Predictive maintenance method and system for lighting device

Abstract

The invention provides a predictive maintenance method and system for a lighting device, and the method comprises the steps: constructing a lighting working rule of the lighting device according to a plurality of pieces of real-time parameter information of the lighting device, and determining a plurality of constant working time periods of the lighting device, respectively predicting the historical working data corresponding to each constant working time period by using an autoregression model, determining the predicted residual life of the lighting device, constructing the real-time working damage of the lighting device by using the real-time parameter information, determining the accumulated damage feature corresponding to each lighting device in the lighting device, and determining the residual life of the lighting device according to the accumulated damage feature. And marking the device total loss moment corresponding to each lighting device in the predicted residual life, determining the device maintenance moment corresponding to each lighting device, carrying out predictive maintenance on the lighting device by monitoring the lighting device, predicting the fault time period of the lamp and the faulty device, so that workers can replace the devices in advance, and the working efficiency is improved. And the risk of lack of illumination is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of lighting maintenance, and particularly relates to a predictive maintenance method and system for lighting devices. Background Art

[0002] In recent years, with the economic growth, the urban lighting technology and scale in China have developed rapidly. Lighting devices are no longer just simple lighting devices. Most of them are attached with lighting device controllers and have functions such as electronic switches and induction dimming. While lighting device controllers are widely used, the monitoring of lighting device controller failures still takes the "lighting" and "not lighting" of the lighting device as the judgment standard, making it difficult to effectively judge the failures of lighting devices. Generally, by organizing specialized personnel to conduct regular manual inspections along the line and manually determining the fault points, the lighting devices can be maintained and repaired. However, this method will result in low work efficiency and high maintenance costs, and it is difficult for maintenance personnel to discover lighting device controller failures in the first place. Even if a failure is discovered, it is impossible to quickly determine the type of failure, such as lighting device failure, lighting device controller failure, or power supply line failure, etc. Often, due to untimely maintenance, the lighting in some areas is missing, causing inconvenience to residents.

[0003] Therefore, the present invention provides a predictive maintenance method and system for lighting devices. Summary of the Invention

[0004] A predictive maintenance method and system for lighting devices according to the present invention monitors lighting devices for predictive maintenance, predicts the failure time period and failed components of the lamps, and workers can replace the components in advance, reducing the risk of lighting loss.

[0005] The present invention provides a predictive maintenance method for lighting devices, including:

[0006] Step 1: According to several real-time parameter information of the lighting device, construct the lighting working rule of the lighting device and determine several normal working time periods of the lighting device;

[0007] Step 2: Use autoregressive models to predict the historical working data corresponding to each normal working time period respectively to obtain the remaining normal working times of the lighting device;

[0008] Step 3: Deduce the predicted remaining life of the lighting device, use the real-time parameter information to construct the real-time working damage of the lighting device, and determine the cumulative damage characteristics corresponding to each lighting component in the lighting device;

[0009] Step 4: Mark the device complete failure moment corresponding to each lighting device in the predicted remaining life according to the cumulative damage characteristics, and determine the device maintenance moment corresponding to each lighting device.

[0010] In an implementable manner,

[0011] Step 1 includes:

[0012] Step 11: Transmit the real-time parameter information of the lighting device through the communication network, perform data encoding on the real-time parameter information in the application layer of the communication network to obtain the real-time application data of the lighting device, identify the first-byte data of the real-time application data, and construct the real-time data feature of the lighting device according to the data content included in the first-byte data;

[0013] Step 12: Identify the second-byte data of the real-time application data, construct the key byte value of the real-time application data according to the first-byte data and the second-byte data, use AI technology to perform type matching on the key byte value to determine the device attribute of the lighting device, and perform data calibration on the real-time application data based on the legal data packet type of the device attribute to obtain the real-time valid data of the lighting device;

[0014] Step 13: Normalize the real-time valid data corresponding to different moments, sort the normalized data according to the time sequence, perform linear analysis on the sorting result to obtain the data linear relationship of the lighting device, and identify several lighting segments of the lighting device in the data linear relationship to construct the lighting working rule of the lighting device;

[0015] Step 14: Map the lighting working rule into the time space to obtain several historical lighting time periods of the lighting device, perform fuzzification processing on the lighting start moment corresponding to each historical lighting time period respectively to obtain the fuzzy lighting start point corresponding to each historical lighting time period, and determine several regular working time periods of the lighting device.

