Open Access
Add to BibSonomy
Abstract
Light-emitting diodes (LEDs) are the 4th plant supplemental lighting source. Pulsed lighting is benefit to increase energy utilization efficiency in greenhouse production. A pulsed-lighting LED luminaire with geometrical optical solution is proposed to overcome the shortcoming with pulse width modulation (PWM) solution. In addition, this luminaire also achieves uniform lighting by designing optical surfaces. In the illumination area, the lighting frequency is 117.6 Hz and the illuminance uniformity is 0.789, which is better than 0.75 (+/- 12.5%). In an actual planting experiment of Brassica chinensis, the average fresh weight of the plants under the pulsed-lighting LED luminaire was 33.1% higher than that under the conventional LED luminaire. The results showed that the energy utilization efficiency of the pulsed-lighting LED luminaire is 22.9% higher than that of the conventional LED luminaire.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Light is the energy source of plant photosynthesis and functions as a signal for plants that affects germination and flowering time. For photosynthesis, plants utilize light mainly between 400 nm to 700 nm spectral range called photosynthetically active radiation (PAR) [1]. In recent years, light-emitting diodes (LEDs) have gradually replaced gaseous discharge-type lamps (high-pressure sodium, HPS) as lighting sources in most greenhouses due to their advantages of high electron conversion efficiency, longevity, and adjustable spectral quality [2]. The electricity costs of supplemental light are 30% of the operation costs of a greenhouse under HPS lighting conditions [3,4]. By transitioning from HPS to LED, the total energy savings are 10–25% [5]. However, the investment costs for LEDs are 5-10 times higher than those for HPS lamps [6]. Although LEDs have higher electron conversion efficiency, the high investment costs limit their application in greenhouses. To solve this problem, it is essential to develop a technology that reduces energy costs while securing the production and quality of crops [7]. Clearly, a more efficient supplemental lighting mode dose increases the sustainability and profitability of greenhouse production. To promote the application of LEDs in greenhouses, LED design needs to focus on reducing cost and improving energy utilization efficiency (EUE) [1].
At present, the efficacy of current LED packages has already reached 80% of the theoretical maximum [8]. Therefore, there is limited opportunity for improvement in the light source efficacy. A proposed alternative method is that pulsed lighting is a more appropriate lighting condition for plant cultivation than continuous lighting, which can improve energy utilization efficiency [9–19]. Pulsed light is intermittent light that can provide light periodically, and the main parameters characterizing pulsed light are light intensity, light frequency and duty cycles or light/dark cycles [11]. Olvera-Gonzalez et al. applied different frequencies of pulsed light and continuous light in tomato plants with LEDs. Their results showed a higher photosynthetic rate under pulsed light than under continuous light, and the frequencies with the highest values in the φPSII parameter were 0.1 Hz and 1 Hz [13–15]. Miliauskiene et al. evaluated the effect of pulsed light (0.5 and 1 kHz) on the growth of lettuce. The growth parameters of lettuce, in terms of leaf area, fresh and dry biomass showed up to 32%, 36% and 48% higher values in pulsed light treatments than in continuous light treatments [12]. Some research has analysed the enhancement of photosynthesis under pulsed light from physiological mechanisms [11,17]. Under high pulsed lighting frequencies, the amount of adenosine-triphosphate (ATP) generated in the electron transfer chain under the light phase is sufficient to support the Calvin cycle under the dark phase. The photodynamic damage is reduced since the pulsed light exposure is too short to cause damage or the periodic dark intervals facilitate the repair of the damage, which can improve the photosynthetic rate.
In general, pulsed light has two types: pulsed light generated by electrical or optical methods, as shown in Fig. 1. Pulsed light is generally generated by pulse width modulation (PWM) which provides periodic electrical pulses in the time dimension [17]. However, the photon flux of pulsed light is determined by duty cycles in this way, which limits the maximum photon flux of LEDs. This shortcoming causes the LEDs to not be utilized sufficiently, which increases the costs per unit photon flux of LEDs. At present, few studies have been performed on pulsed light generated by optical methods, which generate pulsed light by periodically switching the position of the light spot on the target plane. In this way, the LEDs remain in a persistent switch-on state, which guarantees that the LEDs are utilized sufficiently and that the illumination mode is still pulsed lighting. Under this type of pulsed light, the supplemental lighting area can be expanded without lowering crop yields, which decreases the number of LEDs per unit area in the greenhouse and reduces the investment costs of LEDs in greenhouse production.
