A load switching event detection method based on the AlexNet model

Abstract

This application discloses a load switching event detection method based on the AlexNet model. The method includes: acquiring periodic power information of the power system; acquiring step power information based on the periodic power information; acquiring a trained AlexNet model; and inputting the step power information into the AlexNet model to obtain suspected switching event information from the step power information. This AlexNet-based load switching event detection method addresses the phenomena of "switching-like disturbance signals" and "pseudo-switching signals" caused by power fluctuations. Through transient signal step-based image preprocessing, it transforms the load switching identification problem into an image recognition problem, and utilizes deep learning technology to achieve accurate identification of load switching events in complex household power environments.

Description

Technical Field

[0001] This application relates to the field of power load analysis technology, specifically to a load switching event detection method and a load switching event detection device based on the AlexNet model. Background Technology

[0002] Due to normal fluctuations in mains power, noise, and the inherent characteristics of electrical appliances, "switching-like disturbance signals" and "pseudo-switching signals" are common phenomena in smart power consumption. These two situations exhibit transient processes distinct from other loads in terms of current amplitude, harmonics, and other information changes, particularly showing certain unique characteristics in the load power fluctuation curve. Furthermore, given the recent advancements in image recognition, transforming the load switching identification problem into an image recognition scientific problem provides a new approach for load switching detection in complex power consumption environments.

[0003] Existing technologies typically employ the CUSUM algorithm, which has played a certain role in online monitoring and identification of electricity-consuming targets, but the accuracy of this method is not ideal. Summary of the Invention

[0004] The purpose of this invention is to provide a load switching event detection method based on the AlexNet model to overcome or at least mitigate one of the above-mentioned defects in the prior art.

[0005] One aspect of the present invention provides a load switching event detection method based on the AlexNet model, the load switching event detection method based on the AlexNet model comprising:

[0006] Obtain periodic power information of the power system;

[0007] Step power information is obtained based on the periodic power information of the power system;

[0008] Obtain the trained AlexNet model;

[0009] The step power information is input into the AlexNet model to obtain suspected switching event information from the step power information.

[0010] Optionally, the periodic power information of the power system can be obtained using the following formula:

[0011] in,

[0012] I c To intercept current; V c This is the voltage data for the current cycle.

[0013] Optionally, before obtaining the step power information based on the periodic power information of the power system, the load switching event detection method based on the AlexNet model further includes:

[0014] The acquired periodic power information of the power system is filtered and / or anomaly removal is performed.

[0015] Optionally, the load switching event detection method based on the AlexNet model further includes:

[0016] The authenticity of the suspected drop event information is determined to obtain the drop event information.

[0017] This application also provides a load switching event detection device based on the AlexNet model, the load switching event detection device based on the AlexNet model comprising:

[0018] A power system periodic power information acquisition module, wherein the power system periodic power information acquisition module is used to acquire power system periodic power information;

[0019] A step power information acquisition module, wherein the step power information acquisition module is used to acquire step power information based on the periodic power information of the power system;

[0020] AlexNet model acquisition module, which is used to acquire a trained AlexNet model;

[0021] A suspected switching event acquisition module is used to input step power information into the AlexNet model to acquire suspected switching event information in the step power information.

[0022] Optionally, the load switching event detection device based on the AlexNet model further includes:

[0023] A filtering module is used to filter the acquired periodic power information of the power system.

[0024] Optionally, the load switching event detection device based on the AlexNet model further includes:

[0025] An anomaly removal module is used to remove anomalies from the acquired periodic power information of the power system.

[0026] Optionally, the load switching event detection device based on the AlexNet model further includes:

[0027] The authenticity determination module is used to determine the authenticity of the acquired suspected switching event information, thereby obtaining the switching event information.

[0028] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the load switching event detection method based on the AlexNet model as described above.

[0029] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to implement the load switching event detection method based on the AlexNet model as described above.

[0030] Beneficial effects

[0031] The load switching event detection method based on the AlexNet model in this application targets the phenomena of "switching disturbance signals" and "pseudo-switching signals" caused by power fluctuations. By preprocessing transient signal step images, the load switching identification problem is transformed into an image recognition problem. Deep learning technology is used to achieve accurate identification of load switching events in complex household power environments. Attached Figure Description

[0032] Figure 1 This is a flowchart of a load switching event detection method based on the AlexNet model according to an embodiment of this application.

