Internet of Things energy management method, system and device and medium

Abstract

The invention discloses an Internet of Things energy management method, system and device and a medium. The method comprises the steps of obtaining initial energy management data of all nodes in a target Internet of Things; performing exception processing on the initial energy management data of each node to obtain first energy management data of the corresponding node; performing denoising processing on the first energy management data of each node to obtain second energy management data of the corresponding node; calculating energy management characteristics of the target Internet of Things according to the second energy management data of each node; constructing a global objective function of the target Internet of Things according to the energy management characteristics of the target Internet of Things; and updating the initial node strategy of each node based on the global objective function to obtain the optimal energy management strategy of each node, thereby realizing real-time dynamic coordination of the node local energy strategy and the global network energy efficiency, and improving the energy management optimization efficiency.

Description

Technical Field

[0001] The present application relates to the technical field of Internet of Things energy management, and in particular, to an Internet of Things energy management method, system, device and medium. Background Art

[0002] With the deep penetration of Internet of Things technology into fields such as industrial monitoring, smart cities and smart homes, a large number of Internet of Things devices are widely distributed and unattended. Since most of these devices are deployed in areas without power grid coverage (such as production line sensors, monitoring nodes in remote areas), their energy supply highly depends on batteries and renewable energy such as solar / wind energy, and such energy is significantly affected by weather and environmental factors, having defects such as unstable supply and strong volatility.

[0003] Current Internet of Things energy management technologies adopt fixed optimization strategies and lack the ability to dynamically adapt to environmental changes, resulting in poor adaptability of algorithms in dynamic energy environments, poor adaptability to sudden environmental changes, difficulty in timely adjusting energy scheduling strategies, and lack of cooperation among nodes, resulting in one-sided optimization results, difficulty in achieving true global optimization, prone to coexistence of local energy surplus and shortage, and lack of integration of multi-objective constraints such as energy consumption cost, energy storage risk and energy volatility, resulting in low optimization efficiency. Summary of the Invention

[0004] The following is an overview of the subject matter described in detail in this article. This overview is not intended to limit the scope of protection of the claims.

[0005] The main purpose of the embodiments of the present disclosure is to propose an Internet of Things energy management method, system, device and storage medium, which can realize real-time dynamic coordination of node local energy strategies and global network energy efficiency, thereby improving the energy management optimization efficiency.

[0006] The first aspect of the embodiments of the present application provides an Internet of Things energy management method for a central controller, and the method includes: Obtain the initial energy management data of all nodes in the target Internet of Things; Perform anomaly processing on the initial energy management data of each node to obtain the first energy management data of the corresponding node; Perform denoising processing on the first energy management data of each node to obtain the second energy management data of the corresponding node; Calculate the energy management characteristics of the target Internet of Things according to the second energy management data of each node; Construct a global objective function of the target Internet of Things according to the energy management characteristics of the target Internet of Things; Based on the global objective function, update the initial node strategy of each node to obtain the optimal energy management strategy of each node.

[0007] In some embodiments of the present application, the abnormal processing of the initial energy management data of each node to obtain the first energy management data of the corresponding node includes: Calculating the median according to the initial energy management data of all nodes; Calculating the median absolute deviation according to the median; Based on the median and the median absolute deviation, performing abnormal judgment on the initial energy management data of each node, and correcting the abnormal values according to the judgment results to obtain the first energy management data of the corresponding node.

[0008] In some embodiments of the present application, the denoising processing of the first energy management data of each node to obtain the second energy management data of the corresponding node includes: Performing multi-scale decomposition on the first energy management data of each node to obtain corresponding multi-scale coefficients; Using a soft threshold function to perform noise reduction processing on each multi-scale coefficient to obtain the denoised multi-scale coefficient; Performing inverse transformation on the denoised multi-scale coefficient to obtain the second energy management data of the corresponding node.

[0009] In some embodiments of the present application, the calculation of the energy management characteristics of the target Internet of Things according to the second energy management data of each node includes: Extracting the key features of each node from the second energy management data of each node; Fusing the key features of all nodes to obtain the energy management characteristics of the target Internet of Things.

[0010] In some embodiments of the present application, the updating of the initial node strategy of each node based on the global objective function to obtain the optimal energy management strategy of each node includes: Calculating the initial node strategy and the local objective function of the corresponding node according to the initial energy management data of each node and the global objective function; Calculating the dual variable of each node according to the local objective function of each node; Updating the initial node strategy of the corresponding node according to the dual variable of each node to obtain the optimal energy management strategy of each node.

[0011] In some embodiments of the present application, the updating of the initial node strategy of the corresponding node according to the dual variable of each node to obtain the optimal energy management strategy of each node includes: Calculating the subgradient of each node according to the local objective function of each node; Update the dual variable of each corresponding node according to the sub-gradient of each said node; Use the consensus mechanism to iteratively update the initial node strategy of each said node to obtain the optimal energy management strategy of each said node.

[0012] In some embodiments of the present application, the calculation formula for calculating the median based on the initial energy management data of all nodes includes:

[0013] where, is the median at time of the initial energy management data of all nodes, is the node at time of the initial energy management data; The calculation formula for calculating the median absolute deviation based on the median includes:

[0014] where, is the median absolute deviation at time of all nodes, is the node at time of the initial energy management data.

[0015] To achieve the above object, a second aspect of the embodiments of the present invention provides an Internet of Things energy management system, and the system includes: An acquisition module, configured to acquire the initial energy management data of all nodes in the target Internet of Things; A correction module, configured to perform anomaly processing on the initial energy management data of each said node to obtain the first energy management data of the corresponding node; A noise reduction module, configured to perform denoising processing on the first energy management data of each said node to obtain the second energy management data of the corresponding node; A calculation module, configured to calculate the energy management characteristics of the target Internet of Things according to the second energy management data of each said node; A construction module, configured to construct a global objective function of the target Internet of Things according to the energy management characteristics of the target Internet of Things; An update module, configured to update the initial node strategy of each said node based on the global objective function to obtain the optimal energy management strategy of each said node.

[0016] To achieve the above object, a third aspect of the embodiments of the present invention provides an electronic device, including: at least one control processor and a memory communicatively connected to the at least one control processor; the memory stores instructions executable by the at least one control processor, and when the instructions are executed by the at least one control processor, the at least one control processor is enabled to execute the above-mentioned Internet of Things energy management method.

[0017] To achieve the above object, a fourth aspect of the embodiments of the present invention provides a computer-readable storage medium storing computer-executable instructions for causing a computer to execute the above-mentioned Internet of Things energy management method.

