Multi-blade angle self-adaptive adjusting system based on illumination intensity feedback

Abstract

The invention relates to the technical field of automatic control, in particular to a multi-blade angle self-adaptive adjusting system based on illumination intensity feedback, which comprises a data acquisition module for capturing a physical field illumination intensity sequence in real time and determining space-time reference data of a boundary; the reference model module is used for constructing a theoretical reference model by using an astronomical algorithm and generating a standard illumination intensity reference sequence; the interference decoupling module is used for generating a global interference model and a local interference model based on knowledge driving, calculating residual features through synthetic analysis, and comparing topological similarity of spatial distribution to solve an interference coupling judgment result; the self-adaptive execution module responds to the judgment result, distinguishes global environment change and local part abnormity, and generates an angle adjustment instruction or a self-inspection maintenance instruction; the theoretical reference model is introduced to serve as a third-dimensional reference system, local noise camouflage is effectively broken, the oscillation effect of array malfunction is avoided, and the system robustness is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of automatic control technology, specifically to a multi-blade angle adaptive adjustment system based on light intensity feedback. Background Technology

[0002] In multi-blade angle adjustment applications based on illumination feedback, the system needs to rely on illumination intensity data collected by the sensor array to control the blade attitude in real time. Existing solutions generally adopt a traditional closed-loop control architecture that relies solely on the deviation between setpoints and measured values, lacking a reference for comparing the theoretical true values ​​of the physical environment. Although such solutions can respond to changes in light intensity, they cannot effectively distinguish between global environmental evolution caused by cloud cover and local noise interference caused by sensor dust accumulation, foreign object obstruction, or device failure when facing complex lighting environments. This lack of judgment on the attributes of interference sources makes the control system highly susceptible to triggering erroneous overall adjustment commands due to abnormal data from individual sensors, leading to array oscillation effects. Furthermore, it cannot adaptively calibrate for systematic errors caused by equipment aging or environmental pollution. Therefore, how to introduce a physical reference model to decouple global and local disturbances, thereby improving the system's control robustness, regulation accuracy, and self-healing ability in complex environments, has become an urgent technical problem to be solved. Summary of the Invention

[0003] To solve the above-mentioned technical problems, the present invention provides a multi-blade angle adaptive adjustment system based on light intensity feedback. Specifically, the technical solution of the present invention includes: The data acquisition module is used to collect real-time light intensity sequence data reflecting the physical field state of the target, and to obtain the current spatiotemporal reference data of the system. The reference model module is communicatively connected to the data acquisition module and is used to construct a theoretical reference model for the current moment based on the spatiotemporal reference data and preset system calibration parameters, and generate a standard light intensity reference sequence. The interference decoupling module is communicatively connected to the reference model module and the data acquisition module. It is used to generate a global environmental interference model and a local sensor interference model based on knowledge-driven generation, and to calculate the interference coupling decision result of the root cause of the current system state deviation by combining the light intensity sequence data and the standard light intensity reference sequence. An adaptive execution module, which is communicatively connected to the interference decoupling module, is used to receive the interference coupling decision result and generate control commands for adaptively adjusting the multi-blade angle based on the decision result, or to generate self-test and maintenance commands for the data acquisition module. The interference decoupling module analyzes the residual characteristics of real data and model prediction data, and compares the topological similarity of the residual characteristics with the preset interference mode in spatial distribution to distinguish between systematic shifts caused by global environmental changes and local noise caused by local component anomalies.

[0004] Preferably, the data acquisition module includes: The sensing array unit consists of multiple light sensors distributed in space according to a predetermined topology, used to synchronously or asynchronously acquire the light intensity sequence data; A spatiotemporal reference unit is used to provide the system with spatiotemporal reference data containing location information and precise time information; The actuator feedback unit is used to collect the physical angle state of the angle actuator of each blade in real time and feed it back to the adaptive execution module to form a closed-loop control circuit.

[0005] Preferably, the reference model module includes: The parameter calculation unit is used to calculate and generate theoretical scenario parameters based on the spatiotemporal reference data and the system's historical operation data. The reference generation unit, connected to the parameter calculation unit, is used to call the preset system calibration parameters according to the theoretical scene parameters, and to generate the standard light intensity reference sequence expected to be sensed at each sensor position in the sensing array unit under an ideal state without interference.

[0006] Preferably, the interference decoupling module includes: The simulation engine unit is used to run the global environmental interference model and the local sensor interference model in parallel. The global environmental interference model is used to simulate interference patterns that have a global, gradual, or consistent impact on the sensor array unit, so as to generate a first theoretical interference data sequence. The local sensor interference model is used to simulate interference patterns that produce sudden, step-like effects on only one or a few sensors in the sensor array unit, in order to generate a second theoretical interference data sequence.

