A method for pre-scheduling control of a three-tower one-machine light and heat mirror field across towers under cloud shadow disturbance

Abstract

The application discloses a kind of cloud shadow disturbance under three towers one machine light and heat mirror field cross-tower pre-scheduling control method, belong to solar thermal power generation technical field.The method includes: obtaining three towers one machine mirror field basic data;Predict cloud shadow coverage area and local DNI attenuation;Calculate the future available power matrix of sub-group heat absorbing tower and each tower predicted power gap;Based on the heat absorber energy flow, temperature, temperature rise rate and molten salt state Comprehensive acceptance margin and screening candidate acceptance tower;Determine cross-tower pre-acceptance mirror group and target tower;Generate pre-scheduling instruction, so that the acceptance mirror group completes pre-switch before cloud shadow arrives;Cloud rolling correction;After cloud, according to the slope of recovery normal scheduling after switching back.This application realizes the smooth transfer of power between towers through pre-scheduling before cloud, makes full use of the complementary ability of multiple towers, reduces thermal power fluctuation and light rejection, avoids heat shock of heat absorber, improves the operation safety and light and heat utilization efficiency of photovoltaic power station.

Description

Technical Field

[0001] This invention belongs to the field of solar thermal power generation technology, and specifically relates to a method for cross-tower pre-scheduling control of a three-tower, one-machine solar thermal mirror field under cloud shadow disturbance. Background Technology

[0002] Tower solar thermal power plants use heliostats to reflect solar radiation to the receiver at the top of the tower, achieving photothermal conversion. For single-tower systems, cloud cover can cause drastic fluctuations in receiver energy flow and outlet temperature, resulting in thermal shock. For three-tower-one-unit (or multi-tower-one-unit) systems, the problem is more complex due to the presence of multiple receiver towers and shared mirror areas; partial cloud cover can cause varying thermal power deficits in each tower.

[0003] Existing technologies have disclosed solutions such as cloud shadow recognition, direct normal irradiance (DNI) prediction, heliostat aiming adjustment, and multi-receiver architecture. For example, all-sky imagers identify cloud clusters and predict local DNI changes to guide the number of heliostats deployed; or linear programming determines the energy allocation ratio of subgroups in a multi-receiver system; or heliostats are redirected to different receivers based on system conditions (weather, temperature, etc.). However, these solutions typically have the following shortcomings: First, most solutions are geared towards single-tower or single-receiver control, failing to fully utilize the complementary support capabilities of different sub-mirror fields and shared mirror areas in a multi-tower, single-machine system. When the service mirror area of ​​a certain tower is blocked by cloud shadows, the power margin of the unblocked towers cannot be utilized in a timely manner.

[0004] Secondly, most solutions only respond passively after cloud shadows occur, lacking a mechanism to formulate a cross-tower pre-response plan before the cloud shadows arrive. Since switching heliostats takes a certain amount of time, passive adjustments after the arrival of cloud shadows are often delayed, leading to a sudden drop or rise in receiver power.

[0005] Third, existing methods only adjust the number of heliostats based on the average DNI or the proportion of local shading, without systematically calculating the future power gap of the shaded towers and the thermal capacity of the unshaded towers, which can easily lead to excessive load reduction (curtailment) or excessive load (thermal shock).

[0006] Fourth, although multi-receiver technology allows heliostats to point at multiple receivers, it lacks joint optimization of cloud shadow prediction, receiving capacity, target tower thermal safety, and action cost, making it difficult for ordinary scheduling algorithms to achieve smooth cross-tower power transfer.

[0007] Therefore, in a three-tower, one-unit solar thermal power plant, there is an urgent need for a control method that can predict cloud shadows in time and space and proactively plan the cross-tower power transfer before the arrival of cloud shadows, so as to reduce thermal power fluctuations, reduce curtailment, and improve system thermal safety. Summary of the Invention

[0008] To address the aforementioned technical problems, this invention provides a cross-tower pre-scheduling control method for a three-tower, one-machine solar thermal mirror field under cloud shadow disturbance, aiming to proactively plan power transfer between the blocked tower and the receiving tower before the arrival of cloud shadow, reduce thermal power fluctuations, reduce light curtailment, and improve the thermal safety of the receiver.