[0033] Figure 2 It is an electronic device used to achieve Figure 1 The method for detecting load switching events based on the AlexNet model is shown.

[0034] Figure 3 This is a schematic diagram of the load power discretization process.

[0035] Figure 4 This is a schematic diagram of the step transformation process for periodic power signals.

[0036] Figure 5 This diagram illustrates the comparison of filtering effects under different parameters.

[0037] Figure 6 This is a schematic diagram illustrating the filtering effect on the load power signal.

[0038] Figure 7 This is a schematic diagram illustrating the effect of a step change in load power.

[0039] Figure 8 This is a schematic diagram for identifying suspected load switching events.

[0040] Figure 9 This is a schematic diagram illustrating the normalization effect of the load power image.

[0041] Figure 10 A schematic diagram of the switching event image obtained from the computer switch using the FSP preprocessing algorithm.

[0042] Figure 11 A schematic diagram of five types of load onset event images obtained by the FSP preprocessing algorithm. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some, but not all, embodiments of this application. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0044] In the description of this application, it should be understood that the terms "center", "longitudinal", "lateral", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting the scope of protection of this application.

[0045] Figure 1 This is a flowchart of a load switching event detection method based on the AlexNet model according to an embodiment of this application.

[0046] like Figure 1 The load switching event detection method based on the AlexNet model shown includes:

[0047] Step 1: Obtain periodic power information of the power system;

[0048] Step 2: Obtain step power information based on the periodic power information of the power system;

[0049] Step 3: Obtain the trained AlexNet model;

[0050] Step 4: Input the step power information into the AlexNet model to obtain the suspected switching event information in the step power information.

[0051] The load switching event detection method based on the AlexNet model in this application targets the phenomena of "switching disturbance signals" and "pseudo-switching signals" caused by power fluctuations. By preprocessing transient signal step images, the load switching identification problem is transformed into an image recognition problem. Deep learning technology is used to achieve accurate identification of load switching events in complex household power environments.

[0052] In non-intrusive monitoring of power systems, this application uses changes in active power to realize the detection process of switching events.

[0053] Because sensors cannot collect data at standard equal time intervals and sampling rates are limited, the accuracy of calculating periodic power using the expected sampling rate may be compromised. However, due to the periodic nature of alternating current, this invention determines and extracts data based on the period, calculating the power of each period as a single point.

[0054] See Figure 3 In this embodiment, the current is set to I, the voltage to V, and the real-time power to P. t Periodic power P p Due to the periodic operation of alternating current, the fundamental frequency f is defined. b A single period is defined as a unit, constituting a power sequence P. p To perform switching event detection and segment the periodic current and voltage signals, zero-crossing detection is used as the basis. The first zero-crossing is defined as v. a The second zero crossing is v b When v is satisfied a ≤0, v b <0 and v b+1 When I ≥ 0, the current I and voltage V are extracted from a to b as the current period data, and I is used as the basis for each period. c and V c This is indicated, and power-related calculations are performed. Among them, the real-time power P... t The calculation formula is P t =I c *V c Current cycle power P p The calculation formula is as follows:

[0055] in,

[0056] I c To intercept current; V c This represents the voltage data for the current cycle, where N represents the number of data points in the current cycle.

[0057] This application proposes a filtered-step data preprocessing method (FSP) that can sensitively extract data suspected of switching operations. The FSP method includes filtering and denoising the original periodic power signal and transforming the power signal into a step pattern. Its basic process is as follows: Figure 4 As shown.

[0058] In this embodiment, before obtaining step power information based on the periodic power information of the power system, the load switching event detection method based on the AlexNet model further includes:

[0059] The acquired periodic power information of the power system is filtered and / or anomaly removal is performed.

[0060] For the calculated periodic power P p ={P(i)}, i = 1, 2, ... Since it contains a large amount of noise and weak power fluctuation information from the power line, directly applying it to determine suspected switching points will generate a large amount of data, increasing the training difficulty of the AlexNet-based network model. Therefore, firstly, the periodic power P... p Filtering and outlier removal are performed to generate the filtered periodic power P′. p ={p′(i)}, i = 1, 2, ..., then the filtered periodic power P′ p The periodic power P″ converted to a step state p ={p″(i)}, i = 1, 2, ..., and finally, the time slice information of suspected switches is extracted based on the periodic power of the step pattern to determine the authenticity of the switch event.

