An abnormal use alarm system for LED floor lamp
Through the abnormal alarm system of LED floor lamp, the data acquisition and processing module is used to monitor temperature and brightness changes, and combined with phase space reconstruction and prediction network, abnormal detection and early warning of LED floor lamps is realized, and the high temperature of LED floor lamps is solved The material damage problem ensures stability of lighting requirements and extended equipment life.
Patent Information
- Application Number
- CN202210461758.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-28
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2042-04-28
AI Technical Summary
During long-term use, Led floor lamps are damaged due to high temperatures, which affects lighting needs and service life, and it is difficult for the existing technology to effectively monitor and early warning.
An abnormal alarm system for using LED ground lamps was designed. The temperature and brightness data were collected through the data acquisition module, and the temperature and brightness change trends were obtained by using the data processing module. Combined with the phase space reconstruction method and the abnormal prediction network, abnormal data were screened out and Training is performed to generate prediction networks to achieve real-time alarms.
It realizes fast and accurate abnormality detection and early warning of LED floor lamps, ensures the meeting of lighting needs, and extends the service life of LED floor lamps.
Smart Images

Figure CN114925753B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of data processing, and in particular to an abnormal use alarm system for an LED floor lamp. Background Art
[0002] LED floor lamps are widely used in daily life. For some scenes with high lighting demand, such as the entrance of a community or a shopping mall, LED floor lamps need to work all night. During operation, a large amount of heat will be generated, which will cause damage to the materials of the LED floor lamps and shorten the life of the LED lamps. If the quality of the LED lamps is damaged, it will affect the lighting demand in the scene and affect the user experience. Summary of the invention
[0003] In order to solve the above technical problems, the purpose of the present invention is to provide an abnormal use alarm system for LED floor lamps, and the technical solutions adopted are as follows:
[0004] The present invention proposes an abnormal use alarm system for LED floor lamps, the system comprising:
[0005] The data acquisition module is used to collect the temperature of the energy-saving LED floor lamp, the temperature outside the lampshade, and the brightness of the LED floor lamp;
[0006] The first data processing module is used to obtain the temperature change trend of the LED floor lamp based on the relationship between the temperature and time of each LED floor lamp; take the temperature change trend of the LED floor lamp on the first night of use as the standard temperature change trend; similarly obtain the brightness change trend and the standard brightness change trend; obtain an abnormal evaluation index according to the difference between the temperature change trend and the standard temperature change trend every night, and the difference between the brightness change trend and the standard brightness change trend, and obtain an abnormal evaluation index sequence;
[0007] The second data processing module is used to use the phase space reconstruction processing method to obtain the phase space convergence domain for the abnormal evaluation index sequence; when the convergence domain distribution probability obtained by the phase space tracking value is greater than the preset threshold, it means that the LED floor lamp data at this time is normal data; otherwise, it is abnormal data; obtain a normal data set and an abnormal data set;
[0008] The third data processing module is used to obtain a second abnormal index according to the temperature difference between the LED floor lamp in the abnormal data set and the temperature outside the lampshade; abnormal data whose second abnormal index is greater than a preset abnormal index threshold is used as abnormal training data, and the abnormal training data and the normal data set are used as training data;
[0009] The network generation module is used to obtain an abnormal prediction network through training according to the training data; the real-time temperature of the LED floor lamp, the temperature outside the lampshade, and the brightness of the LED floor lamp are included in the abnormal prediction network, and the corresponding real-time abnormal evaluation indicators are output; and whether to alarm is determined according to the real-time abnormal evaluation indicators.
[0010] Furthermore, obtaining the temperature change trend of each LED floor lamp based on the relationship between the temperature and time of each LED floor lamp includes:
[0011] Set the optimized function model y=ax b , where y is the center temperature of the LED array, x is the time corresponding to the temperature, and a and b are fitting parameters;
[0012] Set the loss function and make it approach 0 to get the optimal function model. The loss function includes:
[0013]
[0014] in, For a combination of a and b, we get the temperature corresponding to the time in the model, y n is the actual temperature corresponding to the time.
