Lamp maintenance device and method

Abstract

The present invention provides a predictive maintenance device and method for lamps, the device comprising: an electrical box, the electrical box comprising: a digital lightning arrester, a lighting information collector, an AC contactor, an AC / DC power supply; the lighting information collector comprising: a temperature and humidity sensor, an inclination sensor, a first NTC and a first thermocouple; the LED light source comprising: a lighting module, a light intensity sensor, a temperature and humidity sensor, a second NTC and a second thermocouple. The predictive maintenance device and method for lamps of the present invention are used to monitor various parameter data of the lamps, power grids and use environment in real time, send them to a cloud server, use artificial intelligence technology to perform data analysis, create high-quality, multi-dimensional lighting data sets, combine expert experience with data-driven, form industrial intelligent algorithm innovations such as abnormality detection, fault diagnosis, and life prediction, and form capabilities such as precise illumination control, rapid and precise maintenance, and predictive maintenance.

Description

Technical Field

[0001] The present invention relates to the technical field of lamp maintenance, and in particular to a lamp maintenance device and method. Background Art

[0002] With the improvement of living standards and quality, the requirements for lighting are getting higher and higher. The traditional patrol inspection street lamp maintenance mode: that is, the mode of discovering, repairing and replacing lamps after they fail can no longer meet the requirements. Even if some single lamp controllers remotely monitor in real time instead of patrol inspection, they can only alarm for damaged faulty lamps. Especially in some occasions with high requirements for lighting continuity, people are more eager to get early warnings for lamps that are about to be damaged or whose illumination is about to fail to meet the requirements, and automatically inform users when and what the problems are, so as to prepare spare parts in advance and formulate maintenance plans.

[0003] For example, at night, street lighting largely determines the safety of ground vehicles and pedestrians. The current street lighting system failure not only affects the stability of the system, but also some high-altitude or coastal street lighting systems. The lines and lamps are exposed to harsh outdoor environments, subjected to a lot of wind and sun, cold and hot shocks or salt spray corrosion, and are prone to accelerated aging of equipment and lines. At the same time, lamps are susceptible to lightning strikes, haze and acid rain and other severe weather tests. When subjected to large-scale mechanical equipment activities, they are prone to vibration, tilt or leakage and other failures. At present, the maintenance of street lighting systems is mostly combined with daily inspections, weekly inspections, monthly inspections and irregular inspections. The inspection of the lighting system mainly relies on manual labor. The traditional manual inspection method is time-consuming and labor-intensive, and cannot effectively respond to the operation and maintenance management needs of lighting equipment. Similarly, sudden failures of lighting fixtures at airports, ports and terminals may cause aircraft to fail to take off and land normally, cargo to be loaded and unloaded in time, and ships to be unable to enter and exit ports normally, seriously affecting airport safety and port terminal throughput, causing major life safety and economic losses.

[0004] The commonly used maintenance methods and their shortcomings in the market are as follows:

[0005] 1. Restorative maintenance: Fix the lamps when they break. Unplanned downtime requires a large number of spare parts, expensive emergency repairs, high risk of catastrophic failures, and collateral damage.

[0006] 2. Preventive maintenance: Perform regular maintenance based on the recommendations of the lighting manufacturer and the experience of on-site personnel. Failures are also possible during the maintenance interval, the equipment utilization rate is relatively low, and the maintenance rate is high, which may cause failures during maintenance.

[0007] 3. Condition-based maintenance: Based on real-time monitoring of the equipment's operating status, combined with the equipment mechanism model, maintenance is performed before a failure occurs. It requires accurate fault detection / diagnosis methods, immediate measures must be taken when a fault is detected, and it is difficult to determine reasonable threshold parameters.

[0008] Therefore, it is hoped to solve how to monitor lamps in real time and how to achieve predictive maintenance of lamps. Summary of the invention

[0009] In view of the above-mentioned shortcomings of the prior art, an object of the present invention is to provide a lamp maintenance device and method for solving the problems of how to monitor lamps in real time and how to achieve predictive maintenance of lamps in the prior art.