[0009] To achieve the above objectives, the technical solution of the present invention is as follows: A method for cross-tower pre-scheduling control of a three-tower, one-machine optical thermal mirror field under cloud shadow disturbance includes the following steps: S1. Obtain the basic data of the solar thermal mirror field of the three towers and one machine. The basic data includes the location of the three heat-absorbing towers, the parameters of the heat-absorbing panel, the location of the heliostat, the division result of the mirror field subgroup, the set of target towers that each heliostat can point to, and the operating constraints of each heat-absorbing tower. S2. Acquire cloud shadow data and meteorological data, and predict the coverage area of ​​cloud shadow on the mirror field plane in the future time domain based on the cloud shadow data and meteorological data. S3. Calculate the local DNI attenuation coefficient of each heliostat or each mirror field subgroup based on the cloud shadow coverage area, and calculate the available projection power of each mirror field subgroup when pointing to each heat absorber tower in the future time domain, and generate the subgroup-heat absorber tower future available power matrix. S4. Based on the future available power matrix of the subgroup-heat absorption tower and the target thermal power of each heat absorption tower, calculate the predicted power gap of each heat absorption tower in the future time domain. S5. Based on the energy flow, temperature, temperature rise rate, molten salt flow rate and molten salt inlet and outlet temperatures of each heat absorber, calculate the comprehensive load-bearing margin of each heat absorber and screen candidate heat absorber towers. S6. Based on the comprehensive acceptance margin of the candidate receiving tower, the available projection power of the candidate mirror field subgroup, and the uncertainty of cloud shadow prediction, determine the cross-tower pre-receiving mirror group and its target heat-absorbing tower. S7. Generate a cloud-front pre-acceptance boundary and a pre-switching command, so that the cross-tower pre-acceptance mirror group completes the pre-switching to the target heat absorption tower before the cloud shadow arrives. The cloud-front pre-acceptance boundary includes the pre-acceptance mirror group, the target heat absorption tower, the acceptance ratio, the switching start time, and the switching completion time. S8. During cloud shadow disturbance, rolling corrections are made to the predicted power gap, comprehensive load-bearing margin, and cloud front pre-load-bearing plan based on real-time DNI, receiver energy flow, infrared temperature, and molten salt operation data. S9. After the cloud shadow leaves and the local DNI meets the recovery criterion, the cloud shadow is gradually switched back according to the recovery waiting time and the state switching threshold, and the recovered cloud front switching boundary is used as the candidate boundary input for subsequent algorithms.

[0010] In the above scheme, the cloud shadow coverage area is calculated using the three-dimensional position of the cloud cluster and the direction of solar incidence, and the cloud shadow position satisfies: ; in, For the location of the cloud shadow, Location of the cloud cluster The height of the cloud cluster. Let be the unit vector of the direction of solar incidence. for The vertical component.

[0011] In the above scheme, the local DNI attenuation coefficient is calculated based on the proportion of the heliostat covered by cloud shadows and the light transmittance of the cloud cluster: ; in, For the first Heliostat predicts time The attenuation coefficient, For the first The proportion of the heliostat covered by cloud shadows. The light transmittance of the cloud.

[0012] In the above scheme, the elements in the future available power matrix of the subgroup heat absorption towers are: ; in, For the first Each camera subgroup at the predicted time Pointing to the The maximum available projected power of the heat absorption tower For the first The set of heliostats in each mirror field subgroup For the first Predicted local DNI at the position of the heliostat For the first The reflective area of ​​a heliostat. For the first heliostat facing the first The overall optical efficiency projected by the heat-absorbing tower. For the first heliostat facing the first Feasibility indicator for the projection of a heat absorption tower.

[0013] In the above scheme, the overall bearing capacity of the heat absorption tower is the minimum value among the power margin, panel energy flow margin, panel temperature margin, and molten salt operating margin: ; in, For the first The heat absorption tower at the predicted time The overall capacity to accommodate, To provide a margin of capacity based on the power limit, To accommodate the upper limit of panel power flow, This is based on the tolerance margin of the panel's upper temperature limit. This represents the tolerance margin based on the operating status of the molten salt.

[0014] In the above scheme, the cross-tower pre-acceptance mirror group is determined through acceptance priority scoring, which is calculated as follows: ; in, For the first The first camera subgroup towards the first Priority scoring for the scheduling of heat absorption towers. Normalized optical efficiency; To normalize the carry-over margin; To normalize, the stability of direct normal irradiance can be used; To normalize the cost of switching actions; To normalize the handling of tower heat risks; To reduce the uncertainty in normalized cloud shadow prediction; to These are the weighting coefficients.

[0015] In the above scheme, the objective function of the cross-tower pre-scheduling control command is as follows: ; The constraints include: ; ; ; ; ; ; in, For discrete prediction steps, The underlined variable is the normalized value; The power ratio allocated to the j-th receiver tower for the m-th mirror field subgroup in the n-th prediction step; To predict the power gap; To mitigate thermal risks, dimensionless; Cost of subgroup switching action; For cloud shadow prediction uncertainty; to These are the weighting coefficients; The maximum allowable power of the j-th heat absorption tower; To ensure sufficient overall capacity; Let this be the predicted time for the cloud shadow to reach the m-th subgroup. The time required for the subgroup to switch to the target tower. Allow for a safe lead time; For the first The baseline of the mirror field projection power that the heat absorption tower has received at the current moment; This refers to the time when the pre-scheduling instruction is issued; The first camera subgroup towards the first The subgroup-level feasibility flags projected by the heat-absorbing tower are determined by the subgroup that meets the requirements. The proportion of the heliostat is determined by the heliostat.