[0018] An embodiment of the present application provides an Internet of Things energy management method, which includes: obtaining initial energy management data of all nodes in a target Internet of Things; performing anomaly processing on the initial energy management data of each node to obtain first energy management data of the corresponding node; performing denoising processing on the first energy management data of each node to obtain second energy management data of the corresponding node; calculating energy management characteristics of the target Internet of Things according to the second energy management data of each node; constructing a global objective function of the target Internet of Things according to the energy management characteristics of the target Internet of Things; and based on the global objective function, updating the initial node strategy of each node to obtain an optimal energy management strategy for each node, which can realize real-time dynamic coordination between the local energy strategy of the node and the global network energy efficiency, thereby improving the efficiency of energy management optimization.

[0019] It can be understood that the beneficial effects of the above-mentioned second aspect to the fourth aspect compared with the related art are the same as those of the first aspect compared with the related art. For the relevant descriptions, please refer to the relevant descriptions in the first aspect, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The above and / or additional aspects and advantages of the present application will become apparent and easier to understand from the following description of the embodiments in conjunction with the accompanying drawings, where: Figure 1 is a flowchart of an Internet of Things energy management method provided by an embodiment of the present application; Figure 2 is a schematic structural diagram of an Internet of Things energy management training system provided by an embodiment of the present application; Figure 3 is a schematic hardware structure diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] Embodiments of the present application will be described in detail below. Examples of the embodiments are shown in the accompanying drawings, where like or similar reference numerals denote like or similar elements or elements having like or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application and should not be construed as a limitation of the present application.

[0022] In the description of the present application, if the first, second, etc. are described only for the purpose of distinguishing technical features, they should not be construed as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features or implicitly indicating the sequence of the indicated technical features.

[0023] In the description of the present application, it should be understood that for the orientation description, such as up, down, etc., the orientation or positional relationship indicated is based on the orientation or positional relationship shown in the accompanying drawings. It is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present application.

[0024] In the description of the present application, it should be noted that unless otherwise clearly defined, words such as setting, installation, connection, etc. should be understood in a broad sense, and those skilled in the art can reasonably determine the specific meanings of the above words in the present application in combination with the specific content of the technical solution.

[0025] With the rapid development of the Internet of Things technology, more and more devices are widely used in various fields such as industrial production, smart home, smart city, and transportation. These Internet of Things devices are huge in number, widely distributed in different geographical locations, and usually in an unattended state. Since most of the Internet of Things nodes are deployed in areas far from stable power supply, the energy supply problem has gradually become an important bottleneck restricting the stable operation of the Internet of Things system. For example, in the field of industrial monitoring, a large number of sensors and actuators are installed at various positions on the production line to monitor the device status and optimize production scheduling. These nodes often cannot directly obtain stable power supply from the power grid and can only rely on batteries or renewable energy sources such as solar energy and wind energy for power supply. However, these energy sources have a high degree of uncertainty and volatility. For example, solar energy is affected by weather, and wind energy is difficult to supply stably with the change of environmental wind speed. At present, the existing energy management technologies perform poorly in dealing with the volatility and uncertainty of renewable energy sources (such as solar energy and wind energy). When the environmental conditions change rapidly, such as the sharp fluctuation of solar power generation due to weather changes and the unstable supply of wind energy affected by drastic changes in wind speed, the existing energy management technologies usually have difficulty in responding quickly, resulting in decision-making lag and being unable to adapt to the changes in energy supply in real time. This lag not only reduces the energy utilization efficiency but also may lead to energy waste or device operation interruption, seriously affecting the system stability.

[0026] In addition, existing distributed energy optimization methods often handle several key factors such as node energy consumption cost, energy storage risk, and energy supply uncertainty in isolation, lacking an effective coordination mechanism. This isolated approach ignores the internal connections and interactions among various factors, which may lead to one-sided optimization results and make it difficult to achieve true global optimization. For example, when a device only focuses on reducing short-term energy consumption costs, it may neglect the energy storage capacity risk, resulting in an increased risk of energy depletion and instead increasing the overall operating cost.

[0027] Moreover, the current distributed optimization methods have a high computational complexity. Especially in the Internet of Things network environment with a large number of nodes and a wide distribution range, the system has huge computational overhead and is difficult to meet the real-time requirements. In actual operation, due to the limitations of computing and communication resources, the goal of real-time optimization is difficult to achieve, and the timeliness and accuracy of decision-making are greatly reduced, restricting the effective application of distributed optimization methods in actual scenarios.

[0028] At the same time, traditional distributed optimization technologies usually adopt fixed optimization strategies and lack the ability to dynamically adapt to environmental changes, resulting in poor adaptability of the algorithm in a dynamic energy environment and being unable to flexibly respond to real-time changing operating states. In addition, when existing methods are designed and implemented, they usually lack an effective robust optimization mechanism and do not fully consider abnormal data and interference factors in the actual operating environment. Once data anomalies or external interferences occur, the algorithm performance may drop significantly or even fail to operate normally, further exacerbating the uncertainty and instability of the energy management system.

[0029] Based on this, the embodiments of the present application provide an Internet of Things energy management method, system, electronic device, and medium, aiming to be able to achieve real-time dynamic coordination between node local energy strategies and global network energy efficiency, thereby improving the energy management optimization efficiency.

[0030] The Internet of Things energy management method, system, electronic device, and medium provided by the embodiments of the present application are specifically described through the following embodiments. First, the Internet of Things energy management method in the embodiments of the present application is described.

[0031] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is a theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results.

[0032] The basic technologies of artificial intelligence generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation / interaction systems, and mechatronics. The software technologies of artificial intelligence mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0033] The Internet of Things energy management method provided by the embodiments of this application relates to the technical field of Internet of Things energy management. The Internet of Things energy management method provided by the embodiments of this application can be applied to terminals, can also be applied to the server side, or can be software running on the terminal or the server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, can also be configured as a server cluster or a distributed system composed of multiple physical servers, or can also be configured as a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the Internet of Things energy management method, etc., but is not limited to the above forms.

[0034] This application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0035] It should be noted that in each specific embodiment of the present application, when it comes to relevant processing based on data related to the user's identity or characteristics, such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first. Moreover, the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiments of the present application need to obtain the user's sensitive personal information, the user's separate permission or separate consent will be obtained through pop-up windows or by jumping to a confirmation page. After clearly obtaining the user's separate permission or separate consent, the necessary user-related data for the normal operation of the embodiments of the present application will be obtained.