[0016] In an implementable manner,

[0017] It further includes:

[0018] Determine the lighting duration of the corresponding historical lighting time period according to each lighting start moment, and determine the historical average lighting duration of the lighting device;

[0019] Determine the adjustable duration corresponding to each historical lighting time period based on the historical average lighting duration;

[0020] Perform a fuzzification process on the corresponding lighting start point within the adjustable duration range to obtain a fuzzy lighting start point corresponding to the historical lighting time period.

[0021] In an implementable manner,

[0022] Step 2 includes:

[0023] Step 21: Perform autocorrelation calculations on each of the historical working data using the autoregressive model to determine the current order threshold of the autoregressive model, and adjust the initial window length according to the current order threshold to obtain the effective window function of the lighting device;

[0024] Step 22: Perform windowing processing on each of the historical working data using the window function to obtain several optimized lighting time periods of the lighting device, and feedback the optimized lighting time periods into the autoregressive model for model fitting to obtain the autoregressive coefficients of the lighting device;

[0025] Step 23: Perform regression prediction on the historical working data according to the autoregressive coefficients to obtain several predicted lighting moments of the lighting device, determine the failure lighting time period of the lighting device according to the predicted lighting effect characteristics corresponding to each predicted lighting moment, and determine the remaining normal working times of the lighting device.

[0026] In an implementable manner,

[0027] It further includes:

[0028] After obtaining the autoregressive coefficients, perform residual identification on the autoregressive model to obtain the residuals of the autoregressive model;

[0029] When there is an autocorrelation relationship between the residuals and the historical working data, use the residuals to correct the current order of the autoregressive model to obtain an updated autoregressive model;

[0030] Use the updated autoregressive model to perform model fitting on the optimized lighting time periods and update the autoregressive coefficients according to the fitting results.

[0031] In an implementable manner,

[0032] Step 3 includes:

[0033] Step 31: Map the remaining normal working times into the lighting working pattern, expand the lighting working pattern according to the mapping positions corresponding to each remaining normal working time, and determine the predicted remaining life of the lighting device according to the expansion length;

[0034] Step 32: Restore the real-time lighting state of the lighting device according to the real-time parameter information, and construct the cumulative lighting state of the lighting device. Deduce the historical cumulative damage of the lighting device according to the cumulative lighting state, and determine the additional damage to the historical cumulative damage according to the real-time lighting state;

[0035] Step 33: Identify and count the number of damages with the same attribute included in the historical cumulative damage according to the damage attribute of the additional damage, and use the number of damages with the same attribute to enhance the damage attribute to determine the real-time working damage of the additional damage to the lighting device;

[0036] Step 34: Perform damage simulation on the lighting device using the real-time working damage, determine the damage position corresponding to each lighting moment and the damage value corresponding to each damage position, and determine the cumulative damage characteristics corresponding to each lighting device in the lighting device.

[0037] In an implementable manner,

[0038] Step 4 includes:

[0039] Step 41: Construct a damage lighting model of the lighting device according to the initial device parameters corresponding to each lighting device and the cumulative damage characteristics, and run the damage lighting model to determine the device operation parameters corresponding to each lighting device;

[0040] Step 42: Deduce the total loss moment of the device corresponding to the lighting device using the device operation parameters, mark the total loss moment in the predicted remaining life, and determine the total loss consequence corresponding to each lighting device based on the device attribute corresponding to each lighting device;

[0041] Step 43: Set a first device sorting for the lighting devices according to the order of the total loss consequences from high to low, set a second device sorting for the lighting devices according to the order of the total loss moments from near to far, and re-sort the lighting devices according to the first device sorting and the second device sorting;

[0042] Step 44: Feed the sorting result back into the predicted remaining life to determine the device maintenance moment corresponding to each lighting device, and generate a corresponding maintenance list and display it.