[0016] In the above scheme, during the cloud shadow disturbance process, the predicted power gap of the blocked tower is taken over by the candidate receiving tower in the following manner: ; ; in, For the obscured tower The predicted power gap is due to the receiving tower The amount of power it bears; For receiving tower The overall capacity to accommodate such a large volume of passengers; For the obscured tower The predicted power gap; For the candidate receiving tower set; To prevent the power from being too small, resulting in a denominator of zero; This represents the untapped remaining power gap.

[0017] In the above scheme, during the cloud back-cut process, the rate of change of the received power of each heat absorption tower does not exceed the preset power ramp limit: ; in, Let be the mirror field projection power obtained by the j-th heat-absorbing tower at the predicted time; For scheduling step size; To preset the power change rate limit, This is the preset power ramp limit.

[0018] In the above scheme, the uncertainty in cloud shadow prediction is represented by the local DNI prediction standard deviation and used to conservatively calculate the available projection power. ; in, DNI can be used conservatively for the position of the i-th heliostat; Let be the mean of the predicted DNI at the position of the i-th heliostat; Let be the standard deviation of the predicted DNI at the position of the i-th heliostat; is the confidence level coefficient.

[0019] Through the above technical solution, the cross-tower pre-scheduling control method for three towers and one machine photothermal mirror field under cloud shadow disturbance provided by the present invention has the following beneficial effects: 1. Transforming passive response into proactive pre-scheduling. This invention predicts the local DNI attenuation of each mirror field subgroup and the power gap of each heat absorber tower before the arrival of cloud shadows, and generates cross-tower takeover commands in advance, so that the takeover mirror group can complete the switching before the arrival of cloud shadows. This avoids the lag problem of passive adjustment after the heat absorber heat power drops sharply after the occurrence of cloud shadows, and effectively reduces the amplitude of heat power fluctuation.

[0020] 2. Fully utilize the complementary power-carrying capacity of the multi-tower system. When a local cloud shadow only obscures the service area of ​​a certain heat-absorbing tower, this invention only needs to mobilize the relevant subgroups and shared mirror groups to transfer power to the unobstructed towers, without the need for unified load reduction across the entire site, thereby reducing unnecessary light curtailment and improving the efficiency of solar thermal utilization.

[0021] 3. Simultaneously ensuring power delivery and thermal safety. This invention not only calculates the predicted power shortfall of the blocked tower, but also comprehensively considers the power upper limit, panel energy flow upper limit, temperature upper limit, temperature rise rate, and molten salt operating status of the receiving tower. The comprehensive delivery margin is used as the safety boundary of the delivery capacity to ensure that the receiving tower will not experience thermal safety problems such as overheating or overflow due to receiving too much power.

[0022] 4. Smooth control process, avoiding thermal shock. This invention uses power change rate constraints, subgroup switching ramp constraints, and cloud back-cut ramp strategies to make the heliostat switching and power transfer process smooth and controllable, avoiding local thermal shock to the receiver caused by a large number of instantaneous switching or sudden reactivation of heliostats, thus extending the receiver's lifespan.

[0023] 5. Strong engineering adaptability and easy deployment. The cloud shadow prediction data of this invention can come from all-sky imagers, cloud shadow cameras, satellite cloud images or multi-source fusion data. The optimization model can be implemented in various ways such as linear programming, model predictive control, and heuristic algorithms, and can be flexibly adapted to the existing control system architecture of different three-tower-one-machine solar thermal power plants.

[0024] 6. Complementary to existing scheduling methods. This invention is specifically designed for cloud shadow disturbance conditions and can work in conjunction with conventional scheduling methods for three towers and one machine mirror field (such as steady-state linear programming scheduling): under normal operating conditions, conventional scheduling is used; when cloud shadow disturbance is triggered, it switches to the pre-scheduling method of this invention; after the cloud shadow leaves, conventional scheduling is restored, thus taking into account both steady-state efficiency and dynamic safety.

[0025] 7. Expandable to multi-tower single-unit and multi-working-medium systems. The technical solution of this invention is not only applicable to three-tower single-unit systems, but can also be extended to multi-tower single-unit systems consisting of two or four or more heat-absorbing towers, and is applicable to heat absorbers with various heat transfer media such as molten salt, water / steam, air, and particles, and has broad application prospects. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0027] Figure 1 This is a schematic diagram of a cross-tower pre-scheduling control method for a three-tower, one-machine photothermal mirror field under cloud shadow disturbance, as disclosed in an embodiment of the present invention. Figure 2 A schematic diagram of the cloud shadow disturbance and cross-tower support scenario for the three towers and one machine mirror field.