[0015] Furthermore, the abnormal evaluation index obtained according to the difference between the temperature change trend and the standard temperature change trend, and the difference between the brightness change trend and the standard brightness change trend every night includes:
[0016]
[0017] Among them, r is the abnormal evaluation index, PPMCC() is the Pearson coefficient calculation function, S standard is the standard temperature variation trend, S actual is the temperature change trend, B actual is the brightness change trend, B standard is the standard brightness change trend, STD() is the standard deviation calculation function, and DTW() is the DTW distance calculation function.
[0018] Furthermore, the phase space convergence domain is obtained by using the phase space reconstruction processing method for the abnormal evaluation index sequence, including:
[0019] Obtain the abnormal evaluation index sequence S = {s1,…,s N};
[0020] According to the embedding theorem, the reconstructed phase space can be obtained as:
[0021] Where: Nm is the number of phase points in the reconstructed phase space, and Nm = N-(m-1)τ; m is the embedding dimension; τ is the delay time.
[0022] The delay time τ is selected by the autocorrelation function method, and the autocorrelation function is selected:
[0023] The τ corresponding to the first zero crossing point of the autocorrelation function is taken as the delay time, where j = 1, ..., m-1;
[0024] Calculate the phase point at the next moment in the phase space: Calculate the Euclidean distance di between each phase point and the central phase point FS1 in the phase space = || FS i -FS C ||, find out FS C The reference vector set RS(C) = {RS1,…,RS N}, the phase point FS′ at the next moment in the phase space is calculated by RS(C) using the following formula: C :
[0025]
[0026] Among them, d min is the spatial distance and minimum distance from each point in the neighborhood to the center point, that is, min{di=||FS i -FS C ||}, L is the prediction adjustment parameter, generally L≥1;
[0027] And the phase point FS′ at the next moment C The phase difference is obtained by subtracting the true value as the tracking index p, which is put into the normal distribution model to obtain the convergence domain
[0028] Further, obtaining a second abnormality index according to the difference between the temperature of the LED floor lamp in the abnormal data set and the temperature outside the lampshade includes:
[0029]
[0030] Among them, u is the second abnormal index, D actural is the LED lamp temperature set, T actural is the temperature set outside the lampshade; the distance() function calculates the difference between the two sets in sequence, one by one, to obtain the difference set.
[0031] Furthermore, the training according to the training data to obtain the abnormality prediction network includes:
[0032] Use the training data to train the TCN network.
[0033] The present invention has the following beneficial effects:
[0034] The embodiment of the present invention obtains the temperature change trend and brightness change trend of the LED floor lamp every night through a data fitting method, and obtains the abnormal evaluation index every night based on the difference from the standard change trend. The phase space reconstruction method is further used to determine the abnormal data and normal data in the abnormal evaluation index sequence. The abnormal data is further screened according to the temperature information and brightness information corresponding to the abnormal data, and training data with strong reference is obtained for training the abnormal prediction network. According to the abnormal prediction network, the predicted abnormal evaluation index of the future moment can be quickly and accurately obtained, and the staff can be warned in time according to the predicted abnormal evaluation index, so that the LED floor lamp can be replaced in sufficient time to meet the lighting needs every night. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0036] Figure 1 A block diagram of an abnormal use alarm system for an LED floor lamp provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0037] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation, structure, features and effects of an abnormal use alarm system of an LED floor lamp proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures, or characteristics in one or more embodiments may be combined in any suitable form.
[0038] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
[0039] The specific scheme of the abnormal use alarm system of the LED floor lamp provided by the present invention is described in detail below with reference to the accompanying drawings.
[0040] See also Figure 1 , which shows a block diagram of an abnormal use alarm system for a LED floor lamp provided by an embodiment of the present invention; the system includes: a data acquisition module, a first data processing module, a second data processing module, a third data processing module and a network production module.
[0041] The data acquisition module is used to collect the temperature of the energy-saving LED floor lamp, the temperature outside the lampshade, and the brightness of the LED floor lamp. Specifically includes:
[0042] 1) Collect the temperature of each LED floor lamp of the energy-saving lamp.
