A drone battery monitoring method and system for smart hangar

Abstract

The invention discloses a method and system for monitoring a battery of an unmanned aerial vehicle (UAV) for an intelligent hangar, and belongs to the technical field of battery monitoring. The method for monitoring a battery of an unmanned aerial vehicle (UAV) comprises: inputting the real-time charge and discharge times of a battery into an empirical degradation model to obtain an initial value of a battery health assessment; inputting the real-time operating condition information of the battery of the UAV into an error compensation model to obtain a battery health error assessment value; superimposing the battery health error assessment value with the initial value of the battery health assessment to obtain a battery health assessment value; acquiring real-time operation data of a power system, and inputting the real-time operation data of the power system into a pre-built UAV scheduling capacity prediction model to obtain a UAV scheduling capacity prediction value; generating a UAV scheduling strategy based on the battery health assessment value and the UAV scheduling capacity prediction value; and allocating tasks by considering the health status of the UAV battery, thereby ensuring that the battery status matches the task undertaken, and avoiding mission failure or UAV damage due to battery health problems.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery monitoring, and in particular relates to a method and system for monitoring the battery of an unmanned aerial vehicle (UAV) used in an intelligent hangar. Background Art

[0002] At present, some power companies use drones to inspect power lines, substations, etc., to promptly discover and deal with potential safety hazards, improve the stability and safety of the power system, and at the same time establish drone hangars for storing drones at different data collection sites.

[0003] When calling a drone to perform a task through the drone hangar control system, the battery health status of each drone in the drone hangar is different in actual applications. In the prior art, when calling multiple drones to perform tasks, the task allocation does not fully consider the battery health status of the drones. The battery status of the drones does not match the tasks they undertake, which may cause the task execution to fail due to the poor battery health of some drones, or even cause some drones to be damaged. Summary of the invention

[0004] The present invention provides a method and system for monitoring the battery of an unmanned aerial vehicle (UAV) for an intelligent hangar, so as to solve the technical problem that the battery status of the UAV does not match the task it undertakes, resulting in the failure of the UAV to perform the task.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] A first aspect of the present invention provides a method for monitoring a battery of a drone in a smart hangar, comprising:

[0007] Obtain the real-time charge and discharge times of the drone battery and the real-time working condition information of the drone battery;

[0008] Input the real-time charge and discharge times of the battery into a preset empirical degradation model to obtain an initial value for battery health assessment;

[0009] The real-time operating condition information of the drone battery is input into the trained error compensation model to obtain the battery health error assessment value;

[0010] The battery health error evaluation value is superimposed on the battery health evaluation initial value to obtain the battery health evaluation value;

[0011] Acquire the real-time operation data of the power system, and input the real-time operation data of the power system into the pre-built UAV dispatch capacity prediction model to obtain the UAV dispatch capacity prediction value;

[0012] Generate a drone scheduling strategy based on the battery health assessment value and the drone scheduling capacity prediction value.

[0013] Furthermore, the construction process of the empirical degradation model includes:

[0014] Obtain the battery charge and discharge record data of the drone, and use the convex optimization algorithm to smooth the charge and discharge record data to obtain the charge and discharge standardized data;

[0015] The empirical degradation model of the UAV battery is constructed based on the standardized charge and discharge data and the corresponding charge and discharge times, and the parameter weights of the empirical degradation model are identified using the least squares method.

[0016] Inputting the number of charge and discharge times to be predicted into the empirical degradation model to obtain a battery health assessment fitting value; and calculating a battery life prediction value based on the battery health assessment fitting value;

[0017] Get the real value of the drone battery capacity, and calculate the real value of the battery life based on the real value of the drone battery capacity;

[0018] The remaining life prediction error is calculated based on the predicted battery life value and the actual battery life value. The expression formula is:

[0019] ;

[0020] In the formula, Represents the predicted value of battery life; Indicates the actual value of battery life; It is expressed as the remaining life prediction error;

[0021] The regularization parameter in the convex optimization algorithm is adjusted based on the remaining life prediction error, and the construction process of the empirical degradation model is iterated repeatedly until the remaining life prediction error converges, and the final empirical degradation model is output.

