Battery discharge duration determination model training method, device and apparatus
By using real-time voltage feature maps during the period from battery power supply to power outage and training with generative adversarial networks, the problem of inaccurate determination of battery discharge duration in computer rooms was solved, achieving efficient and low-cost battery discharge duration monitoring.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA MOBILE COMM GRP CHONGQING CO LTD
- Filing Date
- 2021-08-26
- Publication Date
- 2026-08-04
AI Technical Summary
Existing methods for determining the discharge time of data center batteries are inaccurate, lacking sufficient sample data during battery power outages and service shutdowns, leading to inaccurate results.
The model is trained using real-time voltage feature maps from battery power supply to power outage. It is then trained in conjunction with a generative adversarial network (GAN). Based on the adversarial principle of the GAN, the number of training samples is reduced, thus ensuring the accuracy of the model.
This improved the accuracy of the battery discharge duration determination model, reduced training costs, and enabled efficient battery discharge duration monitoring.
Smart Images
Figure CN115730502B_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of battery performance testing, and in particular relates to a training method, apparatus and equipment for determining a battery discharge duration model. Background Technology
[0002] The discharge performance of data center batteries refers to the duration for which the batteries in a data center can support the continued operation of equipment within the data center during an AC mains power outage. This performance is crucial for guiding power generation maintenance during power outages, battery resource replacement and allocation, and intelligent power saving in data centers, making it an extremely important capability metric for network operations and maintenance. Currently, operators' understanding of the battery discharge performance of most data centers remains vague and inaccurate, necessitating the use of relevant methods to obtain more accurate data.
[0003] Existing methods for determining the discharge time of data center batteries require a large amount of battery sample data from actual data center power outages and service outages. However, actual data center power outages and service outages are relatively rare, which is insufficient to provide a large amount of battery sample data from actual data center power outages and service outages, resulting in inaccurate methods for determining the discharge time of data center batteries. Summary of the Invention
[0004] This application provides a training method, apparatus, and device for a battery discharge duration determination model, which can solve the problem of inaccurate determination methods for computer room battery discharge duration.
[0005] In a first aspect, embodiments of this application provide a training method for a battery discharge duration determination model, comprising:
[0006] Obtain training samples; the training samples are real time-series voltage feature maps corresponding to the period from battery power supply to power outage and depletion of power.
[0007] The real time-series voltage feature map is input into the generative adversarial network for training, so as to update the parameters of the generative adversarial network.
[0008] When the preset training stopping condition of the generative adversarial network is met, the battery discharge duration is determined to determine the model.
[0009] Furthermore, in one embodiment, obtaining training samples includes:
[0010] Acquire AC power alarm information, DC power alarm information, and voltage monitoring information;
[0011] The true timing voltage characteristic diagram is determined based on AC alarm information, DC alarm information, and voltage monitoring information.
[0012] Furthermore, in one embodiment, the generative adversarial network includes: a generative model and a discriminative model;
[0013] The real-time voltage feature maps are input into the generative adversarial network (GAN) for training, and the parameters used to update the GAN include:
[0014] Randomly generate a random feature map of the same size as the actual time-series voltage feature map;
[0015] The random feature map is input into the generation model for feature expansion, which is used to generate pseudo-time series voltage feature maps;
[0016] The real and pseudo time series voltage feature maps are input into the discrimination model to distinguish between them and obtain the true and false discrimination result features.
[0017] The parameters of the generative adversarial network are updated based on the characteristics of the true / false discrimination results.
[0018] Furthermore, in one embodiment, when a preset generative adversarial network training stopping condition is met, a battery discharge duration determination model is established, including:
[0019] When the preset training stopping condition of the generative adversarial network is met, the generative model is determined as the battery discharge duration determination model.
[0020] Secondly, embodiments of this application provide a method for determining battery discharge duration using a battery discharge duration determination model. The battery discharge duration determination model is trained using a training method for battery discharge duration determination models, the method including:
[0021] Obtain a reference sample, which is an incomplete timing voltage feature map corresponding to the period from battery power supply to power supply stop and before the power is depleted;
[0022] The reference sample is input into the battery discharge duration determination model for feature expansion, resulting in a complete time-series voltage feature map of the battery.