[0007] Preferably, the interference decoupling module further includes: The residual analysis unit is used to calculate the difference between the light intensity sequence data and the standard light intensity reference sequence to generate a real residual vector, and to calculate the difference between the first theoretical interference data sequence, the second theoretical interference data sequence and the standard light intensity reference sequence to generate corresponding theoretical residual vectors. The pattern matching unit, connected to the residual analysis unit, is used to calculate the similarity measure between the actual residual vector and each of the theoretical residual vectors in spatial distribution, and output the interference coupling decision result based on a predetermined matching threshold and decision logic.

[0008] Preferably, the adaptive execution module includes: The global control unit is used to determine that a systematic adjustment is needed when the interference coupling decision result indicates that the current state offset matches the global environmental interference model, and to generate the control command to uniformly adjust all blades to the new target angle. The fault-tolerant control unit is used to determine that the interference coupling decision result matches the local sensor interference model as local noise interference, and generate instructions to freeze or reduce the weight of the interfered sensor data while maintaining or smoothly adjusting the current blade angle.

[0009] Preferably, the fault-tolerant control unit is also used for: When local noise interference is detected, a marking and maintenance alarm is triggered for suspected abnormal sensors in the data acquisition module. Meanwhile, based on the weight information provided by the interference coupling decision result, a reliable state feedback signal is reconstructed using the effective data from the other sensors through a data fusion algorithm, so that the global control unit can continue to generate continuous and stable control commands.

[0010] Preferably, the system also includes a self-calibration function: When there is a systematic deviation between the light intensity sequence data continuously acquired by the data acquisition module and the standard light intensity reference sequence, and the interference coupling decision result continuously indicates that there is no specific interference model matching, the reference model module starts the self-calibration process to refit or correct the preset system calibration parameters. The adaptive execution module updates its control law parameters based on the corrected system calibration parameters.

[0011] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention constructs an idealized target position model based solely on physical laws, detached from external environmental interference; a theoretical reference model and an interference decoupling module based on topological similarity, achieving accurate judgment of system state deviation; effectively distinguishing between global environmental changes caused by clouds and local noise caused by obstruction; compared to the traditional closed-loop control in the prior art that relies solely on the deviation between setpoint and measured values ​​for feedback, this invention introduces the theoretical truth value as a third-dimensional reference system, solving the oscillation problem of the entire array malfunctioning due to the obstruction or failure of individual sensors, and significantly improving the robustness of the system in complex lighting environments; 2. This invention, by employing spatially distributed sensor array units in conjunction with a synthetic analysis method, extends illumination data from a single scalar to a vector field with spatial dimensions; it achieves the effect of capturing spatial gradient information of illumination and identifying shadow movement characteristics; compared with the single-point acquisition schemes in the prior art that lack spatial perception capabilities, this invention can identify the disguise of local noise based on the topological similarity of residual features in spatial distribution, avoiding misadjustment caused by local point interference being misjudged as overall environmental changes, and ensuring the accuracy of the shading system's operation; 3. This invention implements a weighted logic isolation and data reconstruction strategy through a fault-tolerant control unit, achieving self-healing of the control signal; it achieves the effect of maintaining high-precision continuous control even when some sensors fail; compared with the shortcomings of existing technologies that directly lead to feedback distortion or system shutdown when sensor failure occurs, this invention utilizes the effective data of the remaining healthy sensors, reconstructs the state feedback signal through a spatial correlation algorithm and injects it into the control loop, solving the problem of global control stability being destroyed by single-point failure, and greatly extending the service life of mechanical components; 4. This invention introduces a self-calibration function that includes least squares optimization, enabling adaptive evolution of system parameters throughout their entire lifecycle. It achieves the effect of automatically correcting transmittance attenuation parameters caused by sensor aging or dust accumulation on glass curtain walls. Compared with existing technologies that require frequent manual calibration or whose control accuracy decreases over time due to aging, this invention can dynamically update the global system gain coefficient based on long-term observation data, ensuring that the theoretical model always accurately reflects the current physical state. This effectively reduces the system's full-cycle operation and maintenance costs and ensures the long-term stability of the indoor lighting environment. Attached Figure Description

[0012] The present invention will be further explained below with reference to the accompanying drawings and embodiments: Figure 1 This is a structural diagram of the system of the present invention. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0014] Example 1: Please see Figure 1 A multi-blade angle adaptive adjustment system based on light intensity feedback includes: a data acquisition module, used to acquire real-time light intensity sequence data reflecting the physical field state of the target, and to obtain the current spatiotemporal reference data of the system; The reference model module communicates with the data acquisition module and is used to construct a theoretical reference model for the current moment based on spatiotemporal reference data and preset system calibration parameters, and to generate a standard light intensity reference sequence. The interference decoupling module communicates with the reference model module and the data acquisition module. It is used to generate a global environmental interference model and a local sensor interference model based on knowledge-driven generation, and to calculate the interference coupling decision result of the root cause of the current system state deviation by combining the light intensity sequence data and the standard light intensity reference sequence. The adaptive execution module is connected in communication with the interference decoupling module. It is used to receive the interference coupling decision result and generate control commands for adaptively adjusting the multi-blade angle based on the decision result, or generate self-test and maintenance commands for the data acquisition module. The interference decoupling module analyzes the residual characteristics of real data and model prediction data, and compares the topological similarity of the residual characteristics with the preset interference mode in spatial distribution to distinguish between systematic shifts caused by global environmental changes and local noise caused by local component anomalies.

