Real-time simulation method of radiant energy density distribution on a gpu-based power plant receiver

Abstract

The application discloses a kind of real-time simulation method of radiation energy density distribution on power station receiver based on GPU, the method is by calculating heliostat is shadowed, the gravity of the effective reflection area of heliostat is calculated, then the relevant parameters of heliostat reflection light spot on imaging plane are calculated, including the total energy of radiation energy, the variance matrix corresponding to elliptical Gaussian distribution obeyed by radiation energy density distribution;Then parallel computing each heliostat reflection on the radiation energy density distribution of each place on receiver and add the radiation energy density distribution generated by all heliostats on receiver, obtain the simulation result of mirror field reflection on the radiation energy density distribution of receiver.Using parallel computing method, the radiation energy density distribution of each heliostat reflection on the place of receiver is solved in parallel, the calculation speed of mirror field simulation is improved under the premise of basically not affecting the calculation precision.

Description

Technical Field

[0001] This invention relates to the field of energy, and in particular to a GPU-based method for real-time simulation of radiative energy density distribution on a power plant receiver. Background Technology

[0002] Tower solar thermal power generation systems are a new generation of solar thermal systems, characterized by low cost per kilowatt-hour and support for large-scale grid-connected power generation. These systems utilize heliostat fields to concentrate sunlight. Simulation of the radiation density distribution within the heliostat field is fundamental to research on annual power generation estimation, heliostat field layout optimization, and heliostat focusing scheme design for tower solar thermal power generation systems.

[0003] Existing fast simulation algorithms for mirror field radiation energy density distribution have problems such as being unsuitable for simulating non-planar receivers, inaccurate calculation of shadow occlusion, and slow calculation speed. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides a GPU-based real-time simulation method for radiated energy density distribution on a power plant receiver, comprising:

[0005] S1: Set a mirror field to be simulated, which contains a specified number of heliostats, and then calculate the shading of the heliostats and the centroid of the effective reflection area of ​​the heliostats;

[0006] S2: Calculate the relevant parameters of the heliostat reflected light spot on the imaging plane, including the total radiant energy and the variance matrix corresponding to the elliptical Gaussian distribution followed by the radiant energy density distribution;

[0007] S3: Based on GPU, combined with pre-lookup table, it calculates in parallel the distribution of radiant energy density reflected by each heliostat at various points on the receiver;

[0008] S4: By reducing and summing, the radiation energy density distribution generated by all heliostats on the receiver is accumulated to obtain the simulation results of the radiation energy density distribution of the mirror field reflection on the receiver.

[0009] Further, S1 specifically involves: calculating the effective reflection area of ​​the heliostat using the BMSS algorithm; uniformly discretizing the heliostat surface into several sub-heliostats; emitting two probe rays from the centroid of the sub-heliostat along the ideal incident and reflected ray directions, detecting the voxels they pass through in the mirror field using the 3D-DDA algorithm, and determining their intersection with the heliostats within the voxels; if neither of these two rays intersects with other heliostats, then the sub-heliostat is considered to be unaffected by shadows, and the unaffected sub-heliostats constitute the effective reflection area of ​​the heliostat; calculating the centroid of this area and the proportion of this area to the area of ​​the heliostat.

[0010] Furthermore, in S2, the imaging plane is a virtual plane passing through the heliostat's focal point and perpendicular to the direction of the ideal reflected light ray. Using a parameterized model of the radiant energy density distribution defined on the imaging plane by an elliptical Gaussian distribution, such as the HFLCAL model, iHFLCAL model, or bHFLCAL model, the total energy of the reflected light spot and the corresponding radiant energy density distribution F at various points on the imaging plane are calculated. image (x, y):

[0011]

[0012] Where (x, y) represents the two-dimensional coordinates at any point on the imaging plane, P h Let f(x, y) represent the total energy of the radiation spot reflected by the heliostat, σ1 and σ2 be the variances of the elliptic Gaussian distribution, and f(x, y) be the probability density function of the elliptic Gaussian distribution.

