Compact shelf load self-balancing method and system and storage medium

Abstract

The invention provides a compact shelf load self-balancing method and system and a storage medium, and relates to the technical field of intelligent storage equipment, and the method comprises the steps: obtaining multi-source real-time data of a compact shelf, the multi-source real-time data comprises time sequence data, space stress data and environment parameter data of the compact shelf; performing prediction according to the time sequence data, the space stress data and the environment parameter data through an LSTM model to obtain future load data of the compact shelving in a preset future time period; generating an adjustment strategy of the compact shelving according to the future load data; and controlling the compact shelf to perform self-balancing according to the adjustment strategy. The load of the compact shelf is monitored in real time, accurately predicted and dynamically adjusted, self-balancing control of the supporting arms is optimized, and therefore the storage effect of the compact shelf is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent warehousing equipment, and in particular, to a method, a system, and a storage medium for self-balancing the load of a compact rack. Background Art

[0002] As a core storage device, the load management of an intelligent compact rack directly affects storage safety and space utilization rate. In order to ensure that the compact rack can achieve dynamic balance between storage density and rack load, it is necessary to adjust the layer spacing of the compact rack in real time according to the load condition.

[0003] In the related art, pressure sensors arranged discretely are usually used to monitor the load condition of the rack body, and the layer spacing of the compact rack is adjusted according to the load condition through artificially preset rules. However, since the pressure sensors cannot achieve full coverage of the compact rack, the detection of the load of the compact rack will be incomplete, and the adjustment of the compact rack has a delay due to excessive dependence on artificially preset rules. Finally, the compact rack is difficult to handle the scenario of load mutation, which causes the support arms of the compact rack to oscillate, seriously affecting the safety of the compact rack. Summary of the Invention

[0004] The problem solved by the present invention is how to improve the storage effect of the compact rack.

[0005] To solve the above problems, the present invention provides a method, a system, and a storage medium for self-balancing the load of a compact rack.

[0006] In a first aspect, the present invention provides a method for self-balancing the load of a compact rack, including:[[]] acquiring multi-source real-time data of the compact rack, where the multi-source real-time data includes time series data, spatial stress data, and environmental parameter data of the compact rack; predicting, through an LSTM model, according to the time series data, the spatial stress data, and the environmental parameter data to obtain future load data of the compact rack within a preset future time period; generating an adjustment strategy for the compact rack according to the future load data; controlling the compact rack to perform self-balancing according to the adjustment strategy.

[0007] Optionally, the acquiring of the multi-source real-time data of the compact rack includes:[[]] acquiring, from the database of the compact rack, the file access time stamps, operation durations, and operation interval periods of the compact rack within a preset historical time period, and using the file access time stamps, the operation durations, and the operation interval periods as the time series data; Through the stress sensors installed on the compact shelving, the compact shelving is monitored in real time to obtain the strain value and strain gradient tensor of the compact shelving, and the strain value and the strain gradient tensor are used as the spatial stress data; Through environmental sensors, the temperature and humidity, vibration acceleration, and light intensity of the environment where the compact shelving is located are obtained, and the temperature and humidity, the vibration acceleration, and the light intensity are used as the environmental parameter data.

[0008] Optionally, through the LSTM model, predictions are made based on the time series data, the spatial stress data, and the environmental parameter data to obtain the future load data of the compact shelving within a preset future time period, including: Through interpolation, the time series data, the spatial stress data, and the environmental parameter data are aligned according to a preset time reference, and the aligned time series data, spatial stress data, and environmental parameter data are successively subjected to feature engineering and normalization processing to obtain the input data of the LSTM model; The input data is input into the LSTM model for prediction to obtain the load probability heat map, deformation risk index, and adjustment priority sequence of the compact shelving within the preset future time period; The load probability heat map, the deformation risk index, and the adjustment priority sequence are used as the future load data.

[0009] Optionally, the step of inputting the input data into the LSTM model for prediction to obtain the load probability heat map, deformation risk index, and adjustment priority sequence of the compact shelving within the preset future time period includes: The input data is input into the LSTM model, and feature extraction is performed through the hidden layer of the LSTM model to obtain spatio-temporal features; Attention allocation is performed on the spatio-temporal features through the attention mechanism layer of the LSTM model to obtain key spatio-temporal node features; The key spatio-temporal node features are subjected to dimensionality reduction processing and feature fusion through the fusion layer of the LSTM model to obtain the load probability heat map, deformation risk index, and adjustment priority sequence within the preset future time period.

[0010] Optionally, generating an adjustment strategy for the compact shelving according to the future load data includes: Planning the adjustment path of the support arms of the rack body of the compact shelving according to the load probability heat map to obtain the future path planning of the support arms; Determining the support mode of the support arms according to the deformation risk index; Determine the adjustment sequence of the rack according to the adjusted priority sequence; Generate the adjustment strategy of the mobile rack according to the support mode, the adjustment sequence, and the future path planning of the support arms corresponding to each rack.

[0011] Optionally, the planning of the adjustment path of the support arms of the mobile rack according to the load probability heat map to obtain the future path planning of the support arms includes: Determine the high-load probability area of the mobile rack according to the load probability heat map; Determine the target position of the rack of the mobile rack according to the high-load probability area; Obtain the target position of the support arm according to the path planning algorithm in combination with the target position and the constraint conditions of the support arm corresponding to the rack; Obtain the future path planning according to the initial position and the target position of the support arm.