[0023] The battery discharge duration is determined based on the complete timing voltage characteristic diagram.
[0024] Furthermore, in one embodiment, the size of the incomplete timing voltage feature map is N×N;
[0025] Determining the battery discharge duration based on the complete time-series voltage characteristic map includes:
[0026] The characteristic curves in the complete time-series voltage characteristic map are determined by traversing the complete time-series voltage characteristic map.
[0027] The battery discharge time is determined using the following formula:
[0028] T = X / N × M
[0029] Where T is the battery discharge time, X is the maximum time corresponding to the characteristic curve, and M is the preset maximum battery discharge time.
[0030] Furthermore, in one embodiment, obtaining a reference sample includes:
[0031] Acquire AC power alarm information, DC power alarm information, and voltage monitoring information;
[0032] The incomplete timing voltage characteristic map is determined based on AC alarm information, DC alarm information, and voltage monitoring information.
[0033] Thirdly, embodiments of this application provide a training apparatus for a battery discharge duration determination model, comprising:
[0034] The acquisition module is used to acquire training samples; the training samples are the real time-series voltage feature maps corresponding to the period from battery power supply to power outage and battery depletion.
[0035] The input module is used to input real time-series voltage feature maps into the generative adversarial network for training, so as to update the parameters of the generative adversarial network.
[0036] The determination module is used to determine the battery discharge duration determination model when the preset generative adversarial network training stopping conditions are met.
[0037] Furthermore, in one embodiment, the acquisition module is specifically used for:
[0038] Acquire AC power alarm information, DC power alarm information, and voltage monitoring information;
[0039] The true timing voltage characteristic diagram is determined based on AC alarm information, DC alarm information, and voltage monitoring information.
[0040] Furthermore, in one embodiment, the generative adversarial network includes: a generative model and a discriminative model;
[0041] The input module is specifically used for:
[0042] Randomly generate a random feature map of the same size as the actual time-series voltage feature map;
[0043] The random feature map is input into the generation model for feature expansion, which is used to generate pseudo-time series voltage feature maps;
[0044] The real and pseudo time series voltage feature maps are input into the discrimination model to distinguish between them and obtain the true and false discrimination result features.
[0045] The parameters of the generative adversarial network are updated based on the characteristics of the true / false discrimination results.
[0046] Furthermore, in one embodiment, the determining module is specifically used for:
[0047] When the preset training stopping condition of the generative adversarial network is met, the generative model is determined as the battery discharge duration determination model.
[0048] Fourthly, embodiments of this application provide an apparatus for determining battery discharge duration using a battery discharge duration determination model. The battery discharge duration determination model is trained using a training device for the battery discharge duration determination model. The apparatus includes:
[0049] The acquisition module is used to acquire reference samples, which are incomplete time-series voltage feature maps corresponding to the period from battery power supply to power supply stop and before the power is depleted;
[0050] The input module is used to input reference samples into the battery discharge duration determination model for feature expansion, thereby obtaining the complete time-series voltage feature map of the battery.
[0051] The determination module is used to determine the battery discharge duration based on the complete timing voltage characteristic map.
[0052] Furthermore, in one embodiment, the size of the incomplete timing voltage feature map is N×N;
[0053] The module is specifically used for:
[0054] The characteristic curves in the complete time-series voltage characteristic map are determined by traversing the complete time-series voltage characteristic map.
[0055] The battery discharge time is determined using the following formula:
[0056] T = X / N × M
[0057] Where T is the battery discharge time, X is the maximum time corresponding to the characteristic curve, and M is the preset maximum battery discharge time.
[0058] Furthermore, in one embodiment, the acquisition module is specifically used for:
[0059] Acquire AC power alarm information, DC power alarm information, and voltage monitoring information;
[0060] The incomplete timing voltage characteristic map is determined based on AC alarm information, DC alarm information, and voltage monitoring information.
[0061] Fifthly, embodiments of this application provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program implements the above-described method when executed by the processor.