[0061] Specifically, let the length of the sliding window be l. w Assume the current time is t+l w / 2, with a step size of 1, truncate P p Window data:

[0062]

[0063] The judgment function in the filtering part is defined as follows:

[0064] x m1 =median(X)

[0065] x m2 =c|Xx m1 |

[0066] x m3 =median(|X(t)-x) m1 |)

[0067] In the formula, c represents the degree of filtering of data noise. The smaller the value of c, the higher the signal-to-noise ratio of the filtered signal. However, this does not mean that the smaller the value of c, the better. When c reaches a certain critical point, its noise filtering effect is difficult to improve further. Experiments have verified that setting this value to 0.1 is more reasonable.

[0068] The statistical function g1 in the filtering section is defined as follows:

[0069]

[0070] In the formula, d is the length control parameter of the smoothing filter window, and the value g1 obtained by the statistical function is the filtered value p′(t) obtained at the current time. After the filtering process, the periodic power signal P′ with outliers and noise interference removed can be obtained. p The effect is as follows Figure 6 As shown.

[0071] The original periodic power signal, after filtering, achieves good noise reduction and outlier removal. However, due to significant power fluctuations during actual load switching and operation, the filtered periodic power signal is still insufficient for detecting suspected switching points using simple threshold methods. Existing switch detection methods also suffer from filtering out normal switching events with large thresholds. To further expand the power range of identifiable switches and simplify the suspected switching point detection process, a step-based transformation of the filtered data is proposed.

[0072] In this embodiment, obtaining step power information based on the periodic power information of the power system includes:

[0073] With a step size of 0.5l w Capture window data Y:

[0074] Y={y(1),y(2),...,y(t)},i=1,2,...,l w

[0075] Calculate the median y of window Y m1 :

[0076] y m1 =median(Y)

[0077] The statistical function g2 in the step morphology part is defined as follows:

[0078]

[0079] In the formula, h represents the power threshold for filtering, indicating that power fluctuations within h will not be identified. Setting h = 20 means that switching events above 20 watts can be identified. i represents the power index in window Y. The statistical function g2 is applied to the data in window Y to sequentially determine and assign values, resulting in array Y′.

[0080] Y′={y′(1),y′(2),...,y′(t)},i=1,2,...,l w

[0081] In practical applications, it has been observed that power systems experience short-term power spikes or drops followed by a rapid return to their original levels. This phenomenon is difficult to suppress using filtering algorithms, and experiments confirm that this process often occurs during load switching. This invention treats power changes with a duration shorter than the window length as outliers and processes them accordingly. The statistical function g3 is defined as follows:

[0082]

[0083]

[0084] In the formula, 'e' represents the control parameter for determining the step mode. A larger value indicates a more lenient requirement for the step mode and may result in a smooth transition at the power jump. Experiments have shown that a value of 0.1 is reasonable. The statistical function g3 is applied to sequentially determine and assign values ​​to the data in window Y′, obtaining the array Y″ and updating it to set P″. p The power change curve that transforms into a step-like state can be obtained from this, with the effect as follows: Figure 7 As shown.

[0085] As can be seen from the figure, the step power conversion successfully transforms the filtered power into a step signal. Experiments have verified that the filtered power and the step power have good synchronous change capability at the moment of actual load switching. By monitoring the changes in the step signal, data of suspected switching points in all power ranges can be extracted, and the AlexNet model can be used to train the model for judging the authenticity of switching events.

[0086] Based on the requirements of filtering and step-change processing, the processed point is 1.5l backwards from the actual acquisition point. w Data points. Based on the detection of suspected switch points, up to l data points of equal length need to be extracted from front to back. s Each point is used as sample data. Therefore, the actual latency of this algorithm in data processing is 1.5 points. w +l s With a reasonable window length, event detection capability with minimal latency can be achieved. Let the point to be judged be t′(i), t′(i) = tl. w -0.5ls If i∈N, then the corresponding suspected switch determination function is:

[0087]

[0088] Depend on Figure 8 It can be seen that the FSP method has significant advantages in detecting suspected switching point problems, as it can effectively detect the location of the switching point while filtering noise. Furthermore, by setting a reasonable threshold h, it can detect low-power events over a wider range, demonstrating superior performance compared to existing rule-based judgment and switching point detection methods.