[0012] Optionally, the determination of the support mode of the support arm according to the deformation risk index includes: Determine the support mode according to the magnitude relationship between the deformation risk index and the preset risk threshold; Wherein, when the deformation risk index is greater than or equal to the preset risk threshold, the support mode is set to the rigid mode; When the deformation risk index is less than the preset risk threshold, the support mode is set to the compliant mode, and the damping coefficient of the mobile rack is set according to the preset requirements.

[0013] Optionally, the control of the self-balancing of the mobile rack according to the adjustment strategy includes: Perform mode switching on the mobile rack according to the support mode; Control the support arms corresponding to the rack to perform angle adjustment according to the adjustment sequence with the future path planning until all the support arms of the mobile rack reach the target position.

[0014] In a second aspect, the present invention provides a mobile rack load self-balancing system, including: A monitoring unit for acquiring multi-source real-time data of the mobile rack, where the multi-source real-time data includes time series data, spatial stress data, and environmental parameter data of the mobile rack; A prediction unit for predicting, through an LSTM model, according to the time series data, the spatial stress data, and the environmental parameter data, to obtain future load data of the mobile rack within a preset future time period; An analysis unit for generating an adjustment strategy for the compact storage rack according to the future load data; A control unit for controlling the self - balancing of the compact storage rack according to the adjustment strategy.

[0015] In a third aspect, the present invention provides a computer - readable storage medium, including a computer - readable storage medium storing a computer program and a processor. When the computer program is read and run by the processor, the method for self - balancing the load of the compact storage rack as described above is implemented.

[0016] The method, system and storage medium for self - balancing the load of the compact storage rack of the present invention realize the comprehensive monitoring of the state of the compact storage rack by acquiring multi - source real - time data of the compact storage rack, including time - series data, spatial stress data and environmental parameter data, providing accurate data support for subsequent prediction and adjustment, enabling the system to timely perceive the state of the compact storage rack, providing a data basis for improving the storage effect of the compact storage rack, and ensuring that the adjustment strategy is formulated based on accurate information. Then, the LSTM model is used to analyze and predict the multi - source real - time data to obtain the future load data of the compact storage rack within a preset future time period, determining in advance the load situation and potential risks faced by the compact storage rack in the future, converting the traditional passive response into active prevention, providing a key basis for generating a scientific and reasonable adjustment strategy in the follow - up, which is the core link for improving the storage effect and can effectively prevent structural problems caused by abnormal loads. Then, an adjustment strategy is generated according to the future load data, and the compact storage rack is controlled to perform self - balancing accordingly, realizing the dynamic and refined management of the compact storage rack, ensuring its reliable operation under various working conditions. According to the adjustment strategy, the linear motor is driven by the electric control system to adjust the angle of the support arm, so that the compact storage rack performs self - balancing, realizing the dynamic balance adjustment of the compact storage rack, enhancing its stability and safety, improving the space utilization rate, reducing the demand for safety redundant space, ensuring the safety and reliability of stored items, effectively enhancing the stability and safety of the compact storage rack, preventing structural damage caused by uneven loads, ensuring the safety and reliability of stored items, improving the storage density and space value of the compact storage rack, realizing a good balance between storage efficiency and safety, and effectively improving the overall storage effect of the compact storage rack. Description of the Drawings

[0017] Figure 1 It is a flowchart of the method for self - balancing the load of the compact storage rack according to an embodiment of the present invention; Figure 2 It is a structural block diagram of the system for self - balancing the load of the compact storage rack according to an embodiment of the present invention. Detailed Embodiments

[0018] To make the above objects, features, and advantages of the present invention more apparent and understandable, the following provides a detailed description of specific embodiments of the present invention in conjunction with the accompanying drawings. Although certain embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments described herein. Instead, these embodiments are provided to more thoroughly and completely understand the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not used to limit the protection scope of the present invention.

[0019] It should be understood that the various steps described in the method embodiments of the present invention can be executed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this regard.

[0020] The term "including" and its variants used herein are open-ended, that is, "including but not limited to"; the term "based on" means "at least partially based on"; the term "optionally" means "optional embodiments". The relevant definitions of other terms will be given in the following description.

[0021] It should be noted that the modifications of "one" and "multiple" mentioned in the present invention are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly specified in the context, it should be understood as "one or more".

[0022] The names of the messages or information exchanged between multiple devices in the embodiments of the present invention are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0023] Combined Figure 1 As shown, the present invention provides a method for self-balancing the load of a compact storage rack, including: Obtaining multi-source real-time data of the compact storage rack, where the multi-source real-time data includes time series data, spatial stress data, and environmental parameter data of the compact storage rack.

[0024] Specifically, by collecting the time series data of the compact storage rack (such as file access timestamps, operation durations, and operation interval periods), spatial stress data (strain values and strain gradient tensors obtained by stress sensors), and environmental parameter data (including temperature and humidity, vibration acceleration, and light intensity, etc.), real-time monitoring of the full range and multi-dimensions of the compact storage rack is achieved, providing a detailed and accurate data basis for subsequent load prediction and adjustment strategy generation, enabling the system to timely and accurately perceive the current state of the compact storage rack, ensuring that the formulated adjustment strategies can be carried out based on the latest, true, and reliable information, and thus laying a solid data support for improving the storage effect of the compact storage rack.