[0062] Sixthly, embodiments of this application provide a computer-readable storage medium storing an information transmission implementation program, which, when executed by a processor, implements the above-described method.
[0063] The battery discharge duration determination model training method, apparatus, and device of this application embodiment uses real time-series voltage feature maps corresponding to the period from battery power supply to power outage and depletion of power as training samples. It is trained using a generative adversarial network (GAN). Based on the adversarial principle of GAN, a small number of training samples are required. Using real time-series voltage feature maps as training samples ensures the accuracy of the trained battery discharge duration determination model. The method of determining battery discharge duration using the battery discharge duration determination model in this application embodiment is trained using the same method, ensuring the accuracy of the determined battery discharge duration. Attached Figure Description
[0064] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0065] Figure 1 This is a flowchart illustrating a training method for a battery discharge duration determination model provided in an embodiment of this application.
[0066] Figure 2 This is a schematic diagram of the principle of a generative adversarial network provided in an embodiment of this application;
[0067] Figure 3 This is a flowchart illustrating a method for determining battery discharge time using a battery discharge time determination model, as provided in an embodiment of this application.
[0068] Figure 4 This is a schematic diagram of the structure of a training device for a battery discharge duration determination model provided in one embodiment of this application;
[0069] Figure 5 This is a schematic diagram of the structure of a device for determining battery discharge time using a battery discharge time determination model provided in one embodiment of this application;
[0070] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0071] The features and exemplary embodiments of various aspects of this application will now be described in detail. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only configured to explain this application and are not configured to limit this application. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples of this application.
[0072] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.
[0073] Existing methods for determining the discharge time of data center batteries require a large amount of battery sample data from actual data center power outages and service outages. However, actual data center power outages and service outages are relatively rare, which is insufficient to provide a large amount of battery sample data from actual data center power outages and service outages, resulting in inaccurate methods for determining the discharge time of data center batteries.
[0074] To address the problems of existing technologies, this application provides a training method, apparatus, and device for a battery discharge duration determination model. This application uses real-time voltage feature maps corresponding to the period from battery power supply to power depletion and complete battery exhaustion as training samples. It employs a generative adversarial network (GAN) for model training. Based on the adversarial principle of GANs, a small number of training samples is required. Using real-time voltage feature maps as training samples ensures the accuracy of the trained battery discharge duration determination model. The method for determining battery discharge duration using this application's battery discharge duration determination model is also described below. The applied battery discharge duration determination model is trained using the training method of this application, ensuring the accuracy of the determined battery discharge duration.
[0075] Figure 1This diagram illustrates a flowchart of a training method for a battery discharge duration determination model provided in one embodiment of this application. Figure 1 As shown, the method may include the following steps:
[0076] S110, Obtain training samples.
[0077] The training samples are real time-series voltage feature maps corresponding to the period from battery power supply to power outage and depletion of power.
[0078] In one embodiment, S110 may include:
[0079] Acquire AC power alarm information, DC power alarm information, and voltage monitoring information.
[0080] The true timing voltage characteristic diagram is determined based on AC alarm information, DC alarm information, and voltage monitoring information.
[0081] During operation, the equipment room generates environmental alarm data, including AC and DC power alarms. An AC power supply alarm indicates mains power supply, while an AC power outage alarm indicates a mains power failure, at which point the battery is activated. A DC power outage alarm indicates that the battery can no longer provide power, causing the battery to shut down and all equipment in the equipment room to shut down. A DC power supply alarm indicates that the battery is still supplying power. Therefore, based on the AC and DC power alarm information, the battery startup and shutdown times can be obtained. The base station equipment records the equipment room's voltage monitoring information. By plotting curves based on the battery startup and shutdown times and the voltage monitoring information, the true time-series voltage characteristic map can be determined.
[0082] This application embodiment can automatically generate a real time-series voltage characteristic map by using AC power alarm information, DC power alarm information, and voltage monitoring information, without the need for manual data collection.
[0083] S120: Input the real time-series voltage feature map into the generative adversarial network for training, so as to update the parameters of the generative adversarial network.
