LED spectrum distribution optimization design method and LED light source

Abstract

This invention discloses an LED spectral distribution optimization design method and an LED light source. The design method includes the following steps: S1, establishing a basis function library based on the Gaussian emission spectral function of phosphor materials; S2, defining the full spectral distribution function and optimization variables of the phosphor mixture; S3, constructing a comprehensive evaluation function and constraints based on color temperature, spectral efficacy, and color rendering index; S4, applying simulated annealing algorithm for optimization, aiming to minimize the comprehensive evaluation function, and globally optimizing the basis function weight vector w; S5, adjusting the comprehensive evaluation function and constraints and iteratively optimizing. This design method is applicable to the spectral control of white LEDs excited by multi-color phosphors, and can achieve synergistic optimization of color rendering index, color temperature accuracy, and luminous efficacy over a wide color temperature range, significantly improving the overall light quality and design efficiency of the LED light source.

Description

Technical Field

[0001] This invention relates to the field of semiconductor light-emitting device technology, and in particular to an LED spectral distribution optimization design method, and to an LED light source obtained by the design method. Background Technology

[0002] White LEDs (Light Emitting Diodes) are widely used in general lighting, special lighting, and display backlighting as a high-efficiency and environmentally friendly solid-state lighting source. Their performance is primarily measured by three key indicators: Color Rendering Index (CRI / Ra), which evaluates the light source's ability to reproduce the true colors of objects; Color Temperature (CCT), which determines the warm or cool tone of the light source; and Luminous Efficacy (LE), which reflects the efficiency with which the light source converts electrical energy into visual luminous flux. In high-end lighting scenarios (such as museum lighting, medical lighting, and film shooting), light sources are often required to simultaneously possess high color rendering index (Ra > 90, or even higher for special color rendering indices like R9), precisely adjustable color temperature (such as stepless dimming and color adjustment within the range of 2700K-6500K), and minimize luminous efficacy loss due to the pursuit of high light color quality.

[0003] Currently, the mainstream technical solutions for achieving spectral modulation of white LEDs mainly include:

[0004] Blue LED chip-excited phosphor scheme: This method uses a single blue LED chip to excite YAG phosphor to produce yellow light, which is then mixed to form white light. This method has a simple structure and high luminous efficiency, but its spectral continuity and richness are insufficient, resulting in a low color rendering index (typically Ra < 80), a narrow color temperature control range, and limited precision, making it difficult to simultaneously meet the requirements of high color rendering and high-precision color temperature control.

[0005] Multi-color phosphor excitation scheme: This method uses blue or ultraviolet chips to excite multiple phosphors, such as red, green, and blue phosphors, to broaden the spectrum. While this method can improve the color rendering index, the differences in efficiency, temperature characteristics, and attenuation coefficients among different phosphors cause color temperature drift with changes in driving current or temperature, making precise color temperature control difficult to guarantee. Furthermore, the energy conversion of multiple phosphors involves cumulative Stokes losses, which significantly reduce the overall luminous efficiency of the system.

[0006] Multi-chip (RGB or RGBW) hybrid scheme: This scheme integrates red, green, blue, and even white LED chips into a single package, synthesizing the desired white light by independently controlling the drive current of each chip. This approach offers advantages in color temperature control range and precision. However, because each chip's spectrum has a narrow peak spectrum, the spectral continuity of the synthesized white light is poor, resulting in a generally low color rendering index, particularly in the reproduction of red and saturated colors (e.g., R9 value). Furthermore, the luminous efficacy and photoelectric properties of different color chips vary significantly. During the color mixing process, the overall luminous efficacy of the system fluctuates greatly with changes in the mixing ratio, making it difficult to maintain high luminous efficacy over a wide color temperature range.

[0007] In summary, existing technologies exhibit significant interdependencies in achieving the three core performance indicators of high color rendering index (CRI), high-precision wide-range color temperature (CCT) control, and high luminous efficacy (LE), often referred to as the "light-color-electric iron triangle dilemma." Specifically, improving color rendering and expanding color temperature control often comes at the cost of luminous efficacy; conversely, pursuing high luminous efficacy makes it difficult to simultaneously achieve excellent light and color quality and precise control.

[0008] Therefore, there is an urgent need in this field for an innovative LED spectral design and optimization method that can systematically and collaboratively optimize the spectral distribution of the light source from the fundamental level, thereby breaking through the existing technical bottlenecks and achieving the simultaneous optimization of high color rendering index, high-precision wide-range color temperature control and high luminous efficacy. Summary of the Invention

[0009] The primary technical problem to be solved by this invention is to provide an LED spectral distribution optimization design method, which is based on spectral basis function modeling and simulated annealing algorithm to obtain an LED spectral distribution that has high color rendering index, high precision color temperature control and high luminous efficacy.

[0010] Another technical problem to be solved by the present invention is to provide an LED light source.

[0011] To achieve the above-mentioned technical objectives, the present invention adopts the following technical solution:

[0012] An LED spectral distribution optimization design method, applicable to spectral modulation of white LEDs excited by multicolor phosphors, includes the following steps:

[0013] S1. Using the Gaussian emission spectrum function of phosphors excited by light of a specific wavelength as the basis function, establish a basis function library consisting of Gaussian emission spectrum functions corresponding to various phosphors; where, for a known dominant wavelength λ... p The phosphor with a full width at half maximum (FWHM) Δλ has a Gaussian emission spectral function S. p (λ) can be expressed as: Where λ represents the excitation wavelength of the phosphor, λp λ is the center wavelength of the Gaussian emission spectral function corresponding to the phosphor, and Δλ is the full width at half maximum (FWHM) of the Gaussian emission spectral function corresponding to the phosphor.