[0010] To achieve the above-mentioned purpose and other related purposes, the present invention provides a lamp maintenance device, the device comprising: an electrical box, the electrical box comprising: a digital lightning arrester, a lighting information collector, an AC contactor, and an AC / DC power supply; the digital lightning arrester is used to detect the type of lightning strike, the size of the surge current, and the frequency of lightning strikes; the digital lightning arrester is connected to the lighting information collector to provide AC power to the lighting information collector; the lighting information collector is connected to the AC contactor to provide AC power to the AC contactor; the AC / DC power supply is connected to the AC contactor to convert AC power into DC power; the lighting information collector is connected to the AC / DC power supply to receive the DC power provided by the AC / DC power supply; the lighting information collector comprises: a first temperature and humidity sensor, an inclination sensor, a first NTC, and a first thermocouple; the first temperature and humidity sensor is used to detect the temperature and humidity of the electrical box environment; the inclination sensor is used to detect the inclination angle and vibration value of the lamp and the lamp pole; the first NTC is used to detect the ambient temperature of the lamp use environment; the first thermocouple is used to detect the T of the AC / DC power supply C point temperature; the LED light source is connected to the electrical box, the LED light source comprises: a lighting module, a light intensity sensor, a second temperature and humidity sensor, a second NTC and a second thermocouple; the second temperature and humidity sensor is used to detect the ambient temperature of the LED light source; the second NTC is used to detect the T of the lighting module p point temperature; the second thermocouple is used to detect the surface temperature of the radiator of the lighting module; the lighting information collector also includes: a wireless module and a relay switch; the wireless module is used to transmit the data detected by the sensor to the cloud server, and the cloud server uses artificial intelligence technology to analyze the data and monitor the health status of the lamp in real time. At the same time, it receives instructions from the cloud server to perform corresponding operations.

[0011] To achieve the above-mentioned purpose, the present invention also provides a lamp maintenance method, which is applied to a lamp maintenance device, wherein the lamp maintenance device comprises: an electrical box, a lighting information collector, and an LED light source; the electrical box comprises: a digital lightning arrester, a lighting information collector, an AC contactor, and an AC / DC power supply; the lighting information collector comprises: a first temperature and humidity sensor, an inclination sensor, a first NTC, and a first thermocouple; the LED light source comprises: a lighting module, a light intensity sensor, a second temperature and humidity sensor, a second NTC, and a second thermocouple; the method comprises: detecting the type of lightning strike, the size of the surge current, and the frequency of lightning strikes based on the digital lightning arrester; connecting the digital lightning arrester to the lighting information collector, and providing AC power to the lighting information collector based on the digital lightning arrester. The lighting information collector; connecting the lighting information collector with the AC contactor, providing AC power to the AC contactor based on the lighting information collector; connecting the AC / DC power supply with the AC contactor, converting the AC power into DC power based on the AC / DC power supply; connecting the lighting information collector with the AC / DC power supply, receiving the DC power provided by the AC / DC power supply based on the lighting information collector; detecting the temperature and humidity of the electrical box environment based on the first temperature and humidity sensor; detecting the inclination angle and vibration value of the lamp and the lamp pole based on the inclination sensor; detecting the ambient temperature of the lamp use environment based on the first NTC; detecting the T of the AC / DC power supply based on the first thermocouple C point temperature; connecting the LED light source to the electrical box, detecting the ambient temperature and humidity of the LED light source based on the second temperature and humidity sensor; detecting the T of the lighting module based on the second NTC p point temperature; based on the second thermocouple, the surface temperature of the lighting module radiator is detected; the lighting information collector also includes: a wireless module and a relay switch; based on the wireless module, the data detected by the sensor is transmitted to the cloud server, and the cloud server uses artificial intelligence technology to perform data analysis and monitor the health status of the lamp in real time. At the same time, the cloud server receives instructions from the cloud server to perform corresponding operations.

[0012] As described above, a lamp maintenance device and method of the present invention have the following beneficial effects: it is used to monitor various parameter data of the lamp cloud server, power grid and usage environment in real time, send them to the cloud server, use artificial intelligence technology to perform data analysis, create a high-quality, multi-dimensional lighting data set, combine expert experience with data-driven, form industrial intelligent algorithm innovations such as anomaly detection, fault diagnosis, and life prediction, and form the ability to accurately control illumination, quickly and accurately repair, and predictive maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 Shown is a schematic structural diagram of a lamp maintenance device according to an embodiment of the present invention;

[0014] Figure 2a Shown is a flow chart of a lamp maintenance method in one embodiment of the present invention;

[0015] Figure 2b Shown is a schematic diagram of the health status of equipment in an embodiment of the lamp maintenance method of the present invention;

[0016] Figure 2c Shown is a Fourier transform diagram of a lamp maintenance method according to an embodiment of the present invention;

[0017] Figure 2d Shown is a schematic diagram of wavelet transformation in an embodiment of the lamp maintenance method of the present invention;

[0018] Figure 2e Shown is a deep learning schematic diagram of a lamp maintenance method in one embodiment of the present invention.