[0028] In the diagram, 1 is heat absorption tower one; 2 is heat absorption tower two; 3 is heat absorption tower three; 4 is the service mirror area of ​​heat absorption tower one; 5 is the service mirror area of ​​heat absorption tower two; 6 is the service mirror area of ​​heat absorption tower three; 7 is the shared mirror area of ​​heat absorption tower one and two; 8 is the shared mirror area of ​​heat absorption tower two and three; and 9 is the shared mirror area of ​​heat absorption tower one and three. Detailed Implementation

[0029] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0030] This invention provides a method for cross-tower pre-scheduling control of a three-tower, one-machine optical thermal mirror field under cloud shadow disturbance, and specific embodiments are as follows: I. System Hardware Configuration and Basic Data This embodiment uses a three-tower, one-machine tower-type solar thermal power plant as an example for illustration. The power plant includes three heat-absorbing towers. , , The three receiver towers share a single generator set, thermal storage system, and steam generation system. The total number of heliostats in the field is 12,000, divided into 60 sub-fields based on spatial location and optical characteristics. This includes dedicated service subgroups for each heliostat and shared mirror area subgroups between adjacent sub-mirror fields. The coordinates, reflective area, optical efficiency curve, and the set of heliostats that can be pointed to by each heliostat (whether within mechanical and optically feasible limits) have all been pre-entered into the database of the mirror field master control system.

[0031] Absorber parameters: Each absorber tower is equipped with 24 panels (circumferentially distributed), each panel having a maximum permissible energy flux density, maximum wall temperature, and maximum temperature rise rate. The molten salt loop is equipped with flow meters, inlet and outlet temperature sensors, and pressure sensors to monitor the molten salt operating status.

[0032] Predictive and control parameters: scheduling step size Seconds, pre-scheduled time domain length Seconds (i.e., the next 30 steps). Cloud shadow warning time. Seconds, safety lead time Seconds. Power deficit threshold. MWt. Conditions for cloud recovery and stabilization: and Duration Second.

[0033] To avoid ambiguity regarding variables, the main variables and their units in this invention are explained as follows: This represents the j-th heat absorption tower. Let represent the m-th mirror field subgroup, i represent the heliostat number, j represent the receiver tower number, k represent the receiver panel number, and M represent the total number of mirror field subgroups. This indicates the number of panels in the j-th heat absorption tower.

[0034] Indicates the scheduling step size, in seconds; Indicates the discrete prediction step. ; This represents the local DNI of the i-th heliostat, in W / m. 2 ; This represents the reflective area of ​​a heliostat, expressed in m². 2 ; , , , and All are calculated in MWt. , , and It is a dimensionless quantity.

[0035] To avoid mixing symbols, This represents a small positive number in the power distribution formula, with the unit being MWt; This represents a small positive number in the DNI uncertainty formula, with units of W / m. 2 .

[0036] II. Cross-tower pre-scheduling control methods, such as Figure 1 As shown.

[0037] S1: Establishing a basic data model for the three towers and one machine lens field The coordinates of the three heat-absorbing towers, the absorber panel number, the coordinates of the heliostats and their subgroup affiliations, the set of target towers that each heliostat can point to, the range of the shared mirror area, the maximum power of the absorber, the upper limit of panel energy flow, the upper limit of temperature, and the molten salt operation constraints are stored in the control system database.

[0038] S2: Collect cloud shadow and meteorological data, and predict the spatiotemporal trajectory of cloud shadows. Three all-sky imagers were deployed at the power plant site (located on the extended line connecting the three receiver towers, 300m from the nearest tower) and a DNI sensor array (arranged around the perimeter and center of the imager field). The all-sky imagers acquired all-sky images every 10 seconds, and image processing identified cloud boundaries, cloud movement speed, and cloud transmittance. The DNI sensor array measured the actual DNI in real time and estimated it using a clear-sky model. It is used to calibrate cloud transmittance models. Simultaneously, the system can access meteorological satellite cloud images as a supplementary data source.

[0039] like Figure 2 As shown, the three-tower, one-machine mirror field includes receiver tower 1, receiver tower 2, receiver tower 3, and their respective service mirror areas 4, 5, and 6, and shared mirror areas 7, 8, and 9. The dashed lines represent the cloud shadow coverage area, and the arrows indicate the direction of cloud shadow movement. In the diagram, the service mirror area of ​​receiver tower 1 is currently obscured by the cloud shadow, while receiver towers 2 and 3, and part of their shared mirror area, are not obscured. Some of the unobscured mirror groups can be directed across towers towards receiver towers 2 or 3 to absorb the system heat power shortfall caused by the cloud shadow obstruction of receiver tower 1. The remaining unobscured mirror groups within each mirror area still point towards their respective towers. Let the position of the centroid of a certain cloud cluster in three-dimensional space be denoted as . ,in, , , These are the three-dimensional position components of the cloud center in the site coordinate system.

[0040] The unit vector of the direction of solar incidence is ,in, This represents the vertical component corresponding to the solar altitude angle.

[0041] The projection position of the cloud cluster on the ground mirror field plane is: ; The above formula is based on This represents the unit vector of the direction of solar incidence. If the unit vector pointing from the ground to the sun is used in the engineering implementation, the direction sign in the projection formula will be adjusted accordingly based on coordinate conventions.