[0043] a) Since LED lamps are composed of light-emitting and heat-dissipating components, the main heat dissipation path of the lamp is active heat dissipation from the heat dissipation components. The heat dissipation components conduct the temperature inside the LED lamp out of the lamp.
[0044] b) The floor lamp LEDs are arranged in an array. The heat concentration center is in the center of the array, that is, the temperature in the center is the highest temperature of the entire LED lamp. Therefore, the probe is placed in the center of the array to monitor the core temperature.
[0045] c) Obtain the temperature information of the center of the LED floor lamp array, with a sampling frequency of once per second.
[0046] 2) Collect the temperature outside the lampshade.
[0047] a) Ambient temperature can be expressed as the temperature outside the lampshade. The purpose of collecting ambient temperature is to observe the effect of temperature changes in the environment on heat dissipation. In order to be waterproof, the lampshade has a better sealing effect. At the same time, the floor lamp cannot dissipate heat through the air like a normal LED lamp. Therefore, the heat dissipation effect of the floor lamp is relatively poor and can only rely more on the radiator for heat dissipation. Therefore, the failure rate of the floor lamp is relatively high.
[0048] b) However, if the temperature difference between the inside and outside of the lampshade is large, heat dissipation can be performed. Heat conduction inside the lampshade is carried out, and the heat dissipation process exists. The heat dissipation process will increase as the temperature difference between the inside and outside of the lampshade increases. Therefore, it is necessary to collect the temperature difference between the inside and outside of the lampshade.
[0049] c) Install a temperature sensor near the floor lamp and collect ambient temperature information t with a sampling frequency of once per second.
[0050] 3) Collect the brightness of LED lights.
[0051] a) Since the brightness of the LED floor lamp is determined by the quality of each electronic component, the quality change of each electronic component is due to the different thermal expansion coefficients of the LED lamp components. When the temperature continues to rise, the temperature affects the thermal expansion coefficient of the original parts, which can easily cause bending or damage of a single software, and then cause small cracks at the connection, affecting the overall use effect of the LED lamp due to the rapid temperature rise.
[0052] b) Too fast a temperature rise causes the packaging equipment to age faster, and the continuous high temperature changes the properties of the material itself, reducing the original degree of adhesion, and then brittle, unable to provide a good power barrier effect, affecting the output of the LED light source. Once the material changes, the high temperature generated will continue to rise. Under the influence of long-term high temperature, other internal components will continue to expand. After the movement of thermal expansion and contraction, the properties of the material itself will also change, affecting the overall performance and greatly shortening the service life.
[0053] c) A brightness sensing module is installed in the lampshade of the LED floor lamp. The module is composed of a brightness sensor to monitor the real-time brightness of the LED lamp. The brightness of the LED lamp is obtained with a sampling frequency of once per second.
[0054] Since the effect of lamp temperature on brightness has no research significance within a time point, the unit time is set to every night, that is, the working time of the floor lamp every night. First of all, the temperature change of LED lamps is similar under the condition of constant quality, so it is of practical significance to study the linear similarity of the temperature change of LED floor lamps every night.
[0055] The first data processing module is used to obtain the temperature change trend of each LED floor lamp based on the relationship between the temperature and time of each LED floor lamp; take the temperature change trend of the first night of use of the LED floor lamp as the standard temperature change trend; similarly obtain the brightness change trend and the standard brightness change trend. Specifically including:
[0056] Set the optimized function model y=ax b , where y is the center temperature of the LED array, x is the time corresponding to the temperature, and a and b are fitting parameters;
[0057] Set the loss function and make it approach 0 to get the optimal function model. The loss function includes:
[0058]
[0059] in, For a combination of a and b, we get the temperature corresponding to the time in the model, y n is the actual temperature corresponding to the time.
[0060] For iterative training, the iterative loss function is set to approach 0, that is, when the loss function is equal to 0, the optimal combination of a and b is obtained, corresponding to the optimal function model, that is, corresponding to the temperature change trend or the brightness change trend.
[0061] i) The advantages of this fitting method are:
[0062] 1) First, the optimal change law of the temperature data of an LED lamp working overnight can be determined.