[0084] In one embodiment, a generative adversarial network (GAN) may include a generative model and a discriminative model, and the principle of a GAN is as follows: Figure 2 As shown, S120 may include:
[0085] A random feature map of the same size as the real time-series voltage feature map is randomly generated; the random feature map is input into the generation model for feature expansion to generate a pseudo time-series voltage feature map; the real time-series voltage feature map and the pseudo time-series voltage feature map are input into the discrimination model for authenticity discrimination to obtain the authenticity discrimination result features; the parameters of the generative adversarial network are updated based on the authenticity discrimination result features.
[0086] The pseudo-time-series voltage feature map can be generated by a pixel grayscale image generator. The random feature map contains a short pixel curve, which is intended to accurately correspond to the time-series voltage. The random feature map is expanded by generating random pixels in opposite directions, starting from the two endpoints of the pixel curve, to expand the pixel curve in the random feature map, thus generating the pseudo-time-series voltage feature map. A random 128×128 pixel grayscale image generator p(z) is defined, where 128×128 pixels represents the size of the real time-series voltage feature map and the random feature map. Let the generation model be G and the discriminator model be D. Training begins:
[0087] Iteration 1: Generative Adversarial Network training iterations n = 1, ..., N;
[0088] Iteration 2: Discriminant model training iterations k = 1, ..., K;
[0089] Step 1. Randomly generate m samples {z(1), z(2), ..., z(m)} from p(z) to form a bundle; take m samples from the real time series voltage feature map to form a bundle {(batch), {x(1), x(2), ..., x(m)}.
[0090] Step 2. Fix the parameters of the generator model G, and optimize the parameters of the discriminant model D using stochastic gradient descent (SGD) according to the following formula.
[0091]
[0092] in, The expression represents the application of gradient descent to optimize the parameters of the discriminant network D, and is the symbol for mathematical gradient calculation. In the formula, m is the number of samples extracted in this iteration, D is the discriminant network model, x represents the true time-series voltage feature map, i is the identifier of the 1st to mth samples, G is the generative model, and z represents the pseudo-time-series voltage feature map. The entire formula means optimizing the parameters through gradient descent: D(x) is the discriminant network, and x is the true sample; therefore, the value of D(x) input to the discriminant network should be as close to 1 as possible, i.e., ... The value of should be as small as possible, while G is the generator network that generates fake samples to confuse the discriminator network. G(zi) is a generated "fake sample" that is input into the D model, and the result should be as close to 0 as possible. To minimize the sum of the two squares, we input the samples with a batch size of m into the above formula and find the parameters of the discriminant network that minimize the sum of the two squares. This will improve the discriminant network's ability to make a decision.
[0093] Step 3. When k = K, iteration 2 ends;
[0094] Step 4. Randomly select m samples {z(1), z(2), ..., z(m)} from p(z) to form a bundle.
[0095] Step 5. Fix the discriminant model D, and update the parameters of the generator model G according to the following formula;
[0096]
[0097] in, This indicates that gradient calculation is used to optimize the parameters of the discriminant network G.
[0098] Step 6. When n = N, end iteration 1 and training is complete.
[0099] In this embodiment, the actual time-series voltage feature map corresponding to the period from battery power supply to power outage and power depletion is used as training sample. The model is trained in conjunction with a generative adversarial network. Based on the adversarial principle of generative adversarial network, the number of training samples is required to be small. The use of actual time-series voltage feature map as training sample ensures the accuracy of the trained battery discharge duration determination model.
[0100] S130, when the preset generative adversarial network training stopping condition is met, determine the battery discharge duration determination model.
[0101] In one embodiment, S130 may include:
[0102] When the preset training stopping condition of the generative adversarial network is met, the generative model is determined as the battery discharge duration determination model.
[0103] The preset stopping condition for generative adversarial network training can be set to the number of model training iterations. For example: Set iteration 1: number of generative adversarial network training iterations n = 1, ..., N; Set iteration 2: number of discriminant model training iterations k = 1, ..., K; When k = K, end iteration 2; when n = N, end iteration 1, training is complete, and the generative model is determined as the battery discharge time determination model.