[0042] Calculate the cloud movement speed based on multiple consecutive frames of images. and acceleration Then the future moment The area covered by the cloud shadow is: ; in, This indicates the area covered by the cloud shadow.

[0043] For the A heliostat, whose projection area on the ground is The proportion covered by cloud shadows is: ; in, Indicates the first The projection area of ​​the heliostat on the ground. .

[0044] S3: Calculate the local DNI attenuation coefficient and generate the future available power matrix of the subgroup-endothermic tower. Assume the cloud's light transmittance is (Values ​​range from 0 to 1, where 0 represents completely opaque and 1 represents completely transparent), then the first... The local DNI attenuation coefficient of the heliostat is: ; Therefore, we obtain the first... Predicted local DNI at the position of the heliostat: in, The clear sky DNI is calculated using the clear sky model, or it is obtained by correcting the actual DNI measured at the weather station with the clear sky model.

[0045] For each mirror field subgroup Define the intensity of cloud shadow disturbance: ; in, The predicted local DNI at the position of the i-th heliostat is... This is the reference value for direct normal irradiance under clear sky conditions. Let be the reflective area of ​​the i-th heliostat.

[0046] when (Pick Furthermore, the arrival time of the cloud shadow is less than the warning time. At a certain time, the pre-scheduled state of the subgroup is triggered.

[0047] No. The individual camera subgroups in the future Pointing to the The maximum available projected power when using a heat absorption tower is: ; In the formula, Indicates the first The heliostat at time Does it have a directional heat absorption tower? The feasibility, if feasible ,otherwise Feasibility conditions include whether the target tower is within the heliostat's projection range, whether the tracking angle meets mechanical limits, whether there is severe obstruction, and whether it is under maintenance or experiencing communication abnormalities. For subgroup-level control, further definitions can be made. Subgroup For heat absorption tower Equivalent feasibility; when the proportion of heliostats meeting the feasibility requirements within a subgroup is not lower than a preset proportion threshold, take... Otherwise take .

[0048] ; in, This is the feasible proportion threshold for the subgroup. Subsequent constraints will use... This is to avoid distributing power to non-directional heat-absorbing towers.

[0049] To achieve comprehensive optical efficiency (including cosine efficiency, atmospheric attenuation, occlusion efficiency, cutoff efficiency, specular reflectivity, and tracking error correction coefficient), in this embodiment, for the shared mirror subgroup, Multiple target towers are allowed to be set to 1 simultaneously.

[0050] The calculated future available power matrix is ​​60×3, with each element being a time series (30 steps).

[0051] S4: Calculate the predicted power deficit for each heat absorber tower. Let the first The heat absorption tower at the predicted time The target thermal power is (Given by grid dispatch instructions or thermal storage strategies), the actual projected power of the mirror field is: ; in, To optimize variables, represent subgroups Distributed to the heat absorption tower The power ratio. The predicted power gap is: ; The total thermal power demand of a three-tower, one-generator system can be given by the generator load, thermal storage charging power, or heating load command: .

[0052] At a certain moment in this embodiment, cloud projection prediction shows that the northern service mirror area and part of the shared mirror area will be covered within the next 180 seconds. The calculated available projection power will decrease from the current 45 MWt to ​​37 MWt (a gap of 8 MWt). and The target power is 40MWt and 38MWt respectively, while the current actual power is 39MWt and 38MWt respectively, with a power gap of 0 for both.

[0053] S5: Calculate the overall load-bearing capacity of each heat absorption tower and screen candidate heat absorption towers. Calculate the four tolerance margins separately: Power margin: ; in, The maximum allowable absorption power of the heat absorption tower (in this embodiment) Both are 55MWt).

[0054] Energy flow margin: ; in, and The unit is kW / m 2 , The area of ​​the k-th panel is expressed in m². 2 , This is a correction factor for converting energy flow margin to available heat power; in the formula... Used to convert kW to MWt, It is unified with other bearing margins as MWt.

[0055] Temperature margin: ; in, The conversion factor representing the temperature margin to the power capacity can be estimated from the absorber's heat capacity, heat transfer coefficient, and molten salt flow rate. In this embodiment, , .

[0056] Molten salt operating margin: ; in, For the first Molten salt flow rate of the heat absorption tower Inlet temperature, For the outlet temperature, To prevent freezing and maintain a safe temperature, To allow for temperature differences.

[0057] The molten salt operating margin is calculated based on a combination of molten salt flow rate, antifreeze temperature, and allowable temperature difference. The output unit is MWt, and it is consistent with the power margin, energy flow margin, and temperature margin.

[0058] The overall capacity margin is: ; when At that time, the heat absorption tower It can be used as a receiving tower; when If the thermal risk exceeds the threshold, the tower shall not be used as a receiving tower.

[0059] Further calculation of risk indicators: ; In this embodiment, take Risk limits .satisfy and The heat absorption tower was listed as a candidate receiving tower.