[0043] In an implementable manner,

[0044] It further includes:

[0045] Obtain the real-time device state corresponding to each lighting device respectively;

[0046] Screen the target lighting devices with abnormal real-time device status, define the sorting of the target lighting devices as the first, generate an emergency maintenance warning and issue a corresponding alarm.

[0047] The present invention provides a predictive maintenance system for a lighting device, including:

[0048] A rule analysis module for constructing the lighting working rule of the lighting device according to a plurality of real-time parameter information of the lighting device, and determining a plurality of regular working time periods of the lighting device;

[0049] A prediction execution module for respectively predicting the historical working data corresponding to each of the regular working time periods by using an autoregressive model to obtain the remaining regular working times of the lighting device;

[0050] A damage derivation module for deriving the predicted remaining life of the lighting device, constructing the real-time working damage of the lighting device by using the real-time parameter information, and determining the cumulative damage characteristics corresponding to each lighting device in the lighting device;

[0051] A maintenance assistance module for marking the device complete loss moment corresponding to each lighting device in the predicted remaining life according to the cumulative damage characteristics, and determining the device maintenance moment corresponding to each lighting device.

[0052] In an implementable manner,

[0053] The rule analysis module includes:

[0054] A first analysis unit for transmitting the real-time parameter information of the lighting device through a communication network, performing data encoding on the real-time parameter information in the application layer of the communication network to obtain the real-time application data of the lighting device, identifying the first byte data of the real-time application data, and constructing the real-time data feature of the lighting device according to the data content included in the first byte data;

[0055] A second analysis unit for identifying the second byte data of the real-time application data, constructing the key byte value of the real-time application data according to the first byte data and the second byte data, using AI technology to perform type matching on the key byte value to determine the device attribute of the lighting device, and performing data calibration on the real-time application data based on the legal data packet type of the device attribute to obtain the real-time valid data of the lighting device;

[0056] A third analysis unit is configured to normalize the real-time valid data corresponding to different moments, sort the normalized data according to the time sequence, perform a linear analysis on the sorting result to obtain the data linear relationship of the lighting device, identify several lighting segments of the lighting device in the data linear relationship, and construct the lighting working rule of the lighting device;

[0057] A fourth analysis unit is configured to map the lighting working rule into the time space to obtain several historical lighting time periods of the lighting device, perform a fuzzification process on the lighting start moment corresponding to each historical lighting time period to obtain a fuzzy lighting start point corresponding to each historical lighting time period, and determine several regular working time periods of the lighting device.

[0058] The achievable beneficial effects of the above technical solution are as follows: In order to improve the efficiency of workers in maintaining the lighting device and improve the effectiveness of the lighting device during lighting, real-time parameter information of the lighting device is collected daily, so as to deduce the lighting working rule of the lighting device, determine the regular working time periods of the lighting device, and then use the autoregressive model to analyze the historical working data, and use the prediction function of the natural regression model to determine the remaining regular working times of the lighting device, so as to deduce the predicted remaining life of the lighting device in combination with the known lighting working rule. At the same time, analyze the real-time working damage of the lighting device according to the real-time parameter information, deduce the total damage moment of the lighting device by determining the cumulative damage characteristics of each lighting device in the lighting device, and finally determine the device maintenance moment of each lighting device. In this way, it can assist workers to quickly determine the location and type of faults that are about to occur in the lighting device, so as to perform preventive maintenance in advance, reduce the cost of large-scale maintenance in the later stage, and bring long-term convenience to the residents and pedestrians within the lighting range.

[0059] Other features and advantages of the present invention will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in the written specification and the drawings.

[0060] The technical solution of the present invention will be further described in detail below through the drawings and embodiments. Description of the Drawings

[0061] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention. In the drawings:

[0062] Figure 1 It is a schematic diagram of the working process of a predictive maintenance method for a lighting device in an embodiment of the present invention;

[0063] Figure 2 This is a schematic diagram of the composition of a predictive maintenance system for a lighting device in an embodiment of the present invention. Detailed implementation manners

[0064] The following describes the preferred embodiments of the present invention with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used for illustrating and explaining the present invention, and are not used to limit the present invention.