[0025] Predict according to the time series data, the spatial stress data, and the environmental parameter data through the LSTM model to obtain the future load data of the compact shelving within a preset future time period.

[0026] Specifically, based on the collected multi-source real-time data, utilize the powerful time series data processing and prediction capabilities of the LSTM model to accurately predict the load condition of the compact shelving within a preset future time period, and obtain future load data such as a load probability heat map, a deformation risk index, and an adjustment priority sequence. Through the LSTM model, transform the traditional passive reactive management into active preventive management, enabling the system to anticipate in advance the upcoming load change trends and potential risk points faced by the compact shelving. In this way, plan countermeasures in advance, providing a key basis for generating a scientific, reasonable, and forward-looking adjustment strategy in the subsequent process. This is the core link for realizing intelligent prediction in the entire method, effectively avoiding structural problems caused by abnormal load changes, thereby enhancing the stability and storage safety of the compact shelving.

[0027] Generate an adjustment strategy for the compact shelving according to the future load data.

[0028] Specifically, based on the future load data predicted by the LSTM model, including the load probability heat map, the deformation risk index, and the adjustment priority sequence, generate an adjustment strategy that matches the future load condition of the compact shelving. By reasonably planning the adjustment path of the support arms, determining the support mode, and clarifying the adjustment sequence of the rack body, ensure that the compact shelving can make accurate and effective adjustments before the load changes. The generation of this dynamic adjustment strategy based on prediction data in the present invention not only improves the response speed and adaptability of the compact shelving to load changes, but also realizes the refined and dynamic management of the compact shelving. It is an important link connecting the prediction link and the actual control link, directly related to the stable operation and reliable performance of the compact shelving under various working conditions, and provides clear and scientific guidance for subsequent actual control operations.

[0029] Control the compact shelving to perform self-balancing according to the adjustment strategy.

[0030] Specifically, by means of driving a linear motor through an electric control system to adjust the angle of the support arm, etc., the compact storage rack is self-balanced, and the adjustment strategy formulated based on data monitoring and intelligent prediction in the previous steps is put into practice. Through this real-time and precise dynamic balance adjustment, the compact storage rack can effectively enhance its own stability and safety, reduce the risks of mechanical wear and structural damage caused by uneven load distribution, and extend the service life of the equipment. At the same time, reasonable self-balancing adjustment also helps to improve the space utilization rate of the compact storage rack, reduce unnecessary safety redundancy space reservation, improve the storage density and space value on the premise of ensuring the safety and reliability of stored items, achieve a good balance between storage efficiency and safety, and significantly improve the overall storage effect and performance of the compact storage rack.

[0031] The load self-balancing method of the compact storage rack of the present invention realizes the comprehensive monitoring of the state of the compact storage rack by obtaining multi-source real-time data of the compact storage rack, including time series data, spatial stress data, and environmental parameter data, provides accurate data support for subsequent prediction and adjustment, enables the system to timely perceive the state of the compact storage rack, provides a data basis for improving the storage effect of the compact storage rack, and ensures that the adjustment strategy is formulated based on accurate information. Then, the LSTM model is used to analyze and predict the multi-source real-time data to obtain the future load data of the compact storage rack within a preset future time period, determines in advance the load situation and potential risks faced by the compact storage rack in the future, converts the traditional passive response into active prevention, provides a key basis for generating a scientific and reasonable adjustment strategy in the future, is the core link for improving the storage effect, and can effectively prevent structural problems caused by abnormal loads. Then, an adjustment strategy is generated according to the future load data, and the compact storage rack is controlled accordingly to perform self-balancing, realizing the dynamic and refined management of the compact storage rack and ensuring its reliable operation under various working conditions. According to the adjustment strategy, a linear motor is driven through an electric control system to adjust the angle of the support arm, so that the compact storage rack performs self-balancing, realizes the dynamic balance adjustment of the compact storage rack, enhances its stability and safety, improves the space utilization rate, reduces the demand for safety redundancy space, ensures the safety and reliability of stored items, effectively improves the stability and safety of the compact storage rack, prevents structural damage caused by uneven loads, ensures the safety and reliability of stored items, improves the storage density and space value of the compact storage rack, realizes a good balance between storage efficiency and safety, and effectively improves the overall storage effect of the compact storage rack.

[0032] Optionally, the obtaining of the multi-source real-time data of the compact storage rack includes: From the database of the compact storage rack, obtain the file access and storage timestamps, operation durations, and operation interval periods of the compact storage rack within a preset historical time period, and use the file access and storage timestamps, the operation durations, and the operation interval periods as the time series data; By means of the stress sensors arranged on the compact shelving, the compact shelving is monitored in real time to obtain the strain value and strain gradient tensor of the compact shelving, and the strain value and the strain gradient tensor are used as the spatial stress data; By means of the environmental sensors, the temperature and humidity, vibration acceleration and light intensity of the environment where the compact shelving is located are obtained, and the temperature and humidity, the vibration acceleration and the light intensity are used as the environmental parameter data.