[0014] S2, using the multiple Gaussian emission spectral functions S corresponding to the multiple phosphor materials in step S1. p (λ) is used as a basis function, with the weight coefficient w of each basis function. i As variables, the full-spectral distribution function of the phosphor mixture is established through the superposition of multiple basis functions: Where λ represents the excitation wavelength of the phosphor mixture, n is the number of basis functions, and λ p,i Δλ is the center wavelength of the i-th Gaussian emission spectral function. i Let be the full width at half maximum (FWHM) of the i-th Gaussian emission spectral function;

[0015] S3. Based on the full-spectrum distribution function in step S2, construct the color temperature evaluation function f. CCT Spectral visual effect evaluation function f LER and the display index evaluation function f Ra The three evaluation functions are summed to construct a comprehensive evaluation function, Score, for the phosphor mixture: ; S4. The simulated annealing algorithm is applied for optimization. With the goal of minimizing the comprehensive evaluation function Score, the basis function weight vector w is globally optimized and the optimal weight vector is output, thereby obtaining the optimal spectral distribution scheme.

[0016] S5. If the optimized spectral distribution deviates from the target, adjust the weight coefficient vector w in the comprehensive evaluation function and repeat step S4 until the LED spectral distribution that meets the requirements is obtained.

[0017] Preferably, in step S3, the expression for the color temperature evaluation function is: Among them, w CCT This is a weighting constant for color temperature in optimization, typically ranging from 10. 4 <w CCT <10 5 T target Let CCT(S) be the target color temperature, CCT(S) be the color temperature corresponding to the current spectral distribution S calculated using the Robertson iterative method, and Duv(S) be the distance of the chromaticity coordinates corresponding to the current spectral distribution S from the blackbody locus. Duv This is the weighting constant for color cast Duv(S), typically in the range of 10. 6 <w Duv<10 7 .

[0018] Preferably, in step S3, the expression f of the spectral visual effect evaluation function is... LER for: Among them, w LER This is a weighting constant for spectral visual effects in the optimization, with a value range of 10. 3 <w LER <10 4 LER min The minimum value for lumens visual effect is given by δ, which is a constant factor with a value range of 10 < δ < 10. 2 ;

[0019] LER(S) is the spectral visual effect function of the current spectral distribution S, and its calculation formula is as follows: Where V(λ) is the human eye's visual sensitivity function.

[0020] Preferably, in step S3, the expression f of the color rendering index evaluation function is... Ra for: Among them, w Ra This is the weighting constant of the color rendering index in the optimization, with a value range of 10. 4 <w Ra <10 5 Ra min Ra(S) represents the minimum required color rendering index, where Ra(S) is the color rendering index of the current spectral distribution S, and γ is a constant factor with a value range of 10 < γ < 10. 2 .

[0021] Preferably, step S3 further includes the following sub-step: using the normalized second-order difference operator χ as a smoothing constraint, the calculation formula for the second-order difference operator χ is as follows: ; Among them, w i λ represents the basis function weights. i The wavelength is the center wavelength of the basis functions. By minimizing the sum of squares of all difference operators, the energy distribution of adjacent basis functions is forced to conform to physical continuity, eliminating spurious peaks in mathematical fitting and ensuring the physical continuity of the synthesized spectrum.

[0022] Preferably, in step S4, the specific process of optimizing using the simulated annealing algorithm includes:

[0023] S4.1 Adaptive estimation of initial temperature: Calculate the initial value S0 of the comprehensive evaluation function based on the initial weight vector w0. Apply multiple small-range random perturbations to the initial weight vector w0 while maintaining non-negative weights. Calculate the comprehensive evaluation function value corresponding to the perturbation sample, and calculate the absolute difference between the comprehensive evaluation function value corresponding to the perturbation sample and the initial value of the comprehensive evaluation function. Take the median of the difference as the typical scale estimate ΔS. t And calculate the initial temperature T0 based on this: ;

[0024] S4.2 Algorithm Parameter Configuration and Iterative Optimization: Configure key parameters of the simulated annealing algorithm, including: initial temperature T0, maximum number of iterations iter max The cooling coefficient α is optimized using the following algorithm iteratively:

[0025] (1) Generate a new solution: under the current temperature, optimize the current weight vector w current Candidate solutions are generated by applying a normally distributed random perturbation while maintaining non-negative weights;

[0026] (2) Evaluation and Metropolis criterion acceptance: Calculate the difference between the comprehensive evaluation function value corresponding to the candidate solution and the current solution respectively. If the difference is less than 0, the candidate solution is directly accepted. If the difference is greater than or equal to 0, it is accepted with a probability of P=exp(-ΔS / T), where ΔS is the function difference and T is the current temperature.

[0027] (3) Cooling and convergence judgment: After completing the predetermined number of iterations at the current temperature, the temperature is reduced according to the cooling coefficient, and the iterative optimization steps (2) are repeated until the temperature drops to the termination threshold or the optimal solution has no improvement for multiple consecutive generations;

[0028] (4) Output the optimal solution: When the algorithm terminates, it outputs the final optimal weight vector w. opt and the corresponding minimum value S of the comprehensive evaluation function min .