[0019] Component number description

[0020] 1 Electrical box

[0021] 11 Digital lightning arrester

[0022] 12 Lighting Information Collector

[0023] 121 First Temperature and Humidity Sensor

[0024] 122 Tilt sensor

[0025] 123 First NTC

[0026] 124 First thermocouple

[0027] 125 AC metering sensor

[0028] 126 DC Metering Sensor

[0029] 127 Salt spray sensor

[0030] 128 pH sensor

[0031] 13 AC contactor

[0032] 14 AC / DC power supply

[0033] 2 LED light source

[0034] 21 Lighting module

[0035] 22 Light sensor

[0036] 23. Second temperature and humidity sensor

[0037] 24 Second NTC

[0038] 25 Second thermocouple

[0039] 3 Access door

[0040] 31 Leakage sensor

[0041] 32 Water sensor DETAILED DESCRIPTION

[0042] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0043] It should be noted that the illustrations provided in the following embodiments are only used to illustrate the basic concept of the present invention in a schematic manner. Therefore, the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0044] The lamp maintenance device and method of the present invention are used to monitor the health status of the lamp in real time and perform corresponding operations in a timely manner based on cloud server instructions.

[0045] like Figure 1 As shown, in one embodiment, the lamp maintenance device of the present invention comprises: an electrical box 1, wherein the electrical box 1 comprises: a digital lightning arrester 11, a lighting information collector 12, an AC contactor 13, and an AC / DC power supply 14;

[0046] The digital lightning arrester 11 is used to detect the type of lightning strike, the size of the surge current and the frequency of lightning strikes; the digital lightning arrester 11 is connected to the AC metering sensor 125, and the AC metering sensor 125 is arranged on the lighting information collector 12. The digital lightning arrester 11 is used to detect the type of lightning strike, the size of the surge current and the frequency of lightning strikes; the AC (alternating current) metering sensor is used to detect the input voltage, input current and power factor of the power grid. According to the type of lightning strike, the size of the surge current, the power grid fluctuation and power factor of the lightning strike frequency, the actual lightning strike resistance of the digital lightning arrester 11 and the experience data, laboratory data, historical data and manufacturer recommended data of the digital lightning arrester 11, the cloud server calculates the failure time of the digital lightning arrester 11 in the current actual use environment based on the preset algorithm, so as to achieve the timely replacement of the digital lightning arrester 11.

[0047] The digital lightning arrester 11 is connected to the lighting information collector 12 and is used to provide alternating current to the lighting information collector 12 .

[0048] The lighting information collector 12 is connected to the AC contactor 13 to provide AC power to the AC contactor 13 .

[0049] The AC / DC (alternating current / direct current) power source is connected to the AC contactor 13 and is used to convert alternating current into direct current.

[0050] The lighting information collector 12 is connected to the AC / DC power supply 14, and is used to receive the direct current provided by the AC / DC power supply 14; the lighting information collector 12 includes: a first temperature and humidity sensor 121, a tilt sensor 122, a first NTC 123 and a first thermocouple 124; NTC: the abbreviation of Negative Temperature CoeffiCient, which is a negative temperature coefficient thermistor. The first temperature and humidity sensor 121 is used to detect the temperature and humidity of the electrical box; the tilt sensor is used to detect the tilt angle and vibration value of the lamp and the lamp pole; the first NTC 123 is used to detect the ambient temperature of the lamp use environment 1; the first thermocouple 124 is used to detect the T of the AC / DC power supply. C Point temperature. C The point refers to the point that is most relevant to the PN node and has the most thermal conductivity stability. Specifically, the inclination sensor 122 is used to detect the inclination angle and vibration value of the lamp and the lamp pole, and the first temperature and humidity sensor 121 is used to detect the temperature and humidity of the environment of the electrical box 1. According to the inclination angle and vibration value of the lamp pole 1, the temperature and humidity of the environment of the electrical box 1, and the structural materials, surface treatment processes, and fixed connection methods used by the known lamps, based on the empirical data, laboratory data, historical data, and manufacturer recommended data of the known lamp structural parts, the cloud server calculates the estimated time that the lamp can work normally under the current inclination angle, vibration value, temperature and humidity based on the preset algorithm, so as to arrange the maintenance time of the lamp in time. Specifically, the first NTC123 is used to detect the ambient temperature of the environment in which the lamp is used; the first thermocouple 124 is used to detect the T of the AC / DC power supply. C Point temperature, the T C Point refers to the point of the AC / DC power supply similar to the performance thermal mark. And the first temperature and humidity sensor 121 is used to detect the temperature and humidity of the electrical box environment; according to the actual detected ambient temperature, device temperature and temperature and humidity of the detection environment of the electrical box 1 and the known waterproof level of the electrical box 1; the cloud server calculates the damage time of the lamp under the current conditions based on the preset algorithm based on the empirical data, laboratory data and historical data, so as to arrange the maintenance time of the lamp in time.