[0015] This embodiment details the overall architecture and core logic of a multi-blade angle adaptive adjustment system based on light intensity feedback; the system activates a data acquisition module, which acts as a sensing front end to capture real-time light intensity sequence data from the physical world. And spatiotemporal reference data containing latitude and longitude coordinates and high-precision timestamps; this step aims to establish the true state boundaries of the physical field; Based on the acquired spatiotemporal reference data, the reference model module calls upon internally preset system calibration parameters, utilizes astronomical algorithms to construct a theoretical reference model for the current moment, and generates a standard illumination intensity reference sequence. Subscript This represents a theoretical reference, or a theoretical benchmark; corresponding to the aforementioned subscript The measured data represents the actual measured data; the sequence represents the theoretical true value anchor point that the sensor should output under ideal conditions of absolutely clear sky, no clouds and no obstruction; the interference decoupling module runs the global environmental interference model and the local sensor interference model in parallel, and uses the synthetic analysis method to calculate the residual characteristics between the actual data and the reference sequence. Based on this, by comparing the topological similarity of the residual characteristics with the preset interference mode in spatial distribution, the interference coupling decision result is calculated; the adaptive execution module responds to the decision result, and if it is determined to be a global environmental change, it generates a control command to adjust the blade angle; if it is determined to be a local interference, it generates a self-check and maintenance command for the data acquisition module. This embodiment introduces a theoretical reference model as a third-dimensional reference system, breaking the limitation of traditional closed-loop control that relies solely on the deviation between the setpoint and the measured value. By utilizing the topological similarity analysis of residual features, the system can effectively identify the disguise of local noise, avoiding the oscillation effect of the entire array acting incorrectly due to the obstruction of individual sensors, and significantly improving the robustness of the system in complex lighting environments.

[0016] Example 2: The data acquisition module includes: a sensor array unit, which consists of multiple light sensors distributed in space according to a predetermined topology, for synchronously or asynchronously acquiring light intensity sequence data; The spatiotemporal reference unit is used to provide the system with spatiotemporal reference data containing location information and precise time information; The actuator feedback unit is used to collect the physical angle state of the angle actuator of each blade in real time and feed it back to the adaptive execution module to form a closed-loop control loop.

[0017] This embodiment further defines the hardware architecture of the data acquisition module, aiming to support the accurate acquisition of high-dimensional data; the sensor array unit is configured as follows: A matrix composed of light sensors, for example The topology is spatially distributed according to a predetermined topology; the design aims to capture spatial gradient information of illumination, expanding a single light intensity scalar into illumination field data with spatial dimensions. Meanwhile, the spatiotemporal reference unit integrates a GPS / BeiDou receiver module and an RTC real-time clock, providing real-time output including precise longitude. ,latitude and time The data comes from satellite positioning systems and its physical meaning is the spatiotemporal boundary conditions for calculating solar altitude and azimuth angles, where latitude... The definition is consistent with the calculation parameters in the subsequent reference model module, that is... Latitude The use of longitude ensures the consistency of the physical meaning throughout the text; In addition, the actuator feedback unit uses Hall sensors or rotary encoders to collect the physical angle status of each blade in real time. The feedback is then sent to the adaptive execution module. This feedback mechanism ensures the closed-loop execution of control commands and can effectively compensate for errors caused by mechanical transmission backlash or motor step loss. In this embodiment, an illumination matrix is ​​constructed through a spatially distributed sensor array, which enables the system to have the physical basis for identifying the direction of shadow movement and the coverage of clouds, providing necessary data support for distinguishing between global uniform changes and local point-like disturbances.

[0018] Example 3: The reference model module includes: a parameter calculation unit, used to calculate and generate theoretical scene parameters based on spatiotemporal reference data and historical system operation data; and a reference generation unit, connected to the parameter calculation unit, used to call preset system calibration parameters according to the theoretical scene parameters, and to generate a standard light intensity reference sequence expected to be sensed at each sensor position in the sensor array unit under an ideal state without interference.