[0033] Specifically, the archive access timestamps, operation durations and operation interval periods within a preset historical time period are obtained from the database of the compact shelving and used as time series data. For example, the operation durations and operation interval periods are arranged in sequence according to the archive storage timestamps as time series data. These data can reflect the dynamic change rules of archive access. At the same time, the strain value and strain gradient tensor of the rack body are monitored in real time by the stress sensors arranged on the compact shelving to obtain spatial stress data. The sampling frequency of the stress sensors is 100 Hz, and the spatial resolution is a 10 cm × 10 cm grid, which can accurately reflect the stress distribution of the rack body. The spatial stress data at multiple different times can correspond to the archive storage timestamps. In addition, the environmental parameter data such as the temperature and humidity, vibration acceleration and light intensity of the environment where the compact shelving is located are obtained by using the environmental sensors. These data help to evaluate the influence of the external environment on the load of the compact shelving. The environmental parameter data can be collected in real time by MEMS sensors, and the environmental parameter data at multiple different times can correspond to the archive storage timestamps. In this way, the comprehensive monitoring of the state of the compact shelving is realized, providing accurate data support for subsequent prediction and adjustment.

[0034] In a preferred embodiment of the present invention, in the input data of the LSTM model, the time window of the time series data is the past 24 hours, the operation durations and operation interval periods during archive access are collected at a sampling frequency of 1 Hz, and then the spatial stress data is collected by 16 FBG sensors. The strain value can be [50, 55, 60,..., 100] με, which is used to simulate the strain values of 16 sensors at a certain moment, and the strain gradient tensor is , where the technical specification of the spatial stress data can be a grid with a spatial resolution of 10 cm × 10 cm. For the environmental parameter data, the temperature is 25 °C, the humidity is 40%RH, the vibration acceleration is 0.5 g, and the light intensity is 300 lux. These data can be collected in real time by MEMS sensors.

[0035] In this alternative embodiment, by obtaining time series data such as file access timestamps, operation durations, and operation interval periods, the dynamic pattern of file access can be understood, providing a basis in the time dimension for predicting future load changes. Real-time monitoring of spatial stress data such as strain values and strain gradient tensors can accurately grasp the stress distribution of the rack body and promptly detect potential stress concentration areas. At the same time, obtaining environmental parameter data such as temperature and humidity, vibration acceleration, and light intensity helps evaluate the impact of the external environment on the compact storage rack, further improving the accuracy of prediction. The comprehensive collection of this data ensures that the system can timely and accurately perceive the real-time state of the compact storage rack, enabling subsequent adjustment strategies to be formulated based on the latest, true, and reliable information, thereby effectively improving the storage effect of the compact storage rack and enhancing its stability and security.

[0036] Optionally, through the LSTM model, predicting according to the time series data, the spatial stress data, and the environmental parameter data to obtain future load data of the compact storage rack within a preset future time period, including: By using the interpolation method, align the time series data, the spatial stress data, and the environmental parameter data according to a preset time reference, and successively perform feature engineering and normalization processing on the aligned time series data, spatial stress data, and environmental parameter data to obtain the input data of the LSTM model; Input the input data into the LSTM model for prediction to obtain a load probability heat map, a deformation risk index, and an adjustment priority sequence of the compact storage rack within the preset future time period; Use the load probability heat map, the deformation risk index, and the adjustment priority sequence as the future load data.

[0037] Specifically, first, interpolation method is used to align the time series data, spatial stress data, and environmental parameter data in terms of time, ensuring that all data have a unified time benchmark with an error controlled within 10 ms, providing a synchronized data basis for subsequent analysis. Then, feature engineering and normalization processing are sequentially performed on the aligned data. Features such as the low-frequency component (0 - 0.1 Hz) of the Fourier transform of the operation frequency, the first moment (center of gravity position) and the second moment (distribution dispersion) of the strain field are extracted, and the data is normalized to obtain a data format suitable for input to the LSTM model. After that, the processed data is input into a dual-channel LSTM model for prediction. The temporal channel of the model processes dynamic features, and the spatial channel processes real-time stress distribution data. The attention mechanism is used to focus on key nodes, and finally, a load probability heat map for the next 15 minutes (resolution 10 cm × 10 cm, probability value 0 - 1), a deformation risk index (scalar value 0 - 1, with a warning threshold of 0.3), and an adjustment priority sequence (rack unit numbers arranged in descending order of risk level) are output. For example, the model can predict the load probability distribution of each area of the mobile rack in the future, and identify high-risk areas and the rack units to be adjusted first.

[0038] In a preferred embodiment of the present invention, the normalization process can be achieved by first using the interpolation method to align the time series data, spatial stress data, and environmental parameter data in terms of time, ensuring that all data sources are on a unified time benchmark with an error controlled within less than 10 milliseconds, so as to realize data synchronization. This process is crucial for integrating information from different sensors and data sources, thereby allowing the model to compare and analyze different types of data at the same time point. After alignment, feature engineering is performed on these data to extract key features such as the low-frequency component (0 - 0.1 Hz) of the Fourier transform of the operation frequency and the first and second moments of the strain field. These features help to reveal the patterns and trends behind the data. Subsequently, the following formula is used to normalize these features

[0039] ; Through the above formula, x is the value of the original data point, is the value of the data point after normalization, and the data is scaled to the interval [-1, 1]. Here, μ represents the mean of the data, and σ represents the standard deviation of the data. Normalization ensures that different features are on the same scale, which helps to improve the training efficiency and prediction accuracy of the LSTM model. After completing these steps, the obtained data will be used as the input of the LSTM model to predict the future load situation of the mobile rack.