[0089] In this embodiment, the load switching event detection method based on the AlexNet model further includes:

[0090] The authenticity of the suspected drop event information is determined to obtain the drop event information.

[0091] Specifically, when a suspected switching event is detected, it is necessary to extract valid current, voltage, and power information from the preceding and / or following moments of the suspected switching event to determine the authenticity of the event. Given the decision point t′(i), define a truncated data segment Q with length l. s To prevent other suspected switching points from being captured in the current data segment, a Boolean function g4 for data capture and a Boolean function g5 for data capture are defined:

[0092]

[0093]

[0094]

[0095]

[0096] The function z in the above formula records the positions of other suspected switching points in the data segment Q being extracted. Boolean functions g4 and g5 are used to iterate through the data Q to generate arrays K and M respectively. When the Boolean function value is 1, the actual value at the corresponding position is retrieved. When the Boolean function value is 0, the backward process retrieves z1 to record the data corresponding to the position, and the forward process retrieves z2 to record the data corresponding to the position. Similarly, the current data Q is extracted according to the above logic. i Voltage data Q v Original power Q p1 Filter power Q p2 and step mode power Q p3 This enables the interception of sample data when a suspected switching event is detected, as well as the handling of anomalies when other suspected switching events exist.

[0097] In this embodiment, if multiple loads are already in normal operation before load switching, the data intercepted during the operation of the detected load includes information about the existing operating loads. Therefore, the intercepted power sequence needs to be processed to improve the model training accuracy. This application will use the calculated power sequence as sample data. To intuitively demonstrate the processing process, the data processing principle and effect of the power sequence will be explained in conjunction with data visualization.

[0098] In residential homes, the maximum power of a single appliance can reach around 3000W, while the minimum power of a identifiable switchable appliance is approximately 20W, resulting in a significant power difference. To accelerate the convergence speed of the neural network and avoid numerical problems such as neuron output saturation, the power sequence is normalized, limiting the data information to the range [-1, 1].

[0099]

[0100] In the formula Q u The maximum power of a single appliance is considered. Furthermore, during load switching, data closer to the switching moment is more correlated with the current load, while data from more distant moments has less reference value. To strengthen the data in the sequence Q concerning the moment of load switching and weaken the influence of other data on the model, a Gaussian function is applied to weight the sequence.

[0101]

[0102] In the formula G a G is the peak value of the Gaussian curve. b G represents the coordinates of the peak center. c σ represents the standard deviation and the width. Since data Q is truncated around the step point of a step-like pattern, 0.5l... s Corresponding to the coordinates G of the Gaussian curve peak center b The maximum weight is defined as 1, i.e., G. a =1, in order to controllable weight values ​​at the boundaries of the truncated data segments and define them as G d G can be obtained by inverse solving of the Gaussian function. c :

[0103]

[0104] Applying this method to the original power Q p1 Filter power Q p2 and step mode power Q p3 Example effect as follows Figure 7 As shown.

[0105] Data visualization readily demonstrates that the above process weights the sequence data, making the data in region A, which is closer to the load switching instant, stand out, while weakening the data in regions B and C, which are farther away from the switching instant. Its advantage lies in mitigating interference from external factors beyond the load switching transient, thus improving the accuracy of the judgment.

[0106] The following examples further illustrate this application in detail. It is understood that these examples do not constitute any limitation on this application.

[0107] The experiment mainly verifies the image recognition capabilities of the FSP preprocessing algorithm and the AlexNet model from two aspects: single load switching event identification and multi-load switching event identification.

[0108] 2.1 Identification and Verification of Single Load Switching Events

[0109] In the single-load switching event detection experiment, power data from six typical household appliances—computers, refrigerators, microwave ovens, air conditioners, electric kettles, and oxygen concentrators—were still used. A simulated household electricity environment was created, with each load being switched on and off 100 times. The on / off events of the six loads were preprocessed using the FSP algorithm to obtain real switching images, switch-like disturbance images, and pseudo-switching images.