[0022] Furthermore, the training method of the error compensation model includes:

[0023] According to the given battery charge and discharge times, the corresponding battery health prediction initial value is read from the empirical degradation model;

[0024] Read the actual value of the battery health assessment corresponding to the given battery charge and discharge times, and calculate the battery health prediction error value based on the initial value of the battery health prediction and the corresponding actual value of the battery health assessment;

[0025] Construct error compensation model based on BP neural network;

[0026] The historical operating condition information of the drone battery is input into the error compensation model to obtain the battery health assessment training error value; the MAPE training loss value is calculated based on the battery health assessment training error value and the battery health prediction error value;

[0027] The parameters of the error compensation model are optimized based on the MAPE training loss value, and iterations are repeated until the MAPE training loss value converges, and the trained error compensation model is output.

[0028] Furthermore, the expression of the empirical degradation model is as follows:

[0029] ;

[0030] In the formula, It is represented as the battery health assessment fitting value in the cth charge and discharge cycle; It is expressed as the rate of charging and discharging of the drone battery in the cth charging and discharging cycle; It represents the initial capacity of the battery of the drone to be tested; , , and Represents setting parameter weights.

[0031] Furthermore, the expression for calculating the battery health prediction error value based on the battery health prediction initial value and the corresponding battery health assessment true value is as follows:

[0032] ;

[0033] In the formula, It is represented as the battery health prediction error value of the mth UAV; Represented as the true value of the battery health assessment of the mth drone; Represented as the initial value of the battery health prediction of the m-th drone.

[0034] Furthermore, the MAPE training loss value is calculated based on the battery health assessment training error value and the battery health assessment prediction error value. The expression formula is:

[0035] ;

[0036] In the formula, n is the total number of charge and discharge cycles of the drone battery; m is the drone serial number; c is the charge and discharge cycle serial number of the drone battery; It is represented as the battery health prediction error value in the cth charge and discharge cycle of the drone battery; It is expressed as the battery health training error value in the cth charge and discharge cycle of the drone battery. Expressed as MAPE training loss value.

[0037] Furthermore, the real-time operation data of the power system is input into the pre-built UAV dispatch capacity prediction model to obtain the UAV dispatch capacity prediction value, including:

[0038] Inputting the real-time operation data of the power system into a pre-built UAV dispatch capacity prediction model, wherein the UAV dispatch capacity prediction model includes a real-time dispatch capacity prediction unit, a distributed file unit and a day-ahead dispatch capacity prediction unit;

[0039] The real-time dispatch capacity prediction unit outputs the temporary dispatch capacity prediction value of the UAV according to the real-time operation data of the power system and stores it in the distributed file unit;

[0040] The day-ahead scheduling capacity prediction unit reads the temporary scheduling capacity prediction value of the UAV from the distributed file unit, predicts the future scheduling capacity prediction value of the UAV based on the temporary scheduling capacity prediction value of the UAV, and performs weighted summation of the temporary scheduling capacity prediction value of the UAV and the future scheduling capacity prediction value of the UAV to obtain the UAV scheduling capacity prediction value.

[0041] Furthermore, the drone scheduling strategy includes drone mission scheduling instructions and drone maintenance scheduling instructions;

[0042] Generate a drone scheduling strategy based on the battery health assessment value and the drone scheduling capacity prediction value, including: when the battery health assessment value is greater than the drone scheduling capacity prediction value, read the corresponding drone number to obtain a reserve drone set; retrieve the drone number from the reserve drone set to generate a drone task scheduling instruction;

[0043] When the battery health assessment value is less than the battery maintenance threshold, the corresponding drone number is retrieved, and a drone maintenance scheduling instruction is generated based on the retrieved drone number.

[0044] Furthermore, the construction of the UAV scheduling capacity prediction model includes:

[0045] A real-time scheduling capacity prediction unit is constructed based on the spark framework, wherein the real-time scheduling capacity prediction unit includes a Spilt segmentation layer, a Map function, and a ReduceByKey function in sequence; a day-ahead scheduling capacity prediction unit is constructed based on a parallel decision tree algorithm;

[0046] Obtain historical operation data of the power system to build a resilient distributed data set;

[0047] Inputting the elastic distributed data set into the UAV scheduling capacity prediction model and outputting the scheduling capacity training prediction value;

[0048] The MSE training loss value is calculated based on the scheduling capacity training prediction value and the actual scheduling capacity value. The parameters of the drone scheduling capacity prediction model are optimized according to the MSE training loss value. The iteration is repeated until the MSE training loss value converges to output the trained drone scheduling capacity prediction model.