[0040] In another preferred embodiment of the present invention, the time-series dynamic features include the moving average of the operation frequency in the past 1 hour and the load change rate in the adjacent 15 minutes, which are used to reflect the trend of the access operation intensity, and the weights are 0.18 and 0.15 respectively. The spatial stress data includes the ratio of the maximum principal stress to the material yield strength and the proportion of the area of the high strain gradient region, which are used to evaluate the plastic deformation risk and identify local stress concentration, and the weights are 0.22 and 0.19 respectively. The environmental parameter data includes the temperature-humidity-stress correlation coefficient and the cumulative vibration energy value, which are used to quantify the influence of the environment on the mechanical properties of the material and judge the influence of external disturbances on the structural stability, and the weights are 0.12 and 0.08 respectively. The time series data includes the time-lagged cross-correlation between the operation frequency and the strain change, which is used to reveal the relationship between human operation and structural response, and the weight is 0.06. The above data together constitute the input of the LSTM model, which is used to predict the future load condition of the compact shelving to achieve self-balancing control.

[0041] In this alternative embodiment, through data alignment and preprocessing, multi-source heterogeneous data can be converted into a format recognizable by the model, ensuring the quality and consistency of the data and laying a foundation for accurate prediction. The dual-channel LSTM model combined with the attention mechanism is adopted to fully exploit the spatio-temporal features and key information in the data, effectively improving the prediction accuracy and reliability. The future load data such as the load probability heat map, deformation risk index, and adjustment priority sequence obtained in advance enable the system to anticipate the load change trend and potential risks of the compact shelving in advance, transforming the traditional passive response into a pro-active prevention, providing a key basis for generating scientific and reasonable adjustment strategies subsequently, thus effectively improving the storage effect of the compact shelving, enhancing its stability and safety, reducing structural problems caused by abnormal loads, and at the same time optimizing the space utilization rate to achieve a good balance between storage efficiency and safety.

[0042] Optionally, inputting the input data into the LSTM model for prediction to obtain the load probability heat map, deformation risk index, and adjustment priority sequence of the compact shelving within the preset future time period includes: Inputting the input data into the LSTM model, and extracting features through the hidden layer of the LSTM model to obtain spatio-temporal features; Performing attention allocation on the spatio-temporal features through the attention mechanism layer of the LSTM model to obtain key spatio-temporal node features; Performing dimensionality reduction processing and feature fusion on the key spatio-temporal node features through the fusion layer of the LSTM model to obtain the load probability heat map, deformation risk index, and adjustment priority sequence within the preset future time period.

[0043] Specifically, this step inputs the processed input data into the LSTM model, and the feature extraction of the hidden layer can deeply mine the spatiotemporal features in the data, capture the dynamic trend of load changes and stress distribution patterns. The attention mechanism layer focuses on key spatiotemporal node features, improves the model's attention to important information, and enhances the accuracy of prediction. The dimensionality reduction processing and feature fusion of the fusion layer further optimize the feature expression, remove redundancy, and highlight key information. The load probability heat map, deformation risk index, and adjustment priority sequence finally obtained can reveal in advance the load risks that the compact rack may face in the future and the areas that need to be focused on. In a preferred embodiment of the present invention, the data is feature extracted through the hidden layer of the model (2 layers × 128 units of the time series channel, 1 layer × 64 units of the space channel), and the spatiotemporal features containing time series trends and spatial stress distribution are obtained. These features reflect the dynamic laws of load changes and structural stress patterns of the compact rack. Next, the attention mechanism layer, i.e., the 4-head attention mechanism, allocates attention to the spatiotemporal features, identifies and highlights the key spatiotemporal node features, such as time periods with drastic load changes in the future or spatial areas of stress concentration. Finally, through the fusion layer of 256, 128, and 64 dimensions, that is, the fully connected network, the key spatiotemporal node features are reduced in dimension and fused to remove redundant information and enhance the expressiveness and discrimination of the features, thereby obtaining the load probability heat map in the preset future time period, where the resolution is 10cm×10cm, the probability value is 0-1 floating point type, the deformation risk index is a scalar value 0-1, and the adjustment priority sequence is used to provide accurate prediction information for the subsequent adjustment strategy generation.

[0044] In this optional embodiment, by accurately predicting the future load conditions of the compact shelving, structural problems caused by abnormal loads can be effectively prevented, the stability and safety of the compact shelving can be improved, mechanical wear can be reduced, and the service life can be extended. At the same time, space utilization can be optimized to achieve a benign balance between storage efficiency and safety.

[0045] Optionally, generating an adjustment strategy for the compact rack according to the future load data includes: Planning the adjustment path of the support arm of the rack of the compact rack according to the load probability heat map to obtain the future path planning of the support arm; determining a support mode of the support arm according to the deformation risk index; Determining the adjustment order of the racks according to the adjustment priority sequence; The adjustment strategy of the compact shelving is generated according to the support mode, the adjustment sequence and the future path planning of the support arm corresponding to each of the racks.

[0046] Specifically, first, the adjustment path of the support arms of the mobile shelving body is planned according to the load probability heat map. Based on the grid cells of the load probability heat map and the load probability of the corresponding time slice, high-risk areas are identified to plan the future path for the support arms, ensuring that they move to areas with increasing future loads to provide better support. Secondly, the support mode is determined according to the deformation risk index. Among them, in the preferred embodiment of the present invention, the support mode of the adjustable support arms can be determined according to the magnitude of the deformation risk index. Finally, the adjustment order of the shelving body is determined according to the adjustment priority sequence, and the shelving unit numbers can be arranged in descending order of risk level according to the adjustment priority sequence, thus providing a clear adjustment priority order for the actuator. Combining this information, the system generates an adjustment strategy for the mobile shelving that includes the support mode, adjustment order, and future path planning to achieve precise and dynamic load self-balancing.