[0110] Computer operation data is labeled with switching events. The switching events are preprocessed using the FSP algorithm to obtain images of real switching events, switch-like perturbations, and pseudo-switching events, such as... Figure 8 As shown, blue represents the real-time power curve, yellow represents the switching step signal, and cyan represents the average power. Six types of load signals were labeled, and a total of 1200 real switching images were obtained. Data showing obvious air conditioner switching phenomena were used as the real switching data. Simultaneously, during each load switching process, one set of switching disturbance images and one set of pseudo-switching images were obtained, totaling 1200 images. Load switching detection was performed based on the AlexNet-based algorithm. The training and test sets were divided in a 7:3 ratio, and the experimental statistics for the test set are shown in Table 1.

[0111] Table 1. Detection results of single load switching events

[0112] Table 1. Detection results of single load switching events

[0113]

[0114] Among them, false detection events N fThis represents the number of switch disturbance images or pseudo-switch images identified as real switch switching events. Statistical data shows that the AlexNet-based load switching detection algorithm still has a relatively ideal detection effect on resistive loads. The overall accuracy of switching identification is not much different from the load switching detection algorithm based on adaptive threshold sliding window, but it shows a significant improvement compared to the CUSUM algorithm.

[0115] 2.2 Identification and Verification of Multi-Load Switching Events

[0116] The multi-load switching event detection experiment still uses power data from five types of loads operating simultaneously: refrigerator, microwave oven, humidifier, electric kettle, and desktop computer. Simulating a household electricity environment, 100 sets of switching data for simultaneous load operation were collected, and the power data were labeled according to time sequence.

[0117] The FSP algorithm is applied to preprocess the switching events, obtaining real switching events, switch-like disturbances, and pseudo-switching images of five loads: computer, refrigerator, microwave oven, kettle, and humidifier. The transient processes of each load switching event are shown below. Figure 11 As shown in Table 2, a total of 1580 real switch images and 2420 corresponding switch-like perturbation images and pseudo-switch images were obtained in the experiment. The training set and test set were divided in a 7:3 ratio. The load switching detection algorithm based on AlexNet was used for switching detection. The experimental statistics of the test set are shown in Table 2.

[0118] Table 2 Detection results of multi-load switching events

[0119] Table 2 Detection results of multi-load switching events

[0120]

[0121] Comparative experiments show that the AlexNet-based load switching detection algorithm is 14.65% better than the CUSUM detection algorithm, demonstrating a significant improvement. Therefore, it is feasible to use this load event switching detection algorithm for non-intrusive load monitoring.

[0122] This application also provides a load switching event detection device based on the AlexNet model. The AlexNet-based load switching event detection device includes a power system periodic power information acquisition module, a step power information acquisition module, an AlexNet model acquisition module, and a suspected switching event acquisition module. The power system periodic power information acquisition module is used to acquire power system periodic power information; the step power information acquisition module is used to acquire step power information based on the power system periodic power information; the AlexNet model acquisition module is used to acquire a trained AlexNet model; and the suspected switching event acquisition module is used to input the step power information into the AlexNet model to acquire suspected switching event information from the step power information.

[0123] In this embodiment, the load switching event detection device based on the AlexNet model further includes a filtering module, which is used to filter the acquired periodic power information of the power system.

[0124] In this embodiment, the load switching event detection device based on the AlexNet model further includes an anomaly removal module, which is used to perform anomaly removal processing on the acquired periodic power information of the power system.

[0125] In this embodiment, the load switching event detection device based on the AlexNet model further includes a authenticity determination module, which is used to determine the authenticity of the acquired suspected load switching event information, thereby obtaining the load switching event information.

[0126] Beneficial effects

[0127] The load switching event detection method based on the AlexNet model in this application targets the phenomena of "switching disturbance signals" and "pseudo-switching signals" caused by power fluctuations. By preprocessing transient signal step images, the load switching identification problem is transformed into an image recognition problem. Deep learning technology is used to achieve accurate identification of load switching events in complex household power environments.

[0128] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to implement the load switching event detection method based on the AlexNet model as described above.

[0129] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the load switching event detection method based on the AlexNet model as described above.

[0130] Figure 2 This is an exemplary structural diagram of an electronic device capable of implementing the load switching event detection method based on the AlexNet model provided in one embodiment of this application.