[0065] Embodiment 1

[0066] This embodiment provides a predictive maintenance method for a lighting device, as Figure 1 shown, including:

[0067] Step 1: According to a plurality of real-time parameter information of the lighting device, construct the lighting working rule of the lighting device, and determine a plurality of regular working time periods of the lighting device;

[0068] Step 2: Use an autoregressive model to predict the historical working data corresponding to each regular working time period respectively, and obtain the remaining regular working times of the lighting device;

[0069] Step 3: Deduce the predicted remaining life of the lighting device, use the real-time parameter information to construct the real-time working damage of the lighting device, and determine the cumulative damage characteristics corresponding to each lighting device in the lighting device;

[0070] Step 4: Mark the device complete damage moment corresponding to each lighting device in the predicted remaining life according to the cumulative damage characteristics, and determine the device maintenance moment corresponding to each lighting device.

[0071] In this example, the real-time parameter information includes: information composed of the working current, working voltage, working temperature, working duration, and working humidity of each lighting device at the current moment;

[0072] In this example, the regular working time period represents the time period when the lighting device often conducts lighting. For example: the lighting device starts lighting at 19:00 every day and ends lighting until 06:30 the next day;

[0073] In this example, the lighting working rule represents the lighting working rule presented by the lighting device within a 30*24-hour time period;

[0074] In this example, the remaining regular working times represent the number of times the lighting device can complete lighting during the regular working time period;

[0075] In this example, the real-time working damage represents the damage caused to the lighting device due to external environmental influence and its own work;

[0076] In this example, the moment of complete device failure represents the moment when the lighting device is scrapped;

[0077] In this example, the device maintenance moment represents any moment before the moment of complete failure of the lighting device and not within the normal working time period.

[0078] The working principle and beneficial effects of the above technical solution: In order to improve the efficiency of workers in maintaining lighting devices and the effectiveness of lighting devices during lighting, real-time parameter information of lighting devices is collected daily to deduce the lighting working rules of lighting devices, determine the normal working time period of lighting devices, and then use autoregressive models to analyze historical working data, use the prediction function of natural regression models to determine the remaining normal working times of lighting devices, so as to deduce the predicted remaining life of lighting devices in combination with the known lighting working rules. At the same time, analyze the real-time working damage of lighting devices according to real-time parameter information, deduce the moment of complete failure of lighting devices by determining the cumulative damage characteristics of each lighting device in the lighting device, and finally determine the device maintenance moment of each lighting device. In this way, it can assist workers in quickly determining the positions and types of upcoming failures in lighting devices, so as to carry out preventive maintenance in advance, reduce the cost of large-scale maintenance in the later stage, and bring long-term convenience to residents and pedestrians within the lighting range.

[0079] Embodiment 2

[0080] Based on Embodiment 1, for the predictive maintenance method for a lighting device, Step 1 includes:

[0081] Step 11: Transmit the real-time parameter information of the lighting device through a communication network, perform data encoding on the real-time parameter information in the application layer of the communication network to obtain the real-time application data of the lighting device, identify the first-byte data of the real-time application data, and construct the real-time data feature of the lighting device according to the data content included in the first-byte data;

[0082] Step 12: Identify the second-byte data of the real-time application data, construct the key byte value of the real-time application data according to the first-byte data and the second-byte data, use AI technology to perform type matching on the key byte value to determine the device attribute of the lighting device, and perform data calibration on the real-time application data based on the legal data packet type of the device attribute to obtain the real-time valid data of the lighting device;

[0083] Step 13: Normalize the real-time valid data corresponding to different times, sort the normalized data according to the time sequence, perform linear analysis on the sorting result to obtain the data linear relationship of the lighting device, identify several lighting segments of the lighting device in the data linear relationship, and construct the lighting working rule of the lighting device;

[0084] Step 14: Map the lighting working rule into the time space to obtain several historical lighting time periods of the lighting device, perform fuzzification processing on the lighting start time corresponding to each historical lighting time period respectively to obtain the fuzzy lighting start point corresponding to each historical lighting time period, and determine several regular working time periods of the lighting device.