[0036] For this reason, referring to Figure 1 , an Internet of Things energy management method is provided in an embodiment of the present application. This method is applied to a central controller. The controller can be a server, an electronic device, a mobile terminal, etc., and no specific limitation is made here. The method includes the following steps S110 to S160: Step S110: Obtain the initial energy management data of all nodes in the target Internet of Things; Step S120: Perform anomaly processing on the initial energy management data of each node to obtain the first energy management data of the corresponding node; Step S130: Perform denoising processing on the first energy management data of each node to obtain the second energy management data of the corresponding node; Step S140: Calculate the energy management characteristics of the target Internet of Things according to the second energy management data of each node; Step S150: Construct a global objective function of the target Internet of Things according to the energy management characteristics of the target Internet of Things; Step S160: Based on the global objective function, update the initial node strategy of each node to obtain the optimal energy management strategy of each node.

[0037] In this step, the initial energy management data refers to the original energy information collected by the node, which can be specifically realized by real-time monitoring of parameters such as the remaining battery power, renewable energy input power, and device energy consumption rate through voltage sensors, current sensors, or power measurement modules, providing a basic input for subsequent processing and solving the optimization deviation problem caused by low data quality in traditional methods.

[0038] Specifically, anomaly processing refers to identifying and correcting energy data that deviates from the normal range through the median and absolute deviation method. Preferably, the median of all node data is statistically calculated using a sliding time window, and the anomaly points are judged in combination with a preset deviation threshold and replaced with the mean or interpolation data at adjacent times, so as to eliminate the noise interference caused by sensor failures or environmental mutations and improve the data reliability.

[0039] Specifically, the denoising process refers to eliminating high-frequency noise in data through signal decomposition techniques. Preferably, wavelet transform and empirical mode decomposition methods are used to perform multi-scale decomposition on energy data. After filtering out the noise coefficients using a soft threshold function, signal reconstruction is carried out to retain the effective components in energy fluctuations and enhance data availability.

[0040] Specifically, the energy management features refer to the key indicators reflecting the overall energy state of the Internet of Things. Preferably, parameters such as the mean residual energy, charge-discharge efficiency, and energy fluctuation variance of each node are extracted through principal component analysis, and then a comprehensive feature vector is generated through weighted fusion to provide a quantitative basis for constructing a global optimization model.

[0041] Specifically, the global objective function refers to a mathematical model that integrates energy utilization rate, energy storage device loss, and energy supply-demand balance. Preferably, a linear weighting method can be used to transform multi-objective constraints into a single-objective optimization problem, and a mathematical expression including an energy consumption cost function, an energy storage risk function, and an energy fluctuation penalty term is defined to achieve a unified quantitative evaluation of multi-dimensional optimization objectives.

[0042] Specifically, the optimal energy management strategy refers to a dynamic decision-making scheme generated through a distributed optimization algorithm. Preferably, the dual decomposition method is combined with a consensus mechanism, and during the iterative process, the charge-discharge strategy, device start-stop plan, and energy distribution ratio are adjusted according to the node energy state to achieve collaborative optimization among nodes and avoid local resource imbalance.

[0043] In this step, first, the initial energy management data of all nodes in the target Internet of Things are obtained, and these data include the energy state information of each node; then, the initial data of each node are processed for anomalies to correct possible outliers to obtain the first energy management data; then, the first energy management data are denoised to filter out interference signals to obtain more accurate second energy management data; furthermore, based on the processed second energy management data, the energy management features of the target Internet of Things are calculated to reflect the energy distribution status of the entire network through these features, and then based on the energy management features, a global objective function of the target Internet of Things is constructed, comprehensively considering the overall energy optimization objectives of the network; finally, based on the global objective function, the initial strategy of each node is updated to obtain the optimal energy management strategy, solving the problem of uneven energy distribution among nodes, thereby achieving the improvement of data quality through multi-step data processing, establishing a global objective function by integrating multi-node features, adopting a distributed optimization mechanism to dynamically generate an energy management strategy, and realizing the adaptive collaborative management of the Internet of Things system in a dynamic energy environment.

[0044] In some embodiments, in step S120, the initial energy management data of each node are processed for anomalies to obtain the first energy management data of the corresponding node, including the following steps S210 to S230: Step S210: Calculate the median based on the initial energy management data of all nodes; Step S220: Calculate the median absolute deviation based on the median; Step S230: Based on the median and the median absolute deviation, perform an outlier judgment on the initial energy management data of each node, and correct the outlier according to the judgment result to obtain the first energy management data of the corresponding node.

[0045] In this embodiment, the median is calculated by sorting the values at each time point and taking the middle value, which is suitable for scenarios where energy data fluctuates dynamically; the median absolute deviation is calculated by finding the median of the absolute distances between each node's data and the median, avoiding the interference of outliers on the deviation estimation. Specifically, in a dynamic energy environment, energy data may exhibit intermittent extreme fluctuations. Taking the initial energy data of each node at a certain moment as an example, sorting the data and taking the middle value as the global median can eliminate the interference of high-energy-consuming nodes or low-power nodes on the data center. By calculating the absolute deviation between each node's data and the median and then taking the median again, the degree of data dispersion can be quantified, and this metric is insensitive to outliers.

[0046] In some embodiments, when a certain node generates abnormally high energy consumption data due to a sudden failure, its deviation from the median will be significantly higher than that of other nodes. However, the calculation process of the median absolute deviation automatically excludes the influence of such extreme deviations. After setting a dynamic threshold based on the median and the median absolute deviation, the data of each node is detected point by point in the time dimension, and the data outside the threshold range is marked as an outlier. During the correction process, the outlier is replaced with the threshold boundary value or a reasonable value is generated through linear interpolation to ensure that the corrected data not only retains the true fluctuation trend but also eliminates the abnormal interference. Thus, in the case of unstable energy supply, the robustness of anomaly detection is effectively improved, providing a high-quality data basis for subsequent denoising processing and policy optimization, and avoiding the interference of abnormal data on the overall energy management decision-making.

[0047] In some embodiments, the calculation formula for calculating the median based on the initial energy management data of all nodes includes:

[0048] where, is the median of the initial energy management data of all nodes at time is the initial energy management data of node at time ; is the initial energy management data at time The calculation formula for calculating the median absolute deviation based on the median includes:

[0049] Among them, is the median absolute deviation of all nodes at time ; is the node at time initial energy management data.