[0051] The LED light source 2 is connected to the electrical box 1, and the LED light source 2 includes: a lighting module 21, a light intensity sensor 22, a second temperature and humidity sensor 23, a second NTC 24, and a second thermocouple 25; the second temperature and humidity sensor 23 is used to detect the ambient temperature and humidity of the LED light source 2; the second NTC 24 is used to detect the T of the lighting module 21. p Point temperature; the Tp point refers to the point of the lighting module that is similar to the performance thermal mark. The second thermocouple 25 is used to detect the surface temperature of the heat sink of the lighting module 21. Specifically, the lighting module 21 is used for lighting, and the illuminance sensor 22 is used to detect the illuminance of the lighting module 21. Specifically, the second temperature and humidity sensor 23 is used to detect the ambient temperature and humidity of the LED light source 2; the second NTC 24 is used to detect the ambient temperature of the lighting module 21; the second thermocouple 25 is used to detect the surface temperature of the heat sink of the lighting module 21; the lighting information collector 12 also includes: a DC (direct current) metering sensor, the DC metering sensor 126 is connected to the LED light source 2, and is used to monitor the current and voltage output to the LED light source 2 in real time. Based on the current and voltage of the LED light source 2, the ambient temperature of the lighting module 21, the device temperature of the lighting module 21, the illuminance of the lighting module 21, and the chip, packaging method, and bracket of the lighting module 21, combined with the life curve, laboratory data, and manufacturer recommended data of the lighting module 21, the failure time of the lighting module 21 under the actual use environment conditions is calculated based on the cloud server. In this way, the failure time of the lighting module 21 can be predicted, and the lighting module 21 can be replaced in time.

[0052] The lighting information collector 12 also includes: a wireless module and a relay switch; the wireless module is used to transmit the data detected by the sensor to the cloud server, so that the cloud server uses artificial intelligence technology to perform data analysis and monitor the health status of the lamps in real time. At the same time, it receives instructions from the cloud server to perform corresponding operations. The operations include: turning on and off the lights, dimming, and powering off.

[0053] Specifically, it also includes an inspection door 3, and the inspection door 3 includes: a leakage sensor 31 and a water sensor 32; the leakage sensor 31 is used to detect whether there is leakage, and the water sensor 32 is used to detect whether it is flooded. The leakage sensor 31 is used to detect whether the lamp power supply line has leakage, and the water sensor 32 is used to detect whether the set position of the lamp pole is flooded. The data detected by the leakage sensor 31 and the water sensor 32 are uploaded to the cloud server, so that the cloud server monitors the data of the leakage sensor 31 and the water sensor 32 in real time. When the data of the leakage sensor 31 and the water sensor 32 exceed the preset threshold, the cloud server sends a power-off command to the lamp power supply end to cut off the power to the lamp and the power supply line.

[0054] Specifically, the lighting information collector 12 also includes: MCU, which is used to control the operation of each sensor module. The lighting information collector 12 also includes: EEPROM (Electrically Erasable Programmable Read-Only Memory), which is an electrically erasable programmable read-only memory - a storage chip that does not lose data after power failure. EEPROM can erase existing information on the lighting information collector 12 and reprogram it, and is generally plug-and-play.