[0019] This embodiment details the specific process by which the reference model module generates truth anchor points based on first principles; the parameter calculation unit receives spatiotemporal reference data and uses solar position algorithms such as the SPA algorithm to solve for theoretical scene parameters, including the solar zenith angle. and solar azimuth The formula for calculating the zenith angle is as follows: Based on this, in order to obtain the specific angle value required for the target angle calculation in subsequent control logic, this embodiment explicitly performs the inverse cosine operation step, and its calculation formula is as follows: To ensure the completeness of the projection calculation, this embodiment specifies the solar azimuth angle. The solution logic is designed to avoid situations where the sun is at its zenith (i.e., This embodiment uses a two-parameter arctangent function to calculate the solar azimuth angle, which automatically resolves the issue of calculation overflow caused by a denominator of zero. The calculation formula is as follows: in, The output range of the function is This ensures the continuity of computing around the clock; The local geographical latitude is derived from a spatiotemporal reference unit. The solar declination angle is derived from date calculations; to ensure the feasibility of the algorithm, this embodiment defines its calculation formula as follows: In this formula, the constant 360 represents a value in degrees. Before substituting it into the sine function for calculation, it needs to be converted to radians. It is the accumulated days of the year, with January 1st being 1; To ensure accurate code-level reproduction and eliminate mean solar time errors, this embodiment introduces mean time difference correction logic: calculating intermediate variables. The mean time difference is calculated using the following formula: The unit is minutes; ultimately combined with longitude. From Example 2: Spatiotemporal reference unit and time zone standard longitude In this embodiment, the acquisition logic is defined as follows: based on longitude Automatically calculates the longitude of the center of the current time zone, i.e. Alternatively, it can be calculated by directly reading the preset local timezone configuration value. The calculation formula is as follows: This embodiment explicitly points out the time parameter in the formula. The unit is hours, that is, Decimal Hours using the 24-hour clock. For example, 2:30 PM corresponds to... The value is 14.5; if the original time data contains minutes and seconds, it needs to be converted to a decimal hour value and substituted into the calculation. To ensure the feasibility of the technical solution, this embodiment specifically states: the formulas mentioned above involve... That is, declination angle, Instantaneous angle and That is, parameters such as latitude are all in degrees; the above trigonometric functions are implemented using computer programming languages ​​such as C++ and Python. When calculating, the angle value must first be multiplied by Convert to radians, inverse cosine function The calculation results also need to be converted back to angles to avoid errors in the calculation dimensions; The reference generation unit calls preset system calibration parameters, including atmospheric transparency coefficient. Sensor mounting plane tilt angle Array azimuth angle and global system gain coefficient A reference sequence is generated by combining the clear sky model and projection geometry. To address the issue that a single DNI model cannot characterize the actual light intensity received by the tilt sensor, this step employs a component synthesis method. The normal direct irradiance is then calculated. The formula is as follows: in, The solar constant is approximately ; The Earth orbital eccentricity correction factor is calculated using the following formula: ; Calculate the light rays at the 1st Angle of incidence on the surface of each sensor And calculate the direct component. The calculation formula is as follows: Furthermore, to eliminate the systematic negative bias caused by considering only direct light, this embodiment introduces an isotropic scattering model to calculate the sky scattering component, and finally multiplies it by the global system gain coefficient. To generate theoretical benchmark values ​​that encompass energy across the entire wavelength range. The calculation formula is as follows: For atmospheric mass, the Rosenberger correction formula is used; specifically, its calculation formula is: The tilt angle of the sensor mounting plane; The atmospheric transparency coefficient is set as an empirical constant in this embodiment, with a value ranging from 0.65 to 0.75, and a preferred value of 0.7, to characterize the atmospheric attenuation characteristics under standard clear sky conditions; The scattering scaling factor is preset to 0.1-0.2 and is used to characterize the background scattering intensity under clear sky conditions. Sky visibility factor; ~ This is the global system gain coefficient, with an initial default value of 1.0. This parameter is a system calibration parameter preset in the embodiment, specifically used to characterize the overall attenuation of optical transmittance caused by dust accumulation on the glass curtain wall or sensor aging. Its value is dynamically updated by the self-calibration process in embodiment 8. This embodiment introduces a complete physical model that includes projection geometry and scattering corrections, and explicitly defines calibrable gain parameters. A digital twin with adaptive evolutionary capabilities was constructed to ensure that the theoretical value It can accurately reflect the expected total irradiance of the sensor under a specific installation posture and current aging condition.

[0020] Example 4: The interference decoupling module includes a simulation engine unit for running a global environmental interference model and a local sensor interference model in parallel. The global environmental interference model is used to simulate interference patterns that have a global, gradual, or consistent impact on the sensor array unit to generate a first theoretical interference data sequence. The local sensor interference model is used to simulate interference patterns that have a sudden, step-like impact on only one or a few sensors in the sensor array unit to generate a second theoretical interference data sequence.