[0085] In this example, the communication network refers to the network used to transmit real-time parameter information to the remote terminal;

[0086] In this example, the real-time application data refers to the real-time parameter information processed by the application layer;

[0087] In this example, the first-byte data refers to the data contained in the first byte of the real-time application data, and the second-byte data refers to the data contained in the second byte of the real-time application data;

[0088] In this example, the real-time data feature refers to the feature presented by the data content contained in the first-byte data;

[0089] In this example, the device attribute refers to the attribute presented when the lighting device performs lighting work;

[0090] In this example, the legal data packet type refers to the types of data packets that can be recognized by and can be generated by this device attribute;

[0091] In this example, the historical lighting time period refers to the time period when the lighting device performs lighting work;

[0092] In this example, the fuzzy lighting start point refers to the result of expanding the moment of the lighting start point within the historical lighting time period.

[0093] Working principle and beneficial effects of the above technical solution: To better maintain the lighting device, it is necessary to first determine the working law of the lighting device and analyze whether it will malfunction based on the working law. First, the application layer of the communication network is used to encode the real-time parameter information to obtain the real-time application data of the lighting device. Then, the first-byte data and the second-byte data in the real-time application data are analyzed and calibrated multiple times to deduce the real-time effective data of the lighting device. Further, the real-time effective data corresponding to different moments is normalized and sorted, and several lighting segments of the lighting device are determined according to the data linear relationship shown by the sorting result, generating the lighting working law of the lighting device. Since the lighting time period of the lighting device may change, the lighting start moment of the lighting device is fuzzified, and finally several regular working time periods of the lighting device are determined. In this way, not only can the data generated by the lighting device be reasonably processed, but also the regular working time periods of the lighting device are determined, improving the efficiency and quality of subsequent prediction and maintenance.

[0094] Embodiment 3

[0095] Based on Embodiment 2, the predictive maintenance method for a lighting device further includes:

[0096] Determine the lighting duration of the corresponding historical lighting time period according to each lighting start moment, and determine the historical average lighting duration of the lighting device;

[0097] Determine the adjustable duration corresponding to each historical lighting time period based on the historical average lighting duration;

[0098] Fuzzify the corresponding lighting start point within the adjustable duration range to obtain the fuzzy lighting start point corresponding to the historical lighting time period.

[0099] Working principle and beneficial effects of the above technical solution: To avoid too high a deviation between the fuzzified lighting start point and the original lighting start point, the historical average lighting duration is used to determine the adjustable duration of each historical lighting time period, and then fuzzification is performed to ensure the effectiveness of the simulated lighting start point.

[0100] Embodiment 4

[0101] Based on Embodiment 1, in the step 2 of the predictive maintenance method for a lighting device, it includes:

[0102] Step 21: Use the autoregressive model to perform autocorrelation calculations on each historical working data respectively, determine the current order threshold of the autoregressive model, and adjust the initial window length according to the current order threshold to obtain the effective window function of the lighting device;

[0103] Step 22: Use the window function to perform windowing processing on each piece of the historical working data to obtain several optimized lighting time periods of the lighting device, and feedback the optimized lighting time periods into the autoregressive model for model fitting to obtain the autoregressive coefficients of the lighting device;

[0104] Step 23: Perform regression prediction on the historical working data according to the autoregressive coefficients to obtain several predicted lighting moments of the lighting device, determine the failed lighting time periods of the lighting device according to the predicted lighting effect characteristics corresponding to each predicted lighting moment, and determine the remaining normal working times of the lighting device.

[0105] In this example, autocorrelation calculation represents the process of analyzing the logical relationship between time series;

[0106] In this example, the current order threshold represents the value of the current order of the autoregressive model;

[0107] In this example, the initial window length represents the window length of the initial window function.