[0049] In a second aspect, the present invention provides a drone battery monitoring system, comprising:

[0050] The acquisition module is used to obtain the real-time charge and discharge times of the drone battery and the real-time working condition information of the drone battery;

[0051] The battery evaluation module is used to input the real-time charge and discharge times of the battery into a preset empirical degradation model to obtain an initial value of the battery health evaluation; input the real-time working condition information of the drone battery into the trained error compensation model to obtain a battery health error evaluation value; and superimpose the battery health error evaluation value with the initial value of the battery health evaluation to obtain a battery health evaluation value;

[0052] A prediction module is used to obtain real-time operation data of the power system, and input the real-time operation data of the power system into a pre-built UAV dispatch capacity prediction model to obtain a UAV dispatch capacity prediction value;

[0053] The scheduling module is used to generate a UAV scheduling strategy based on the battery health assessment value and the UAV scheduling capacity prediction value.

[0054] In a third aspect, the present invention provides an electronic device, including a storage medium and a processor; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the drone battery monitoring method for a smart hangar described in the first aspect.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] The present invention obtains real-time operation data of the power system, inputs the real-time operation data of the power system into a pre-built UAV scheduling capacity prediction model to obtain a UAV scheduling capacity prediction value; generates a UAV scheduling strategy based on a battery health assessment value and a UAV scheduling capacity prediction value; and assigns tasks by considering the health status of the UAV battery, thereby ensuring that the battery status matches the task undertaken, thereby avoiding mission failure or UAV damage due to battery health problems.

[0057] The present invention inputs the real-time charge and discharge times of the battery into a preset empirical degradation model to obtain an initial value of battery health assessment; inputs the real-time operating condition information of the UAV battery into the trained error compensation model to obtain a battery health error assessment value; and superimposes the battery health error assessment value with the initial value of the battery health assessment to obtain a battery health assessment value; the error compensation model can effectively supplement the uncertainty in the battery degradation process, thereby enhancing the adaptability and prediction accuracy of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A flowchart of a method for monitoring a drone battery in a smart hangar provided in Embodiment 1 of the present invention;

[0059] Figure 2 A flowchart of a method for monitoring a drone battery in an intelligent hangar provided in Embodiment 2 of the present invention;

[0060] Figure 3 A flowchart of constructing an empirical degradation model provided in Example 2 of the present invention;

[0061] Figure 4 A flowchart of a training error compensation model provided in Example 2 of the present invention;

[0062] Figure 5 A comparison chart of experimental data of the empirical degradation model provided in Example 2 of the present invention;

[0063] Figure 6 A comparison chart of experimental data predicted by the battery health estimation model provided in Example 2 of the present invention;

[0064] Figure 7 A structural diagram of a UAV scheduling capacity prediction model provided in Example 2 of the present invention;

[0065] Figure 8 This is a structural diagram of the real-time scheduling capacity prediction unit provided in Example 2 of the present invention. DETAILED DESCRIPTION

[0066] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0067] Example 1

[0068] like Figure 1 As shown, this implementation provides a drone battery monitoring method for a smart hangar, including:

[0069] Obtain the real-time charge and discharge times of the drone battery and the real-time working condition information of the drone battery;

[0070] Input the real-time charge and discharge times of the battery into a preset empirical degradation model to obtain an initial value for battery health assessment;

[0071] The real-time operating condition information of the drone battery is input into the trained error compensation model to obtain the battery health error assessment value;

[0072] The battery health error evaluation value is superimposed on the battery health evaluation initial value to obtain the battery health evaluation value;

[0073] Acquire real-time operation data of the power system, which includes power transmission path, real-time environmental data and power system fault data; the power transmission path is used to confirm the flight mission duration and mission flight path of the UAV; input the real-time operation data of the power system into a pre-built UAV dispatch capacity prediction model to obtain the UAV dispatch capacity prediction value;

[0074] A drone dispatch strategy is generated based on the battery health assessment value and the drone dispatch capacity prediction value, that is, to dispatch drones for battery replacement and repair, or to dispatch drones to perform power grid path monitoring tasks; by assigning tasks considering the health status of the drone battery, it can be ensured that the battery status matches the task undertaken, avoiding mission failure or drone damage due to battery health problems.