[0021] This embodiment details the operation mechanism of the simulation engine unit in the interference decoupling module, which actively generates interference modes using a knowledge-driven approach. The simulation engine unit loads and runs two types of physical models in parallel. Among them, the global environment interference model constructs a spatial plane-based attenuation field to cover the global, gradual, or uniform effects of the embodiment. This model not only simulates uniform attenuation but also introduces a spatial gradient vector. Simulate the gradual changes at the edge of clouds; To cover the unknown direction of cloud movement, the simulation engine unit performs a full-parameter spatial scan based on a preset gradient search step size, such as... Discretized decay gradient vector Direction angle ,Right now The gradient magnitude is discretized with a preset step size, such as 0.1. , construct containing Global disturbance candidate space of group hypothesis Simultaneously set The upper limit of the value is 2.0. This upper limit is set to ensure that the gradual change characteristics of the cloud edge do not degenerate into a step signal within the small scale of the sensor array, thereby ensuring the orthogonality of the global interference model and the local sensor fault model in terms of mathematical characteristics. For each set of parameter assumptions in this space, the corresponding first theoretical interference data sequence is generated. subscript The representative is generated through simulation. The representative is the global model. For sensor indexes, their mathematical expression is: in, The basic cloud attenuation coefficient is set to a preset empirical scan value, such as 0.3 to 0.7; The attenuation gradient magnitude is used to characterize the rate of change of cloud thickness; For the first The horizontal and vertical coordinates of each sensor in the normalized coordinate system; Explicitly introduce the internal here The function is used to enforce the non-negativity of the attenuation coefficient, preventing non-physical light intensity amplification (i.e., negative attenuation) in gradient extrapolation calculations and ensuring that the model strictly represents the attenuation characteristics; at the same time, an external function is introduced. The function is used to enforce the non-negativity of physical light intensity, preventing negative values ​​in simulation data that violate physical principles due to excessively large attenuation gradients. Meanwhile, regarding the gradient direction angle involved in the formula When performing trigonometric function operations, the angle values ​​need to be converted to radians to match the input requirements of the simulation engine's underlying mathematical library. To ensure the uniqueness and accuracy of the calculations, this embodiment explicitly defines the normalized coordinate system: with the geometric center of the sensor array unit as the origin. The X and Y axes are parallel to the row and column directions of the array, respectively, and the coordinate values ​​are based on the physical grid spacing of the array. Normalization is performed; the specific normalization calculation formula is as follows: in, For the first The physical coordinates of each sensor The coordinates of the array's geometric center are... This represents the physical grid spacing between sensors; it is calculated using this formula, for example, for... Array, coordinate set is ; Meanwhile, the local sensor interference model simulates foreign object obstruction or circuit failure. This model iterates through every possible location of the faulty sensor in the array. ,in , generate containing The second theoretical interference data sequence set of the group hypothesis Among them, for the first A sequence of sensor failures Expressed as: in, The hypothetical fault sensor index is automatically generated by the simulation engine through polling. The occlusion level coefficient is derived from preset fault modes, such as... This indicates severe obstruction; This embodiment actively constructs a global candidate space that includes multi-angle spatial gradient features. With discrete local fault sets This method transforms the abstract problem of fault diagnosis into a specific waveform matching search problem. Instead of relying on exhaustive black-box training data, it is based on explicit physical knowledge, ensuring that the system has interpretable reasoning ability for complex weather conditions, such as rapidly moving cloud shadows and single-point faults.

[0022] Example 5: The interference decoupling module further includes: a residual analysis unit, used to calculate the difference between the light intensity sequence data and the standard light intensity reference sequence to generate a real residual vector, and to calculate the difference between the first theoretical interference data sequence, the second theoretical interference data sequence and the standard light intensity reference sequence to generate corresponding theoretical residual vectors; and a pattern matching unit, connected to the residual analysis unit, used to calculate the similarity measure between the real residual vector and each theoretical residual vector in spatial distribution, and to output the interference coupling decision result based on a predetermined matching threshold and decision logic.

[0023] This embodiment further describes the core steps of interference detection using residual analysis and pattern matching; the residual analysis unit calculates the actual residual vector. Defined as real-world collected data With reference sequence The difference; at the same time, the candidate space generated by this unit based on the aforementioned simulation model. and Batch calculation of the corresponding global theoretical residual vector set With local theoretical residual vector set The pattern matching unit introduces a numerically stable cosine similarity algorithm to iterate and calculate the spatial similarity between the actual residuals and each theoretical residual. The calculation formula is as follows: in, The actual residual vector; The theoretical residual vector is taken from... or Any one of the candidates; For example, the numerical stability constant. This is used to prevent division by zero errors when the residual vector magnitude is close to zero, i.e., the ideal state of no system error. The value is determined based on being less than the minimum quantization error of the analog-to-digital converter in the data acquisition module, ensuring that it will not substantially interfere with the similarity calculation results within the effective resolution range of the physical signal. The system executes competitive decision logic based on maximum likelihood search; the pattern matching unit searches for the maximum similarity in the global candidate space. And search for the maximum similarity in the local candidate space. The final judgment logic is as follows: If and This is determined to be a global environmental change, and corresponding changes will be generated. The gradient parameters are output as the current environment state; if and It was determined to be a local sensor malfunction, and will generate Corresponding index The location of the faulty sensor is locked; if both are below the threshold, it is determined that there is no matching interference. Regarding the matching threshold The method of obtaining it is clearly stated in this embodiment. It was not arbitrarily set, but obtained through statistical calibration during the system initialization phase; the specific process was as follows: under manually confirmed clear sky and undisturbed conditions, data was collected. For example, with 1000 sets of background noise residual data, calculate the maximum similarity distribution between them and the theoretical model, and fit a Gaussian distribution curve. and set Or set the false alarm rate based on the ROC curve. The operating point, typically 0.85, establishes the statistical boundary for distinguishing random noise from structured interference. This embodiment utilizes a full-space search and competitive decision mechanism to keenly capture the topological features of the residuals. Even when the overall light intensity is weak or the residuals are small, the system can still effectively distinguish whether it is a cloud edge that has passed by, i.e., a global feature, or a single point of bird droppings that has been obscured, i.e., a local feature, by comparing the relative size of the similarity. This avoids the ambiguity that may be caused by a single threshold judgment.