[0029] Preferably, in step S4, the simulated annealing algorithm is implemented using the simulannealbnd function.

[0030] A computer program product includes a computer program that, when executed by a processor, implements the steps of the above-described LED spectral distribution optimization design method.

[0031] A computer-readable storage medium storing a computer program that, when executed, implements the steps of the above-described LED spectral distribution optimization design method.

[0032] An LED light source comprising multiple phosphors, wherein the weight percentage of each phosphor is equal to the weight of each phosphor in the optimization results obtained by the above design method.

[0033] This invention proposes a global optimization method for LED spectra based on spectral basis function modeling and simulated annealing algorithm, applicable to white LED spectral modulation using multi-color phosphor excitation schemes. This method uses the Gaussian emission spectrum of common phosphors as basis functions, with the weights of each basis function as optimization variables, to construct a full-spectrum distribution and comprehensive evaluation function, and then performs global optimization using a simulated annealing algorithm. The optimized comprehensive evaluation function comprehensively considers high-precision color coordinates and color temperature control, luminous efficacy weighted by the visibility function, and color rendering index calculation based on ray tracing, thereby achieving synergistic optimization of color rendering index, color temperature accuracy, and luminous efficacy over a wide color temperature range. This invention provides an efficient spectral design solution for high-end lighting, full-spectrum healthy light sources, and intelligent dimming systems, significantly improving the overall light quality and design efficiency of LED light sources. Attached Figure Description

[0034] Figure 1 This is a flowchart of a global optimization design algorithm for LED spectral distribution based on simulated annealing;

[0035] Figure 2 It is the iterative result of the comprehensive evaluation function corresponding to the optimized design of the spectrum with a color temperature of 4000K, Ra≥98, and LER≥250;

[0036] Figure 3 It is a spectral distribution with a color temperature of 4000K, Ra≥98, and LER≥250 achieved through optimized design;

[0037] Figure 4 It is the iterative result of the comprehensive evaluation function corresponding to the optimized design of the spectrum with a color temperature of 3000K, Ra≥98, and LER≥280;

[0038] Figure 5 It is a spectral distribution with a color temperature of 3000K, Ra≥98, and LER≥280 achieved through optimized design;

[0039] Figure 6 It is the iterative result of the comprehensive evaluation function corresponding to the optimized design of the spectrum with a color temperature of 5000K, Ra≥98, and LER≥250;

[0040] Figure 7 It is a spectral distribution with a color temperature of 5000K, Ra≥98, and LER≥250 achieved through optimized design;

[0041] Figure 8 It is the iterative result of the comprehensive evaluation function corresponding to the optimized design of the spectrum with a color temperature of 4000K, Ra≥97, and LER≥300;

[0042] Figure 9 It is a spectral distribution with a color temperature of 4000K, Ra≥97, and LER≥300 achieved through optimized design. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0044] It should be noted that if the embodiments of the present invention involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0045] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0046] This invention provides a spectral optimization design method for LED light sources, which can systematically and collaboratively optimize the spectral distribution of the light source from the fundamental level, thereby breaking through the existing technical bottlenecks and achieving the simultaneous optimization of the performance of high color rendering index (CRI / Ra), high-precision wide-range color temperature (CCT) control and high luminous efficacy (LE).

[0047] This invention proposes a global optimization method for LED spectra based on simulated annealing algorithm and spectral basis function modeling, applicable to white LED spectral modulation using multi-color phosphor excitation schemes. This method uses the Gaussian emission spectral function of common phosphors as basis functions, and the weighting coefficients of each basis function as optimization variables to construct a multi-dimensional weighted comprehensive evaluation function, which is then globally optimized using a simulated annealing algorithm. The optimized comprehensive evaluation function comprehensively considers high-precision color coordinates and color temperature control, luminous efficacy weighted by the visual function, and color rendering index calculation based on ray tracing, thereby achieving optimal synergy between color rendering index (Ra), color temperature accuracy, and luminous efficacy over a wide color temperature range. This invention provides an efficient spectral design solution for high-end lighting, full-spectrum healthy light sources, and intelligent dimming systems, significantly improving the overall light quality and design efficiency of LED light sources.

[0048] like Figure 1 The diagram shows the flowchart of the global optimization design algorithm for LED spectral distribution based on simulated annealing provided by this invention. The following is based on... Figure 1 The flowchart shown provides a detailed description of the algorithm steps of the LED spectral distribution optimization design method provided by this invention.

[0049] Step S1: Establish a basis function library for Gaussian emission spectral functions based on phosphor materials.

[0050] As is well known to those skilled in the art, phosphor materials, when excited by light of a specific wavelength, exhibit a spectral distribution with a Gaussian curve shape determined by their fluorescent properties. In optical engineering applications, the spectrum of phosphors is typically approximated as an asymmetric Gaussian distribution. For a known dominant wavelength λ... p The relative spectral energy distribution (i.e., Gaussian emission spectral function) of a phosphor with a half-width at half-maximum Δλ, S. p (λ) can be expressed as: (1) Where λ is the excitation wavelength of the phosphor, λ p λ is the center wavelength of the Gaussian emission spectral function corresponding to the phosphor material, and Δλ is the full width at half maximum (FWHM) of the Gaussian emission spectral function corresponding to the phosphor.