[0075] The construction process of the empirical degradation model and the error compensation model includes:

[0076] Obtain the charging and discharging record data of the drone battery, and fit the charging and discharging record data and the battery charging and discharging times to obtain an empirical degradation model;

[0077] Reading a corresponding battery health prediction initial value from the empirical degradation model according to a given battery charge and discharge times; reading a battery health assessment true value corresponding to a given battery charge and discharge times, and calculating a battery health prediction error value according to the battery health prediction initial value and the corresponding battery health assessment true value;

[0078] Construct error compensation model based on BP neural network;

[0079] The historical operating condition information of the drone battery is input into the error compensation model to obtain the battery health assessment training error value; the MAPE training loss value is calculated based on the battery health assessment training error value and the battery health prediction error value;

[0080] The parameters of the error compensation model are optimized based on the training loss value, and the iteration is repeated until the training loss value converges, and the trained error compensation model is output; the error compensation model can be used to effectively supplement the uncertainty in the battery degradation process, thereby enhancing the adaptability and prediction accuracy of the model.

[0081] The UAV scheduling capacity prediction model construction process includes:

[0082] A real-time scheduling capacity prediction unit is constructed based on the spark framework, wherein the real-time scheduling capacity prediction unit includes a Spilt segmentation layer, a Map function, and a ReduceByKey function in sequence; a day-ahead scheduling capacity prediction unit is constructed based on a parallel decision tree algorithm;

[0083] Obtain historical operation data of the power system to build a resilient distributed data set;

[0084] Inputting the elastic distributed data set into the UAV scheduling capacity prediction model, and outputting the scheduling capacity training prediction value;

[0085] The MSE training loss value is calculated based on the scheduling capacity training prediction value and the actual scheduling capacity value. The parameters of the drone scheduling capacity prediction model are optimized according to the MSE training loss value. The iteration is repeated until the MSE training loss value converges to output the trained drone scheduling capacity prediction model.

[0086] The UAV dispatch capacity prediction model can be used to predict in advance the number of UAVs that need to be dispatched. Not only can tasks be assigned based on the health status of the UAV batteries, but the smart hangar's charging strategy for UAV batteries can also be adjusted to ensure that there are enough UAVs with sufficient power.

[0087] Example 2

[0088] like Figures 2 to 8 As shown, this implementation provides a drone battery monitoring method, including:

[0089] like Figure 3 As shown, the charging and discharging record data of the drone battery is obtained, and the charging and discharging record data and the number of battery charging and discharging times are fitted to obtain an empirical degradation model, including:

[0090] Obtain the battery charge and discharge record data of the drone, and use the convex optimization algorithm to smooth the charge and discharge record data to obtain the charge and discharge standardized data;

[0091] The empirical degradation model of the UAV battery is constructed based on the standardized charge and discharge data and the corresponding charge and discharge times, and the parameter weights of the empirical degradation model are identified by the least squares method; the empirical degradation model is expressed as follows:

[0092] ;

[0093] In the formula, It is represented as the battery health assessment fitting value in the cth charge and discharge cycle; It is expressed as the rate of charging and discharging of the drone battery in the cth charging and discharging cycle; It represents the initial capacity of the battery of the drone to be tested; , , and Represents setting parameter weights.

[0094] Inputting the number of charge and discharge times to be predicted into the empirical degradation model to obtain a battery health assessment fitting value; and calculating a battery life prediction value based on the battery health assessment fitting value;

[0095] Get the real value of the drone battery capacity, and calculate the real value of the battery life based on the real value of the drone battery capacity;

[0096] The remaining life prediction error is calculated based on the predicted battery life value and the actual battery life value. The expression formula is:

[0097] ;

[0098] In the formula, Represents the predicted value of battery life; Indicates the actual value of battery life; It is expressed as the remaining life prediction error;

[0099] The regularization parameter in the convex optimization algorithm is adjusted based on the remaining life prediction error, and the construction process of the empirical degradation model is iterated repeatedly until the remaining life prediction error converges, and the final empirical degradation model is output.

[0100] like Figure 5 As shown in the figure, the comparison results of the predicted value of the empirical degradation model, the convex optimization denoising result and the actual value of the battery health; the empirical degradation model uses historical data to predict the initial value of the battery health prediction, and adopts the convex optimization algorithm to denoise the battery aging capacity data to obtain smoother data that is closer to the actual situation, thereby improving the accuracy of the prediction.