[0055] Specifically, the lighting information collector 12 further includes: a salt fog sensor 127, which is used to detect the salt fog concentration of the environment in which the lamp is used. The salt fog concentration detected by the salt fog sensor 127 is uploaded to the cloud server, so that the cloud server calculates the failure time under the current salt fog concentration based on the detected salt fog concentration and the salt fog resistance concentration data of the lamp, so as to replace the lamp housing and other structural parts in time.

[0056] Specifically, the lighting information collector 12 further includes: a pH sensor 128, which is used to detect the pH of the environment in which the lamp is used. The pH detected by the pH sensor 128 is uploaded to the cloud server, so that the cloud server calculates the failure time under the current pH based on the detected pH and the pH resistance data of the lamp, so as to replace the lamp housing and other structural parts in time.

[0057] Specifically, the lighting information collector 12 further includes: a dimming module, which is connected to the AC / DC power supply 14 and is used to adjust the output current of the AC / DC power supply 14 .

[0058] Specifically, the lighting information collector 12 further includes: a DC metering sensor 126 , which is connected to the AC / DC power supply 14 and the LED light source 2 and is used to detect the current and voltage output from the AC / DC power supply 14 to the LED light source 2 .

[0059] Specifically, the lighting information collector 12 further includes: a relay switch, and the relay switch is also used to control the on and off of the AC contactor 13 to increase the flow rate and avoid damage to the relay contacts.

[0060] Specifically, the lighting information collector 12 further includes: a wireless module, which is used to upload the data detected by the above-mentioned various sensors to the cloud server and receive instructions from the cloud server.

[0061] like Figure 2a As shown, in one embodiment, the lamp maintenance method of the present invention is applied to a lamp maintenance device, the lamp maintenance device comprising: an electrical box, a lighting information collector, and an LED light source; the electrical box comprises: a digital lightning arrester, a lighting information collector, an AC contactor, and an AC / DC power supply; the lighting information collector comprises: a first temperature and humidity sensor, an inclination sensor, a first NTC, and a first thermocouple; the LED light source comprises: a lighting module, a light intensity sensor, a second temperature and humidity sensor, a second NTC, and a second thermocouple; the method comprises the following steps:

[0062] Step S21: Detecting the lightning strike type, surge current size and lightning strike frequency based on the digital lightning arrester.

[0063] Step S22: connecting the digital lightning arrester to the lighting information collector, and providing alternating current to the lighting information collector based on the digital lightning arrester.

[0064] Step S23: connecting the lighting information collector to the AC contactor, and providing AC power to the AC contactor based on the lighting information collector.

[0065] Step S24: connecting the AC / DC power supply to the AC contactor, and converting the AC power into DC power based on the AC / DC power supply.

[0066] Step S25, connecting the lighting information collector to the AC / DC power supply, receiving the direct current provided by the AC / DC power supply based on the lighting information collector; detecting the temperature and humidity of the electrical box environment based on the first temperature and humidity sensor; detecting the inclination angle and vibration value of the lamp and the lamp pole based on the inclination sensor; detecting the ambient temperature of the lamp using environment based on the first NTC; detecting the T of the AC / DC power supply based on the first thermocouple; CStep S26: Connect the LED light source to the electrical box, detect the ambient temperature and humidity of the LED light source based on the second temperature and humidity sensor; detect the T of the lighting module based on the second NTC. p Point temperature; based on the second thermocouple detecting the surface temperature of the heat sink of the lighting module.

[0067] Step S27, the lighting information collector also includes: a wireless module and a relay switch; based on the wireless module, the data detected by the sensor is transmitted to the cloud server, and the cloud server uses artificial intelligence technology to perform data analysis and monitor the health status of the lamps in real time. At the same time, it receives instructions from the cloud server to perform corresponding operations.

[0068] Specifically, it also includes pre-processing edge computing on the data detected by the sensor. Edge computing refers to analyzing and processing data near the data source, without data flow, thereby reducing network traffic and response time. Data without abnormalities is uploaded to the cloud server at a regular interval after weighting and other processing; data with abnormalities is uploaded to the cloud server in real time, and at the same time, based on the abnormal classification, it is decided whether to immediately cut off the power supply of the lamp and power supply line. Specifically, it also includes the cloud server calculating whether the lamp maintenance device needs maintenance or overhaul based on the data combined with a preset algorithm. When maintenance or overhaul is required, the cloud server sends a maintenance or overhaul instruction to the lamp maintenance device. The preset algorithms include statistical algorithms, deep learning algorithms, and machine learning algorithms.