[0051] Therefore, by using multiple discrete Gaussian emission spectral functions corresponding to common phosphor materials as basis functions, and optimizing the weighting coefficients (corresponding to weight percentages) of the basis functions, a spectral distribution with engineering feasibility can be achieved by superimposing multiple discrete Gaussian emission spectral functions. This is a basic idea behind the optimized design of the spectrum in this invention. To this end, the Gaussian emission spectra of various phosphors under specific wavelength illumination are used as basis functions to construct a basis function library composed of discrete Gaussian emission spectral functions corresponding to various phosphors.

[0052] Step S2: Define the full-spectrum distribution function and optimization variables for the phosphor mixture.

[0053] Multiple discrete Gaussian emission spectral functions corresponding to multiple phosphor materials are used as basis functions, with the weighting coefficient w of each basis function as the basis function. i As variables, the superposition of multiple basis functions can be used to construct the full-spectrum distribution function of the phosphor mixture: (2) Where λ is the excitation wavelength of the phosphor mixture, and n is the number of basis functions (the number of phosphors involved in the design). p,i Δλ is the center wavelength of the Gaussian emission spectral function corresponding to the i-th phosphor material. i Let w be the full width at half maximum (FWHM) of the i-th Gaussian emission spectral function. i These are the weighting coefficients for each basis function. For example, the spectrum constructed in this invention has a distribution range between 380 nm and 780 nm.

[0054] Step S3: Construct a comprehensive evaluation function and constraints based on color temperature, spectral visual effect, and color rendering index.

[0055] To achieve precise optimization of the synthesized spectral distribution S(λ), this embodiment constructs a multidimensional weighted comprehensive evaluation function, transforming colorimetric indices, physical efficacy, and colorimetric quality into a unified mathematical minimum problem. The specific construction process is as follows:

[0056] (1) Construction of color temperature evaluation function

[0057] The color temperature evaluation function aims to constrain the chromaticity positions of the synthesized spectrum to approximate a preset target color temperature T. target And it falls precisely near the blackbody locus (BBL). Color temperature evaluation formula f CCT The expression is as follows: (3) Among them, w CCT This is a weighting constant for color temperature in optimization, typically ranging from 10. 4 <w CCT <10 5 T target Let CCT(S) be the target color temperature, CCT(S) be the color temperature corresponding to the current spectral distribution S calculated using the Robertson iterative method, and Duv(S) be the distance of the chromaticity coordinates corresponding to the current spectral distribution S from the blackbody locus. Duv This is the weighting constant for color cast Duv(S), typically in the range of 10. 6 <w Duv <10 7The first term of the formula uses the square of the relative error to ensure consistency in the optimization weights under different color temperature standards; the second term introduces a weight w. Duv As a hard constraint, the spectrum is forced to fall within the visually comfortable white light region, avoiding a greenish or purplish tint.

[0058] (2) Construction of Spectral Visual Effect Evaluation Function

[0059] The spectral visual effect evaluation function primarily targets luminous efficacy, specifically lumen efficiency (LER), ensuring that high-quality light color is achieved without sacrificing energy efficiency. To guarantee the continuity of the optimization process, this embodiment constructs a spectral visual effect evaluation function f based on the "penalty function method". LER : (4) Among them, w LER This is a weighting constant for spectral visual effects in the optimization, typically in the range of 10. 3 <w LER <10 4 LER min The minimum value for the desired spectral visual effect, where δ is a constant factor, typically in the range of 10 < δ < 10. 2 LER(S) is the spectral efficacy function (i.e., lumen efficiency) of the current spectral distribution S. Those skilled in optics should know that its calculation formula is: (5) Where V(λ) is the human eye's visual sensitivity function.

[0060] According to formulas (4) and (5), it can be seen that the spectral visual effect function f based on the "penalty function method" is... LER A "one-sided penalty term" is constructed using the max function. When LER exceeds a set threshold, ... min When the value is 300 lm / W, this contribution is 0; once it falls below the threshold, the error increases dramatically with the square of the deviation, forcing the optimizer to find an energy distribution that is more in line with visual efficiency.

[0061] (3) Construction of the color rendering index evaluation function

[0062] The color rendering index (CRI) evaluation function is used to ensure the light source's ability to reproduce the colors of objects. For high-quality lighting requirements, this embodiment focuses on constraining the general CRI Ra, and its evaluation function f... Ra Defined as: (6) Among them, w Ra This is the weighting constant of the color rendering index in the optimization, typically in the range of 10. 4 <w Ra <10 5Ra min Ra(S) represents the minimum value of the desired colorimetric index, where Ra(S) is the colorimetric index corresponding to the spectral distribution S, and γ is a constant factor, typically in the range of 10 < γ < 10. 2 .

[0063] In this invention, the color rendering index is calculated using a direct ray tracing method. As is known to those skilled in the art, a known spectral distribution can be input into optical simulation software, and the color rendering index corresponding to that spectrum can be obtained through non-sequential ray tracing.