[0024] Example 6: The adaptive execution module includes: a global control unit, which determines that systematic adjustment is needed when the interference coupling decision result indicates that the current state deviation matches the global environmental interference model, and generates a control command to uniformly adjust all blades to the new target angle; and a fault-tolerant control unit, which determines that local noise interference occurs when the interference coupling decision result indicates that the local sensor interference model matches the local sensor interference model, and generates a command to freeze or reduce the weight of the interfered sensor data while maintaining or smoothly adjusting the current blade angle.

[0025] This embodiment details the differentiated control strategies of the adaptive execution module for different decision results; the global control unit monitors the decision results in real time; and responds to the decision results by indicating the current state deviation and matching it with the global environmental disturbance model. The system confirmed that the current decrease in sunlight was a real meteorological change caused by clouds; based on this, the unit generated unified control commands to drive all blades to adjust to the new target angle adapted to the scattered light. The target angle no longer relies solely on direct light tracking, but switches to scattering enhancement mode. The specific calculation formula is as follows: in, This is the theoretical direct tracking angle corresponding to the current solar altitude angle. It is strictly defined as the projection profile angle of the ray vector on the blade's rotating section, and its calculation formula is: in, This is the blade normal azimuth angle, whose value is set during system initialization to be the same as the azimuth angle of the sensor mounting plane in Embodiment 3. Equal, that is To prevent the denominator in the formula from being... Approaching zero can cause computational overflow; the system presets a minimum value. ,like ;when At that time, take directly The value at the previous time step or based on The positive and negative values ​​are set as The limiting angle; For trigonometric function operations in the formula, it is necessary to clarify , and All data are in degrees; they need to be converted to radians for calculation. The output needs to be converted back to angles to match the control command format; and These are defined as the arithmetic mean of all valid sensor readings in the sensor array unit and the arithmetic mean of the corresponding theoretical reference value, respectively. The ratio Used to characterize cloud thickness; The scattering compensation gain coefficient is preset to . ; outside the formula and Combining these elements to form physical travel limiting logic; Meanwhile, this embodiment clearly defines For example, the minimum physical limit angle of the blade's mechanical structure. That is, a fully closed state. For example, the maximum physical limit angle. That is, the fully open state, to ensure that the output control commands are always within the effective physical travel range of the actuator; to prevent the denominator from being affected in dawn, dusk or extremely dark environments. Approaching zero can lead to computational overflow or control oscillations; therefore, a computational threshold is set in the system. ,like Only when The above formula is executed when the cloud cover is present; otherwise, the angle from the previous moment is maintained. This strategy aims to increase the blade opening to maximize diffused light intake when cloud cover is present. Conversely, the fault-tolerant control unit determines the match between the decision result and the local sensor interference model, i.e. At that time, confirm the sensor The data is unreliable; at this point, the unit executes a freeze strategy, specifically including freezing the sensor. The weight in the control loop is reduced to zero, and the current blade angle is maintained. The system remains unchanged, or is only finely adjusted based on the data from the remaining health sensors. This embodiment achieves the elimination of false data at the control level. When local interference occurs, the system exhibits strong stability and will not cause the entire system to malfunction due to dirt or obstruction of individual sensors. This greatly extends the service life of mechanical components and ensures the stability of the indoor light environment.

[0026] Example 7: The fault-tolerant control unit is also used to: trigger the marking and maintenance alarms for suspected abnormal sensors in the data acquisition module when the interference is determined to be local noise interference; at the same time, based on the weight information provided by the interference coupling decision result, reconstruct a reliable state feedback signal using the effective data of the other sensors through a data fusion algorithm, so that the global control unit can continue to generate continuous and stable control commands.