[0060] In this embodiment, the calculation is obtained Overall capacity to accommodate ,risk ; Overall capacity to accommodate ,risk Both are candidate receiving towers.

[0061] S6: Determine the cross-tower pre-acceptance mirror group and its target heat absorption tower. For a heat-absorbing tower affected by cloud shadows Let its future power gap be... The candidate receiving tower set is defined as follows: ; The candidate receiving mirror group set is defined as: ; in, Indicates the mirror field subgroup Can it point to the receiving tower? .

[0062] That is, the subgroup with low cloud shadow disturbance intensity that can be directed towards the candidate receiving tower. In this embodiment, the subgroup that is about to be blocked... Some subgroups within the sub-mirror field (with high perturbation intensity) do not participate in the reception, while sharing the mirror area and , Some low-perturbation subgroups were included in the candidate set.

[0063] To prioritize mirror clusters with high efficiency, low operational costs, low thermal risk, and low cloud shadow uncertainty, a priority score is calculated for each candidate subgroup-receiving tower pair: ; The cost of a subgroup switching action can be expressed as: ; in, and The first Switching from heliostat to heat absorber tower The required azimuth and elevation angles and For the current tracking angle, This represents the weight of the action cost.

[0064] S7: Generate cloud pre-acceptance boundary and complete pre-switching before cloud shadow arrives. This invention establishes a rolling time-domain pre-acceptance decision model to determine the subgroup-heat absorber tower power allocation variables over several future steps. The optimization time domain is defined as the next 30 steps (300 seconds), and the optimization variables are... (Power allocation ratio of each subgroup to the three towers). The objective function is in normalized form: ; The constraints include: 1. Subgroup allocation ratio constraints: ; ; 2. Maximum heat absorption power constraint of the heat absorption tower: ; 3. Panel energy flow constraints: ; 4. Panel temperature rise rate constraint: ; 5. Molten salt temperature safety constraints: ; 6. Subgroup switching ramp constraints: ; 7. Switch constraints before the cloud shadow arrives: Assume the cloud shadow reaches the first The prediction time for each subgroup is The time required for the subgroup to switch to the receiving tower is Safety lead time is Then it needs to satisfy: ; 8. Capacity constraints: ; in, For discrete prediction steps, The underlined variable is the normalized value; The power ratio allocated to the j-th receiver tower for the m-th mirror field subgroup in the n-th prediction step; The power deficit is predicted in MWt. To mitigate thermal risks, dimensionless; Cost of subgroup switching action; For cloud shadow prediction uncertainty; to These are the weighting coefficients; The maximum allowable power of the j-th heat absorption tower is expressed in MWt. The overall capacity margin is expressed in MWt. Let this be the predicted time for the cloud shadow to reach the m-th subgroup. The time required for the subgroup to switch to the target tower. Allow for a safe lead time; For the first The baseline of the mirror field projection power received by the heat-absorbing tower at the current moment, in MWt; The time of issuance of the pre-scheduling instruction is expressed in seconds. , and The units are all in seconds (s); For the first The first camera subgroup towards the first The subgroup-level feasibility flags projected by the heat-absorbing tower are determined by the subgroup that meets the requirements. The proportion of the heliostat is determined by the heliostat.

[0065] In this embodiment, .

[0066] This embodiment uses a quadratic programming solver (such as OSQP) for online solving, with a single optimization calculation time of approximately 0.3 seconds, meeting the real-time requirements.

[0067] The optimized result The process is converted into a pre-acceptance boundary in front of the cloud, including the pre-acceptance mirror group, target heat absorber tower, acceptance ratio, switching start time, switching completion time, and cloud-front switching boundary. Subsequent mirror field scheduling algorithms use the output boundary of this invention to further generate specific heliostat execution instructions; specific target point allocation, fine-grained scheduling solution, and mirror-by-mirror execution control are not considered the core of this invention.

[0068] This invention sets up a three-stage control logic.

[0069] The first stage is the pre-switching stage before the cloud (corresponding to S7), the second stage is the rolling transition stage in the cloud (corresponding to S8), and the third stage is the back-switching stage after the cloud (corresponding to S9).

[0070] During cloud shadow disturbances, the predicted power gap of the blocked tower is taken over by the candidate receiving tower in the following manner: ; ; in, For the obscured tower The predicted power gap is due to the receiving tower The amount of power it bears; For receiving tower The overall capacity to accommodate such a large volume of passengers; For the obscured tower The predicted power gap; For the candidate receiving tower set; To prevent the power from being too small, resulting in a denominator of zero; This represents the untapped remaining power gap. The corresponding applicability ratio is... When the total capacity of candidate receiving towers is insufficient, the unreceived power gap is used as the candidate boundary input for subsequent algorithms, which are then processed by the subsequent algorithms in conjunction with load reduction, thermal storage, or power generation coordination strategies.