[0064] (4) Comprehensive evaluation function

[0065] Finally, by summing the above three evaluation functions, a comprehensive evaluation function for the entire phosphor mixture system is constructed: (7) This function uses simulated annealing optimization iteration, the simulated annealing optimization process is detailed below, to find the basis function coefficient weight vector w that minimizes the comprehensive evaluation function Score, thereby outputting the optimal weight vector w and obtaining the optimal spectral distribution scheme.

[0066] (5) Construct a normalized smoothing constraint based on wavelength domain gradient.

[0067] Considering the non-uniformity of the basis function distribution along the wavelength axis, and to eliminate spurious peaks generated by mathematical fitting (where multiple Gaussian functions may overlap, potentially producing local spikes that are detrimental to practical applications), this invention introduces a normalized second-order difference operator χ: (8) Where w i Let λ be the basis function weight vector. i The wavelength is the center wavelength of the basis functions. By minimizing the sum of squares of all difference operators, the energy distribution of adjacent basis functions is forced to conform to physical continuity, thereby obtaining a synthetic spectrum with no spurious peaks and a smooth shape.

[0068] This smoothing technology is not only applicable to optimizing the ratio of LED phosphors, preventing abrupt changes in the spectral shape during dimming, thus ensuring a smooth transition of color rendering index and visual brightness when switching color temperatures; in addition, it enhances color stability in LED engineering fabrication: the spectrum obtained through second-order difference smoothing constraints has stronger color robustness to fluctuations in phosphor dispensing amount in actual production.

[0069] Step S4: Apply simulated annealing algorithm for optimization.

[0070] The simulated annealing global optimization algorithm is used to optimize the vector w of the n weight coefficients representing the spectral basis functions in formula (2). iThe algorithm seeks to minimize the overall evaluation function, Score. The implementation process includes core steps such as initial temperature estimation, iterative perturbation, probability acceptance, and annealing cooling. The specific steps are as follows:

[0071] (1) Adaptive estimation of initial temperature

[0072] To improve algorithm efficiency and robustness, an adaptive temperature initialization method based on the scaling of a comprehensive evaluation function is adopted. Specifically:

[0073] 1. Using the initial weight vector w0 as a reference, calculate the initial value S0 of the corresponding comprehensive evaluation function.

[0074] 2. Apply multiple random perturbations to w0 to generate a set of perturbation sample weight vectors {w k |k=1, 2, ..., n}, where N is the number of samplings (usually 20~30 times), and the perturbation amplitude is controlled within ±8% of the weight each time, and the weight after perturbation is kept non-negative.

[0075] 3. Calculate the comprehensive evaluation function value {S} corresponding to all disturbance samples. k}, and calculate the absolute difference ΔS between it and S0. k =|S k −S0|.

[0076] 4. Take all ΔS k The median is used as a typical scale estimate of the change in the comprehensive evaluation function, ΔS. t And based on this, the initial temperature T0 is calculated, and the calculation formula is: (9) This formula ensures that, in the initial stage of optimization, the variation of the comprehensive evaluation function value is approximately ΔS. t The probability of accepting a candidate solution is about 50%, thus achieving a good balance between exploration and convergence.

[0077] (2) Algorithm parameter configuration and iterative optimization

[0078] Configure key parameters for the simulated annealing algorithm, including: initial temperature set to T0, maximum number of iterations iter max The cooling coefficient (less than 1, such as 0.95) is used. The core iterative process of the algorithm is as follows:

[0079] 1. Generate a new solution: At the current temperature T, for the current optimal weight vector w current Perform random perturbation to generate candidate solutions w new The current optimal weight vector w currentThis refers to the initial weight vector w0 and the latest spectral optimized weight vector generated in subsequent iterations. The perturbation method involves applying a small random change Δw to one or more components of the weight vector, which can be expressed as: (10) Where ξ is a small random quantity that follows a normal distribution, and it is necessary to ensure that all weights w after perturbation are within acceptable limits. new,i ≥0.

[0080] 2. Evaluation and acceptance according to Metropolis criteria: Calculate the comprehensive evaluation function value S corresponding to the candidate solution. new and the function value S of the current solution. current By comparison, the difference ΔS = S is obtained. new -S current .

[0081] If ΔS < 0, it indicates that the candidate solution is better, then w is accepted. new As the new current solution.

[0082] If ΔS≥0, then the candidate solution is accepted with probability P=exp(-ΔS / T). This mechanism gives the algorithm a chance to escape local optima.

[0083] 3. Cooling and Convergence Judgment: After completing the predetermined number of iterations at the current temperature, reduce the temperature according to the preset cooling plan, for example, T = α*T, where α is a cooling coefficient less than 1 (e.g., 0.93). Repeat steps 1-2 until the temperature drops to the termination threshold or other convergence conditions are met (e.g., the optimal solution has not improved for multiple consecutive generations).

[0084] 4. Output the optimal solution.

[0085] When the algorithm terminates, it outputs the final optimal weight vector w. opt and the corresponding minimum value S of the comprehensive evaluation function min w opt This characterizes the optimal contribution ratio of each phosphor basis function in the synthesized spectrum when the target color temperature, high color rendering index, and high luminous efficacy are combined to achieve the best overall effect.

[0086] The simulated annealing algorithm can be implemented by calling the optimization toolbox in the programmable mathematical calculation tool (such as the simulannealbnd function), and the optimization process (such as the change of the optimal function value, the temperature drop curve, etc.) can be recorded and visualized in real time through a custom output function, which is convenient for monitoring and debugging.