[0063] 2) This process will ignore the impact of individual function data on the overall function, ensuring that the model is consistent with the actual temperature changes that night.
[0064] 3) The model expresses the overall temperature change of the LED that night.
[0065] The first data processing module obtains an abnormal evaluation index according to the difference between the temperature change trend and the standard temperature change trend every night, and the difference between the brightness change trend and the standard brightness change trend. The abnormal evaluation index sequence obtained specifically includes:
[0066]
[0067] Among them, r is the abnormal evaluation index, PPMCC() is the Pearson coefficient calculation function, S standard is the standard temperature variation trend, S actual is the temperature change trend, B actual is the brightness change trend, B standard is the standard brightness change trend, STD() is the standard deviation calculation function, and DTW() is the DTW distance calculation function.
[0068] Since the fitting of the temperature change trend is holistic, the points on the change trend may not come from the actual temperature value, so the high-precision change similarity of the Pearson coefficient can reduce the accuracy loss of the model. Further, the DTW distance is used in the second item. The advantage of DTW is that it can ignore the small-range fluctuations in brightness caused by voltage instability, and this small-range fluctuation does not affect the final evaluation value.
[0069] Each night corresponds to an abnormal evaluation index, and the abnormal evaluation indicators of multiple nights constitute an abnormal evaluation indicator sequence.
[0070] The second data module is used to obtain the phase space convergence domain by using the phase space reconstruction processing method for the abnormal evaluation index sequence, which specifically includes:
[0071] Obtain the abnormal evaluation index sequence S = {s1,…,s N};
[0072] According to the embedding theorem, the reconstructed phase space can be obtained as:
[0073] Where: Nm is the number of phase points in the reconstructed phase space, and Nm = N-(m-1)τ; m is the embedding dimension; τ is the delay time.
[0074] The delay time τ is selected by the autocorrelation function method, and the autocorrelation function is selected:
[0075] The τ corresponding to the first zero crossing point of the autocorrelation function is taken as the delay time, where j = 1, ..., m-1;
[0076] Calculate the phase point at the next moment in the phase space: Calculate the Euclidean distance di between each phase point and the central phase point FS1 in the phase space = || FS i -FS C ||, find out FS C The reference vector set RS(C) = {RS1,…,RS N}, the phase point FS′ at the next moment in the phase space is calculated by RS(C) using the following formula: C :
[0077]
[0078] Among them, d min is the spatial distance and minimum distance from each point in the neighborhood to the center point, that is, min{di=||FS i -FS C ||}, L is the prediction adjustment parameter, and L≥1;
[0079] And the phase point FS′ at the next moment C The phase difference is obtained by subtracting the true value as the tracking index p, which is put into the normal distribution model to obtain the convergence domain
[0080] When the convergence domain distribution probability obtained by the phase space tracking value is greater than the preset threshold, it means that the LED floor lamp data at this time is normal data; otherwise, it is abnormal data. A normal data set and an abnormal data set are obtained.
[0081] In the embodiment of the present invention, the distribution probability threshold of the convergence domain is set to 97%, that is, when the convergence domain distribution probability obtained by the phase space tracking value is greater than 97%, it means that the LED floor lamp set has not suffered a large loss and can still be used normally, and the corresponding data is normal data. Otherwise, it means that the LED lamp may have quality problems, and the corresponding data is abnormal data.
[0082] Based on the obtained phase space grouping, when the convergence domain distribution probability of the phase space tracking value obtained the previous night is less than 97%, it is necessary to distinguish a situation, that is, when the temperature difference between the inside and outside is too small, the heat dissipation efficiency is reduced. At this time, there is no problem with heat dissipation. Therefore, even if the convergence domain is large, it cannot be considered that the LED floor lamp group has a large loss at this time. Therefore, it is necessary to use the third data processing module to further screen the abnormal data.