[0071] S8: Rolling revision of the cloud-based pre-acceptance plan Because the speed, thickness, and transmittance of cloud shadows are uncertain, this invention introduces robust corrections. Let the... The predicted mean of the future local DNI of the heliostat is The prediction standard deviation is Conservative can be defined using DNI as: ; in, This represents the confidence level coefficient. When predicting the power gap, the following is used: Calculate the available power of the shaded tower to avoid underestimating the power gap; when performing the safety check of the receiving tower, a higher recovery DNI condition can be used to check the upper limit of energy flow to avoid overheating caused by the sudden departure of the cloud shadow.

[0072] The forecast uncertainty index can be expressed as: ; When the threshold is exceeded, conservative scheduling is adopted for the subgroup, including reducing the switching ratio, increasing the safety lead time, or reserving backup candidate boundary inputs.

[0073] This method operates in a rolling closed-loop manner. In each control cycle, the system re-acquires all-sky imagery, DNI array data, infrared temperature, receiver energy flow, molten salt flow rate, and outlet temperature to update cloud shadow predictions, coverage margin, and power gap. During actual cloud shadow coverage, the system re-acquires DNI array data, receiver energy flow, infrared temperature, and molten salt operating data every 10 seconds. If the actual local DNI deviates from the predicted value... Exceeding the set threshold (e.g., 50W / m) 2 If so, the DNI prediction for the next cycle will be revised: ; Pick In this embodiment, to predict the correction gain, If the actual power of the heat absorption tower deviates from the target power... If the demand exceeds 0.5MWt, the required capacity will be adjusted accordingly. ; Pick The revised prediction is then re-introduced into the optimized model to update the subsequent time-domain allocation scheme. In this embodiment, when the cloud shadow actually arrives, due to slight changes in cloud velocity, the actual occlusion is approximately 8 seconds later than predicted, and the DNI decay is slightly greater than predicted. The system promptly switches some of the originally planned occlusions to [the predicted time domain] through rolling corrections. The mirror group was adjusted to ,make The power gap decreased from the predicted 8 MWt to ​​the actual 2.5 MWt, with the total system power decreasing by only 2.1 MWt (compared to approximately 9.5 MWt without pre-scheduling). This closed-loop correction avoids erroneous cross-tower scheduling caused by single cloud shadow prediction errors.

[0074] S9: Cloud back cutoff, restore normal mirror field scheduling When the cloud shadow moves away, the system detects that the local DNI attenuation coefficient of the occluded subgroup meets the requirements. ,and After a 30-second recovery waiting time, the cloud back-off process is initiated. The back-off process is not instantaneous; rather, it gradually restores the original allocation state based on the minimum hold time, recovery waiting time, and state transition threshold. Specifically, the target allocation ratio for back-off is calculated in each control cycle to satisfy the following conditions: ; in, Let be the mirror field projection power obtained by the j-th heat-absorbing tower at the predicted time; For scheduling step size; To preset the power change rate limit, This is the preset power ramp limit.

[0075] In this embodiment, the back-cut lasts for about 120 seconds. During this period, the power of each heat absorption tower changes slowly, and there are no obvious temperature fluctuations or energy flow shocks.

[0076] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for cross-tower pre-scheduling control of a three-tower, one-machine photothermal mirror field under cloud shadow disturbance, characterized in that, Includes the following steps: S1. Obtain the basic data of the solar thermal mirror field of the three towers and one machine. The basic data includes the location of the three heat-absorbing towers, the parameters of the heat-absorbing panel, the location of the heliostat, the division result of the mirror field subgroup, the set of target towers that each heliostat can point to, and the operating constraints of each heat-absorbing tower. S2. Acquire cloud shadow data and meteorological data, and predict the coverage area of ​​cloud shadow on the mirror field plane in the future time domain based on the cloud shadow data and meteorological data. S3. Calculate the local DNI attenuation coefficient of each heliostat or each mirror field subgroup based on the cloud shadow coverage area, and calculate the available projection power of each mirror field subgroup when pointing to each heat absorber tower in the future time domain, and generate the subgroup-heat absorber tower future available power matrix. S4. Based on the future available power matrix of the subgroup-heat absorption tower and the target thermal power of each heat absorption tower, calculate the predicted power gap of each heat absorption tower in the future time domain. S5. Based on the energy flow, temperature, temperature rise rate, molten salt flow rate and molten salt inlet and outlet temperatures of each heat absorber, calculate the comprehensive load-bearing margin of each heat absorber and screen candidate heat absorber towers. S6. Based on the comprehensive acceptance margin of the candidate receiving tower, the available projection power of the candidate mirror field subgroup, and the uncertainty of cloud shadow prediction, determine the cross-tower pre-receiving mirror group and its target heat-absorbing tower. S7. Generate a cloud-front pre-acceptance boundary and a pre-switching command, so that the cross-tower pre-acceptance mirror group completes the pre-switching to the target heat absorption tower before the cloud shadow arrives. The cloud-front pre-acceptance boundary includes the pre-acceptance mirror group, the target heat absorption tower, the acceptance ratio, the switching start time, and the switching completion time. S8. During cloud shadow disturbance, rolling corrections are made to the predicted power gap, comprehensive load-bearing margin, and cloud front pre-load-bearing plan based on real-time DNI, receiver energy flow, infrared temperature, and molten salt operation data. S9. After the cloud shadow leaves and the local DNI meets the recovery criterion, the cloud shadow is gradually switched back according to the recovery waiting time and the state switching threshold, and the recovered cloud front switching boundary is used as the candidate boundary input for subsequent algorithms.