[0050] In this embodiment, the median calculation step eliminates the influence of extreme values on the statistic by taking the median value of the energy data of all nodes at time . The median absolute deviation calculation step constructs a robustness metric for the dynamic data distribution range through the median of the absolute deviations between each node's data and the median. The introduction of the time dimension parameter enables the calculation process to track data fluctuations moment by moment and adapt to the dynamic scenario of unstable energy supply. Specifically, during the execution process, first, the node's initial energy management data at time is sorted by value and the median value is extracted to generate a global median benchmark . Further, all node data is traversed, the absolute deviation between each data point and the benchmark value is calculated, and the median of the new data set is extracted again as , so as to maintain the stability of the statistic when there are intermittent mutations or regional anomalies in the energy data, and avoid the problem that the overall judgment threshold is distorted due to a single-point anomaly in the traditional mean-standard deviation method. Moreover, through the dynamically adjusted and , the population distribution characteristics of node data can be sensed in real time, providing an accurate benchmark reference for subsequent outlier correction, ensuring that the data quality before denoising meets the multi-objective optimization requirements, and helping to improve the accuracy and robustness of anomaly detection.

[0051] In some embodiments, in step S130, denoising processing is performed on the first energy management data of each node to obtain the second energy management data of the corresponding node, including the following steps S310 to S330: Step S310: Perform multi-scale decomposition on the first energy management data of each node to obtain corresponding multi-scale coefficients; Step S320: Use a soft threshold function to perform noise reduction processing on each multi-scale coefficient to obtain the denoised multi-scale coefficients; Step S330: Perform inverse transformation on the denoised multi-scale coefficients to obtain the second energy management data of the corresponding node.

[0052] In this embodiment, preferably, multi-scale decomposition decomposes the data into coefficients in different frequency ranges through wavelet transform; the soft threshold function sets a dynamic threshold, which zeros the coefficient when its absolute value is lower than the threshold and reduces it by the difference when it is higher than the threshold, and the threshold is set to 1.5 times the standard deviation of the noise in each layer; the inverse transform reconstructs the signal using the wavelet basis function corresponding to the decomposition process and retains the coefficients in the energy concentration region.

[0053] Specifically, the data is decomposed into three scales: high frequency, medium frequency, and low frequency. The high-frequency coefficients correspond to short-term energy mutation noise, and the low-frequency coefficients reflect long-term energy trend fluctuations. The high-frequency coefficients are processed by soft thresholding, and the coefficients with absolute values greater than 0.5 are retained to eliminate impulse noise. At the same time, the low-frequency coefficients are compressed by 10% in amplitude to suppress the baseline drift caused by environmental factors. During the reconstruction process, the denoised coefficients are superimposed according to the scale weights. For example, the weight of the high-frequency layer is set to 0.3, and the weight of the low-frequency layer is set to 0.7, balancing the requirements of noise suppression and feature retention, so that the energy data can achieve adaptive noise reduction in the time-frequency domain, eliminate abnormal fluctuations caused by weather mutations, and retain the effective energy consumption characteristics generated by equipment state switching. Thus, the noise interference in the first energy management data can be effectively removed, the reliability and accuracy of the data can be improved, and further the performance and efficiency of the entire Internet of Things energy management system can be improved.

[0054] In some embodiments, in step S140, the energy management features of the target Internet of Things are calculated according to the second energy management data of each node, including the following steps S410 to step S420: Step S410: Extract the key features of each node from the second energy management data of each node; Step S420: Fuse the key features of all nodes to obtain the energy management features of the target Internet of Things.

[0055] In this embodiment, the key feature extraction is preferably implemented by using time-domain statistics, frequency-domain energy distribution, or principal component analysis, and the feature fusion is preferably implemented by using weighted average, linear superposition, or neural network embedding. Among them, the key feature extraction step includes a sliding window mechanism, and a dynamic weight allocation mechanism is introduced in the feature fusion process. The weight coefficient is dynamically adjusted by the ratio of the node energy volatility to the energy storage capacity to match the node state changes in real time. Specifically, after completing multi-scale decomposition and noise reduction processing, key features are extracted from the second energy management data. Preferably, the second energy management data is input into the feature extraction module, and continuous time series are intercepted through a sliding window, and the mean, variance, and kurtosis indexes within each window are calculated to form a time-domain feature vector.

[0056] Further perform a fast Fourier transform on the window data, extract the amplitudes of the first three main frequency components as frequency domain features, and perform dimensionality reduction on the time-frequency mixed features through principal component analysis, retaining the first three principal components with a cumulative contribution rate reaching 90%. Then, input the principal component vectors of all nodes into the fusion layer, calculate the fusion weights according to the proportion of the remaining battery capacity of each node. The weights of nodes with a capacity higher than 60% are increased by 20%, and the weights of nodes with a capacity lower than 30% are decreased by 15%. Finally, generate a global feature vector containing the energy distribution differences and dynamic correlation features between nodes through weighted summation. The dimension of this vector is compressed to 12%-18% of the original data, significantly reducing the computational complexity of subsequent objective function construction, while retaining the spatial correlation of the cross-node energy state, providing a feature input with high information density for global optimization, thereby realizing the effective extraction and fusion of the energy management features of the Internet of Things system, comprehensively reflecting the energy status of the entire Internet of Things system, obtaining more global and comprehensive energy management features, and avoiding the one-sidedness that may be caused by relying only on the data of a single node.

[0057] In some embodiments, in step S160, based on the global objective function, update the initial node strategy of each node to obtain the optimal energy management strategy of each node, including the following steps S510 to step S530: Step S510: Calculate the initial node strategy and the local objective function of the corresponding node according to the initial energy management data of each node and the global objective function; Step S520: Calculate the dual variable of each node according to the local objective function of each node; Step S530: Update the initial node strategy of the corresponding node according to the dual variable of each node to obtain the optimal energy management strategy of each node.

[0058] In this embodiment, the calculation of the initial node strategy needs to simultaneously fuse the real-time parameters of the global objective function and the node's initial energy management data. Among them, the global objective function includes the weight factors of energy cost, energy storage risk, and volatility constraints, and the initial energy management data includes historical energy consumption, current energy storage capacity, and energy supply prediction values; the dual variable is generated by the Lagrange multiplier method or the alternating direction multiplier method, and is used to characterize the constraint relationship between the node strategy and the global objective; the subgradient is calculated according to the first-order derivative or subgradient of the local objective function with respect to the node strategy, and its update step size is negatively correlated with the energy fluctuation amplitude; the consensus mechanism adopts a distributed optimization algorithm to synchronously exchange the dual variables of adjacent nodes in each iteration.

[0059] Specifically, when calculating the initial node strategy, input the historical energy consumption data and the current energy storage capacity of the node into the local objective function. Among them, the historical energy consumption data is processed by a sliding window mechanism, and at the same time, the multi-objective constraints in the global objective function are converted into penalty terms and added to the calculation, thereby generating an initial strategy that matches the dynamic characteristics of the node.

