Design Indoor PV-Battery Systems for Smart Motorized Roller Shades Bo Hu 1 , Shuang Yi 2 , Daniyal Rajput 3 , Yizhong Huang 4 , and Huaizhong Li 5 Abstract— We present an indoor solar irradiance model, which can accurately estimate the actual solar energy passing through window’s glazing materials at any time and geograph- ical locations of the earth under the clear sky. The data from this model are further utilized to compute the energy harvested by photovoltaic (PV) panels, which is stored in Li-ion batteries as power supplies of smart motorized roller shades. Our experiments deduce one energy generation model of one PV panel with maximum power point tracking circuit, which is a function of solar irradiance and temperature. The loss of load probability (LLP) curves are obtained for multiple PV areas and battery capacities, which achieve the self-sufficiency of the off-grid motorized roller shades. Our experiments show that those models correctly predict indoor solar irradiance and PV energy generation. The designed PV-battery systems are proved by uninterrupted operations of ten thousand off-grid motorized roller shades installed inside buildings worldwide over one year. keyword: Indoor solar irradiance, smart motorized roller shade, photovoltaic, Li-ion battery I. INTRODUCTION Smart motorized roller shades bring enormous oppor- tunities of increasing the comfort of the occupants and decreasing building energy usage [1][2][3]. Buildings con- sume 40 percent of global energy to maintain the oper- ations for occupant’s comfort [4]. The control operations include temperature, humidity, lighting, acoustic comfort, air ventilation, and home appliances. Primarily, the build- ings’ windows dissipate considerable energy [5][6][7]. In the United States, it is estimated that 10-25% of heat is lost owing to those windows [8]. The unshaded glass heat transfer is a hundred times higher than the insulated walls of buildings. As a solution, modern glazing materials are employed to control solar insolation and heat transfer through windows, such as silica aerogel glazing material, low-emissivity coatings, and reflective glass[9]. However, those new glazing materials are expensive, semi-transparent, and difficult to be retrofitted. In contrast, smart motorized roller shades have significant advantages, including dynamic heat control, remote operation via wireless IoT connec- tion, and easy installation[10][11][12][13][14]. Moreover, the buildings’ energy consumption can be further decreased by 1 Bo Hu is with Innovation & Design Center, Rollease Acmeda Pty Ltd, Melbourne, Australia john.bo.hu@gmail.com 2 Shuang Yi is with Department of Precision In- strument, Tsinghua University, Beijing, 100084, China yishuang@mail.tsinghua.edu.cn 3 Daniyal Rajput is with Innovation & Design Center, Rollease Acmeda Pty Ltd, Australia daniyal.rajput@rolleaseacmeda.com 4 Yizhong Huang is with School of Materials Science & Eng., Nanyang Technological University Singapore y.zhuang@ntu.edu.sg 5 Huaizhong Li is with School of Engineering & Built Environment, Griffith University, Gold Coast Australia h.li@griffith.edu.au interior photovoltaic (PV) arrays behind windows, which harvest solar energy and store it in batteries for the roller shade’s uninterrupted operation [15][16]. For easy installa- tion, smart motorized roller shades are off-grid charged by the PV-battery system. The self-sufficiency brings challenges for the PV-battery design of motorized roller shades. The off-grid PV arrays require batteries for energy stor- age to provide electrical power to motorized roller shades continuously. The first reason is that solar energy generation strongly depends on irradiance availability during the day- time [17]. The intermittent unavailability of solar insolation stops the PV energy generation to off-grid motorized roller shades. The second reason is that the varying insolation sometimes causes a low PV array electrical current, which is less than the threshold of the motor’s operation. With parallel connection, the current of batteries is increased to satisfy the motor current requirement. Many researchers have deeply in- vestigated the PV charging battery system and the method of enhancing its charging efficiency. Gibson utilized a single sil- icon crystalline photovoltaic module to charge li-ion batteries with no intervening electronics directly. It proved the solar charging capability for battery electric vehicles. An efficiency of 14.5% was achieved by matching the battery’s voltage to the PV module’s maximum power point [18]. The efficiency can be further enhanced by adding the maximum power point tracking (MPPT) techniques. The popular algorithms include perturb and observe (P&O), fractional open-circuit voltage, and incremental conductance methods. Killi developed an anti-drift perturb and observe MPPT algorithm [19]. This method checks the change in power ΔP over the voltage change ΔV and detects the change in the current ΔI . If the slopeΔP/ΔV  0, the method increases the voltage until MPPT point. Otherwise, the method decreases the voltage. If both ΔV and ΔI are positive, the method confirms the increase in insolation. A new type of perovskite PV array development is a vivid research area. This PV is also used to charge batteries [20][21].Xu utilized four perovskite solar cells connected in series to charge li-ion batteries. The overall efficiency of 7.80% was achieved for solar energy conversion and battery storage [22]. Gurung used one perovskite solar cell to charge a li-ion battery with a DC-DC voltage boost converter and MPPT method [23]. The converter boosted the low voltage of the perovskite cell to the higher li-ion charging voltage. The overall efficiency of 9.36% was achieved. In addition to the efficiency improvement, the off-grid PV-battery size is optimized to generate sufficient energy for roller shade operations. Multiple factors influence its self-reliability, such as geographical location, ambient tem- perature, time, and meteorological conditions. Due to its