[0087] Step S5: Adjust the comprehensive evaluation function and constraints.

[0088] After the aforementioned optimization, the obtained spectral distribution may still deviate from the target. At this time, the weight coefficient vector w of each item in the comprehensive evaluation function based on color temperature, spectral visual effect and color rendering index can be manually adjusted as appropriate. Then, step S4 is performed to further optimize by applying the simulated annealing algorithm, and finally obtain the spectral distribution result that meets the requirements.

[0089] To illustrate the effectiveness of the LED spectral distribution optimization design method proposed in this application, specific embodiments are given below.

[0090] Example 1: The optimization goal is to obtain a spectral distribution with a color temperature of 4000K, a color rendering index ≥98, and a LER ≥250.

[0091] This embodiment realizes an optimized design example of a 4000K color temperature spectral distribution, which has both high spectral quality and high color rendering index.

[0092] Table 1 below shows the spectral information of the Gaussian function used in the optimization. Figure 2 The curves showing the change of the comprehensive evaluation function during the optimization process are presented. It can be seen that the error value decreases significantly throughout the optimization process, ultimately resulting in a better optimization result. Figure 3 The results of spectral optimization are shown, with a color temperature of 3998K, a color shift (Duv) of -0.0008, a color rendering index (CRI) of 98.04, and a LER of 263. The spectral curves show an overall smooth surface, indicating good spectral distribution quality and engineering feasibility. Table 2 provides the corresponding... Figure 3 The optimal weights were obtained using a simulated annealing algorithm for the 4000K spectrum.

[0093] Table 1. Spectral information of the Gausky function used in spectral optimization Serial Number Wave peak wavelength (nm) Bandwidth (nm) 1 433.3 15 2 446.7 15 3 460.0 20 4 473.3 20 5 486.7 25 6 500.0 25 7 513.3 30 8 526.7 35 9 540.0 40 10 553.3 40 11 566.7 45 12 580.0 45 13 593.3 50 14 606.7 50 15 620.0 50 16 633.3 40 17 646.7 55 18 660.0 60 19 673.3 70 20 686.7 80

[0094] Table 2. Corresponding Figure 3 Optimal weighting of the 4000K spectrum Example 2: The optimization goal is to obtain a spectral distribution with a color temperature of 3000K, a color rendering index ≥98, and a LER ≥280.

[0095] This embodiment realizes an optimized design example of a low color temperature spectral distribution, which has both high spectral quality and high color rendering index.

[0096] This embodiment also uses the Gaussian function spectral information shown in Table 1. Figure 4 The curves showing the change of the comprehensive evaluation function during optimization are presented. It can be seen that the error value decreases significantly throughout the optimization process, ultimately leading to a better optimization result. Figure 5 The results of spectral optimization are shown, with a color temperature of 2996 K, a color shift (Duv) of 0.00, a color rendering index (CRI) of 98.20, and a LER of 281. The spectral curves show an overall smooth surface, indicating good spectral distribution quality and engineering feasibility. Table 3 provides the corresponding... Figure 5 The optimal weights obtained by the 3000K spectral simulated annealing algorithm.

[0097] Table 3. Correspondence Figure 5 Optimal weighting of the 3000K spectrum Example 3: The optimization goal is to obtain a spectral distribution with a color temperature of 5000K, a color rendering index ≥98, and a LER ≥250.

[0098] This embodiment realizes an optimized design example of a high color temperature (5000K) spectral distribution, which has both high spectral quality and high color rendering index.

[0099] This embodiment also uses the Gaussian function spectral information shown in Table 1. Figure 6 The curves showing the change of the comprehensive evaluation function during optimization are presented. It can be seen that the error value decreases significantly throughout the optimization process, ultimately leading to a better optimization result. Figure 7 The results of the spectral optimization are shown, with a color temperature of 4996K, a color shift (Duv) of -0.0035, a color rendering index of 98, and a LER of 258. The spectral curves show that the overall curves are smooth, exhibiting good spectral distribution quality and engineering feasibility.

[0100] Example 4: The optimization goal is to obtain a spectral distribution with a color temperature of 4000K, a color rendering index ≥97, and a LER ≥300, which has high luminous efficacy (LER).

[0101] This embodiment realizes an optimized design example of a spectral distribution with a color temperature of 4000K and high luminous efficacy (LER=301), which has both high spectral quality and high luminous efficacy.

[0102] This embodiment also uses the Gaussian function spectral information shown in Table 1. Figure 8 The curves showing the change of the comprehensive evaluation function during optimization are presented. It can be seen that the error value decreases significantly throughout the optimization process, ultimately leading to a better optimization result. Figure 9The results show the optimized spectrum, with a color temperature of 3993 K, a color shift (Duv) of -0.0062, a color rendering index (CRI) of 97.88, and a light efficiency ratio (LER) of 301. Since LER, color temperature, and CRI are typically mutually exclusive, the color temperature (CCT) and color shift (Duv) are compromised compared to Example 1. The spectral curves show an overall smoothness, indicating good spectral distribution quality and engineering feasibility.