[0069] In order to do a good job of predictable maintenance of lamps, the following steps are mainly used:

[0070] 1. Anomaly detection, that is, is there a problem?

[0071] Anomaly detection is to find anomalies from a large amount of data. In layman's terms, it is to find data that "looks different" from a vast amount of data.

[0072] 2. Fault diagnosis, that is, what is the problem?

[0073] When it is determined that there is a problem with the lamp, the fault diagnosis method is used to diagnose the detailed fault situation of the machine, including the specific location of the fault, the manifestation of the fault, the cause of the fault and the severity of the fault, etc., and maintenance suggestions are given based on the above situations.

[0074] Fault diagnosis mainly includes: (1) fault location, locating the specific component or subsystem of the fault; (2) fault classification, distinguishing between known and unknown faults. If it is a known fault, the type of fault must also be determined.

[0075] 3. Predictive maintenance, i.e. when might there be a problem?

[0076] like Figure 2b As shown in the figure, in lighting fixtures, different modules have different physical degradation mechanisms. For example, LED lamp beads have a certain aging curve under laboratory conditions. The actual application environment is often more changeable and harsh, so changes in the external environment often interfere with the operation of the system, thereby accelerating the life loss of system components. Therefore, simply using the standard curve in the laboratory cannot effectively perform predictive maintenance. It is necessary to adjust the expected changes of the curve in time according to the data collected in actual operation, so as to achieve more accurate and effective predictive maintenance.

[0077] The specific method is:

[0078] 1. First, professional engineers will determine whether there is any abnormality based on the operating mechanism and design margin of the lamp.

[0079] 2. Single-dimensional anomaly detection based on time series modeling: Sort the detected variables by time and look for anomalies.

[0080] (1) Figure 2c As shown, Fourier transform: Any continuously measured time series or signal can be expressed as an infinite superposition of sine wave signals of different frequencies. The Fourier transform algorithm created based on this principle uses the original signal measured directly to calculate the frequency, amplitude and phase of different sine wave signals in the signal in an accumulative manner. The Fourier transform converts the original difficult-to-process time domain signal into an easy-to-analyze frequency domain signal (signal spectrum). The Fourier transform formula is as follows:

[0081]

[0082] The Fourier transform performs spectrum analysis on the function, reflecting the time-frequency spectrum characteristics of the entire signal and revealing the characteristics of the stationary signal better. The defect of the Fourier transform is that since the sine wave is infinitely wide, the signal being analyzed also needs to have meaningful characteristics from negative infinity to positive infinity, so the Fourier transform cannot handle some local signals very well. For example, a function that has non-zero values ​​in a local range and is equal to 0 everywhere else will have a rather chaotic spectrum. At this time, the signal in the frequency domain is not as intuitive as that in the time domain, and spectrum analysis becomes very difficult. In order to overcome these defects of the Fourier transform, several methods of transforming using finite-width basis functions have been developed. These basis functions vary not only in frequency but also in position, and these finite-width waves are called "wavelets".

[0083] (2) Wavelet transform: It inherits and develops the localization idea of ​​short-time Fourier transform, while overcoming the shortcomings of window size not changing with frequency. It can provide a time-frequency window that changes with frequency, and is an ideal tool for signal time-frequency analysis and processing. Its main feature is that it can fully highlight certain aspects of the problem through transformation. The principle diagram of wavelet transform is shown in the figure below. Figure 2d As shown, the formula is as follows:

[0084]

[0085] 3. Multi-dimensional data anomaly detection based on statistical distribution:

[0086] (1) Statistical model: A programmed expression of the relationship between variables in the form of a mathematical equation. Some processes cannot be derived from a model using theoretical analysis methods, but the functional relationship between variables can be obtained through experiments or direct measurement of data from industrial processes using mathematical statistics. This is called a statistical model.

[0087] (2) Deep learning algorithm: Apply artificial neurons, deep neural networks, recursive neural networks, artificial neural networks, convolutional neural networks and other technologies to continuously reduce the misjudgment rate by adding intermediate hidden layers and feedback loops. Figure 2e As shown, this is an example of a deep learning algorithm.