[0101] When the regularization parameter is set too large or too small, the optimization degree will be insufficient or over-optimized. After repeated experiments, when the regularization parameter is set to 100, the optimization degree is moderate, the actual data fluctuates slightly around the convex optimization data, and the average error is 0.0103%.

[0102] The empirical degradation model is compared with the double exponential model and the power function model, and the prediction results are shown in Table 1.

[0103] Table 1. Comparison of prediction model performance

[0104]

[0105] Compared with the double exponential model and the power function model, the prediction method based on the empirical degradation model has higher prediction accuracy. In addition, the existing prediction models are often limited to one set of test data. For different sets of data, each set of data needs to be retrained and parameters identified before prediction can be made. There is no universal model. The empirical degradation model is suitable for battery data with similar standard working conditions. It only needs to collect the initial value of the battery capacity to directly predict the life distribution of the battery, which has good universality.

[0106] According to the given battery charge and discharge times, the corresponding battery health prediction initial value is read from the empirical degradation model;

[0107] Read the actual value of the battery health assessment corresponding to the given battery charge and discharge times, and calculate the battery health prediction error value based on the initial value of the battery health prediction and the corresponding actual value of the battery health assessment. The expression is as follows:

[0108] ;

[0109] In the formula, It is represented as the battery health prediction error value of the mth UAV; Represented as the true value of the battery health assessment of the mth drone; Represented as the initial value of the battery health prediction of the m-th drone.

[0110] like Figure 4 As shown in the figure, an error compensation model is constructed based on the BP neural network; the historical operating condition information of the drone battery is input into the error compensation model to obtain the battery health assessment training error value; the MAPE training loss value is calculated based on the battery health assessment training error value and the battery health assessment prediction error value, and the expression formula is:

[0111] ;

[0112] In the formula, n is the total number of charge and discharge cycles of the drone battery; m is the drone serial number; c is the charge and discharge cycle serial number of the drone battery; It is represented as the battery health prediction error value in the cth charge and discharge cycle of the drone battery; It is expressed as the battery health training error value in the cth charge and discharge cycle of the drone battery. Expressed as MAPE training loss value.

[0113] The parameters of the error compensation model are optimized based on the training loss value, and the iteration is repeated until the training loss value converges, and the trained error compensation model is output.

[0114] like Figure 6 As shown in the figure, compared with the empirical degradation model, the battery health estimation model has higher accuracy and stability. A single empirical degradation model cannot track the impact of the operating environment on battery degradation, nor can it fully describe the degradation characteristics of different groups of batteries. The battery health estimation model comprehensively considers the impact of charging and discharging conditions on battery SOH degradation, and is suitable for different groups of lithium-ion battery systems under actual operating conditions. The model has good universality.

[0115] The process of building a drone scheduling capacity prediction model includes:

[0116] Based on the Spark framework, a real-time scheduling capacity prediction unit is built, such as Figure 8As shown, the real-time scheduling capacity prediction unit includes a Spilt segmentation layer, a Map function and a ReduceByKey function in sequence; a day-ahead scheduling capacity prediction unit is constructed based on a parallel decision tree algorithm; as shown Figure 7 As shown, the UAV scheduling capacity prediction model is constructed by the real-time scheduling capacity prediction unit, the distributed file unit and the day-ahead scheduling capacity prediction unit;

[0117] The historical operation data of the power system is obtained to construct a resilient distributed data set; the resilient distributed data set is input into the UAV dispatch capacity prediction model, and a dispatch capacity training prediction value is output; the MSE training loss value is calculated based on the dispatch capacity training prediction value and the dispatch capacity true value, and the parameters of the UAV dispatch capacity prediction model are optimized according to the MSE training loss value, and the iteration is repeated until the MSE training loss value converges to output the trained UAV dispatch capacity prediction model.

[0118] Obtaining the real-time charge and discharge times of the drone battery and the real-time working condition information of the drone battery; the real-time working condition information of the drone battery includes the actual capacity, initial capacity, total cumulative working time and charge and discharge efficiency of the drone battery; inputting the real-time charge and discharge times of the battery into the empirical degradation model to obtain the initial value of battery health assessment (referred to as SOH assessment initial value);

[0119] The real-time operating condition information of the drone battery is input into the trained error compensation model to obtain the battery health error assessment value (abbreviated as Error evaluation value); superimpose the battery health error evaluation value and the battery health evaluation initial value to obtain the battery health evaluation value;

[0120] Acquire the real-time operation data of the power system, which includes the power transmission path, real-time environmental data and power system fault data; the power transmission path is used to confirm the flight mission duration and mission flight path of the UAV.