[0104] This application embodiment uses real time-series voltage feature maps corresponding to the period from battery power supply to power outage and depletion of power as training samples, and uses generative adversarial networks for model training. Based on the adversarial principle of generative adversarial networks, the number of training samples is required to be small. Using real time-series voltage feature maps as training samples ensures the accuracy of the trained battery discharge duration determination model. Compared with the hardware and software construction and maintenance investment of online monitoring battery discharge performance systems, this application embodiment only requires a very small additional cost to obtain more accurate monitoring data, and the cost-effectiveness ratio is significantly higher.
[0105] Figure 1-2The training method for the battery discharge duration determination model is described below, in conjunction with the appendix. Figure 3 This application describes a method for determining battery discharge duration using a battery discharge duration determination model, as provided in its embodiments. The battery discharge duration determination model is trained using a battery discharge duration determination model training method. Figure 3 A flowchart illustrating a method for determining battery discharge time using a battery discharge time determination model is shown, such as... Figure 3 As shown, the method includes:
[0106] S310, Obtain a reference sample.
[0107] The reference sample is an incomplete timing voltage characteristic map corresponding to the period from battery power supply to power outage without the battery being depleted.
[0108] In one embodiment, S310 may include:
[0109] Acquire AC power alarm information, DC power alarm information, and voltage monitoring information;
[0110] The incomplete timing voltage characteristic map is determined based on AC alarm information, DC alarm information, and voltage monitoring information.
[0111] During operation, the equipment room generates environmental alarm data, including AC and DC power alarms. An AC power supply alarm indicates mains power supply, while an AC power outage alarm indicates a mains power outage, at which point the battery is activated. A DC power outage alarm indicates that the battery can no longer provide power, causing the battery to shut down and the equipment in the equipment room to shut down. A DC power supply alarm indicates that the battery is still supplying power. Therefore, based on the AC and DC power alarm information, the battery startup and shutdown times can be obtained. The base station equipment records the voltage monitoring information of the equipment room. By plotting curves based on the battery startup and shutdown times and the voltage monitoring information, an incomplete time-series voltage characteristic map can be determined.
[0112] This application embodiment can automatically generate an incomplete time-series voltage characteristic map by using AC power alarm information, DC power alarm information, and voltage monitoring information, without the need for manual data collection.
[0113] S320: Input the reference sample into the battery discharge duration determination model to perform feature expansion and obtain the complete time-series voltage feature map of the battery.
[0114] S330 determines the battery discharge duration based on the complete timing voltage characteristic diagram.
[0115] In one embodiment, the size of the incomplete timing voltage characteristic map is N×N; S330 may include:
[0116] The characteristic curves in the complete time-series voltage characteristic map are determined by traversing the complete time-series voltage characteristic map.
[0117] The battery discharge time is determined using the following formula:
[0118] T = X / N × M
[0119] Where T is the battery discharge time, X is the maximum time corresponding to the characteristic curve, and M is the preset maximum battery discharge time.
[0120] First, tools such as medium-range filters and morphological noise removers can be used to denoise the complete time-series voltage feature map. Then, the characteristic curves can be determined through the following steps:
[0121] a. Traverse the pixels in the complete time-series voltage feature map and find the first black point;
[0122] b. Find the next black point within the eight-neighborhood of the first black point;
[0123] c. Save the coordinates of the found point;
[0124] d. If no black point is found in the eight neighboring regions, it means that the search at one end is complete;
[0125] e. Repeat steps a through d to complete the search at the other end of the curve.
[0126] After completing the traversal, a shorter interference curve may appear. The curve containing the most pixels is selected as the feature curve.
[0127] The method for determining battery discharge duration using a battery discharge duration determination model provided in this application embodiment is trained using the battery discharge duration determination model training method of this application embodiment. Since the battery discharge duration determination model uses the real time-series voltage feature map corresponding to the period from battery power supply to power outage and power depletion as training samples, and is trained in conjunction with a generative adversarial network, based on the adversarial principle of generative adversarial network, the number of training samples is required to be small. Using the real time-series voltage feature map as training samples ensures the accuracy of the trained battery discharge duration determination model, thereby ensuring the accuracy of determining the battery discharge duration.