Fig. 1. (a) Pulsed light generated by pulse width modulation; (b) Pulsed light generated by optical methods.
For plant lighting, the number of photons in the 400–700 nm spectral range received by the plant surface per unit area and per unit time is called the photosynthetic photon flux density (PPFD). When the light intensity reaches a certain PPFD, the net photosynthetic rate (Pn) does not increase with increasing light intensity. This phenomenon is called light saturation, and this PPFD is called the light saturation point (LSP) [1]. Meanwhile, non-uniform PPFD distribution will drive differences in individual plant development and decrease the commercial value of the crop in horticultural systems, as shown in Fig. 2(a). As shown in Fig. 2(b), a schematic diagram illustrates the photosynthetic rate of the plants under non-uniform illumination and uniform illumination. A non-uniform PPFD distribution will cause a part of the region’s PPFD to be higher than the LSP with decreasing the EUE of LEDs in this region. At present, there are two main ways to realize uniform lighting. One is designing the pattern layout of the lamps [20], and the other is designing the optical systems of lamps [21]. A total internal reflection (TIR) lens has been applied successfully to even light spots or collimate beams [22–28]. Moreover, a microlens array can also be used to improve the uniformity of the light spot and shape the beams on the surface [29–35].
Fig. 2. (a) The actual plants growth under non-uniform illumination and uniform illumination; (b) The schematic diagram of the photosynthetic rate of the plants under non-uniform illumination and uniform illumination.
Conventional LED luminaires (CLL) provide continuous and nonuniform lighting. In this paper, we propose a pulsed-lighting LED luminaire (PLL) for agricultural lighting, which can provide pulsed and uniform lighting by designing a specific optical system. The specific optical system consists of TIR lenses and a diffuser sheet, which can shape the beam as a rectangular light spot in the target plane. Pulsed and uniform lighting can be generated by rotating the diffuser sheet. A prototype of the luminaire was produced to measure its optical parameters. The actual frequency of the light spot is 117.6 Hz, and the actual uniformity of the light spot is 0.789, which is better than 0.75 (+/- 12.5%). According to the results of the planting experiment, we obtained that the EUE of the PLL is 22.9% higher than that of the CLL, which is beneficial for reducing the investment costs of greenhouse production.
2. System of the pulsed-lighting LED luminaire
The luminaire consists of LEDs, TIR lenses, a motor, and a diffuser sheet, as shown in Fig. 3(a). Moreover, the diffuser sheet is a double-face optical lens array composed of a specific micro cylindrical lens and is fixed on the motor shaft. As shown in Fig. 3(b), the TIR lens collimates the emitted light from the LED source, and then the diffuser sheet redistributes the collimated light. Finally, there will be a rectangular light spot on the target surface. When the motor is in operation, the diffuser sheet will be rotated with the shaft, meanwhile, the rectangular light spot will also be rotated and become pulsed lighting in the illumination area. When the frequency is high enough, we can only see a circular light spot.
Fig. 3. (a) The simple structure drawing of the luminaire; (b) The light path diagram of the luminaire.
Figure 4 shows a schematic diagram of the pulsed-lighting LED luminaire in this study. When the luminaire is turned on, it passes through three states in sequence: the static state, the starting state and the dynamic state. In this study, we designed a new optical system to make the circular light spot uniform in a target area, which means that plants in the illumination area will receive the same number of photons in a period of spot rotation.
2.1 TIR lens
The TIR lens is used to change LED light in the Lambert distribution into parallel light through refraction and reflection. As shown in Fig. 3(b), the TIR lens consists of two basic surfaces: the TIR surface at the edge part and the refractive surface at the middle part. When the LED light’s divergence angle is less than ${\theta _0}$, it will be collimated by the refraction surface. When the LED light’s divergence angle is larger than ${\theta _0}$, the excess will be collimated by the TIR surface.
In this paper, the classic TIR lens as a mature product was chosen. We will use the same method as in D Pan’s work to design the TIR lens [25]. The design process of the TIR lens will not be described in the paper, and the model of the TIR lens will be directly presented.
2.2 Diffuser sheet
To generate pulsed lighting, the static light spot should be an rectangular spot in the target plane. Therefore, we designed a microcylindrical lens as a unit of the diffuser sheet to expand the collimated light in one direction. The material of the diffuser sheet is polymethyl methacrylate (PMMA) with index of 1.4935. However, when the divergence angle of light emitted from the diffuser sheet is required to be too large, some light will be totally reflected inside the diffuser sheet and cannot reach the target surface, which will result in the loss of energy. In fact, the collimated light is not ideal parallel light, which will worsen this problem. To solve this problem, the cylindrical lens was designed with two optical surfaces, including a front surface and a rear surface. Then, the total reflection phenomenon will be solved because the angle of incidence of the rear surface is much smaller than the critical angle.