[0083] The third data processing module is used to obtain a second abnormality index according to the temperature difference between the LED floor lamp in the abnormal data set and the temperature outside the lampshade, specifically including:
[0084]
[0085] Among them, u is the second abnormal index, D actural is the LED lamp temperature set, T actural is the temperature set outside the lampshade; the distance() function calculates the difference between the two sets in order, one by one, to obtain the difference set. The numerator is to obtain the maximum temperature difference of the whole night, representing the best heat exchange efficiency of the night. The denominator is to obtain the stability of the temperature difference throughout the night, and the stability of the heat exchange throughout the night.
[0086] The threshold is set to evaluate the abnormal temperature difference between the inside and outside of the lampshade of the LED lamp that night, which can roughly represent the efficiency of the internal and external heat exchange. Based on such efficiency evaluation, when it is greater than the threshold, the data of that night is not universal and cannot be used as the following training data, and is directly discarded in the life training. When it is less than the threshold, it is considered that the value of that night is universal and has a relatively high reliability. Therefore, the abnormal data whose second abnormal index is greater than the preset abnormal index threshold is used as abnormal training data, and the abnormal training data and the normal data set are used as training data.
[0087] The network generation is used to obtain an abnormal prediction network through training according to the training data. The abnormal prediction network adopts a TCN network structure. In the embodiment of the present invention, the training process includes:
[0088] i) Input the training data into TCN training, use the value at the next moment as the label, and repeat the training.
[0089] ii) The loss function of TCN is mean square error loss
[0090] 1) For the training data samples of the normal data set, the mean square error loss function can be directly used for training.
[0091] 2) For the sample sequence of abnormal training data, since the source of this group of data has been screened multiple times, the normalized second abnormal index can be used as the quality score of the mean square error loss function.
[0092]
[0093] Among them C z is the normalized quality score, used as the loss weight, and loss is the loss function of the sample. is the prediction sample, y i is a feature sample.
[0094] a) The purpose of training is to ensure that the loss function converges. Through continuous training, the loss becomes smaller and the predicted trend is accurate.
[0095] The temperature of the real-time LED floor lamp, the temperature outside the lampshade, and the brightness of the LED floor lamp belong to the abnormal prediction network, and the corresponding predicted abnormal evaluation index is output. The predicted abnormal evaluation index is monitored, and the predicted time is recorded when the predicted abnormal evaluation index reaches the set threshold range. Since the life of an ordinary LED lamp is 10 years, the difference between the predicted service life and the target life is compared as the expected life. The implementer can set the alarm value of the expected life according to his own needs. If the expected life is less than the alarm value, an alarm is issued for maintenance.
[0096] In summary, the embodiment of the present invention collects temperature data and brightness data of the LED floor lamp, and collects temperature information outside the lampshade as ambient temperature data. An abnormal evaluation index is obtained based on the difference between the temperature data and ambient temperature data of the LED floor lamp and the standard data. The evaluation index sequence is processed once according to the phase space reconstruction processing method to obtain the corresponding abnormal data. A second abnormal index is further obtained for the temperature difference between the floor lamp and the environment in the abnormal data, which is used to filter out abnormal training data. The abnormal prediction network is trained based on normal data and abnormal training data. The embodiment of the present invention trains the abnormal prediction network through data with strong reference value, which is used to predict abnormal situations in the future, thereby realizing effective abnormal alarm for the LED floor lamp and facilitating the handling of staff.
[0097] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0098] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.