[0047] In this alternative embodiment, the adjustment path of the support arms is planned in advance through the load probability heat map, enabling the support arms to move to areas with increasing future loads in advance, providing timely and effective support, and preventing structural damage caused by uneven loads. Determining the support mode according to the deformation risk index can enable the compliant mode to be enabled at low risk to reduce energy consumption, and switch to the rigid mode at high risk to ensure structural stability, achieving a balance between energy conservation and safety. Determining the adjustment order of the shelving body according to the adjustment priority sequence ensures that high-risk areas are processed first, improving the adjustment efficiency and avoiding resource waste. To sum up, this step effectively improves the stability and safety of the mobile shelving by generating a scientific and reasonable adjustment strategy, reduces mechanical wear, extends the service life, optimizes the space utilization rate, reduces the demand for safety redundant space, achieves a good balance between storage efficiency and safety, and significantly improves the overall storage effect of the mobile shelving.

[0048] Optionally, the planning of the adjustment path of the support arms of the shelving body of the mobile shelving according to the load probability heat map to obtain the future path planning of the support arms includes: Determine the high-load probability area of the mobile shelving according to the load probability heat map; Determine the target position of the shelving body of the mobile shelving according to the high-load probability area; Obtain the target position of the support arm according to the path planning algorithm in combination with the target position and the constraint conditions of the shelving body corresponding to the support arm; Obtain the future path planning according to the initial position and the target position of the support arm.

[0049] Specifically, first, identify the high-load probability areas of the compact shelves based on the load probability heat map. For example, divide the compact shelves into 10 cm × 10 cm grid cells and provide a load probability (0 - 1) for each cell in the corresponding time slice. By analyzing these probability values, the areas where the future load may increase can be located. Then, determine these high-load probability areas as the target positions of the rack body so that the support arms can provide more effective support. Next, use path planning algorithms such as the A* algorithm or the artificial potential field method, combined with the constraint conditions of the target position and the support arms (such as physical structure limitations, kinematic parameters, etc.), to calculate the target positions of the support arms. The support arms need to reach the target positions at the allowed speed and acceleration within the specified time while avoiding collisions with other components of the compact shelves. Finally, generate the future path planning of the support arms based on the initial positions and the calculated target positions of the support arms to ensure that the support arms can move to the target positions safely and efficiently.

[0050] In this alternative embodiment, by identifying the high-load probability areas and planning the adjustment paths of the support arms, the system can move the support arms to the areas that need to be strengthened before the actual load increases, thus effectively avoiding structural damage caused by uneven loads and enhancing the stability and safety of the compact shelves. Planning the path in advance can also reduce mechanical wear and extend the service life of the compact shelves.

[0051] Optionally, determining the support mode of the support arm according to the deformation risk index includes: Determine the support mode according to the magnitude relationship between the deformation risk index and a preset risk threshold; Among them, when the deformation risk index is greater than or equal to the preset risk threshold, set the support mode to the rigid mode; When the deformation risk index is less than the preset risk threshold, set the support mode to the compliant mode and set the damping coefficient of the compact shelves according to the preset requirements.

[0052] Specifically, first, compare the deformation risk index with the preset risk threshold to determine the support mode of the support arm. Among them, the deformation risk index is calculated based on the predicted maximum strain in the next 15 minutes, the current real-time maximum strain, and the environmental coupling coefficient. When the deformation risk index is greater than or equal to the preset risk threshold, it indicates that the compact storage rack is in a high-risk state. At this time, set the support mode to the rigid mode, that is, the electromagnetic clutch locks the support arm, the magnetorheological damper increases the viscosity, and the vibration suppression rate of the rack body increases to enhance the structural stability and prevent structural damage caused by load changes. On the contrary, when the deformation risk index is less than the preset risk threshold, set the support mode to the compliant mode. At this time, the support arm is finely adjusted at a preset speed, and the magnetorheological damper maintains a low-viscosity state to reduce energy consumption and maintain the flexibility of the structure. At the same time, set the damping coefficient of the compact storage rack according to the preset requirements to ensure the balance between energy saving and stability in the low-risk state.

[0053] In a preferred embodiment of the present invention, assume that the current preset risk threshold is 0.35, and exceeding the preset risk threshold by 0.3, the system marks it as a high risk. Calculate the change rate of the preset risk threshold within the last 30 seconds and find that Δshort = 0.02 / s, exceeding the preset risk threshold of 0.01 / s, triggering the fast response protocol. Analyze the predicted heat map and find that the preset risk threshold will continue to exceed 0.28 within the next 5 minutes, and the current relative humidity is 30%RH, exceeding 20%RH. Therefore, the threshold is increased by 5%, and the new preset risk threshold is adjusted to 0.33; the vibration energy Ev is detected as 0.3g²·s, which is lower than the threshold of 0.5g²·s, so there is no need to adjust the preset risk threshold.

