published studies, the difference between the annual output of electricity of a solar farm predicted by Aladdin and the Global Solar Atlas (GSA) program developed by the World Bank is less than 7% [14], whereas the difference between the annual outputs of a solar farm with a fixed tilt angle and a solar farm driven by azimuth-altitude dual axis trackers predicted by Aladdin and PVWatts Calculator developed by NREL are approximately 4% and 8%, respectively [13]. Similar agreements between Aladdin and yet another simulation program, Photovoltaic Geographical Information System (PVGIS) developed by the Joint Research Centre of the European Commission, on the value for the optimal tilt angle of solar panels were also found. This means that, when students design a solar energy solution with Aladdin, they can expect a degree of trustworthiness comparable with three widely-used professional tools developed by reputable research institutions.
In the following sections, we show how Aladdin can be used to model and design all the four main types of CSP [15], even for real-world projects. We start with parabolic dishes since they are the closest ones to solar cookers.
Like a solar cooker, a parabolic dish for CSP focuses sunlight that strikes it onto a focal point above it, where a receiver absorbs the light energy and turns it into thermal energy. Unlike a solar cooker, a parabolic dish for CSP is motorized and programmed to follow the sun throughout the day to maximize the input. Parabolic dishes are often coupled with Stirling engines operated by the cyclic compression and expansion of a working fluid like air or helium to convert up to 32% of incoming solar energy into electricity.
Fig. 2a shows a single parabolic dish designed using Aladdin. To illuminate the physics concept, we show the optical paths of a number of rays from the sun to the cylinder of the Stirling engine placed at the focal point to receive the concentrated solar radiation. These optical paths change as the dish rotates to follow the sun, allowing students to see how it works to collect solar energy in a daily cycle. Designers can adjust the diameter of the dish rim, the latus rectum, and so on to create dishes of different sizes and shapes [16]. They can also add many dishes to form an array (Fig. 2b). Note that, in the current version of Aladdin, we model the movement of a parabolic dish by changing its azimuth and tilt angles, but we do not explicitly model the mechanism that drives it and the Stirling engine that uses thermal energy to do the mechanical work to generate electricity. Fig. 2b also shows a heatmap that visualizes the distribution of solar irradiance, which includes direct and indirect solar radiation (though the latter does not contribute to the CSP output). Due to the projection effect (the solar irradiance on a surface is inversely proportional to the cosine of the angle of incidence) and the curvature of the paraboloid, an area closer to the center receives stronger irradiance than an area closer the rim, as indicated by the color gradient in the radial direction on the surface of a dish. Heatmap analysis provides a visual tool for students to quickly make sense of their designs. For instance, a larger contrast of color on a deeper dish suggests less efficient use of materials as a larger area near the rim may contribute less to the total output, prompting students to consider adjusting the shape to improve cost effectiveness.
Through the integration with Google Maps, Aladdin allows students to model or design renewable energy projects in the real world. All students need to do is to type the address or coordinates of a site in Google Maps and Aladdin will load a map image into the scene for them to draw a CSP power plant on top of the image. Fig. 2c shows a model of the 1.5MW parabolic dish array at Tooele Army Depot in Utah, created in Aladdin and rendered on top of the satellite image of the site.
Unlike parabolic dishes, parabolic troughs are linear solar collectors with a parabolic cross section. They are typically aligned along the north-south axis and rotate from east to west to track the sun. A receiver tube filled with a thermal fluid such as molten salt is installed along the focal line to absorb the concentrated solar energy. The thermal energy is then passed through a heat exchanger to a turbine engine for generating electricity. Although parabolic troughs are less efficient in focusing sunlight than parabolic dishes, they are the most deployed CSP technology-with about two thirds of the operational capacity as of 2020 [17], largely because of their cost effectiveness. Fig. 3a shows an array of parabolic troughs modeled after Nevada Solar One in Nevada with Aladdin (users can create an array by copy-and-paste). Similar to the case of parabolic dishes, there exists a gradient of solar irradiance in the cross section due to the projection effect, as shown in the heatmap of Fig. 3b. Different from the case of parabolic dishes, though, the effect of shadowing between adjacent rows of troughs
also increases this gradient as the lower side of a trough in the middle of the array is blocked from direct sunlight when the sun is low in the sky.
The linear Fresnel reflectors (LFRs) are a type of CSP that uses long, flat mirrors to focus sunlight onto a fixed absorber pipe mounted at a common focal line of the reflectors, similar to a Fresnel lens (Fig. 4a). There are several design options for an LFR array.
The key factors include the dimension and spacing of the reflectors, the height of the absorber pipe (Fig. 4b), and the azimuth of the reflectors. There can also be multiple absorbers (Fig. 4c). In that case, the reflectors are laid in the same way but they reflect the sunlight alternately to the two absorbers (e.g., #1, #3, #5, ... reflect to the absorber on the east side and #2, #4, #6, ... to that on the west side). Fig. 4d shows the LFR array of the 5MW Sundt Generating Station in Tucson, Arizona. One of the "strange" things that students may notice from the satellite image is that the absorber pipe stretches out a bit at the northern end of the reflector array (as indicated by the arrow in Fig. 4d), whereas it does not at the southern end. The reason that the absorber pipe was designed in such a "strange" way becomes apparent when we show the optical paths. As the sun rays come from the south in the northern hemisphere, the focal point on the absorber pipe shifts towards the north. During most days of the year, the shift decreases when the sun rises from the east to the zenith position at noon and increases as the sun sets to the west. This shift would have resulted in energy losses if the absorber pipe had not extended to the north to allow for the capture of sunlight bounced off the LFR modules near the northern end.
A solar power tower (SPT) is surrounded by numerous mirrors controlled by computers, called heliostats, to follow the sun and reflect the light to a central receiver installed at the top of the tower (Fig. 5a). To some extent, a SPT plant is a gigantic Fresnel lens formed by many discrete reflectors (instead of rings or stripes). Its design involves deciding the layout of the heliostat field, the height of the tower, among other factors.
Unlike the design of PV solar panel arrays in which we only need to consider the shadowing among the panels, in the design of the heliostat field, we also need to ensure that the mirrors of the heliostats do not block the light reflected by others from reaching the central receiver. Fig. 5b shows the heatmaps of the heliostats that account for both shadowing and blocking losses.
A key factor that affects the design of the heliostat field is the so-called cosine efficiency [18]. The heliostats to the south of the tower have lower cosine efficiency because they reflect less sunlight than those to the north of the tower (assuming that the SPT plant is in the northern hemisphere). This is the reason why many SPT plants do not use a simple concentric heliostat field layout that has a north-south symmetry. The cosine efficiency can be clearly visualized in Aladdin (Fig. 5c). Another design factor is the height of the tower.
The taller the tower is, the more sunlight the heliostats reflect (Fig. 5d). This is the reason why engineers build very tall towers for large SPT plants.
We have demonstrated Aladdin as a versatile CAD platform for modeling and designing four types of CSP plants, which can be adopted by teachers as a design tool to connect physics and engineering. In terms of NGSS learning, the tool has the potential to support all the performance expectations set in HS-ETS1 of NGSS. In particular, it represents an appropriate, even needed, response to HS-ETS1-4 that requires students to "use a computer simulation to model the impact of proposed solutions to a complex real-world problem with numerous criteria and constraints on interactions within and between systems relevant to the problem." The modeling, simulation, and designing capabilities of Aladdin demonstrated in this paper can be used to entice students to learn and apply physics to engineer renewable energy solutions, which are critical to the sustainability of our society. In a sense, Aladdin can be used as a tool to enhance, or even replace, solar cooker activities widely used in science education to engage and motivate students.