[0121] Input the real-time operation data of the power system into the pre-built UAV dispatch capacity prediction model to obtain the UAV dispatch capacity prediction value, including:

[0122] Input the real-time operation data of the power system into the pre-built drone dispatch capacity prediction model,

[0123] The real-time scheduling capacity prediction unit outputs the UAV temporary scheduling capacity prediction value and stores it in the distributed file unit; the distributed file unit obtains the UAV temporary scheduling capacity prediction value for each time period and sends it to the day-ahead scheduling capacity prediction unit to obtain the UAV future scheduling capacity prediction value, and the UAV temporary scheduling capacity prediction value and the UAV future scheduling capacity prediction value are weightedly summed to obtain the UAV scheduling capacity prediction value.

[0124] The drone dispatching strategy is generated based on the battery health assessment value and the drone dispatching capacity prediction value. The process includes:

[0125] The UAV scheduling strategy includes UAV task scheduling instructions and UAV maintenance scheduling instructions;

[0126] When the battery health assessment value is greater than the predicted value of the UAV dispatch capacity, the preset number of the UAV is read to obtain the set of reserve UAVs; the numbers of G UAVs in the set of reserve UAVs are retrieved to generate UAV task dispatch instructions; G represents the number of UAVs that perform the task;

[0127] When the battery health assessment value is less than the battery maintenance threshold, the drone number is retrieved and a drone maintenance scheduling instruction is generated based on the read drone number. This embodiment assigns tasks by considering the health status of the drone battery, thereby ensuring that the battery status matches the task undertaken, thereby avoiding mission failure or drone damage due to battery health problems.

[0128] Example 3

[0129] This embodiment provides a drone battery monitoring system for a smart hangar. The drone battery monitoring system described in this embodiment can apply the drone battery monitoring method described in Embodiment 1 and Embodiment 2. The drone battery monitoring system includes:

[0130] The acquisition module is used to obtain the real-time charge and discharge times of the drone battery and the real-time working condition information of the drone battery;

[0131] The battery evaluation module is used to input the real-time charge and discharge times of the battery into a preset empirical degradation model to obtain an initial value of the battery health evaluation; input the real-time working condition information of the drone battery into the trained error compensation model to obtain a battery health error evaluation value; and superimpose the battery health error evaluation value with the initial value of the battery health evaluation to obtain a battery health evaluation value;

[0132] A prediction module is used to obtain real-time operation data of the power system, and input the real-time operation data of the power system into a pre-built UAV dispatch capacity prediction model to obtain a UAV dispatch capacity prediction value;

[0133] The scheduling module is used to generate a UAV scheduling strategy based on the battery health assessment value and the UAV scheduling capacity prediction value.

[0134] The drone battery monitoring system can monitor the battery status in a timely manner, improve the efficiency and response speed of battery management; combined with the empirical degradation model and error compensation model, the system can more accurately evaluate the battery health status; then based on the battery health assessment value and the drone scheduling capacity prediction value, a more scientific and reasonable drone scheduling strategy is generated to improve the success rate and efficiency of mission execution and reduce flight safety risks caused by battery problems.

[0135] Example 4

[0136] In a third aspect, the present invention provides an electronic device, including a storage medium and a processor; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the drone battery monitoring method for a smart hangar described in Example 1 and Example 2.

[0137] By taking the health status of the drone battery into consideration when assigning tasks, electronic devices can ensure that the battery status matches the mission being undertaken, thus avoiding mission failure or drone damage due to battery health issues. At the same time, the storage medium and processor are integrated into one device, simplifying the system architecture and improving the reliability and stability of the system.

[0138] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, etc.) containing computer-usable program code. M , optical storage, etc.).