[0108] The working principle and beneficial effects of the above technical solution: By using the autoregressive model to perform autocorrelation calculation on the historical working data to determine the current order threshold of the autoregressive model, thereby adjusting the window length of the initial window function using the current order threshold, the effective window function of the lighting device is determined. Furthermore, the optimized lighting time periods of the lighting device are obtained by performing windowing processing on the historical working data using the effective window function, and then they are feedback into the autoregressive model for model fitting to obtain the autoregressive coefficients of the lighting device. Further, regression prediction is performed on the historical working data using the autoregressive coefficients to determine several predicted lighting moments of the lighting device and the corresponding predicted lighting effect characteristics, so that the failed lighting time periods of the lighting device can be easily determined, and the remaining normal working times of the lighting device are deduced. In this way, the effective information in the historical working data can be separated, and the prediction function of the autoregressive model and the function of the window function to improve data resolution are used to predict the lighting device. Finally, the remaining normal working times of the lighting device are deduced, achieving the purpose of preliminary prediction.

[0109] Embodiment 5

[0110] Based on Embodiment 4, the predictive maintenance method for a lighting device further includes:

[0111] After obtaining the autoregressive coefficients, perform residual identification on the autoregressive model to obtain the residuals of the autoregressive model;

[0112] When there is an autocorrelation relationship between the residual value and the historical working data, the current order of the autoregressive model is corrected using the residual value to obtain an updated autoregressive model;

[0113] The updated autoregressive model is used to perform model fitting on the optimized lighting time period, and the autoregressive coefficients are updated according to the fitting results.

[0114] In this example, the residual represents the difference between the predicted value and the actual value of the autoregressive model.

[0115] The working principle and beneficial effects of the above technical solution: Since the daily work of the lighting device may be different, there may be deviations in the prediction work. When there is an autocorrelation relationship between the residual in the autoregressive model and the historical working data, the current order of the autoregressive model is corrected, and then the autoregressive coefficients of the lighting device are updated. Through continuous optimization training, the prediction accuracy is improved.

[0116] Embodiment 6

[0117] Based on Embodiment 1, for the predictive maintenance method for a lighting device, Step 3 includes:

[0118] Step 31: Map the remaining normal working times to the lighting working pattern, expand the lighting working pattern according to the mapping position corresponding to each remaining normal working time, and determine the predicted remaining life of the lighting device according to the expansion length;

[0119] Step 32: Restore the real-time lighting state of the lighting device according to the real-time parameter information, and construct the cumulative lighting state of the lighting device. Derive the historical cumulative damage of the lighting device according to the cumulative lighting state, and determine the increased damage of the historical cumulative damage according to the real-time lighting state;

[0120] Step 33: According to the damage attribute of the increased damage, identify and count the number of damages with the same attribute included in the historical cumulative damage, and use the number of damages with the same attribute to enhance the damage attribute to determine the real-time working damage of the increased damage to the lighting device;

[0121] Step 34: Perform damage simulation on the lighting device using the real-time working damage, determine the damage position corresponding to each lighting moment and the damage value corresponding to each damage position, and determine the cumulative damage characteristics corresponding to each lighting device in the lighting device.

[0122] In this example, the real-time lighting state represents the state currently presented by the lighting device, and the cumulative lighting state represents the state presented after multiple real-time lighting states are superimposed;

[0123] In this example, the increased damage represents the effect of the real-time lighting state on the historical cumulative damage;

[0124] In this example, the damage of the same attribute represents the damage in the historical cumulative damage that is consistent with the damage attribute of the increased damage;

[0125] In this example, the cumulative damage feature represents the feature presented by the lighting device after long-term damage.

[0126] The working principle and beneficial effects of the above technical solution: By mapping the remaining normal working times to the lighting working rules to analyze the predicted remaining life of the lighting device, and then using the real-time parameter information to restore the real-time lighting state and cumulative lighting state of the lighting device to deduce the historical cumulative damage of the lighting device, and determining the increased damage of the historical cumulative damage, analyzing the real-time working damage of the lighting device by identifying the number of damages of the same attribute included in the historical cumulative damage. In order to ensure the effectiveness of the deduction process, the damage position and damage value at each lighting moment are analyzed by means of modeling. Finally, the cumulative damage feature of each lighting device is obtained. In this way, the lighting device can be monitored for a long time, and the damage feature of each lighting device is determined, improving the efficiency of later maintenance.