3. Optical design method
The optical system can be divided into TIR lenses and a diffuser sheet. Here we will focus on the design process of the diffuser sheet whose structure can be transformed from 3 dimensions to 2 dimensions. Next, we need to analyse the illuminance distribution of the light spot.
3.1 Illuminance distribution of the light spot
Figure 5(a) shows the ideal illuminance distribution of the rectangular light spot. The ideal illuminance distribution of the circular light spot is uniform as shown in Fig. 5(b). In Fig. 5(c), the rectangular light spots were symmetrically distributed in the y axis and divided into part I and part II. In the part I, a half-rectangular light spot is divided into N parts (G1, G2,$\cdots$, GN). The luminous flux of each small part are ΦS1, ΦS2,$\cdots$, and ΦSN. In part II, the luminous flux of each small part are Φ’S1, Φ’S2,…, and Φ’SN. When a rectangular light spot rotates, each part will also rotate and generate an annular light spot. The luminous flux of each annular light spot are ΦD1, ΦD2,$\cdots$, and ΦDN, as shown in Fig. 5(d). According to the law of energy conservation, the relationship between the luminous flux in the static state and dynamic state can be expressed as
Fig. 5. (a) The ideal illuminance distribution of rectangular light spot in the static state; (b) The ideal illuminance distribution of circular light spot in the dynamic state; (c) The ideal rectangular light spot in the static state; (d) The ideal light circular light spot of dynamic state.
Since the luminous flux through a plane is the product of the illuminance and the area of the plane, we can obtain Eqs. (3)–(5) from Eq. (2):
When the difference between Gxi and Gxi + 1 is small enough, Eq. (5) can be approximated as
With the combination of Eqs. (3)–(6), ESi can be expressed asThe value of $G{x_i}$ can be expressed as
where $\frac{1}{2}$ is the distribution factor of $G{x_i}$.3.2 Design method of the cylindrical lens
Due to the symmetry of the cylindrical lens, we can design only half of the front surface and the rare surface. In the ideal state, the rays emitted from the TIR lens are parallel with nonuniform intensity distribution. However, because of the small size of the cylindrical lens, the intensity distribution of the rays reach each cylindrical lens can be thought to be uniform. As shown in Fig. 6, the light can be divided by N equal parts (e1, e2,$\cdots$, eN) that reach the half cylindrical lens.
Light passing through the medium’s surface obeys Snell’s law, which is expressed as [34]
The process of designing the cylindrical lens can be divided into two parts: the front surface design and the rear surface design. Before designing the cylindrical lens, we need to determine some parameters, which include the reflective index of the material (n1), the distance between point H and the front surface vertex (L1), the half-width of the cylindrical lens (W0), the distance between the target plane and the cylindrical lens (L2) and the radius of the light spot (R).
We use an tangent method to design the surface [22], related method has been described by Jin Jia Chen’s group. In the tangent method, the next point (Ci + 1/Fi + 1) is the intersection point of the incident light vector and the tangent vector of the previous point (Ci/Fi). In addition, the tangent vector of related point can be calculated according to Eq. (10) after determining the unit direction vectors of incident light and emitted light. More details of the design has been described in Supplemental 1.
4. Construction of the pulsed-lighting LED luminaire
4.1 Merit function of the light spot
Before constructing the pulsed-lighting LED luminaire, we provided two functions to evaluate the performance of the luminaire, as shown in the follow [36]:
where U is illuminance uniformity, ${E_{min}}$ is the minimum illuminance value, and ${E_{avg}}$ is the average illuminance value, $\eta $ is optical efficiency of the optical system, ${\mathrm{\Phi }_{LEDs}}$ is the luminous flux of the LEDs light source, ${\mathrm{\Phi }_{Luminaire}}$ is the luminous flux of the luminaire.In analysing of the circular light spot’s uniformity, we put ${E_{Si}}$ (LightTools 8.4.0, USA) into Eq. (6) to calculate ${E_{Di}}$. Finally, we obtained the uniformity of the circular light spot based on Eq. (10).