[0128] Figure 1-3 The method provided in the embodiments of this application is described below, in conjunction with the appendix. Figure 4-6 This application describes the apparatus provided in the embodiments.
[0129] Figure 4 This illustration shows a schematic diagram of the structure of a training device for a battery discharge duration determination model provided in one embodiment of this application. Figure 4 Each module in the device shown has the ability to implement Figure 1 The functions of each step in the process are defined, and their corresponding technical effects are achieved. For example... Figure 4 As shown, the device may include:
[0130] The acquisition module 410 is used to acquire training samples; the training samples are real time-series voltage feature maps corresponding to the period from battery power supply to power supply stop and power depletion;
[0131] Input module 420 is used to input real time-series voltage feature maps into the generative adversarial network for training, so as to update the parameters of the generative adversarial network;
[0132] The determination module 430 is used to determine the battery discharge duration determination model when the preset generative adversarial network training stopping conditions are met.
[0133] In this embodiment, the actual time-series voltage feature map corresponding to the period from battery power supply to power outage and power depletion is used as training sample. The model is trained in conjunction with a generative adversarial network. Based on the adversarial principle of generative adversarial network, the number of training samples is required to be small. The use of actual time-series voltage feature map as training sample ensures the accuracy of the trained battery discharge duration determination model.
[0134] In one embodiment, the acquisition module 410 is specifically used for:
[0135] Acquire AC power alarm information, DC power alarm information, and voltage monitoring information;
[0136] The true timing voltage characteristic diagram is determined based on AC alarm information, DC alarm information, and voltage monitoring information.
[0137] In one embodiment, the generative adversarial network includes: a generative model and a discriminative model;
[0138] Input module 420 is specifically used for:
[0139] Randomly generate a random feature map of the same size as the actual time-series voltage feature map;
[0140] The random feature map is input into the generation model for feature expansion, which is used to generate pseudo-time series voltage feature maps;
[0141] The real and pseudo time series voltage feature maps are input into the discrimination model to distinguish between them and obtain the true and false discrimination result features.
[0142] The parameters of the generative adversarial network are updated based on the characteristics of the true / false discrimination results.
[0143] In one embodiment, the determining module 430 is specifically used for:
[0144] When the preset training stopping condition of the generative adversarial network is met, the generative model is determined as the battery discharge duration determination model.
[0145] This application embodiment uses real time-series voltage feature maps corresponding to the period from battery power supply to power outage and depletion of power as training samples, and uses generative adversarial networks for model training. Based on the adversarial principle of generative adversarial networks, the number of training samples is required to be small. Using real time-series voltage feature maps as training samples ensures the accuracy of the trained battery discharge duration determination model. Compared with the hardware and software construction and maintenance investment of online monitoring battery discharge performance systems, this application embodiment only requires a very small additional cost to obtain more accurate monitoring data, and the cost-effectiveness ratio is significantly higher.
[0146] Figure 5 This illustration shows a schematic diagram of a device for determining battery discharge time using a battery discharge time determination model, according to an embodiment of this application. The battery discharge time determination model is trained using a battery discharge time determination model training device. Figure 5 Each module in the device shown has the ability to implement Figure 3 The functions of each step in the process are defined, and their corresponding technical effects are achieved. For example... Figure 5 As shown, the device may include:
[0147] The acquisition module 510 is used to acquire a reference sample, which is an incomplete timing voltage feature map corresponding to the period from battery power supply to power supply stop and before the power is depleted.
[0148] The input module 520 is used to input the reference sample into the battery discharge duration determination model for feature expansion, so as to obtain the complete time-series voltage feature map of the battery.
[0149] The determination module 530 is used to determine the battery discharge duration based on the complete timing voltage characteristic map.