[0013]

[0014] Furthermore, the elliptic Gaussian distribution of S2 can be transformed into the product of two standard Gaussian distributions in the following way:

[0015]

[0016] Furthermore, the domain of the standard Gaussian distribution is (-∞, +∞). When the radiant energy density distribution is represented by an elliptic Gaussian distribution, the interval [μ1-4σ1, μ1+4σ1]×[μ2-4σ2, μ2+4σ2] contains 99.987% of the radiant energy. The pre-calculated lookup table only stores Gaussian function values ​​in the interval [-4, 4]. Since the probability density function of the Gaussian distribution is an even function, only the function values ​​in the interval [0, 4] are pre-stored.

[0017] Furthermore, in the pre-calculation process, the interval [0, 4] is divided into n equal-length intervals [x...]. i x i+1 ], i = 0, ..., n; the lookup table records the left endpoint x of each interval. i The standard Gaussian probability density function value f at the location N (x i ); Calculate the interval [x] using linear interpolation. i x i+1 [x] i Function value at +Δx:

[0018]

[0019] Further, S4 specifically involves: the receiver surface being discretized into a uniform pixel grid; the radiation energy density generated by each heliostat in the mirror field at the center of each pixel of the receiver being calculated in parallel by threads on the GPU; and then the radiation energy density generated by each heliostat in the mirror field at the same pixel being accumulated by reduction summation to obtain the final radiation energy density distribution result on the receiver.

[0020] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0021] This invention uses a pre-lookup table combined with linear interpolation to solve the radiation energy density distribution function, and employs a parallel computing method to solve the radiation energy density distribution reflected by each heliostat at various points on the receiver in parallel, thereby improving the calculation speed of the mirror field simulation without significantly affecting the calculation accuracy. Attached Figure Description

[0022] Figure 1 A flowchart illustrating the real-time simulation method for radiative energy density distribution on a receiver of a tower solar power station based on GPU parallel computing provided by this invention.

[0023] Figure 2 This is a schematic diagram of the effective reflection area and the centroid of the heliostat after it is partially obscured by shadow in this embodiment.

[0024] Figure 3 This is a schematic diagram illustrating the calculation of the standard normal probability density function using a combination of a pre-lookup table and a linear interpolation method in this embodiment. Detailed Implementation

[0025] Set up a mirror field to be simulated, which contains a specified number N heliostats. The receiver is a cylindrical receiver with a height of 10m and a radius of 5m.

[0026] Regarding this mirror field, such as Figure 1 As shown, a simulation of the radiant energy density distribution on the receiver is performed, including:

[0027] S1: The heliostat is divided into a 32×32 sub-heliostat array. The BMSS algorithm combined with the 3D-DDA algorithm is used to calculate the shading of the heliostats and to calculate the centroid of the effective reflection area of ​​the heliostats. A heliostat partially shaded is shown below. Figure 2 As shown, the black part in the figure is the sub-heliostat that is obscured by shadow, and the white part forms the effective reflection area; the black dots in the figure represent the centroid of the effective reflection area.

[0028] S2: Based on the HFLCAL model and combined with the shadow occlusion information calculated in S1, calculate the relevant parameters of the heliostat reflected light spot on the imaging plane, including the total radiant energy and the variance matrix corresponding to the elliptical Gaussian distribution followed by the radiant energy density distribution.

[0029] S3: Divide the receiver surface into a 0.1m × 0.1m pixel grid. Calculate the radiant energy density distribution at the center of each pixel on the receiver, as reflected by each heliostat, in parallel on the GPU. Calculate the radiant energy density distribution of the heliostat on the imaging plane using the HFLCAL model. Transform the radiant energy density distribution calculation defined by the HFLCAL model into the product of two standard normal probability density functions:

[0030]

[0031] like Figure 3 As shown, the standard normal probability density function is calculated using a pre-lookup table combined with linear interpolation. The interval [0, 4] is divided into 400 equal-length intervals [x...]. i x i+1 ], i = 0, ..., 399. The lookup table records the left endpoint x of each interval. i The standard Gaussian probability density function value f at the location N (x i ) Calculate the interval [x] using linear interpolation. i x i+1 [x] i Function value at +Δx:

[0032]

[0033] S4: By using the reduction summation method, the radiation energy density distribution generated by all heliostats at the same pixel is accumulated for all pixels on the receiver surface, and the simulation results of the radiation energy density distribution of the mirror field reflection on the receiver are obtained.