[0127] Embodiment 7

[0128] Based on Embodiment 1, for the predictive maintenance method for a lighting device, Step 4 includes:

[0129] Step 41: Construct a damage lighting model of the lighting device according to the initial device parameters corresponding to each lighting device in combination with the cumulative damage feature, and run the damage lighting model to determine the device operation parameters corresponding to each lighting device;

[0130] Step 42: Deduce the device complete loss moment corresponding to the lighting device by using the device operation parameters, mark the device complete loss moment in the predicted remaining life, and determine the complete loss consequence corresponding to each lighting device based on the device attribute corresponding to each lighting device;

[0131] Step 43: Set a first device sorting for the lighting devices in the order from high to low of the complete loss consequences, set a second device sorting for the lighting devices in the order from near to far of the complete loss moments, and re-sort the lighting devices according to the first device sorting and the second device sorting;

[0132] Step 44: Feed the sorting result back to the predicted remaining life to determine the device maintenance moment corresponding to each lighting device, and generate a corresponding maintenance list and display it.

[0133] In this example, the first device sorting represents the result of sorting lighting devices from high to low according to the severity of the total loss consequences, and the second device sorting represents the result of sorting from near to far according to the total loss time.

[0134] The working principle and beneficial effects of the above technical solution: Using model technology to analyze the device operation parameters of each lighting device, sorting the lighting devices from multiple angles, determining the device maintenance time of each lighting device, and finally generating a corresponding maintenance list. In this way, technical reference can be provided to workers to assist them in maintaining the lighting device.

[0135] Embodiment 8

[0136] Based on Embodiment 7, the predictive maintenance method for a lighting device further includes:

[0137] Obtaining the real-time device status corresponding to each lighting device respectively;

[0138] Screening the target lighting devices with abnormal real-time device status, defining the sorting of the target lighting devices as the first, generating an emergency maintenance warning and performing corresponding alarms.

[0139] The working principle and beneficial effects of the above technical solution: When a certain lighting device suddenly breaks down, its sorting is raised to the first and corresponding alarm work is performed to remind the worker to give priority to maintaining this lighting device.

[0140] Embodiment 9

[0141] This embodiment provides a predictive maintenance system for a lighting device, as Figure 2 shown, including:

[0142] A rule analysis module, configured to construct the lighting working rule of the lighting device according to a plurality of real-time parameter information of the lighting device, and determine a plurality of normal working time periods of the lighting device;

[0143] A prediction execution module, configured to use an autoregressive model to predict the historical working data corresponding to each normal working time period respectively, and obtain the remaining normal working times of the lighting device;

[0144] A damage derivation module, configured to derive the predicted remaining life of the lighting device, construct the real-time working damage of the lighting device using the real-time parameter information, and determine the cumulative damage characteristics corresponding to each lighting device in the lighting device;

[0145] A maintenance assistance module, configured to mark the device total loss time corresponding to each lighting device in the predicted remaining life according to the cumulative damage characteristics, and determine the device maintenance time corresponding to each lighting device.

[0146] In this example, the real-time parameter information includes: information composed of the working current, working voltage, working temperature, working duration, and working humidity of each lighting device at the current moment;

[0147] In this example, the regular working period represents the period during which the lighting device often conducts lighting. For example, the lighting device starts lighting at 19:00 every day and ends lighting at 06:30 the next day;

[0148] In this example, the lighting working pattern represents the lighting working pattern presented by the lighting device within a 30 * 24-hour period;

[0149] In this example, the remaining regular working times represent the number of times the lighting device can complete lighting during the regular working period;

[0150] In this example, the real-time working damage represents the damage to the lighting device caused by external environmental influences and its own operation;

[0151] In this example, the device complete damage moment represents the moment when the lighting device is scrapped;

[0152] In this example, the device maintenance moment represents any moment before the complete damage moment of the lighting device and not within the regular working period.