[0150] This application embodiment describes a method for determining battery discharge duration using a battery discharge duration determination model. The applied battery discharge duration determination model is trained using the training method described in this application embodiment. Since this battery discharge duration determination model uses real time-series voltage feature maps corresponding to the period from battery power supply to power outage and depletion of power as training samples, and is trained in conjunction with a generative adversarial network, based on the adversarial principle of generative adversarial networks, a small number of training samples are required. Using real time-series voltage feature maps as training samples ensures the accuracy of the trained battery discharge duration determination model, thereby ensuring the accuracy of determining the battery discharge duration.
[0151] In one embodiment, the size of the incomplete timing voltage feature map is N×N.
[0152] Module 530 is specifically used for:
[0153] The characteristic curves in the complete time-series voltage characteristic map are determined by traversing the complete time-series voltage characteristic map.
[0154] The battery discharge time is determined using the following formula:
[0155] T = X / N × M
[0156] Where T is the battery discharge time, X is the maximum time corresponding to the characteristic curve, and M is the preset maximum battery discharge time.
[0157] In one embodiment, the acquisition module 510 is specifically used for:
[0158] Acquire AC power alarm information, DC power alarm information, and voltage monitoring information.
[0159] The incomplete timing voltage characteristic map is determined based on AC alarm information, DC alarm information, and voltage monitoring information.
[0160] This application embodiment describes a method for determining battery discharge duration using a battery discharge duration determination model. The applied battery discharge duration determination model is trained using the training method described in this application embodiment. Since this battery discharge duration determination model uses real time-series voltage feature maps corresponding to the period from battery power supply to power outage and depletion of power as training samples, and is trained in conjunction with a generative adversarial network, based on the adversarial principle of generative adversarial networks, a small number of training samples are required. Using real time-series voltage feature maps as training samples ensures the accuracy of the trained battery discharge duration determination model, thereby ensuring the accuracy of determining the battery discharge duration.
[0161] Figure 6 A schematic diagram of the structure of an electronic device according to an embodiment of this application is shown. Figure 6 As shown, the device may include a processor 601 and a memory 602 storing computer program instructions.
[0162] Specifically, the processor 601 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.
[0163] Memory 602 may include mass storage for data or instructions. For example, and not limitingly, memory 602 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. In one instance, memory 602 may include removable or non-removable (or fixed) media, or memory 602 may be non-volatile solid-state memory. Memory 602 may be internal or external to the integrated gateway disaster recovery device.
[0164] In one instance, memory 602 may be read-only memory (ROM). In one instance, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or flash memory, or a combination of two or more of these.
[0165] The processor 601 reads and executes the computer program instructions stored in the memory 602, using the method provided in the embodiments of this application, and achieves the corresponding technical effects achieved by the method provided in the embodiments of this application. For the sake of brevity, the details will not be repeated here.
[0166] In one example, the electronic device may also include a communication interface 603 and a bus 610. Wherein, as... Figure 6 As shown, the processor 601, memory 602, and communication interface 603 are connected through bus 610 and complete communication with each other.
[0167] The communication interface 603 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.
[0168] Bus 610 includes hardware, software, or both, that couples components of an online data traffic metering device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 610 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, this application contemplates any suitable bus or interconnect.
[0169] The electronic device can execute the training method of the battery discharge duration determination model in the embodiments of this application, thereby achieving the corresponding technical effects of the method described in this application.
[0170] Furthermore, in conjunction with the training method for the battery discharge duration determination model in the above embodiments, this application embodiment can provide a computer storage medium for implementation. This computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the battery discharge duration determination model training methods in the above embodiments.
[0171] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.
[0172] The functional blocks shown in the above-described block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.
[0173] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.
[0174] The aspects of this application have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by dedicated hardware performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.
[0175] The above description is merely a specific implementation of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.