[0139] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0140] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0141] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0142] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for monitoring a drone battery, characterized in that: include: Obtain the real-time charge and discharge times of the drone battery and the real-time working condition information of the drone battery; Input the real-time charge and discharge times of the battery into the pre-built empirical degradation model to obtain the initial value of the battery health assessment; The real-time operating condition information of the drone battery is input into the trained error compensation model to obtain the battery health error assessment value; The battery health error evaluation value is superimposed on the battery health evaluation initial value to obtain the battery health evaluation value; Obtain the real-time operation data of the power system and input the real-time operation data of the power system into the pre-built UAV dispatch capacity prediction model to obtain the UAV dispatch capacity prediction value, including: The UAV scheduling capacity prediction model includes a real-time scheduling capacity prediction unit, a distributed file unit and a day-ahead scheduling capacity prediction unit; the construction of the UAV scheduling capacity prediction model includes: A real-time scheduling capacity prediction unit is constructed based on the spark framework, wherein the real-time scheduling capacity prediction unit includes a Spilt segmentation layer, a Map function, and a ReduceByKey function in sequence; a day-ahead scheduling capacity prediction unit is constructed based on a parallel decision tree algorithm; Obtain historical operation data of the power system to build a resilient distributed data set; Inputting the elastic distributed data set into the UAV scheduling capacity prediction model and outputting the scheduling capacity training prediction value; The MSE training loss value is calculated based on the scheduling capacity training prediction value and the scheduling capacity true value, and the parameters of the UAV scheduling capacity prediction model are optimized according to the MSE training loss value. The iteration is repeated until the MSE training loss value converges to output the trained UAV scheduling capacity prediction model. The real-time operation data of the power system is input into the pre-built UAV dispatch capacity prediction model, and the real-time dispatch capacity prediction unit outputs the UAV temporary dispatch capacity prediction value according to the real-time operation data of the power system and stores it in the distributed file unit; The day-ahead scheduling capacity prediction unit reads the temporary scheduling capacity prediction value of the UAV from the distributed file unit, predicts the future scheduling capacity prediction value of the UAV according to the temporary scheduling capacity prediction value of the UAV, and performs weighted summation of the temporary scheduling capacity prediction value of the UAV and the future scheduling capacity prediction value of the UAV to obtain the scheduling capacity prediction value of the UAV; Generate a drone scheduling strategy based on the battery health assessment value and the drone scheduling capacity prediction value.

2. The method for monitoring the battery of a drone according to claim 1, characterized in that: The construction process of the empirical degradation model includes: Obtain the battery charge and discharge record data of the drone, and use the convex optimization algorithm to smooth the charge and discharge record data to obtain the charge and discharge standardized data; The empirical degradation model of the UAV battery is constructed based on the standardized charge and discharge data and the corresponding charge and discharge times, and the parameter weights of the empirical degradation model are identified using the least squares method. Inputting the number of charge and discharge times to be predicted into the empirical degradation model to obtain a battery health assessment fitting value; and calculating a battery life prediction value based on the battery health assessment fitting value; Get the real value of the drone battery capacity, and calculate the real value of the battery life based on the real value of the drone battery capacity; The remaining life prediction error is calculated based on the predicted battery life value and the actual battery life value. The expression formula is: ； In the formula, Represents the predicted value of battery life; Indicates the actual value of battery life; It is expressed as the remaining life prediction error; The regularization parameter in the convex optimization algorithm is adjusted based on the remaining life prediction error, and the construction process of the empirical degradation model is iterated repeatedly until the remaining life prediction error converges, and the final empirical degradation model is output.

3. The method for monitoring the battery of a drone according to claim 2, characterized in that: The training method of the error compensation model includes: According to the given battery charge and discharge times, the corresponding battery health prediction initial value is read from the empirical degradation model; Read the actual value of the battery health assessment corresponding to the given battery charge and discharge times, and calculate the battery health prediction error value based on the initial value of the battery health prediction and the corresponding actual value of the battery health assessment; Construct error compensation model based on BP neural network; The historical operating condition information of the drone battery is input into the error compensation model to obtain the battery health assessment training error value; the MAPE training loss value is calculated based on the battery health assessment training error value and the battery health prediction error value; The parameters of the error compensation model are optimized based on the MAPE training loss value, and iterations are repeated until the MAPE training loss value converges, and the trained error compensation model is output.

4. The method for monitoring the battery of a drone according to claim 2, characterized in that: The expression of the empirical degradation model is as follows: ； In the formula, It is represented as the battery health assessment fitting value in the cth charge and discharge cycle; It is expressed as the rate of charging and discharging of the drone battery in the cth charging and discharging cycle; It represents the initial capacity of the battery of the drone to be tested; , , and Represents setting parameter weights.