[0099] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. An abnormal use alarm system for LED floor lamps, characterized in that: The system comprises: The data acquisition module is used to collect the temperature of the energy-saving LED floor lamp, the temperature outside the lampshade, and the brightness of the LED floor lamp; The first data processing module is used to obtain the temperature change trend of the LED floor lamp based on the relationship between the temperature and time of each LED floor lamp; take the temperature change trend of the LED floor lamp on the first night of use as the standard temperature change trend; similarly obtain the brightness change trend and the standard brightness change trend; obtain an abnormal evaluation index according to the difference between the temperature change trend and the standard temperature change trend every night, and the difference between the brightness change trend and the standard brightness change trend, and obtain an abnormal evaluation index sequence; The second data processing module is used to use the phase space reconstruction processing method to obtain the phase space convergence domain for the abnormal evaluation index sequence; when the convergence domain distribution probability obtained by the phase space tracking value is greater than the preset threshold, it means that the LED floor lamp data at this time is normal data; otherwise, it is abnormal data; obtain a normal data set and an abnormal data set; The third data processing module is used to obtain a second abnormal index according to the temperature difference between the LED floor lamp in the abnormal data set and the temperature outside the lampshade; abnormal data whose second abnormal index is greater than a preset abnormal index threshold is used as abnormal training data, and the abnormal training data and the normal data set are used as training data; The network generation module is used to obtain an abnormal prediction network through training according to the training data; input the real-time temperature of the LED floor lamp, the temperature outside the lampshade, and the brightness of the LED floor lamp into the abnormal prediction network, and output the corresponding predicted abnormal evaluation index; determine whether to alarm according to the predicted abnormal evaluation index; The abnormal evaluation index obtained according to the difference between the temperature change trend and the standard temperature change trend every night, and the difference between the brightness change trend and the standard brightness change trend, includes: Among them, r is the abnormal evaluation index, PPMCC() is the Pearson coefficient calculation function, S standard is the standard temperature variation trend, S actual is the temperature change trend, B actual is the brightness change trend, B standard is the standard brightness change trend, STD() is the standard deviation calculation function, DTW() is the DTW distance calculation function; The obtaining of the second abnormality index according to the difference between the temperature of the LED floor lamp in the abnormal data set and the temperature outside the lampshade comprises: Among them, u is the second abnormal index, D actural is the LED lamp temperature set, T actural is the temperature set outside the lampshade; the distance() function calculates the difference between the two sets in sequence, one by one, to obtain the difference set.
2. The abnormal use alarm system of LED floor lamp according to claim 1, characterized in that: The method of obtaining the temperature change trend of each LED floor lamp based on the relationship between the temperature and time of each LED floor lamp includes: Set the optimized function model y=ax b , where y is the center temperature of the LED array, x is the time corresponding to the temperature, and a and b are fitting parameters; Set the loss function and make it approach 0 to get the optimal function model. The loss function includes: in, For a combination of a and b, we get the temperature corresponding to the time in the model, y n is the actual temperature corresponding to the time.
3. The abnormal use alarm system of an LED floor lamp according to claim 1, characterized in that: The method of using the phase space reconstruction processing method to process the abnormal evaluation index sequence to obtain the phase space convergence domain includes: Obtain the abnormal evaluation index sequence S = {s1,…,s N }; According to the embedding theorem, the reconstructed phase space can be obtained as: Where: Nm is the number of phase points in the reconstructed phase space, and Nm = N-(m-1)τ; m is the embedding dimension; τ is the delay time; The delay time τ is selected by the autocorrelation function method, and the autocorrelation function is selected: The τ corresponding to the first zero crossing point of the autocorrelation function is taken as the delay time, where j = 1, ..., m-1; Calculate the phase point at the next moment in the phase space: Calculate the FS from each phase point to the central phase point in the phase space C The Euclidean distance between them is di=||FS i -FS C ||, find out FS C The reference vector set RS(C) = {RS1,…,RS N }, the phase point FS at the next moment in the phase space is calculated by RS(C) using the following formula: ′ C : Among them, d min is the spatial distance and minimum distance from each point in the neighborhood to the center point, that is, min{di=||FS i -FS C ||}, L is the prediction adjustment parameter, and L≥1; And the next moment's phase point FS ′ C The phase difference is obtained by subtracting the true value as the tracking index p, which is put into the normal distribution model to obtain the convergence domain 4. The abnormal use alarm system of an LED floor lamp according to claim 1, characterized in that: The training according to the training data to obtain the abnormal prediction network includes: The anomaly prediction network adopts the TCN network structure.
Citation Information
Patent Citations
Turbine set online fault early warning method based on abnormality searching and combination forecasting
CN103793601A
Life prediction method, life prediction program, and life prediction device
CN105849798A
Method for predicting service life of high-power LED lamp based on artificial neural network
CN107944190A
LED life prediction method for BP neural network based on genetic algorithm
CN111814401A
Emergency LED street lamp control method and system for intelligent street lamps
CN111970785A