[0034] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

[0035] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A GPU-based real-time simulation method for radiated energy density distribution on a power plant receiver, characterized in that, include: S1: Set a mirror field to be simulated, which contains a specified number of heliostats, and then calculate the shading of the heliostats and the centroid of the effective reflection area of ​​the heliostats; S2: Calculate the relevant parameters of the heliostat reflected light spot on the imaging plane, including the total radiant energy and the variance matrix corresponding to the elliptical Gaussian distribution followed by the radiant energy density distribution; the elliptical Gaussian distribution is transformed into the product of two standard Gaussian distributions in the following way: ； The domain of the standard Gaussian distribution is When the radiant energy density distribution is represented by an elliptic Gaussian distribution, the interval It contains 99.987% of the radiant energy; and the pre-calculated lookup table only stores... The Gaussian function values ​​within the range; and since the Gaussian probability density function is an even function, only the values ​​are pre-stored. The function value within the interval; During the pre-calculation process, the interval Classified as equal-length intervals The lookup table records the left endpoints of each interval. Standard Gaussian probability density function value at [location] ; Calculate the interval using linear interpolation method middle Function value at: ； S3: Based on GPU, combined with pre-lookup table, it calculates in parallel the distribution of radiant energy density reflected by each heliostat at various points on the receiver; S4: By reducing and summing, the radiation energy density distribution generated by all heliostats on the receiver is accumulated to obtain the simulation result of the radiation energy density distribution of the mirror field reflection on the receiver; specifically: the receiver surface is discretized into a uniform pixel grid, and the radiation energy density generated by each heliostat in the mirror field at the center of each pixel of the receiver is calculated in parallel by threads on the GPU. Then, the radiation energy density generated by each heliostat in the mirror field at the same pixel is accumulated by reducing and summing to obtain the final radiation energy density distribution result on the receiver.

2. The real-time simulation method for radiant energy density distribution on a power plant receiver based on a GPU, as described in claim 1, is characterized in that... S1 specifically involves: calculating the effective reflection area of ​​the heliostat using the BMSS algorithm; uniformly discretizing the heliostat surface into several sub-heliostats; emitting two probe rays from the centroid of the sub-heliostat along the ideal incident and reflected ray directions, detecting the voxels they pass through in the mirror field using the 3D-DDA algorithm, and determining their intersection with the heliostats within the voxels; if neither of these two rays intersects with other heliostats, then the sub-heliostat is considered to be unaffected by shadows, and the unaffected sub-heliostats constitute the effective reflection area of ​​the heliostat, calculating the centroid of this area and the proportion of this area relative to the area of ​​the heliostat.

3. The real-time simulation method for radiant energy density distribution on a power plant receiver based on a GPU, as described in claim 1, is characterized in that... In S2, the imaging plane is a virtual plane passing through the focal point of the heliostat and perpendicular to the direction of the ideal reflected light ray. Using a parameterized model of the radiant energy density distribution defined on the imaging plane by an elliptical Gaussian distribution, such as the HFLCAL model, iHFLCAL model, or bHFLCAL model, the total energy of the reflected light spot and the corresponding radiant energy density distribution at various points on the imaging plane are calculated. : ； in Represents the two-dimensional coordinates of any point on the imaging plane. This represents the total energy of the radiant spot reflected by the heliostat. The variance is an elliptic Gaussian distribution. Let the probability density function be an elliptic Gaussian distribution. 。

Images