5. The drone battery monitoring method according to claim 3, characterized in that: The expression for calculating the battery health prediction error value based on the initial value of the battery health prediction and the corresponding true value of the battery health assessment is as follows: ； In the formula, It is represented as the battery health prediction error value of the mth UAV; Represented as the true value of the battery health assessment of the mth drone; Represented as the initial value of the battery health prediction of the m-th drone.

6. The method for monitoring the battery of a drone according to claim 3, characterized in that: The MAPE training loss value is calculated based on the battery health assessment training error value and the battery health assessment prediction error value. The expression formula is: ； In the formula, n is the total number of charge and discharge cycles of the drone battery; m is the drone serial number; c is the charge and discharge cycle serial number of the drone battery; It is represented as the battery health prediction error value in the cth charge and discharge cycle of the drone battery; It is expressed as the battery health training error value in the cth charge and discharge cycle of the drone battery. Expressed as MAPE training loss value.

7. The method for monitoring the battery of a drone according to claim 1, characterized in that: The UAV scheduling strategy includes UAV task scheduling instructions and UAV maintenance scheduling instructions; Generate a drone scheduling strategy based on the battery health assessment value and the drone scheduling capacity prediction value, including: when the battery health assessment value is greater than the drone scheduling capacity prediction value, read the corresponding drone number to obtain a reserve drone set; retrieve the drone number from the reserve drone set to generate a drone task scheduling instruction; When the battery health assessment value is less than the battery maintenance threshold, the corresponding drone number is retrieved, and a drone maintenance scheduling instruction is generated based on the retrieved drone number.

8. A drone battery monitoring system, characterized in that: include: The acquisition module is used to obtain the real-time charge and discharge times of the drone battery and the real-time working condition information of the drone battery; The battery evaluation module is used to input the real-time charge and discharge times of the battery into a preset empirical degradation model to obtain the initial value of the battery health evaluation; input the real-time working condition information of the drone battery into the trained error compensation model to obtain the battery health error evaluation value; The battery health error evaluation value is superimposed on the battery health evaluation initial value to obtain the battery health evaluation value; A prediction module is used to obtain real-time operation data of the power system, and input the real-time operation data of the power system into a pre-built UAV dispatch capacity prediction model to obtain a UAV dispatch capacity prediction value; A scheduling module, used to generate a UAV scheduling strategy based on the battery health assessment value and the UAV scheduling capacity prediction value; The prediction module inputs the real-time operation data of the power system into the pre-built UAV dispatch capacity prediction model to obtain the UAV dispatch capacity prediction value, including: The UAV scheduling capacity prediction model includes a real-time scheduling capacity prediction unit, a distributed file unit and a day-ahead scheduling capacity prediction unit; the construction of the UAV scheduling capacity prediction model includes: A real-time scheduling capacity prediction unit is constructed based on the spark framework, wherein the real-time scheduling capacity prediction unit includes a Spilt segmentation layer, a Map function, and a ReduceByKey function in sequence; a day-ahead scheduling capacity prediction unit is constructed based on a parallel decision tree algorithm; Obtain historical operation data of the power system to build a resilient distributed data set; Inputting the elastic distributed data set into the UAV scheduling capacity prediction model and outputting the scheduling capacity training prediction value; The MSE training loss value is calculated based on the scheduling capacity training prediction value and the scheduling capacity true value, and the parameters of the UAV scheduling capacity prediction model are optimized according to the MSE training loss value. The iteration is repeated until the MSE training loss value converges to output the trained UAV scheduling capacity prediction model. The real-time operation data of the power system is input into the pre-built UAV dispatch capacity prediction model, and the real-time dispatch capacity prediction unit outputs the UAV temporary dispatch capacity prediction value according to the real-time operation data of the power system and stores it in the distributed file unit; The day-ahead scheduling capacity prediction unit reads the temporary scheduling capacity prediction value of the UAV from the distributed file unit, predicts the future scheduling capacity prediction value of the UAV based on the temporary scheduling capacity prediction value of the UAV, and performs weighted summation of the temporary scheduling capacity prediction value of the UAV and the future scheduling capacity prediction value of the UAV to obtain the UAV scheduling capacity prediction value.

9. An electronic device, comprising a storage medium and a processor; the storage medium is used to store instructions; characterized in that: The processor is used to operate according to the instructions to execute the drone battery monitoring method according to any one of claims 1 to 7.

Images