[0060] Furthermore, in the dual variable update stage, the subgradient direction is dynamically adjusted by comparing the energy utilization rate differences of adjacent node strategies. During the iteration process of the consensus mechanism, each node broadcasts its latest policy parameters at a preset time interval, receives and fuses the policy parameters of the nodes within a preset range, and the fusion weights are calculated according to the similarity of energy supply between nodes. For example, the weights of nodes with a similarity higher than 80% are set to 0.7, and the rest are set to 0.3. Then, after 10 - 15 rounds of consensus update iteration, the strategies of each node converge to a stable state that meets the global optimization goal, realizing the adaptive adjustment of the energy scheduling strategy in a dynamic environment, ensuring the overall performance while taking into account the individual needs of each node. Moreover, by introducing the dual variable and the iterative update mechanism, complex constraint conditions can be effectively handled, the convergence speed and stability of the optimization algorithm can be improved, and the energy utilization efficiency and operation reliability of the system can be enhanced.

[0061] In some embodiments, in step S530, the initial node strategy of each node is updated according to the dual variable of each node to obtain the optimal energy management strategy of each node, including the following steps S610 to S630: Step S610: Calculate the subgradient of each node according to the local objective function of each node; Step S620: Update the dual variable of the corresponding node according to the subgradient of each node; Step S630: Use the consensus mechanism to iteratively update the initial node strategy of each node to obtain the optimal energy management strategy of each node.

[0062] In this embodiment, the subgradient calculation is realized by solving the subdifferential set of the local objective function in the policy space, and each subgradient vector corresponds to a feasible descent direction in the policy space; the dual variable update adopts an iterative formula with a momentum term, and the momentum coefficient is set as a dynamic interval value to balance the convergence speed and stability; the consensus mechanism adopts a distributed average consensus algorithm, and in each round of iteration, the node exchanges policy information with its neighbor nodes and performs a weighted average operation, and the weight matrix satisfies the double stochasticity condition to ensure convergence.

[0063] Specifically, in the subgradient calculation stage, each node determines the subgradient vector based on the left derivative and right derivative of the objective function at the current policy point to point to the direction in which the objective function value decreases fastest. Then, when updating the dual variable, the dual variable value of the previous round of iteration and the current subgradient are linearly combined according to a preset step size coefficient, and the step size coefficient decays as the number of iterations increases to ensure convergence.

[0064] Furthermore, during the execution of the consensus mechanism, each node maintains a local copy of the policy. Under the preset communication topology structure, after a preset number of iterations, the difference rate of the policy copies of each node is converged within the ideal preset value range. Thus, while maintaining the local computational independence of each node, the policy update direction gradually aligns with the global optimal solution, ultimately achieving the Pareto optimal state of the energy management policy, realizing the convergence and stability of the distributed optimization algorithm, avoiding the communication overhead of the centralized algorithm, and improving the scalability and robustness of the algorithm. At the same time, through multiple rounds of iterative optimization, it can adapt to the dynamically changing energy environment and timely adjust the energy management policy.

[0065] In some embodiments, first, an Internet of Things energy system is constructed. Consider an energy network composed of multiple distributed Internet of Things devices deployed at different geographical locations, and the overall system is modeled as a set of nodes ; each node is equipped with local energy storage and renewable energy acquisition capabilities. Time is represented in discrete time periods, denoted as The system needs to perform multi-stage and robust energy optimization management under the constraints of energy supply uncertainty and real-time collaboration.

[0066] Step 1: First, through the computing nodes deployed at the device side or the network edge, the device energy consumption and environmental data (such as temperature, wind speed, sunshine intensity, etc.) are collected in real time, and robust preprocessing operations are performed to identify and eliminate abnormal data during the sensor acquisition process. Combining multi-scale signal processing methods, background noise is effectively removed, and representative and highly credible data information is extracted. Furthermore, various types of data are fused into a unified high-quality input feature, providing accurate and stable support for the subsequent optimization model and providing a solid guarantee for achieving precise energy consumption management.

[0067] Specifically, the device energy consumption and environmental data are collected in real time through the edge nodes, and a technical means combining robust statistics and wavelet transform is used to perform abnormal data elimination, denoising processing, and feature extraction on the original data, finally forming a high-quality and stable input feature vector to provide reliable data support for the subsequent robust optimization model.

[0068] In this embodiment, through the input node energy consumption data sequence the environmental perception data vector the threshold parameter (robustness control), and the wavelet threshold parameter (denoising degree control), the output fused feature vector is obtained: to be used as the input feature vector of the subsequent optimization model.

[0069] Specifically, in the Internet of Things environment, the device energy consumption and environmental data are collected in real time through distributed edge nodes. Suppose there are nodes, and each node is at The energy consumption data at a certain moment is denoted as: ; Meanwhile, the environmental data is represented in vector form and denoted as: ; Among them, represents the measurement value of the th environmental variable (such as solar radiation, wind speed, temperature, load status, etc.), which specifically defines the basic mathematical representation of the energy consumption data and the environmental data, laying a foundation for subsequent data preprocessing and feature extraction.

[0070] Furthermore, a robust statistical method is adopted to conduct preliminary preprocessing on the obtained energy consumption data and environmental data. Taking the energy consumption data as an example: First, calculate the median of the energy consumption data of all nodes at the moment :

[0071] In the formula, is the median of the energy consumption data of all nodes at the moment , is the energy consumption data of node at the moment , thus effectively reflecting the central tendency of the data through the median.

[0072] Furthermore, calculate the median of the absolute deviation:

[0073] In the formula, is the median of the absolute deviation at the moment , is the energy consumption data of node at the moment , thus measuring the robust statistic of the data dispersion degree through the median of the absolute deviation.

[0074] Furthermore, set the threshold factor (usually taken as 3). If it satisfies: ; then is considered an outlier. Furthermore, for the detected outliers, the median or interpolation method is used for correction, and the corrected data is denoted as , so as to effectively eliminate or correct the abnormal data generated by noise or sensor failures and ensure the data quality of subsequent processing.

[0075] Furthermore, the discrete wavelet transform (DWT) is used to perform multi-scale decomposition on the signal to achieve fine denoising. Specifically, for the corrected energy consumption signals of each node perform wavelet decomposition, and its wavelet coefficients are: ; wherein, is the scale index, is the translation index, is the corresponding wavelet basis function, so that the original signal is decomposed into components at different scales through wavelet transform, enabling the noise in the signal to be distinguished from the true signal at different scales.

[0076] Furthermore, for each wavelet coefficient perform noise reduction processing using a soft threshold function: ; wherein, is the threshold parameter, usually determined according to the signal noise level, represents the sign function, so as to effectively weaken the noise component while retaining the important information of the signal.