[0027] This embodiment further enhances the data self-healing function of the fault-tolerant control unit; in response to the determination of local noise interference, the system immediately marks the suspected abnormal sensor as requiring maintenance on the human-machine interface and sends a maintenance alarm containing the location number; simultaneously, to ensure the continuity of the control system, the fault-tolerant control unit dynamically constructs a weight vector based on the interference coupling decision result. The specific logic is as follows: if the judgment result locks the first... If a sensor is identified as an anomaly, its weight will be forcibly adjusted. To achieve logical isolation, the weights of the remaining normal sensors are set to... Based on this, a reliable state feedback signal is reconstructed using effective data from other sensors and an interpolation algorithm based on weighted spatial distance. The calculation formula is as follows: in, The set of all sensors; The summation conditions explicitly excluded sensors that were judged to be abnormal. ; For sensors With the rejected sensor The Euclidean distance between them is based on the sensor normalized coordinates in Example 4. The calculation is performed using the following formula: Here, normalized coordinate distance is used directly instead of physical distance; where, Let be the power exponent of distance decay, and take . This value is set based on the inverse square law characteristic of the spatial correlation of physical fields, that is, the difference in light intensity decreases quadratically with increasing distance; in large arrays where the sensor distribution is relatively sparse, it can be adaptively adjusted to To strengthen the nearest neighbor weights, but in this embodiment with a compact array, This is the optimal empirical value; It is worth noting that after calculating the reconstructed signal... Then, the fault-tolerant control unit performs the virtual sensor injection step: that is, injecting the original acquisition sequence. The first one that was identified as faulty Each data point is forcibly replaced with a reconstructed value. Generate the corrected full-dimensional state vector ; The correction vector The data is transmitted in real time to the global control unit; this step is crucial as it ensures that the global control unit performs the mean calculation as described in Example 6. When, the denominator The total energy remains unchanged, and the numerator includes the energy estimate recovered through spatial correlation, thus avoiding a false decrease in the mean total energy caused by directly removing faulty sensors and preventing the system from erroneously entering the cloud scattering enhancement mode. In addition, to enhance the robustness of the algorithm, the system has a preset degradation protection logic. When the denominator approaches zero, it automatically switches to using the historical reconstructed value of that position at the previous moment. This embodiment, through a clear weight allocation and data injection strategy, enables the system to maintain continuous, stable, and high-precision control even when some sensors fail.

[0028] Example 8: When there is a systematic deviation between the light intensity sequence data continuously acquired by the data acquisition module and the standard light intensity reference sequence, and the interference coupling decision result continuously indicates that there is no specific interference model matching, the reference model module starts the self-calibration process to refit or correct the preset system calibration parameters; the adaptive execution module updates its control law parameters according to the corrected system calibration parameters.

[0029] This embodiment describes the system's self-calibration function, which aims to address the aging drift problem caused by long-term operation. The system continuously monitors the deviation between the actual data and the reference sequence. In response to the continued existence of systematic deviations, such as the measured value at noon being lower than the reference value by a certain percentage for several consecutive days, and the interference coupling decision result continuously indicating that it does not match either the cloud model or the local occlusion model, that is, excluding short-term weather and individual pollution factors, the system determines that the self-calibration process needs to be initiated. The reference model module initiates the least squares optimization algorithm based on a valid observation time window. In this embodiment, it is defined To accumulate the most recent clear sky conditions to One calendar day, or containing at least Using data from a continuous time series of valid sampling points, with the objective of minimizing the sum of squared residuals between the measured and baseline values, the global system gain coefficient in Example 3 is calculated. Recommended correction value The formula is as follows: in, For the discrete index of the effective sampling points, The total number of valid sampling points; among which, For the first time window Measured light intensity at any given moment; For the first time window The uncorrected reference light intensity at that time, i.e., in Example 3 Time calculation ; To minimize the objective function The current best-fit gain obtained; To prevent system parameter oscillations caused by a single abnormal data point, this embodiment introduces first-order hysteresis filtering logic to address the issues in Embodiment 3. Perform a smooth update using the following formula: in, To update the step size, such as 0.05; in addition, the system sets a safety circuit breaker range, which only applies when... Updates are only allowed when conditions are met; otherwise, an error alert is triggered, prompting manual intervention. The adaptive execution module receives the corrected parameters. This parameter is then applied in subsequent control processes to generate the corrected parameters. ; During each power-on initialization, the system preferentially reads the data saved in the previous operating cycle from non-volatile memory. If the read fails or there is no historical data, the value is initialized to the default value of 1.0; at the same time, each calculation yields... Afterwards, all data is written to this memory in real time to ensure that the system gain parameters have power-down retention characteristics; This embodiment endows the system with the ability to evolve throughout its entire life cycle, enabling it to automatically adapt to the decrease in sensitivity caused by sensor aging or changes in the installation environment, such as changes in overall light transmittance caused by dust accumulation on glass curtain walls. This eliminates the need for frequent manual recalibration and significantly reduces the system's full-cycle operation and maintenance costs.