4.2 Simulation and optimization of the luminaire
In the simulation of the luminaire, the LED chip was used with a size of 3.5 mm × 3.5 mm, the single luminous flux is set 100 lm. The TIR lens was designed according to the chip size. Before designing the diffuser sheet, we determined the basic parameters (L1, W0, n1, L2 and R). The thickness of the diffuser sheet is determined by L1 (1.55 mm) and W0 (0.65 mm) to match the length of the motor shaft. The material of the TIR lens and the diffuser sheet is PMMA with a refractive index of 1.49. The target plane is 1500 mm away from the luminaire, and the required radius of the dynamic light spot is 1000 mm. Under above parameters, the initial structure of the diffuser sheet is shown in Fig. 7(a). The LEDs and TIR lens are arranged as a ring array above the diffuser sheet, as shown in Fig. 7(b). The total luminous flux of the array LEDs is 800 lm. By ray tracing in the Monte Carlo method, we obtained illuminance data using 20,000,000 rays, which can be used to evaluate the performance of the luminaire.
Fig. 7. (a) The optical system model of the luminaire; (b) The array configuration of LEDs & TIR lenses.
Actually, LED is an extended light source different from point light source. Therefore, 0.5 as the distribution factor in Eq. (9) is no longer fit. To reduce the difficulty of optimization, we simulated a series of luminaire models with different distribution factors. The values of illuminance uniformity and optical efficiency of these luminaire models are shown in Fig. 8. As the distribution factor decreases, the illuminance uniformity increases and gradually flattens out and the optical efficiency decreases and gradually flattens out.
Fig. 8. The illuminance uniformity and optical efficiency of the rectangular light spot with different distribution factors.
According to the above results, we set 0.002 as the distribution factor to construct a initial luminaire model with illuminance uniformity of 0.814 and optical efficiency of 0.772. The luminous flux of the initial luminaire model was 617.2 lm. Then, we performed the optimization based on the feedback optimization algorithm [37]. The illuminance chart of the the final luminaire model is shown in Fig. 9(a). The illuminance uniformity and the optical efficiency are 0.84 and 0.768, which meet the targets for illuminance uniformity and optical efficiency according to relevant research and documents [38–41]. These research and documents suggest that the illuminance uniformity should be higher than 0.75 (+/-12.5%). As shown in Fig. 9(b), the luminous flux and the angle of divergence of the final luminaire model are 614.4 lm and 77.6°, respectively.
Fig. 9. (a) The illuminance chart of the final luminaire model; (b) The light intensity distribution of final luminaire model.
4.3 Tolerance analysis of the diffuser sheet
The diffuser sheet was manufactured by an injection molding process. Actually, there are position shift tolerances in this process, as shown in Fig. 10. In the Z-axis direction, the tolerance between the actual point and the ideal point is ΔZ. In the Y-axis direction, the tolerance between the actual point and the ideal point is ΔY.
Therefore, it is necessary to analyse the influence of tolerances on illuminance uniformity and optical efficiency in simulations. As shown in Fig. 11(a), the higher the absolute value of ΔZ is, the lower illuminance uniformity with the same ΔY. When the value of ΔZ is the same, the illuminance uniformity increases with decreasing ΔY. In general, the value of ΔZ should be in the range of -40 µm to 20 µm, and the value of ΔY should be in the range of 0 µm to 30 µm. As shown in Fig. 11(b), the higher value of ΔZ corresponds to the higher optical efficiency. When the value of ΔZ is the same, the optical efficiency increases as the value of ΔY decreases. Therefore, to meet the optical efficiency requirement, the value of ΔY should be higher than -80 µm, and the value of ΔZ should be higher than -50 µm.
Fig. 11. (a) The tolerance between the position shift and illuminance uniformity; (b) The tolerance between the position shift and optical efficiency.
5. Experiment for testing energy utilization efficiency
5.1 Measurement of luminaire parameters
To test the performance of the pulsed-lighting LED luminaire, we made a prototype as shown in Fig. 12(a). The rated voltage and rated power of the luminaire are 48 V and 60 W, respectively. Figure 12(b) and Fig. 12(c) show the practical TIR lens and diffuser sheet with the corresponding size marking.
Fig. 12. (a) Three views of the physical prototype; (b) The practical TIR lens and diffuser sheet; (c) The cross section of the diffuser sheet.
The radiant flux and luminous intensity distribution were measured by GO2000A_V1 (Distribution photometer, EVERFINE, China) and HP8000 (Integrating sphere, HOPOOCOLOR, China). The radiant flux and luminous intensity distribution of the LED array, TIR lens, and TIR lens with diffuser sheet are summarized in Table 1. The changes in radiant energy and luminous intensity distribution of light during propagation from the LED to the diffuser sheet is shown in Fig. 13.