[0077] Furthermore, after denoising, restore the denoised signal through inverse wavelet transform: ; Thus, the denoised wavelet coefficients are recombined into a time-domain signal through inverse transform to obtain more stable and lower-noise energy consumption data.

[0078] Meanwhile, preprocess the environmental data. The environmental data may also be affected by anomalies and noise. Therefore, perform similar processing on each environmental variable .

[0079] Specifically, for the th environmental variable, calculate its median and median absolute deviation:

[0080] ; Similar to the energy consumption data, reduce the influence of noise and outliers through robust statistical methods.

[0081] Furthermore, judge outliers. If it satisfies: ; then regard the current judgment value as an outlier and correct it to ensure that abnormal noise is removed from the environmental data before subsequent processing.

[0082] Furthermore, for the processed Similarly, wavelet transform and soft threshold denoising are applied to obtain the denoised environmental data, denoted as .

[0083] Furthermore, key features are extracted from each node data and environmental data and fused to form high-quality inputs for subsequent optimization and solution.

[0084] Specifically, for the denoised energy consumption signal of each node key features (such as mean, variance, peak value, etc.) are extracted, denoted as , and the features of each node are fused using the weighted average method: ; where the weight is determined according to the signal-to-noise ratio or variance of the node data, and the calculation formula for the weight is: ; where represents the noise variance of the node data, so as to effectively integrate the data of each node and reduce the impact of the error of a single node on the overall decision-making.

[0085] Specifically, for each denoised environmental variable features are extracted and fused using a similar weighted strategy as: ; where is the weight of the environmental variable, satisfying: ; The environmental feature fusion also uses the weighted average method to integrate multiple environmental data into a comprehensive index, providing environmental background information for subsequent robust optimization.

[0086] Furthermore, the features after fusing the node energy consumption and environmental data are integrated into the final input feature vector: ; Through the vector comprehensively reflects the current energy consumption status of the system and external environmental factors, so as to obtain fusion feature data with high signal-to-noise ratio and stability, providing high-quality and stable input data for the subsequent multi-objective robust optimization model, and providing a solid data foundation for the subsequent construction of an accurate and robust optimization model.

[0087] Step 2: Based on the high-quality input features obtained in Step 1, construct a multi-objective robust optimization model that simultaneously considers energy consumption cost, energy storage risk, and uncertainty in renewable energy supply.

[0088] First, it is determined that the input is: the feature vector (output by the robust feature fusion method), node set , total energy supply constraint , risk weight coefficient , node uncertainty set , and the preliminary local optimal strategy for each node is obtained as follows: , iteration intermediate variable: dual variable (passed into the consensus cooperation algorithm), where is the total electricity consumption budget or total power supply of all nodes at time , and are both parameters preset to limit the variation range of renewable energy supply.

[0089] Specifically, let the energy consumption cost of each node be and the energy storage risk be , where and are the energy consumption decisions and storage strategies of node respectively, and the overall objective function is constructed as: ; where is the risk weight coefficient. At the same time, to consider the uncertainty of renewable energy supply, assume that the renewable energy input of each node falls within the uncertainty set , and the robust constraint for supply-demand balance can be written as: ; where represents the demand of node to ensure that each node can still meet the demand under the most unfavorable conditions.

[0090] Further, Lagrangian dual decomposition is performed. Especially in the Internet of Things system, there are usually global coupling constraints, such as the supply-demand balance of the whole system, denoted as: ; where is the total system demand.

[0091] To reduce the solution complexity of the global problem, the Lagrange multiplier is introduced to relax the coupling constraint and construct the Lagrangian function: ; Through Lagrangian dual decomposition, the global problem can be decomposed into independent local sub-problems for each node, and at the same time, the global information is transmitted through .

[0092] Furthermore, distributed sub-problems are constructed. Using the Lagrangian function, each node can independently solve the following local sub-problems: ; while satisfying their respective local robustness constraints: ; Thus, through the distributed solution method, each node can perform independent optimization based on local information, reducing the computational complexity of the overall problem.

[0093] Furthermore, for distributed sub-gradient updates, after each node independently solves the local sub-problem, it is necessary to update the global dual variable to coordinate the global supply-demand balance. The distributed sub-gradient algorithm is adopted, and its update rule is: ; where is the step size at the -th iteration, is the decision result of node at the -th iteration, to gradually adjust according to the global supply-demand deviation, ensuring that the system tends to the global optimal solution.

[0094] Furthermore, for the dual linearization of the robustness constraint, first for the constraint containing uncertainty: ; assuming that the uncertainty set is in the form of a polyhedron, that is: ; Using the duality theory, introduce the dual variable and satisfy: ; Through the dual strong duality, the above uncertainty constraint is equivalent to: ; Thus, the original non-linear robustness constraint is transformed into a set of linear constraints, which is beneficial to improving the computational efficiency in the distributed solution framework.

[0095] In step 2, the global optimization problem is decomposed into each node, enabling each node to independently solve based on local data, and at the same time transmitting global information through dual variables. This processing not only reduces the computational complexity but also lays the foundation for the final global coordination.

[0096] Step 3: Based on the solution of local sub-problems, through the distributed sub-gradient calculation and neighborhood consensus update mechanism, the dynamic coordination of local decisions of each node is realized, and then the global optimal energy management strategy is achieved.

[0097] First, determine that the input is the initial strategy (initialized by the dual decomposition method), the neighbor set (network topology), the initial dual variable , the initial step size and the maximum number of iterations , and output the converged cooperative optimization strategy and the final dual variables of each node .

[0098] Specifically, in the distributed scenario, each node uses its own local information to calculate the sub-gradient of the objective function with respect to the decision variable. Let the local objective function of node be: ; where is the dual variable introduced in Step 2. Node at the -th iteration calculates the local sub-gradient: ; The local sub-gradient reflects the descent direction of the local objective function at the current decision point, providing an optimization basis for subsequent variable updates.

[0099] Further, a consensus cooperation mechanism between neighborhoods is adopted. Let the neighborhood set of node be , and a weighted average method is used to achieve consensus update. Its update formula is: ; where, is the weight between nodes, satisfying: ; and is the step size parameter at the -th iteration, which realizes combining the optimization information of neighbor nodes, enabling each node to achieve global cooperation through the exchange of local information and gradually making the overall decision tend to be consistent.

[0100] Further, in the later stage of iteration for stable convergence, a dynamic step size adjustment strategy is introduced, using a decreasing step size, such as: ; where $\alpha_0$ is the initial step size. By using a decreasing strategy for the step size, it can decline rapidly in the initial stage and the convergence speed gradually slows down in the later stage, thereby improving the global convergence and stability.