2. The method according to claim 1, characterized in that, The area covered by the cloud shadow is calculated using the three-dimensional position of the cloud cluster and the direction of solar incidence. The position of the cloud shadow satisfies the following: ； in, For the location of the cloud shadow, Location of the cloud cluster The height of the cloud cluster. Let be the unit vector of the direction of solar incidence. for The vertical component.

3. The method according to claim 1, characterized in that, The local DNI attenuation coefficient is calculated based on the proportion of the heliostat covered by cloud shadows and the light transmittance of the cloud cluster: ； in, For the first Heliostat predicts time The attenuation coefficient, For the first The proportion of the heliostat covered by cloud shadows. The light transmittance of the cloud.

4. The method according to claim 1, characterized in that, The elements in the future available power matrix of the subgroup of heat absorption towers are: ； in, For the first Each camera subgroup at the predicted time Pointing to the The maximum available projected power of the heat absorption tower For the first The set of heliostats in each mirror field subgroup For the first Predicted local DNI at the position of the heliostat For the first The reflective area of ​​a heliostat. For the first heliostat facing the first The overall optical efficiency projected by the heat-absorbing tower. For the first heliostat facing the first Feasibility indicator for the projection of a heat absorption tower.

5. The method according to claim 1, characterized in that, The overall operating margin of the heat absorption tower is the minimum value among the power margin, panel energy flow margin, panel temperature margin, and molten salt operating margin. ； in, For the first The heat absorption tower at the predicted time The overall capacity to accommodate, To provide a margin of capacity based on the power limit, To accommodate the upper limit of panel power flow, This is based on the tolerance margin of the panel's upper temperature limit. This represents the tolerance margin based on the operating status of the molten salt.

6. The method according to claim 1, characterized in that, The cross-tower pre-acceptance mirror group is determined through acceptance priority scoring, which is calculated as follows: ； in, For the first The first camera subgroup towards the first Priority scoring for the scheduling of heat absorption towers. Normalized optical efficiency; To normalize the carry-over margin; To normalize, the stability of direct normal irradiance can be used; To normalize the cost of switching actions; To normalize the handling of tower heat risks; To reduce the uncertainty in normalized cloud shadow prediction; to These are the weighting coefficients.

7. The method according to claim 1, characterized in that, The objective function of the cross-tower pre-scheduling control command is as follows: ； The constraints include: ； ； ； ； ； ； in, For discrete prediction steps, The underlined variable is the normalized value; The power ratio allocated to the j-th receiver tower for the m-th mirror field subgroup in the n-th prediction step; To predict the power gap; To mitigate thermal risks, dimensionless; Cost of subgroup switching action; For cloud shadow prediction uncertainty; to These are the weighting coefficients; The maximum allowable power of the j-th heat absorption tower; To ensure sufficient overall capacity; Let this be the predicted time for the cloud shadow to reach the m-th subgroup. The time required for the subgroup to switch to the target tower. Allow for a safe lead time; For the first The baseline of the mirror field projection power that the heat absorption tower has received at the current moment; This refers to the time when the pre-scheduling instruction is issued; For the first The first camera subgroup towards the first The subgroup-level feasibility flags projected by the heat-absorbing tower are determined by the subgroup that meets the requirements. The proportion of the heliostat is determined by the heliostat.

8. The method according to claim 1, characterized in that, During the cloud shadow disturbance, the predicted power gap of the blocked tower is taken over by the candidate receiving tower in the following manner: ； ； in, For the obscured tower The predicted power gap is due to the receiving tower The amount of power it bears; For receiving tower The overall capacity to accommodate such a large volume of passengers; For the obscured tower The predicted power gap; For the candidate receiving tower set; To prevent the power from being too small, resulting in a denominator of zero; This represents the untapped remaining power gap.

9. The method according to claim 1, characterized in that, During the cloud back-cut process, the rate of change of power received by each heat absorption tower does not exceed the preset power ramp limit: ； in, Let be the mirror field projection power obtained by the j-th heat-absorbing tower at the predicted time; For scheduling step size; To preset the power change rate limit, This is the preset power ramp limit.

10. The method according to claim 1, characterized in that, The uncertainty in cloud shadow prediction is represented by the local DNI prediction standard deviation and used to conservatively calculate the available projection power: ； in, DNI can be used conservatively for the position of the i-th heliostat; Let be the mean of the predicted DNI at the position of the i-th heliostat; Let be the standard deviation of the predicted DNI at the position of the i-th heliostat; is the confidence level coefficient.

Images