Fig. 13. The changes of radiant energy and luminous intensity distribution of light during the propagation form LED to diffuser sheet.
Table 1. Summary of the different luminaires’ parameter detection
Figure 14(a) shows the rectangular light spot and the circular light spot of the prototype. The luminaire was hung about 1.5 m above the ground and used LI-120A and LI-190R (LI-COR, USA) to detect the PPFD value of the light spot. As shown in Fig. 14(b), the results were normalized and plotted. Due to machining errors in the injection molding process, the actual light spot is slightly different from the simulation result. By analysing the detection results, we obtained a PPFD uniformity (UL) of 78.9% within a radius (R) of 90 cm.
Fig. 14. (a) Two different light spots; (b) The normalized PPFD distribution of different light spots at 1.5 m.
5.2 Plant cultivation experiment
5.2.1 Plant material and growth conditions
We performed a cultivation experiment to verify the performance of the pulsed-lighting LED luminaire. The plant cultivation experiment was carried out in a hydroponic growing system (Hefei, Anhui, China) under two lighting treatments for 35 days from 10 October to 13 November 2022. Before the lighting treatments, seeds of Brassica chinensis were cultivated in a growth chamber for 14 days until the second true leaves had just opened. For each lighting treatment, 48 seedlings were placed in a tray (60 cm × 80 cm). A nutrient solution (pH:5.5-6.5, EC:1.2-1.4 mS/cm) was used in the hydroponic growing system. Black reflective plastic plates were used to separate the adjacent lighting systems, while black curtains were used to eliminate the influence of ambient light.
The Brassica chinensis seedlings received pulsed lighting and continuous lighting under the pulsed-lighting LED luminaire (PLL) and the conventional LED luminaire (CLL) respectively. The two types of luminaires both consisted of 13 red LED modules, 9 blue LED modules, and 3 white LED modules (Honglizhihui, C3535F5R3EA&C3535F26B3EA, China). The peak wavelengths of the spectrum are 451.6 nm and 666.8 nm, which match the absorption spectrum of chlorophyll. The photoperiod was set to 15 hours (from 7:00 to 22:00). LI-250A and LI-190R (LI-COR, USA) were used to detect the PPFD on the target plane. The uniformity test was implemented when the luminaires were hung at a height of 67 cm with radius of the illumination area of 30 cm. The relevant parameters of the two lighting systems are presented in Table 2.
Table 2. Relevant parameters of the two types of luminaires
5.2.2 Measurements
After 5 weeks, the plants were harvested. Under each lighting treatment, we selected 10 samples randomly to measure plant height, plant width, leaf length and leaf width. The fresh weight of the plants was measured by a digital balance (FD-C30001, Fengde,China) with a resolution of 0.1 g. To measure the dry weight of plants, the samples were placed in individual paper bags and then dried in a ventilated oven (SN-101X-4B, Sunne, China) at a temperature of 60°C over the course of 120 h. The dry weight of each plant was measured by a digital balance (JT3003D, Lichen, China) with a resolution of 0.001 g.
5.2.3 Results and analysis
The results of the plant height, plant width, leaf length, leaf width, fresh weight and dry weight are shown in Table 3. The total fresh weights of plants under the PLL and the CLL were 693.5 g and 521 g, respectively. The average fresh weight of plants under the pulsed-lighting LED luminaire was 33.1% higher than that under the conventional LED luminaire. The average dry weight of plants under the pulsed-lighting LED luminaire was 28.3% higher than that under the conventional LED luminaire. The growth status of some samples of Brassica chinensis at harvest is shown in Fig. 15.
Table 3. Results of the plants paramaters under two types of luminaires
The experimental results were analysed using Paired-Samples T Test by IBM SPSS Statistics 23.0 (IBM Corp, Armonk, NY, USA). P-value smaller than 0.05 is considered as significant difference for pairwise comparisons. As shown in Fig. 16(a), there is statistically significant difference in the fresh weights of Brassica chinensis grown under the PLL and the CLL with p-value of 0.023. Though there isn’t significant difference in the dry weights with p-value of 0.081, which still closes to 0.05 as shown in Fig. 16(b).
Fig. 16. (a) The boxplot of the fresh weights of Brassica chinensis under two lighting treatments; (b) The boxplot of the dry weights of Brassica chinensis under two lighting treatments (“ns” and “*” indicate non-significant and significant at p < 0.05, respectively).