[0101] Furthermore, the process of the entire distributed subgradient and consensus cooperation algorithm (DSC algorithm) can be described as follows: Through local subgradient calculation, neighborhood consensus update, and dynamic adjustment of the step size, real-time coordination and global optimization of the energy management strategies of each node in a distributed environment are achieved. Ultimately, the overall energy efficiency and robustness of the system are promoted. The process of the entire distributed subgradient and consensus cooperation algorithm (DSC algorithm) is as follows: Each node initializes local decision variables , storage strategies and dual variables ; Each node calculates the local objective function and obtains the subgradient ; Nodes exchange their respective and information through communication, and update the local decision variable according to the consensus update formula. At the same time, each node adjusts based on the global supply and demand situation and the dual update rule in step two. Further, the step size is dynamically adjusted, and the above process is repeated until the preset convergence condition is met. Thus, through real-time local gradient transmission and consensus mechanism, the local optimal solutions of each node are coordinated into a global optimal solution, giving full play to the advantages after the decomposition and solution in step two, and ultimately achieving real-time and efficient cooperative control of the system in a dynamic environment.

[0102] As Figure 2 shown, some embodiments of the present application provide an Internet of Things energy management system. The system includes an acquisition module 210, a correction module 220, a noise reduction module 230, a calculation module 240, a construction module 250, and an update module 260. Specifically: The acquisition module 210 is used to acquire the initial energy management data of all nodes in the target Internet of Things; The correction module 220 is used to perform anomaly processing on the initial energy management data of each node to obtain the first energy management data of the corresponding node; The noise reduction module 230 is used to perform denoising processing on the first energy management data of each node to obtain the second energy management data of the corresponding node; The calculation module 240 is used to calculate the energy management characteristics of the target Internet of Things according to the second energy management data of each node; The construction module 250 is used to construct the global objective function of the target Internet of Things according to the energy management characteristics of the target Internet of Things; An update module 260 for updating the initial node policy of each node based on a global objective function to obtain the optimal energy management policy for each node.

[0103] In some embodiments, the correction module 220 may include: calculating the median based on the initial energy management data of all nodes.

[0104] In some embodiments, the correction module 220 may include: calculating the median absolute deviation based on the median.

[0105] In some embodiments, the correction module 220 may include: performing an outlier determination on the initial energy management data of each node based on the median and the median absolute deviation, and correcting the outliers according to the determination result to obtain the first energy management data of the corresponding node.

[0106] In some embodiments, the noise reduction module 230 may include: performing multi-scale decomposition on the first energy management data of each node to obtain corresponding multi-scale coefficients.

[0107] In some embodiments, the noise reduction module 230 may include: using a soft threshold function to perform noise reduction processing on each multi-scale coefficient to obtain the noise-reduced multi-scale coefficients.

[0108] In some embodiments, the noise reduction module 230 may include: performing an inverse transform on the noise-reduced multi-scale coefficients to obtain the second energy management data of the corresponding node.

[0109] In some embodiments, the calculation module 240 may include: extracting the key features of each node from the second energy management data of each node.

[0110] In some embodiments, the calculation module 240 may include: fusing the key features of all nodes to obtain the energy management features of the target Internet of Things.

[0111] In some embodiments, the update module 260 may include: calculating the initial node policy and the local objective function of the corresponding node according to the initial energy management data of each node and the global objective function.

[0112] In some embodiments, the update module 260 may include: calculating the dual variable of each node according to the local objective function of each node.

[0113] In some embodiments, the update module 260 may include: updating the initial node policy of the corresponding node according to the dual variable of each node to obtain the optimal energy management policy for each node.

[0114] In some embodiments, the update module 260 may include: calculating the sub-gradient of each node according to the local objective function of each node.

[0115] In some embodiments, the update module 260 may include: updating the dual variable of the corresponding node according to the sub-gradient of each node.

[0116] In some embodiments, the update module 260 may include: iteratively updating the initial node policy of each node by using a consensus mechanism to obtain the optimal energy management policy of each node.

[0117] In some embodiments, the correction module 220 may include:

[0118] where is the median of the initial energy management data of all nodes at time and is the initial energy management data of node at time .

[0119] In some embodiments, the correction module 220 may include:

[0120] where is the median absolute deviation of all nodes at time and is the initial energy management data of node at time .

[0121] It should be noted that the Internet of Things energy management system provided in this embodiment and the above-mentioned Internet of Things energy management method are based on the same inventive concept. Therefore, the relevant content of the above-mentioned Internet of Things energy management method also applies to the content of the Internet of Things energy management system. Therefore, it will not be elaborated here.

[0122] For the system, it obtains the initial energy management data of all nodes in the target Internet of Things; performs anomaly processing on the initial energy management data of each node to obtain the first energy management data of the corresponding node; performs denoising processing on the first energy management data of each node to obtain the second energy management data of the corresponding node; calculates the energy management characteristics of the target Internet of Things according to the second energy management data of each node; constructs the global objective function of the target Internet of Things according to the energy management characteristics of the target Internet of Things; and updates the initial node policy of each node based on the global objective function to obtain the optimal energy management policy of each node. In this way, real-time dynamic coordination between the local energy policies of nodes and the global network energy efficiency can be achieved, thereby improving the efficiency of energy management optimization.

[0123] An embodiment of the present application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above-mentioned Internet of Things energy management method is implemented.

[0124] As Figure 3 , Figure 3 is a schematic diagram of the hardware structure of the electronic device provided by an embodiment of the present application. The electronic device includes: At least one battery; At least one memory; At least one processor; At least one program; The program is stored in the memory, and the processor executes at least one program to implement an Internet of Things energy management method implemented by the present disclosure above.

[0125] The electronic device may be any intelligent terminal including a mobile phone, a tablet computer, a personal digital assistant (Personal Digital Assistant, PDA), a vehicle-mounted computer, etc.

[0126] The electronic device of the embodiment of the present application will be introduced in detail below.

[0127] The processor 1600 can be implemented by using a general-purpose central processing unit (Central Processing Unit, CPU), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the technical solutions provided by the embodiments of the present disclosure; The memory 1700 can be implemented in the form of a read-only memory (Read Only Memory, ROM), a static storage device, a dynamic storage device, or a random access memory (Random Access Memory, RAM), etc. The memory 1700 can store an operating system and other application programs. When implementing the technical solutions provided by the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 1700, and the processor 1600 is called to execute an Internet of Things energy management method of the present disclosure.