[0103] In summary, this invention proposes an optimized design method for LED spectral distribution with high color rendering index (CRI), high-precision color temperature control, and high luminous efficacy. This method uses the Gaussian emission spectrum of common phosphors as basis functions, with the weights of each basis function as optimization variables, to construct a full-spectrum distribution and comprehensive evaluation function. Global optimization is then performed using a simulated annealing algorithm. The optimized comprehensive evaluation function comprehensively considers high-precision color coordinates and color temperature control, luminous efficacy weighted by the visual function, and CRI calculation based on ray tracing, thereby achieving optimal synergy between CRI, color temperature accuracy, and luminous efficacy over a wide color temperature range. This invention provides an efficient spectral design solution for high-end lighting, full-spectrum healthy light sources, and intelligent dimming systems, significantly improving the overall light quality and design efficiency of LED light sources.

[0104] This invention not only breaks the physical dilemma of the "light-color-electric-iron triangle," but also has milestone significance in engineering logic and design efficiency:

[0105] (1) Overcoming the gap of “engineering feasibility”, traditional mathematical optimization often yields ideal but unmanufacturable spectra. This patent innovatively uses the Gaussian properties of real phosphors as the basis function, ensuring that each weight coefficient obtained by optimization corresponds to the actual material ratio in the laboratory, thus achieving “what is designed is what is obtained”.

[0106] (2) Mathematical robustness and global optimum guarantee: The introduction of the simulated annealing algorithm (SA), combined with adaptive initial temperature estimation, solves the local minima problem that is common in multi-dimensional chromaticity spaces (CCT, Ra, LER). This shifts the spectral design from "relying on engineers' experience and trial and error" to "deterministic global optimum search".

[0107] (3) Achieving excellent color robustness The normalized second-order difference smoothing constraint proposed in the patent not only solves the problem of spurious peaks in mathematical fitting, but also gives the spectrum "fault tolerance" at the production level. This smooth spectrum has a stronger anti-interference ability for small errors in the amount or ratio of phosphor dispensing, and can significantly improve the color tolerance compliance rate of industrial mass production.

[0108] (4) This optimization design method is not limited to the basis functions provided in Table 1. This means that for the needs of specific phosphors or excitation spectra, users can define their own basis function library. For example, for a specific phosphor library, the basis functions corresponding to specific hyperspectral or blue-violet spectra can be added to the basis function library to achieve spectral optimization design for specific parameter requirements.

[0109] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0110] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0111] This invention also provides an LED light source comprising multiple phosphors, wherein the weight percentage of each phosphor is equal to the weight of each phosphor in the optimization result obtained by the aforementioned LED spectral distribution optimization design method. The LED light source obtained through the aforementioned LED spectral distribution optimization design method achieves optimal synergy between color rendering index, color temperature accuracy, and luminous efficacy over a wide color temperature range. This LED light source has broad market potential in various high-end and special lighting fields, including high-end commercial and museum lighting, medical and surgical lighting, full-spectrum human eye lighting and health light sources, and film and television shooting and ultra-high-definition display backlighting.

[0112] In summary, the LED spectral distribution optimization design method proposed in this invention has the following significant characteristics:

[0113] (1) Strong physical feasibility: The Gaussian emission characteristics of real phosphors are innovatively abstracted into a discrete basis function library, which ensures that the mathematical optimization results can be directly converted into engineering proportioning schemes.

[0114] (2) Coordinated optimization of indicators: By constructing a comprehensive evaluation function that includes Robertson iterative color temperature constraint, unilateral penalty function LER constraint and normalized smoothness constraint based on wavelength domain gradient, the blindness of single indicator optimization is completely broken, and the deep coupling of spectral morphology smoothness and light color quality is realized.

[0115] (3) Global optimal search capability: By utilizing the unique probability acceptance mechanism of the simulated annealing (SA) algorithm, it can effectively escape the local minimum trap caused by complex colorimetric constraints in the comprehensive evaluation function and find the optimal point of the comprehensive evaluation index in the global solution space.

[0116] The above provides a detailed description of the LED spectral distribution optimization design method and LED light source provided by this invention. Any obvious modifications made by those skilled in the art without departing from the essential content of this invention will constitute an infringement of the patent rights of this invention and will incur corresponding legal liability.

Claims

1. A method for optimizing the spectral distribution of LEDs, characterized in that: Spectral modulation of white LEDs excited by multicolor phosphors includes the following steps: S1, the Gaussian emission spectrum function of the fluorescent powder under excitation of light of a specific wavelength is taken as a base function, a base function library composed of Gaussian emission spectrum functions corresponding to multiple fluorescent powders is established; wherein, for a fluorescent powder with a known main wavelength λ p and a half-height width Δλ, the Gaussian emission spectrum function S p (λ) can be expressed as: Where λ represents the excitation wavelength of the phosphor, λ p λ is the center wavelength of the Gaussian emission spectral function corresponding to the phosphor, and Δλ is the full width at half maximum (FWHM) of the Gaussian emission spectral function corresponding to the phosphor. S2, using the multiple Gaussian emission spectral functions S corresponding to the multiple phosphor materials in step S1. p (λ) is used as a basis function, with the weight coefficient w of each basis function. i As variables, the full-spectral distribution function of the phosphor mixture is established through the superposition of multiple basis functions: Where λ represents the excitation wavelength of the mixture of multiple phosphors, n is the number of basis functions, and λ p,i Δλ is the center wavelength of the Gaussian function corresponding to the i-th phosphor material. i w is the full width at half maximum (FWHM) of the i-th Gaussian function. i The weighting coefficients for each basis function; S3. Construct a comprehensive evaluation function and constraints based on color temperature, spectral visual effect, and color rendering index: Construct the color temperature evaluation function f. CCT Spectral visual effect evaluation function f LER and the display index evaluation function f Ra And by summing the above sub-items, a comprehensive evaluation function for the entire system is constructed: ； S4. The simulated annealing algorithm is applied to optimize the overall evaluation function Score. The basis function weight vector w is globally optimized to output the optimal weight vector, thereby obtaining the optimal spectral distribution scheme. S5. If the optimized spectral distribution deviates from the target, adjust the weight coefficient vector w in the comprehensive evaluation function and repeat step S4 until the LED spectral distribution that meets the requirements is obtained.