Before analysising the EUE of the PLL, we provided a function to evaluate the EUE, as shown in the follow
where mfresh is total fresh weight of Brassica chinensis, Welectricity is electricity consumption during the cultivation period.In the Brassica chinensis cultivation period, the amounts of electricity consumption of the PLL and the CLL are 31.5 kW·h and 29.1 kW·h, respectively. Thus, the EUE of the PLL and the CLL are 22.0 g/kW·h and 17.9 g/kW·h, respectively. Therefore, the EUE of the PLL is 22.9% higher than that of the CLL. However, the investment costs of the PLL including an additional motor and secondary optics is ${\$}$43.85 with the rated power of 60 W. Compared with the PLL, the investment costs of the CLL is ${\$}$38.00 with the rated power of 55.5 W. Thus, the investment costs per watt of CLL and PLL, are 0.68 ${\$}$/W and 0.73 ${\$}$/W, respectively. In future work, we will consider how to reduce the investment costs per watt of the pulsed-lighting LED luminaire.
6. Conclusion
In this paper, we proposed a pulsed-lighting LED luminaire for plant supplemental lighting based on designing optical surfaces. Compared with conventional LED luminaires, this luminarie has many advantages, such as pulsed lighting, uniform illuminance and high LED utilization efficiency. The design method and construction approach of this luminaire were depicted in detail in this study with illuminance distribution analysis. We modeled, simulated, and optimized this luminaire, and the final simulation results showed that the illuminance uniformity was 0.84 and the optical efficiency was 0.768. The actual illuminance uniformity of the prototype of this luminaire was 0.789, which is better than 0.75 (+/- 12.5%). Depending on the unique illumination mode, this luminaire can achieve high energy utilization efficiency. The results of Brassica chinensis cultivation showed that the fresh weight under the pulsed-lighting LED luminaire was 33.1% higher than that under conventional LED luminaire. Considering the energy consumption of the two luminaires during the planting period, the energy utilization efficiency of the pulsed-lighting LED luminaire was 22.9% higher than that of the conventional LED luminaire. Therefore, the pulsed-lighting LED luminaire can potentially be implemented in greehouses with higher energy utilization efficiency to replace the HPS lamp.
Funding
Fuyang Municipal Government - Fuyang Normal University Horizontal Project (SXHZ202011); Plan for Anhui Major Provincial Science & Technology Project (202203a06020002); Science & Technology Program of Hebei (22326411D, 22327215D); Students' Innovation and Entrepreneurship Foundation of USTC (XY2022G01).
Acknowledgments
The authors thank Yanan Wu provided and cultivated the seeds of Brassica chinensis in the plant cultivation experiment.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
Supplemental document
See Supplement 1 for supporting content.
References
1. K. F. Toyoki Kozai and Erik S. Runkle, LED Lighting for Urban Agriculture (Springer, Singapore, 2016), pp. 49–51, 173-174.
2. D. Singh, C. Basu, M. Meinhardt-Wollweber, and B. Roth, “LEDs for energy efficient greenhouse lighting,” Renewable Sustainable Energy Rev. 49, 139–147 (2015). [CrossRef]
3. G. M. Weaver, M. W. van Iersel, and J. M. Velni, “A photochemistry-based method for optimising greenhouse supplemental light intensity,” Biosyst. Eng. 182, 123–137 (2019). [CrossRef]
4. W. J. P. Kuijpers, D. Katzin, S. van Mourik, D. J. Antunes, S. Hemming, and M. J. G. van de Molengraft, “Lighting systems and strategies compared in an optimally controlled greenhouse,” Biosyst. Eng. 202, 195–216 (2021). [CrossRef]