[0054] Finally, determine the support mode of the support arm and perform corresponding actions according to the combined conditions of the deformation risk index (DRI) and the time persistence factor (TPF). When the DRI value exceeds 0.3 and the TPF is greater than 0.25, the system determines that it is a confirmed high risk and immediately switches to the rigid mode and triggers an audible and visual alarm to ensure the stability of the compact storage rack structure. If the DRI value exceeds 0.3 but the TPF is less than or equal to 0.25, the system determines that it is a transient high risk. At this time, enhance the compliant mode and increase the stability by increasing the damping coefficient by 50%. When the DRI value is less than or equal to 0.3 but the change rate Δshort of DRI is greater than 0.01 / s, the system believes that the risk is rising. At this time, pre-tighten the electromagnetic clutch and be ready to lock at any time to cope with possible risks. In all other cases, the system determines that it is a low risk and maintains the compliant mode to keep the compact storage rack running normally. This conditional decision-making process ensures that the compact storage rack can adopt the most appropriate support mode at different risk levels, thereby improving the safety and reliability of the entire system.

[0055] In this alternative embodiment, the support mode is flexibly adjusted according to the actual risk state of the compact storage rack to achieve a balance between energy conservation and safety. When switched to the rigid mode under high-risk conditions, the structural stability of the compact storage rack can be significantly enhanced, effectively preventing structural damage caused by sudden load changes or stress concentration and ensuring the safety and reliability of stored items. While in the low-risk state, the compliant mode is maintained, which can not only meet the basic support requirements of the structure but also reduce energy consumption and extend the service life of mechanical components.

[0056] Optionally, according to the adjustment strategy, controlling the self-balancing of the compact storage rack includes: Performing mode switching on the compact storage rack according to the support mode; Controlling the support arms corresponding to the rack body to perform angle adjustment according to the adjustment sequence and the future path planning until all the support arms of the compact storage rack reach the target positions.

[0057] Specifically, performing mode switching on the compact storage rack according to the support mode is the first step in controlling the self-balancing of the compact storage rack. Specifically, when the deformation risk index reaches or exceeds the preset risk threshold (DRI≥0.3), the system switches the support mode of the compact storage rack to the rigid mode. In this mode, the electromagnetic clutch locks the support arms, and the viscosity of the magnetorheological damper is increased to 50 N·s / mm, thereby increasing the vibration suppression rate of the rack body to more than 90% to ensure structural stability. On the contrary, if the deformation risk index is lower than the preset risk threshold (DRI<0.3), the compliant mode is adopted. At this time, the support arms are finely adjusted at a speed of 0.1 mm / s, and the magnetorheological damper maintains a low viscosity state (5 N·s / mm) to reduce energy consumption and maintain the flexibility of the structure. Next, the system controls the support arms corresponding to the rack body to perform angle adjustment according to the adjustment sequence and the future path planning. This process is based on the adjustment strategy generated from the load probability heat map, the deformation risk index, and the adjustment priority sequence to ensure that each support arm can move to the target position according to the predetermined sequence and path. For example, the system may give priority to adjusting the support arms in the high-load probability area to ensure that these areas are supported in a timely manner. Through the electric control system driving the linear motor, the support arms gradually adjust the angle until all the support arms reach the target positions, completing the self-balancing process of the compact storage rack.

[0058] In this alternative embodiment, by precisely controlling the angle adjustment of the support arms, the dynamic balance adjustment of the compact storage rack is achieved, significantly improving the stability and safety of the compact storage rack. The mode switching mechanism ensures that in a high-risk state, the compact storage rack can quickly switch to the rigid mode, enhancing the structural stability and preventing structural damage caused by sudden load changes, ensuring the safety and reliability of the stored items. In a low-risk state, the compliant mode is adopted, which can not only meet the support requirements but also reduce energy consumption and extend the service life of mechanical components. Angle adjustment is carried out according to the adjustment sequence and future path planning to ensure the orderly movement of the support arms, avoid mutual interference, improve the adjustment efficiency, and reduce the adjustment time. This precise control method also reduces mechanical wear, extends the service life of the compact storage rack, optimizes the space utilization rate, reduces waste caused by safety redundant space, achieves a good balance between storage efficiency and safety, significantly improves the overall storage effect of the compact storage rack, and enables it to maintain an efficient and stable operating state under various working conditions.

[0059] Combined with Figure 2 As shown, the present invention also provides a compact storage rack load self-balancing system, including: A monitoring unit for obtaining multi-source real-time data of the compact storage rack, where the multi-source real-time data includes time series data, spatial stress data, and environmental parameter data of the compact storage rack; A prediction unit for predicting, through an LSTM model, future load data of the compact storage rack within a preset future time period according to the time series data, the spatial stress data, and the environmental parameter data; An analysis unit for generating an adjustment strategy for the compact storage rack according to the future load data; A control unit for controlling the compact storage rack to perform self-balancing according to the adjustment strategy.

[0060] The advantages of the compact storage rack load self-balancing system of the present invention compared with the prior art are the same as those of the above-mentioned compact storage rack load self-balancing method compared with the prior art, and will not be elaborated here.

[0061] The present invention also provides a computer-readable storage medium, including a computer-readable storage medium storing a computer program and a processor. When the computer program is read and run by the processor, the above-mentioned compact storage rack load self-balancing method is implemented.

[0062] The advantages of the computer-readable storage medium of the present invention compared with the prior art are the same as those of the above-mentioned compact storage rack load self-balancing method compared with the prior art, and will not be elaborated here.

[0063] Although the present invention is disclosed as above, the scope of protection of the present invention is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and these changes and modifications will all fall within the scope of protection of the present invention.