[0131] like Figure 2 As shown, the electronic device includes an input device 501, an input interface 502, a central processing unit 503, a memory 504, an output interface 505, and an output device 506. The input interface 502, central processing unit 503, memory 504, and output interface 505 are interconnected via a bus 507. The input device 501 and output device 506 are connected to the bus 507 via the input interface 502 and output interface 505, respectively, and thus connected to other components of the electronic device. Specifically, the input device 504 receives input information from the outside and transmits it to the central processing unit 503 via the input interface 502. The central processing unit 503 processes the input information based on computer-executable instructions stored in the memory 504 to generate output information, temporarily or permanently storing the output information in the memory 504, and then transmitting the output information to the output device 506 via the output interface 505. The output device 506 outputs the output information to the outside of the electronic device for user use.

[0132] In other words, Figure 2 The illustrated electronic device may also be implemented as including: a memory storing computer-executable instructions; and one or more processors, which can be coupled when executing the computer-executable instructions. Figure 1 This describes a load switching event detection method based on the AlexNet model.

[0133] In one embodiment, Figure 2 The electronic device shown can be implemented as including: a memory 504 configured to store executable program code; and one or more processors 503 configured to run the executable program code stored in the memory 504 to execute the load switching event detection method based on the AlexNet model in the above embodiments.

[0134] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0135] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0136] Computer-readable media include both permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, DVD or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0137] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0138] Furthermore, it is clear that the word "comprising" does not exclude other units or steps. Multiple units, modules, or devices recited in a device claim may also be implemented by a single unit or overall device through software or hardware. The terms "first," "second," etc., are used to identify names, not to indicate any specific order.

[0139] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code, which includes one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutively marked blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or the overall flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0140] In this embodiment, the processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0141] Memory can be used to store computer programs and / or modules. The processor implements various functions of the device / terminal equipment by running or executing the computer programs and / or modules stored in the memory, and by accessing data stored in the memory. Memory can mainly include a program storage area and a data storage area. The program storage area can store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area can store data created based on the use of the mobile phone (such as audio data, phonebook, etc.). In addition, memory can include high-speed random access memory, and can also include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0142] In this embodiment, if the modules / units integrated into the device / terminal equipment are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction.

[0143] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting load switching events based on the AlexNet model, characterized in that, The load switching event detection method based on the AlexNet model includes: Obtain periodic power information of the power system; Step power information is obtained based on the periodic power information of the power system; Obtain the trained AlexNet model; The step power information is input into the AlexNet model to obtain the suspected switching event information in the step power information; The periodic power information of the power system can be obtained using the following formula: ;in, To intercept the current; This is the voltage data for the current cycle. Before obtaining step power information based on the periodic power information of the power system, the load switching event detection method based on the AlexNet model further includes: The acquired periodic power information of the power system is filtered and / or anomaly removal is performed. The load switching event detection method based on the AlexNet model further includes: The authenticity of the suspected drop event information is determined to obtain the drop event information.

2. A load switching event detection device based on the AlexNet model, characterized in that, The load switching event detection device based on the AlexNet model includes: A power system periodic power information acquisition module, wherein the power system periodic power information acquisition module is used to acquire power system periodic power information; A step power information acquisition module, wherein the step power information acquisition module is used to acquire step power information based on the periodic power information of the power system; AlexNet model acquisition module, which is used to acquire a trained AlexNet model; A suspected switching event acquisition module is used to input step power information into the AlexNet model, thereby acquiring suspected switching event information in the step power information; The load switching event detection device based on the AlexNet model further includes: A filtering module is used to filter the acquired periodic power information of the power system. The load switching event detection device based on the AlexNet model further includes: An anomaly removal module is used to perform anomaly removal processing on the acquired periodic power information of the power system. The load switching event detection device based on the AlexNet model further includes: The authenticity determination module is used to determine the authenticity of the acquired suspected switching event information, thereby obtaining the switching event information. The periodic power information of the power system is obtained through the following formula: ;in, To intercept the current; This is the voltage data for the current cycle.

3. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it can implement the load switching event detection method based on the AlexNet model as described in claim 1.

4. An electronic device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to implement the load switching event detection method based on the AlexNet model as described in claim 1.

Images