5. D. Katzin, L. F. M. Marcelis, and S. van Mourik, “Energy savings in greenhouses by transition from high-pressure sodium to LED lighting,” Appl. Energy 281, 116019 (2021). [CrossRef]
6. J. D. Wacker, M. J. Verheul, I. Righini, H. Maessen, and C. Stanghellini, “Optimisation of supplemental light systems in Norwegian tomato greenhouses-A simulation study,” Biosyst. Eng. 215, 129–142 (2022). [CrossRef]
7. T. Liu, Q. Y. Yuan, and Y. G. Wang, “Hierarchical optimization control based on crop growth model for greenhouse light environment,” Computers and Electronics in Agriculture 180, 105854 (2021). [CrossRef]
8. P. Kusuma, P. M. Pattison, and B. Bugbee, “From physics to fixtures to food: current and potential LED efficacy,” Hortic. Res. 7(1), 56 (2020). [CrossRef]
9. O. V. Avercheva, Y. A. Berkovich, I. O. Konovalova, S. G. Radchenko, S. N. Lapach, E. M. Bassarskaya, G. V. Kochetova, T. V. Zhigalova, O. S. Yakovleva, and I. G. Tarakanov, “Optimizing LED lighting for space plant growth unit: Joint effects of photon flux density, red to white ratios and intermittent light pulses,” Life Sciences in Space Research 11, 29–42 (2016). [CrossRef]
10. L. Carotti, G. Potente, G. Pennisi, K. B. Ruiz, S. Biondi, A. Crepaldi, F. Orsini, G. Gianquinto, and F. Antognoni, “Pulsed LED Light: Exploring the Balance between Energy Use and Nutraceutical Properties in Indoor-Grown Lettuce,” Agronomy 11(6), 1106 (2021). [CrossRef]
11. Said Abu-Ghosh, D. Fixler, Z. Duibinsky, and D. Iluz, “Flashing-light-in-microalgae-biotechnology,” Bioresour. Technol. 203, 357–363 (2016). [CrossRef]
12. J. Miliauskiene, R. F. Karlicek Jr., and E. Kolmos, “Effect of Multispectral Pulsed Light-Emitting Diodes on the Growth, Photosynthetic and Antioxidant Response of Baby Leaf Lettuce (Lactuca sativa L.),” Plants (Basel) 10(4), 762 (2021). [CrossRef]
13. E. Olvera-González, D. Alaniz-Lumbreras, R. Ivanov-Tsonchev, J. Villa-Hernandez, I. de la Rosa-Vargas, I. Lopez-Cruz, H. Silos-Espino, and A. Lara-Herrera, “Chlorophyll fluorescence emission of tomato plants as a response to pulsed light based LEDs,” Plant Growth Regul. 69(2), 117–123 (2013). [CrossRef]
14. E. Olvera-Gonzalez, D. Alaniz-Lumbreras, R. Ivanov-Tsonchev, J. Villa-Hernandez, C. Olvera-Olvera, E. Gonzalez-Ramirez, M. Araiza-Esquivel, V. Torres-Arguelles, and V. Castano, “Intelligent lighting system for plant growth and development,” Comput. Electron. Agriculture 92, 48–53 (2013). [CrossRef]
15. E. Olvera-Gonzalez, D. Alaniz-Lumbreras, V. Torres-Arguelles, E. Gonzalez-Ramirez, J. Villa-Hernandez, M. Araiza-Esquivel, R. Ivanov-Tsonchev, C. Olvera-Olvera, and V. M. Castano, “A LED-based smart illumination system for studying plant growth,” Light. Res. Technol. 46(2), 128–139 (2014). [CrossRef]
16. E. Olvera-Gonzalez, N. Escalante-Garcia, D. Myers, P. Ampim, E. Obeng, D. Alaniz-Lumbreras, and V. Castano, “Pulsed LED-Lighting as an Alternative Energy Savings Technique for Vertical Farms and Plant Factories,” Energies 14(6), 1603 (2021). [CrossRef]
17. P. S. C. Schulze, R. Guerra, H. Pereira, L. M. Schuler, and J. C. S. Varela, “Flashing LEDs for Microalgal Production,” Trends Biotechnol. 35(11), 1088–1101 (2017). [CrossRef]
18. A. Shimada and Y. Taniguchi, “Red and blue pulse timing control for pulse width modulation light dimming of light emitting diodes for plant cultivation,” Journal of Photochemistry and Photobiology B: Biology 104(3), 399–404 (2011). [CrossRef]
19. V. Vaštakaite, A. Virsile, A. Brazaityte, G. Samuoliene, J. Jankauskiene, A. Novickovas, and P. Duchovskis, “Pulsed Light-Emitting Diodes for a Higher Phytochemical Level in Microgreens,” J. Agric. Food Chem. 65(31), 6529–6534 (2017). [CrossRef]
20. K. P. Ferentinos and L. D. Albright, “Optimal design of plant lighting system by genetic algorithms,” Engineering Applications of Artificial Intelligence 18(4), 473–484 (2005). [CrossRef]