Claims

1. A method for self - balancing the load of a compact shelving, characterized in that, Including: Obtain multi-source real-time data of the compact storage rack, where the multi-source real-time data includes time series data, spatial stress data, and environmental parameter data of the compact storage rack; Through the LSTM model, perform prediction based on the time series data, the spatial stress data, and the environmental parameter data to obtain future load data of the compact storage rack within a preset future time period; Generate an adjustment strategy for the compact storage rack according to the future load data; Control the compact storage rack to perform self-balancing according to the adjustment strategy.

2. The method for self-balancing the load of a compact rack according to claim 1, wherein The obtaining of the multi-source real-time data of the compact storage rack includes: From the database of the compact storage rack, obtain the file access timestamp, operation duration, and operation interval period of the compact storage rack within a preset historical time period, and use the file access timestamp, the operation duration, and the operation interval period as the time series data; Through stress sensors arranged on the compact storage rack, perform real-time monitoring on the compact storage rack to obtain the strain value and strain gradient tensor of the compact storage rack, and use the strain value and the strain gradient tensor as the spatial stress data; Through environmental sensors, obtain the temperature and humidity, vibration acceleration, and light intensity of the environment where the compact storage rack is located, and use the temperature and humidity, the vibration acceleration, and the light intensity as the environmental parameter data.

3. The method for self - balancing the load of the compact rack according to claim 1, wherein, The performing of prediction through the LSTM model based on the time series data, the spatial stress data, and the environmental parameter data to obtain future load data of the compact storage rack within a preset future time period includes: Through the interpolation method, align the time series data, the spatial stress data, and the environmental parameter data according to a preset time reference, and perform feature engineering and normalization processing on the aligned time series data, spatial stress data, and environmental parameter data in sequence to obtain the input data of the LSTM model; Input the input data into the LSTM model for prediction to obtain a load probability heat map, a deformation risk index, and an adjustment priority sequence of the compact storage rack within the preset future time period; Use the load probability heat map, the deformation risk index, and the adjustment priority sequence as the future load data.

4. The method for self-balancing the load of a compact shelving according to claim 3, wherein, The inputting of the input data into the LSTM model for prediction to obtain a load probability heat map, a deformation risk index, and an adjustment priority sequence of the compact storage rack within the preset future time period includes: Input the input data into the LSTM model, and perform feature extraction through the hidden layer of the LSTM model to obtain spatio-temporal features; Perform attention allocation on the spatio-temporal features through the attention mechanism layer of the LSTM model to obtain key spatio-temporal node features; Perform dimensionality reduction processing and feature fusion on the key spatio-temporal node features through the fusion layer of the LSTM model to obtain a load probability heat map, a deformation risk index, and an adjustment priority sequence within the preset future time period.

5. The method for self-balancing the load of a compact rack according to claim 3, wherein The generating of the adjustment strategy for the compact storage rack according to the future load data includes: Plan the adjustment path of the support arm of the frame of the compact storage rack according to the load probability heat map to obtain the future path plan of the support arm; Determine the support mode of the support arm according to the deformation risk index; Determine the adjustment order of the frame according to the adjustment priority sequence; Generate the adjustment strategy of the compact storage rack according to the support mode, the adjustment order and the future path plan of the support arm corresponding to each frame; 6. The method for self-balancing the load of the compact shelving according to claim 5, characterized in that, The planning of the adjustment path of the support arm of the frame of the compact storage rack according to the load probability heat map to obtain the future path plan of the support arm includes: Determine the high load probability area of the compact storage rack according to the load probability heat map; Determine the target position of the frame of the compact storage rack according to the high load probability area; Obtain the target position of the support arm according to the path planning algorithm in combination with the target position and the constraint conditions of the support arm corresponding to the frame; Obtain the future path plan according to the initial position and the target position of the support arm; 7. The method for self-balancing the load of the compact rack according to claim 5, characterized in that, The determining the support mode of the support arm according to the deformation risk index includes: Determine the support mode according to the magnitude relationship between the deformation risk index and the preset risk threshold; Wherein, when the deformation risk index is greater than or equal to the preset risk threshold, set the support mode to the rigid mode; When the deformation risk index is less than the preset risk threshold, set the support mode to the compliant mode and set the damping coefficient of the compact storage rack according to the preset requirements; 8. The method for self - balancing the load of the compact shelving according to claim 5, wherein, The controlling the compact storage rack to perform self-balancing according to the adjustment strategy includes: Perform mode switching on the compact storage rack according to the support mode; Control the support arm corresponding to the frame to perform angle adjustment according to the adjustment order with the future path plan until all the support arms of the compact storage rack reach the target position; 9. A load self-balancing system for a compact shelving, characterized in that, including: A monitoring unit for acquiring multi-source real-time data of the compact storage rack, where the multi-source real-time data includes time series data, spatial stress data, and environmental parameter data of the compact storage rack; A prediction unit for predicting through an LSTM model according to the time series data, the spatial stress data, and the environmental parameter data to obtain future load data of the compact storage rack within a preset future time period; An analysis unit for generating an adjustment strategy of the compact storage rack according to the future load data; A control unit for controlling the compact storage rack to perform self-balancing according to the adjustment strategy; 10. A computer-readable storage medium, characterized in that, Including a computer-readable storage medium storing a computer program and a processor, when the computer program is read and run by the processor, the method for self-balancing the load of the compact storage rack as described in any one of claims 1 to 8 is implemented.

Images