Claims
1. A method for determining battery discharge time using a battery discharge time determination model, characterized in that, The method includes: Obtain a reference sample, which is an incomplete time-series voltage feature map corresponding to the period from battery power supply to power supply stop and the power is not depleted, and the size of the incomplete time-series voltage feature map is N×N; The reference sample is input into the battery discharge duration determination model for feature expansion to obtain the complete time-series voltage feature map of the battery. The characteristic curves in the complete time-series voltage characteristic map are determined by traversing the complete time-series voltage characteristic map. The battery discharge time is determined by the following formula: T = X / N × M, where T is the battery discharge time, X is the maximum time corresponding to the characteristic curve, and M is the preset maximum battery discharge time. The training process of the battery discharge duration determination model includes: Obtain training samples; the training samples are real time-series voltage feature maps corresponding to the period from battery power supply to power outage and depletion of power. Randomly generate a random feature map of the same size as the actual time-series voltage feature map; The random feature map is input into the generative model of the generative adversarial network for feature expansion to generate a pseudo-time series voltage feature map. The generative adversarial network includes a generative model and a discriminative model. The true time-series voltage feature map and the pseudo time-series voltage feature map are input into the discrimination model to perform true and false discrimination, and the true and false discrimination result features are obtained. The parameters of the generative adversarial network are updated based on the characteristics of the authenticity discrimination results; When the preset training stop condition of the generative adversarial network is met, the generative model is determined as the battery discharge duration determination model. 2.The method of claim 1, wherein, The acquisition of the reference sample includes: Acquire AC power alarm information, DC power alarm information, and voltage monitoring information; The incomplete timing voltage characteristic map is determined based on the AC power alarm information, the DC power alarm information, and the voltage monitoring information. 3.The method of claim 1, wherein the determining the battery discharge duration using the battery discharge duration determination model comprises: determining the battery discharge duration using the battery discharge duration determination model based on the battery state of charge, the battery temperature, and the battery discharge rate. The acquisition of training samples includes: Acquire AC power alarm information, DC power alarm information, and voltage monitoring information; The true timing voltage characteristic map is determined based on the AC power alarm information, the DC power alarm information, and the voltage monitoring information.
4. A device for determining battery discharge time using a battery discharge time determination model, characterized in that, The device includes: The acquisition module is used to acquire a reference sample, which is an incomplete time-series voltage feature map corresponding to the period from battery power supply to power supply stop and the power is not depleted, and the size of the incomplete time-series voltage feature map is N×N; The input module is used to input the reference sample into the battery discharge duration determination model for feature expansion to obtain the complete time-series voltage feature map of the battery; The determination module is used to determine the characteristic curve in the complete time-series voltage characteristic map by traversing the complete time-series voltage characteristic map; and to determine the battery discharge time using the following formula: T = X / N × M; where T is the battery discharge time, X is the maximum time corresponding to the characteristic curve, and M is the preset maximum battery discharge time. The training process of the battery discharge duration determination model includes: The acquisition module is used to acquire training samples; the training samples are real time-series voltage feature maps corresponding to the period from battery power supply to power supply stop and power depletion; The input module is used to input the real time-series voltage feature map into the generative adversarial network for training, so as to update the model parameters of the generative adversarial network; A determination module is used to randomly generate a random feature map of the same size as the real time-series voltage feature map; input the random feature map into the generative model of a generative adversarial network (GAN) for feature expansion to generate a pseudo-time-series voltage feature map, wherein the GAN includes a generative model and a discriminative model; input the real time-series voltage feature map and the pseudo-time-series voltage feature map into the discriminative model for true / false discrimination to obtain true / false discrimination result features; update the parameters of the GAN based on the true / false discrimination result features; when a preset GAN training stopping condition is met, the generative model is determined as the battery discharge duration determination model.
5. The apparatus for determining the battery discharge duration using the battery discharge duration determination model according to claim 4, wherein, The acquisition module is specifically used for: Acquire AC power alarm information, DC power alarm information, and voltage monitoring information; The incomplete timing voltage characteristic map is determined based on the AC power alarm information, the DC power alarm information, and the voltage monitoring information.
6. The apparatus of claim 4, wherein, The acquisition module is specifically used for: Acquire AC power alarm information, DC power alarm information, and voltage monitoring information; The true timing voltage characteristic map is determined based on the AC power alarm information, the DC power alarm information, and the voltage monitoring information.
7. An electronic device, comprising: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the method as described in any one of claims 1 to 3.
8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an implementation program for information transmission, which, when executed by a processor, implements the method as described in any one of claims 1 to 3.