21. B. S. Wu, Y. Hitti, S. MacPherson, V. Orsat, and M. G. Lefsrud, “Comparison and perspective of conventional and LED lighting for photobiology and industry applications,” Environmental and Experimental Botany 171, 103953 (2020). [CrossRef]
22. J. J. Chen and C. T. Lin, “Freeform surface design for a light-emitting diode-based collimating lens,” Opt. Eng. 49(9), 093001 (2010). [CrossRef]
23. J. J. Chen, T. Y. Wang, K. L. Huang, T. S. Liu, M. D. Tsai, and C. T. Lin, “Freeform lens design for LED collimating illumination,” Opt. Express 20(10), 10984–10995 (2012). [CrossRef]
24. D. L. Ma, Z. X. Feng, and R. G. Liang, “Freeform illumination lens design using composite ray mapping,” Appl. Opt. 54(3), 498–503 (2015). [CrossRef]
25. D. Pan, Y. C. Zhang, X. F. Xie, M. J. Fan, Y. A. Chen, S. H. Yan, and S. J. Yang, “Optical design of LED high mast lamp based on TIR freeform lens for annular lighting,” Light. Res. Technol. 53(6), 569–579 (2021). [CrossRef]
26. S. Zhao, K. Wang, F. Chen, Z. Qin, and S. Liu, “Integral freeform illumination lens design of LED based pico-projector,” Appl. Opt. 52(13), 2985–2993 (2013). [CrossRef]
27. Z. L. Zhao, H. H. Zhang, S. Liu, and X. Z. Wang, “Effective freeform TIR lens designed for LEDs with high angular color uniformity,” Appl. Opt. 57(15), 4216–4221 (2018). [CrossRef]
28. Z. Zhenrong, X. Hao, and X. Liu, “Freeform surface lens for LED uniform illumination,” Appl. Opt. 48(35), 6627–6634 (2009). [CrossRef]
29. H. L. Wen and B. S. Lin, “Improvement of illumination uniformity for LED flat panel light by using micro-secondary lens array,” Opt. Express 20(S6), A788–A798 (2012). [CrossRef]
30. X. H. Lee, I. Moreno, and C. C. Sun, “High-performance LED street lighting using microlens arrays,” Opt. Express 21(9), 10612–10621 (2013). [CrossRef]
31. A. P. Riser, “Analysis of single lens arrays using convolution,” Opt. Eng. 40(5), 805 (2001). [CrossRef]
32. C. C. Sun, X. H. Lee, I. Moreno, C. H. Lee, Y. W. Yu, T. H. Yang, and T. Y. Chung, “Design of LED Street Lighting Adapted for Free-Form Roads,” IEEE Photonics J. 9(1), 1–13 (2017). [CrossRef]
33. D. T. Vu, H. Vu, S. Shin, N. M. Kieu, T. Q. Tien, and N. H. Vu, “A simple design of uniform LED illumination using catadioptric collimator and freeform lenslet array,” Light. Res. Technol. 54(6), 546–562 (2022). [CrossRef]
34. R. M. Wu, H. F. Li, Z. R. Zheng, and X. Liu, “Freeform lens arrays for off-axis illumination in an optical lithography system,” Appl. Opt. 50(5), 725–732 (2011). [CrossRef]
35. X. H. Lee, Y. Y. Chang, and C. C. Sun, “Highly energy-efficient agricultural lighting by B plus R LEDs with beam shaping using micro-lens diffuser,” Opt. Commun. 291, 7–14 (2013). [CrossRef]
36. I. Moreno, “Illumination uniformity assessment based on human vision,” Opt. Lett. 35(23), 4030–4032 (2010). [CrossRef]
37. Y. Luo, Z. X. Feng, Y. J. Han, and H. Y. Li, “Design of compact and smooth free-form optical system with uniform illuminance for LED source,” Opt. Express 18(9), 9055–9063 (2010). [CrossRef]
38. E. Runkle, “The Importance of Light Uniformity,” retrieved https://gpnmag.com/atricle/the-importance-of-light-uniformity/.
39. Z. Zheng, Z. Ma, R. Li, H. Wei, X. Chu, and H. Liu, “Design method for improving illumination distribution uniformity of LED plant growth light source,” in 2021 18th China International Forum on Solid State Lighting & 2021 7th International Forum on Wide Bandgap Semiconductors (SSLChina: IFWS), (2021), 202–210.
40. WESTLAND, “LED plant light uniformity design,” retrieved https://www.west-land.com/Expertise/led-plant-light-uniformity-design.html.
41. ENERGY STAR, “Lunimaires Specificaiton Version 2.2,” U. S. D. o. Energy, ed.