[0030] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A multi-blade angle adaptive adjustment system based on light intensity feedback, characterized in that, The system includes: The data acquisition module is used to collect real-time light intensity sequence data reflecting the physical field state of the target, and to obtain the current spatiotemporal reference data of the system. The reference model module is communicatively connected to the data acquisition module and is used to construct a theoretical reference model for the current moment based on the spatiotemporal reference data and preset system calibration parameters, and generate a standard light intensity reference sequence. The interference decoupling module is communicatively connected to the reference model module and the data acquisition module. It is used to generate a global environmental interference model and a local sensor interference model based on knowledge-driven generation, and to calculate the interference coupling decision result of the root cause of the current system state deviation by combining the light intensity sequence data and the standard light intensity reference sequence. An adaptive execution module, which is communicatively connected to the interference decoupling module, is used to receive the interference coupling decision result and generate control commands for adaptively adjusting the multi-blade angle based on the decision result, or to generate self-test and maintenance commands for the data acquisition module. The interference decoupling module analyzes the residual characteristics of real data and model prediction data, and compares the topological similarity of the residual characteristics with the preset interference mode in spatial distribution to distinguish between systematic shifts caused by global environmental changes and local noise caused by local component anomalies.

2. The multi-blade angle adaptive adjustment system based on light intensity feedback as described in claim 1, characterized in that, The data acquisition module includes: The sensing array unit consists of multiple light sensors distributed in space according to a predetermined topology, used to synchronously or asynchronously acquire the light intensity sequence data; A spatiotemporal reference unit is used to provide the system with spatiotemporal reference data containing location information and precise time information; The actuator feedback unit is used to collect the physical angle state of the angle actuator of each blade in real time and feed it back to the adaptive execution module to form a closed-loop control circuit.

3. The multi-blade angle adaptive adjustment system based on light intensity feedback as described in claim 2, characterized in that, The reference model module includes: The parameter calculation unit is used to calculate and generate theoretical scenario parameters based on the spatiotemporal reference data and the system's historical operation data. The reference generation unit, connected to the parameter calculation unit, is used to call the preset system calibration parameters according to the theoretical scene parameters, and to generate the standard light intensity reference sequence expected to be sensed at each sensor position in the sensing array unit under an ideal state without interference.

4. The multi-blade angle adaptive adjustment system based on light intensity feedback as described in claim 3, characterized in that, The interference decoupling module includes: The simulation engine unit is used to run the global environmental interference model and the local sensor interference model in parallel. The global environmental interference model is used to simulate interference patterns that have a global, gradual, or consistent impact on the sensor array unit, so as to generate a first theoretical interference data sequence. The local sensor interference model is used to simulate interference patterns that produce sudden, step-like effects on only one or a few sensors in the sensor array unit, in order to generate a second theoretical interference data sequence.

5. The multi-blade angle adaptive adjustment system based on light intensity feedback as described in claim 4, characterized in that, The interference decoupling module further includes: The residual analysis unit is used to calculate the difference between the light intensity sequence data and the standard light intensity reference sequence to generate a real residual vector, and to calculate the difference between the first theoretical interference data sequence, the second theoretical interference data sequence and the standard light intensity reference sequence to generate corresponding theoretical residual vectors. The pattern matching unit, connected to the residual analysis unit, is used to calculate the similarity measure between the actual residual vector and each of the theoretical residual vectors in spatial distribution, and output the interference coupling decision result based on a predetermined matching threshold and decision logic.

6. The multi-blade angle adaptive adjustment system based on light intensity feedback as described in claim 5, characterized in that, The adaptive execution module includes: The global control unit is used to determine that a systematic adjustment is needed when the interference coupling decision result indicates that the current state offset matches the global environmental interference model, and to generate the control command to uniformly adjust all blades to the new target angle. The fault-tolerant control unit is used to determine that the interference coupling decision result matches the local sensor interference model as local noise interference, and generate instructions to freeze or reduce the weight of the interfered sensor data while maintaining or smoothly adjusting the current blade angle.

7. The multi-blade angle adaptive adjustment system based on light intensity feedback as described in claim 6, characterized in that, The fault-tolerant control unit is also used for: When local noise interference is detected, a marking and maintenance alarm is triggered for suspected abnormal sensors in the data acquisition module. Meanwhile, based on the weight information provided by the interference coupling decision result, a reliable state feedback signal is reconstructed using the effective data from the other sensors through a data fusion algorithm, so that the global control unit can continue to generate continuous and stable control commands.

8. The multi-blade angle adaptive adjustment system based on light intensity feedback as described in claim 3, characterized in that, The system also includes a self-calibration function: When there is a systematic deviation between the light intensity sequence data continuously acquired by the data acquisition module and the standard light intensity reference sequence, and the interference coupling decision result continuously indicates that there is no specific interference model matching, the reference model module starts the self-calibration process to refit or correct the preset system calibration parameters. The adaptive execution module updates its control law parameters based on the corrected system calibration parameters.

Images