[0088] (3) Machine learning algorithms: Traditional machine learning techniques are divided into two categories: supervised learning and unsupervised learning. Supervised learning uses only labeled sample sets for learning, while unsupervised learning uses only unlabeled sample sets for learning. However, in many practical problems, labeled samples are usually difficult to collect, while unlabeled samples are easy to obtain. Supervised machine learning classification algorithms often require a large number of labeled data samples, and it is difficult to solve the problem of data imbalance (huge difference in the number of positive and negative samples). Therefore, it is difficult to use them in anomaly detection scenarios with a small number of samples. Unsupervised learning has certain limitations when applied to real-world scenarios, resulting in varying degrees of false positives (many normal samples are considered anomalies) or false negatives (only a small number of anomalies are detected). The present invention adopts semi-supervised learning. First, the labeled data is trained with a supervised learning method to obtain an initial classifier, and then the initial classifier is used to predict the unlabeled data to obtain the probability value of the unlabeled data belonging to a certain category. The larger the category probability, the higher the confidence of the classification. Samples with higher confidence are added to the labeled data to obtain a larger and more updated labeled training set. At the same time, the selected high-confidence samples are removed from the unlabeled data set, and the classifier is iteratively trained until all training samples have labels. At this time, the unlabeled training set is an empty set.

[0089] As more data is uploaded to the cloud server, algorithms, machine learning methods and AI models are continuously optimized, and artificial intelligence machine learning technology is used to analyze and model data to enable the maintenance and operation of lamps.

[0090] After the lamps have fault warning, maintenance plans can be made in advance. According to the cause of the fault, non-operating time can be used or non-operating time can be arranged through unified scheduling in advance, so that fast and accurate maintenance can be carried out in advance to minimize or even zero the impact of the fault and ensure the safety of people's lives and property. Predictive maintenance can not only reduce unplanned downtime, minimize production losses during downtime, and optimize spare parts inventory, but also:

[0091] Optimize the design of next-generation products. By comparing the usage data of lamp clients with laboratory data, we can design more reasonable accelerated life test models for similar lamps and simulate the relevant reliability performance of similar lamps in a shorter time. According to the different requirements of different industries and regions for products, we can design and estimate suitable personalized lamps. Provide a basis for optimizing the design of next-generation products.

[0092] Provide decision-making basis for the selection and use of LED lamp beads, mounting brackets and surface treatment of subsequent lamps, and use the real-time data of the increasing number of LED lamp beads, mounting brackets and surface treatment in actual scenes, rather than static data in the laboratory, to select suitable components for lamps in different application scenarios.

[0093] Provide a basis for the formulation of industry standards or national standards. Utilize the increasing measured data from actual scenarios to analyze the different requirements of different industries and regions for products and provide a basis with reference value.

[0094] These actual scenario data can be used for a lot of analysis for customers and manufacturers, and can achieve special customization of lamps for specific industries and regions, which are more in line with local environment and usage habits and better serve practical applications.

[0095] In summary, the lamp maintenance device and method of the present invention are used to monitor various parameter data of the lamp maintenance device in real time and perform corresponding operations in a timely manner based on cloud server instructions. Therefore, the present invention effectively overcomes various shortcomings in the prior art and has high industrial utilization value.

[0096] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A lamp maintenance device, characterized in that: The device comprises: An electrical box, the electrical box comprising: a digital lightning arrester, a lighting information collector, an AC contactor, and an AC / DC power supply; The digital lightning arrester is used to detect lightning strike type, surge current size and lightning strike frequency; The digital lightning arrester is connected to the lighting information collector and is used to provide alternating current to the lighting information collector; The lighting information collector is connected to the AC contactor and is used to provide AC power to the AC contactor; The AC / DC power supply is connected to the AC contactor and is used to convert AC power into DC power; The lighting information collector is connected to the AC / DC power supply and is used to receive the direct current provided by the AC / DC power supply; the lighting information collector includes: a first temperature and humidity sensor, a tilt sensor, a first NTC and a first thermocouple; the first temperature and humidity sensor is used to detect the temperature and humidity of the electrical box; the tilt sensor is used to detect the tilt angle and vibration value of the lamp and the lamp pole; the first NTC is used to detect the ambient temperature of the lamp use environment; the first thermocouple is used to detect the T of the AC / DC power supply C Point temperature; The LED light source is connected to the electrical box, and the LED light source includes: a lighting module, a light intensity sensor, a second temperature and humidity sensor, a second NTC and a second thermocouple; the second temperature and humidity sensor is used to detect the ambient temperature and humidity of the LED light source; the second NTC is used to detect the temperature and humidity of the lighting module. p point temperature; the second thermocouple is used to detect the surface temperature of the heat sink of the lighting module; The lighting information collector also includes: a wireless module, a relay switch, a dimming module, and a DC metering sensor; the wireless module is used to transmit the data detected by the sensor to a cloud server, and the cloud server uses artificial intelligence technology to perform data analysis and monitor the health status of the lamps in real time; at the same time, it receives instructions from the cloud server to perform corresponding operations; the dimming module is connected to the AC / DC power supply and is used to adjust the output current of the AC / DC power supply; the DC metering sensor is connected to the AC / DC power supply and the LED light source and is used to detect the current and voltage output from the AC / DC power supply to the LED light source.