2. The LED spectral distribution optimization design method according to claim 1, characterized in that, In step S3, the expression for the color temperature evaluation function is: Among them, w CCT This is a weighting constant for color temperature in optimization, typically ranging from 10. 4 <w CCT <10 5 T target Let CCT(S) be the target color temperature, CCT(S) be the color temperature corresponding to the current spectral distribution S calculated using the Robertson iterative method, and Duv(S) be the distance of the chromaticity coordinates corresponding to the current spectral distribution S from the blackbody locus. Duv This is the weighting constant for color cast Duv(S), typically in the range of 10. 6 <w Duv <10 7 .

3. The LED spectral distribution optimization design method according to claim 1, characterized in that, In step S3, the expression f of the spectral visual effect evaluation function LER for: Among them, w LER This is a weighting constant for spectral visual effects in the optimization, with a value range of 10. 3 <w LER <10 4 LER min The minimum value for lumens visual effect is given by δ, which is a constant factor with a value range of 10 < δ < 10. 2 ; LER(S) is the spectral visual effect function of the current spectral distribution S, and its calculation formula is as follows: Where V(λ) is the human eye's visual sensitivity function.

4. The LED spectral distribution optimization design method according to claim 1, characterized in that, In step S3, the expression f of the color rendering index evaluation function Ra for: Among them, w Ra This is the weighting constant of the color rendering index in the optimization, with a value range of 10. 4 <w Ra <10 5 Ra min Ra(S) represents the minimum required color rendering index, where Ra(S) is the color rendering index of the current spectral distribution S, and γ is a constant factor with a value range of 10 < γ < 10. 2 .

5. The LED spectral distribution optimization design method according to claim 1, characterized in that, Step S3 further includes the following sub-step: using the normalized second-order difference operator χ as a smoothing constraint, the calculation formula for the second-order difference operator χ is as follows: ； Among them, w i λ represents the basis function weights. i The center wavelength of the basis function; By minimizing the sum of squares of all difference operators, the energy distribution of adjacent basis functions is forced to conform to physical continuity, eliminating spurious peaks in mathematical fitting and ensuring the physical continuity of the synthesized spectrum.

6. The LED spectral distribution optimization design method according to claim 1, characterized in that, In step S4, the specific process of applying the simulated annealing algorithm for optimization includes: S4.1 Adaptive estimation of initial temperature: Calculate the initial value S0 of the comprehensive evaluation function based on the initial weight vector w0. Apply multiple small-range random perturbations to the initial weight vector w0 while maintaining non-negative weights. Calculate the comprehensive evaluation function value corresponding to the perturbation sample, and calculate the absolute difference between the comprehensive evaluation function value corresponding to the perturbation sample and the initial value of the comprehensive evaluation function. Take the median of the difference as the typical scale estimate ΔS. t And calculate the initial temperature T0 based on this: ; S4.2 Algorithm Parameter Configuration and Iterative Optimization: Configure key parameters of the simulated annealing algorithm, including: initial temperature T0, maximum number of iterations iter max The cooling coefficient α is optimized using the following algorithm iteratively: (1) Generate a new solution: under the current temperature, optimize the current weight vector w current Candidate solutions are generated by applying a normally distributed random perturbation while maintaining non-negative weights; (2) Evaluation and Metropolis criterion acceptance: Calculate the difference between the comprehensive evaluation function value corresponding to the candidate solution and the current solution respectively. If the difference is less than 0, the candidate solution is directly accepted. If the difference is greater than or equal to 0, it is accepted with a probability of P=exp(-ΔS / T), where ΔS is the function difference and T is the current temperature. (3) Cooling and convergence judgment: After completing the predetermined number of iterations at the current temperature, the temperature is reduced according to the cooling coefficient, and the iterative optimization steps (2) are repeated until the temperature drops to the termination threshold or the optimal solution has no improvement for multiple consecutive generations; (4) Output the optimal solution: When the algorithm terminates, it outputs the final optimal weight vector w. opt and the corresponding minimum value S of the comprehensive evaluation function min .

7. The LED spectral distribution optimization design method according to claim 1, characterized in that, In step S4, the simulated annealing algorithm is implemented using the simulannealbnd function.

8. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the LED spectral distribution optimization design method as described in any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the LED spectral distribution optimization design method as described in any one of claims 1 to 7.

10. An LED light source, characterized in that: It includes a phosphor mixture composed of multiple phosphors, wherein the weight percentage of each phosphor is equal to the weight of each phosphor in the optimization result obtained by the design method according to any one of claims 1 to 7.

Images