[0128] The input / output interface 1800 is used to implement information input and output; The communication interface 1900 is used to implement communication interaction between this device and other devices, and can implement communication through a wired method (such as USB, network cable, etc.), or can also implement communication through a wireless method (such as a mobile network, WIFI, Bluetooth, etc.); The bus 2000 transmits information among various components of the device (such as the processor 1600, the memory 1700, the input / output interface 1800, and the communication interface 1900); Among them, the processor 1600, the memory 1700, the input / output interface 1800, and the communication interface 1900 are communicatively connected to each other inside the device through the bus 2000.

[0129] The embodiments of the present disclosure also provide a storage medium, which is a computer-readable storage medium. The computer-readable storage medium stores computer-executable instructions for causing a computer to execute the above-mentioned Internet of Things energy management method.

[0130] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor through a network. Examples of the above-mentioned network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0131] The embodiments described in the embodiments of the present disclosure are for more clearly illustrating the technical solutions of the embodiments of the present disclosure, and do not constitute a limitation on the technical solutions provided by the embodiments of the present disclosure. Those skilled in the art will know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present disclosure are equally applicable to similar technical problems.

[0132] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present disclosure, and may include more or fewer steps than those shown in the figures, or combine certain steps, or different steps.

[0133] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0134] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices can be implemented as software, firmware, hardware, and appropriate combinations thereof.

[0135] In the description of this application and the above-mentioned accompanying drawings, terms such as "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of this application described herein can be implemented in an order different from those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that comprises a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0136] It should be understood that in this application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects and indicates that three relationships may exist. For example, "A and / or B" may mean: only A exists, only B exists, and both A and B exist at the same time. Among them, A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one (one) of the following" or its similar expression refers to any combination of these items, including any combination of single items (ones) or plural items (ones). For example, at least one (one) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0137] In several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces. The indirect coupling or communication connection of devices or units can be in electrical, mechanical, or other forms.

[0138] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0139] In addition, in each embodiment of the present application, each functional unit can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0140] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions for causing an electronic device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of the present application. The aforementioned storage medium includes: various media that can store programs such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs.

[0141] The above has specifically described the preferred embodiments of the embodiments of the present application, but the embodiments of the present application are not limited to the above-mentioned implementation manners. Those skilled in the art can also make various equivalent deformations or substitutions without departing from the spirit of the embodiments of the present application, and these equivalent deformations or substitutions are all included within the scope defined by the claims of the embodiments of the present application.

[0142] The above has described the embodiments of the present application in detail with reference to the drawings, but the present application is not limited to the above-mentioned embodiments. Within the scope of knowledge possessed by those of ordinary skill in the art, various changes can also be made without departing from the purpose of the present application.

Claims

1. An Internet of Things energy management method, characterized in that, The method includes: Obtaining the initial energy management data of all nodes in the target Internet of Things; Performing anomaly processing on the initial energy management data of each node to obtain the first energy management data of the corresponding node; Performing denoising processing on the first energy management data of each node to obtain the second energy management data of the corresponding node; Calculating the energy management characteristics of the target Internet of Things according to the second energy management data of each node; Constructing the global objective function of the target Internet of Things according to the energy management characteristics of the target Internet of Things; Based on the global objective function, updating the initial node policies of each node to obtain the optimal energy management policies of each node.

2. The Internet of Things energy management method according to claim 1, characterized in that, The performing anomaly processing on the initial energy management data of each node to obtain the first energy management data of the corresponding node includes: Calculating the median according to the initial energy management data of all nodes; Calculating the median absolute deviation according to the median; Based on the median and the median absolute deviation, performing anomaly judgment on the initial energy management data of each node, and correcting the outliers according to the judgment results to obtain the first energy management data of the corresponding node.

3. The Internet of Things energy management method according to claim 2, characterized in that, The performing denoising processing on the first energy management data of each node to obtain the second energy management data of the corresponding node includes: Performing multi-scale decomposition on the first energy management data of each node to obtain the corresponding multi-scale coefficients; Using the soft threshold function to perform noise reduction processing on each multi-scale coefficient to obtain the denoised multi-scale coefficients; Performing inverse transformation on the denoised multi-scale coefficients to obtain the second energy management data of the corresponding node.

4. The Internet of Things energy management method according to claim 3, characterized in that The calculating the energy management characteristics of the target Internet of Things according to the second energy management data of each node includes: Extracting the key features of each node from the second energy management data of each node; Fusing the key features of all nodes to obtain the energy management characteristics of the target Internet of Things.

5. The Internet of Things energy management method according to claim 1, wherein The updating the initial node policies of each node based on the global objective function to obtain the optimal energy management policies of each node includes: Calculating the initial node policy and the local objective function of the corresponding node according to the initial energy management data of each node and the global objective function; Calculating the dual variable of each node according to the local objective function of each node; Updating the initial node policy of the corresponding node according to the dual variable of each node to obtain the optimal energy management policy of each node.

6. The Internet of Things energy management method according to claim 5, characterized in that The updating the initial node policy of the corresponding node according to the dual variable of each node to obtain the optimal energy management policy of each node includes: Calculating the subgradient of each node according to the local objective function of each node; Updating the dual variable of the corresponding node according to the subgradient of each node; Using the consensus mechanism to iteratively update the initial node policies of each node to obtain the optimal energy management policies of each node.

7. The Internet of Things energy management method according to claim 2, wherein The calculation formula for calculating the median according to the initial energy management data of all nodes includes: wherein, is the median of the initial energy management data of all nodes at time ; and is the initial energy management data of node at time . The calculation formula for calculating the median absolute deviation according to the median includes: Among them, is the moment the median absolute deviation of all nodes, is the node at the moment the initial energy management data.

8. An Internet of Things energy management system, characterized in that, The system includes: An acquisition module, configured to acquire initial energy management data of all nodes in the target Internet of Things; A correction module, configured to perform anomaly processing on the initial energy management data of each of the nodes to obtain first energy management data of the corresponding nodes; A noise reduction module, configured to perform noise reduction processing on the first energy management data of each of the nodes to obtain second energy management data of the corresponding nodes; A calculation module, configured to calculate energy management characteristics of the target Internet of Things according to the second energy management data of each of the nodes; A construction module, configured to construct a global objective function of the target Internet of Things according to the energy management characteristics of the target Internet of Things; An update module, configured to update the initial node policies of each of the nodes based on the global objective function to obtain optimal energy management policies of each of the nodes.

9. An electronic device, characterized in that, Comprising at least one control processor and a memory for communicatively connecting with the at least one control processor; the memory stores instructions executable by the at least one control processor, and the instructions are executed by the at least one control processor so that the at least one control processor can execute an Internet of Things energy management method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions for causing a computer to execute an Internet of Things energy management method according to any one of claims 1 to 7.

Images