2. The lamp maintenance device according to claim 1, characterized in that: It also includes an inspection door, which includes: a leakage sensor and a water sensor; the leakage sensor is used to detect whether the power supply network has leakage, and the water sensor is used to detect whether the set location is flooded.

3. The lamp maintenance device according to claim 1, characterized in that: The lighting information collector further includes: a salt fog sensor, which is used to detect the salt fog concentration of the environment in which the lamp is used.

4. The lamp maintenance device according to claim 1, characterized in that: The lighting information collector further includes: a pH sensor, and the pH sensor is used to detect the pH of the environment in which the lamp is used.

5. A lamp maintenance method, characterized in that: Applied to a lamp maintenance device, the lamp maintenance device comprises: an electrical box, a lighting information collector, and an LED light source; the electrical box comprises: a digital lightning arrester, a lighting information collector, an AC contactor, and an AC / DC power supply; the lighting information collector comprises: a first temperature and humidity sensor, an inclination sensor, a first NTC, and a first thermocouple; the LED light source comprises: a lighting module, a light intensity sensor, a second temperature and humidity sensor, a second NTC, and a second thermocouple; the method comprises: Detect lightning strike type, surge current size and lightning strike frequency based on digital lightning arrester; Connecting the digital lightning arrester to the lighting information collector, and providing alternating current to the lighting information collector based on the digital lightning arrester; Connecting the lighting information collector to the AC contactor, and providing AC power to the AC contactor based on the lighting information collector; Connecting the AC / DC power supply to the AC contactor, and converting AC power into DC power based on the AC / DC power supply; The lighting information collector is connected to the AC / DC power supply, and the lighting information collector receives the direct current provided by the AC / DC power supply; the temperature and humidity of the electrical box are detected based on the first temperature and humidity sensor; the inclination angle and vibration value of the lamp and the lamp pole are detected based on the inclination sensor; the ambient temperature of the lamp use environment is detected based on the first NTC; the T of the AC / DC power supply is detected based on the first thermocouple. C Point temperature; The LED light source is connected to the electrical box, and the ambient temperature and humidity of the LED light source is detected based on the second temperature and humidity sensor; the T of the lighting module is detected based on the second NTC p point temperature; detecting the surface temperature of the heat sink of the lighting module based on the second thermocouple; The lighting information collector also includes: a wireless module and a relay switch; based on the wireless module, the data detected by the sensor is transmitted to the cloud server, and the cloud server uses artificial intelligence technology to perform data analysis and monitor the health status of the lamps in real time; at the same time, it receives instructions from the cloud server to perform corresponding operations.

6. The lamp maintenance method according to claim 5, characterized in that: It also includes performing edge computing on the data detected by the sensor in advance, uploading normal data on a scheduled basis, and triggering the upload of abnormal data after edge computing to the cloud server in real time.

7. The lamp maintenance method according to claim 5, characterized in that: It also includes the cloud server calculating the cloud server based on the data in combination with a preset algorithm, using artificial intelligence technology to perform data analysis, creating a high-quality, multi-dimensional lighting data set, combining expert experience and data-driven to form industrial intelligent algorithm innovations for anomaly detection, fault diagnosis, and life prediction, and forming the ability to accurately control illumination, quickly and accurately repair, and predictive maintenance.

8. The lamp maintenance method according to claim 7, characterized in that: The preset algorithms include: statistical algorithms, deep learning algorithms, and machine learning algorithms.

Images