[0153] The working principle and beneficial effects of the above technical solution: In order to improve the efficiency of workers in maintaining the lighting device and the effectiveness of the lighting device during lighting, the real-time parameter information of the lighting device is collected daily, so as to deduce the lighting working pattern of the lighting device, determine the regular working period of the lighting device, and then use the autoregressive model to analyze the historical working data, and use the prediction function of the natural regression model to determine the remaining regular working times of the lighting device, so as to deduce the predicted remaining life of the lighting device in combination with the known lighting working pattern. At the same time, analyze the real-time working damage of the lighting device according to the real-time parameter information, deduce the complete damage moment of each lighting device in the lighting device by determining the cumulative damage characteristics of each lighting device, and finally determine the device maintenance moment of each lighting device. In this way, it can assist workers in quickly determining the location and type of faults that will occur in the lighting device, so as to carry out preventive maintenance in advance, reduce the cost of large-scale maintenance in the later stage, and bring long-term convenience to the residents and pedestrians within the lighting range.

[0154] Embodiment 10

[0155] Based on Embodiment 9, the predictive maintenance system for a lighting device, the pattern analysis module includes:

[0156] The first analysis unit is used to transmit the real-time parameter information of the lighting device through a communication network, perform data encoding on the real-time parameter information in the application layer of the communication network to obtain the real-time application data of the lighting device, identify the first-byte data of the real-time application data, and construct the real-time data feature of the lighting device according to the data content contained in the first-byte data;

[0157] The second analysis unit is used to identify the second-byte data of the real-time application data, construct the key byte value of the real-time application data according to the first-byte data and the second-byte data, use AI technology to perform type matching on the key byte value to determine the device attribute of the lighting device, and perform data calibration on the real-time application data based on the legal data packet type of the device attribute to obtain the real-time valid data of the lighting device;

[0158] The third analysis unit is used to normalize the real-time valid data corresponding to different moments, sort the normalized data according to the time sequence, perform linear analysis on the sorting result to obtain the data linear relationship of the lighting device, and identify several lighting segments of the lighting device in the data linear relationship to construct the lighting working rule of the lighting device;

[0159] The fourth analysis unit is used to map the lighting working rule into the time space to obtain several historical lighting time periods of the lighting device, perform fuzzification processing on the lighting start moment corresponding to each historical lighting time period respectively to obtain the fuzzy lighting start point corresponding to each historical lighting time period, and determine several regular working time periods of the lighting device.

[0160] In this example, the communication network refers to the network used to transmit the real-time parameter information to the remote terminal;

[0161] In this example, the real-time application data refers to the real-time parameter information processed by the application layer;

[0162] In this example, the first-byte data refers to the data contained in the first byte of the real-time application data, and the second-byte data refers to the data contained in the second byte of the real-time application data;

[0163] In this example, the real-time data feature refers to the feature presented by the data content contained in the first-byte data;

[0164] In this example, the device attribute refers to the attribute presented when the lighting device performs lighting work;

[0165] In this example, the legal data packet type refers to the type of data packets that can be recognized and generated by the device attribute;

[0166] In this example, the historical lighting time period represents the time period during which the lighting device performs lighting work;

[0167] In this example, the fuzzy lighting start point represents the result of expanding the moment of the lighting start point within the historical lighting time period.

[0168] The working principle and beneficial effects of the above technical solution: In order to better maintain the lighting device, it is necessary to first determine the working law of the lighting device, analyze whether it will malfunction according to the working law. First, the application layer of the communication network encodes the real-time parameter information to obtain the real-time application data of the lighting device. Then, the first-byte data and the second-byte data in the real-time application data are analyzed and calibrated multiple times to deduce the real-time effective data of the lighting device. Further, the real-time effective data corresponding to different moments are normalized and sorted, and several lighting segments of the lighting device are determined according to the data linear relationship shown by the sorting result, generating the lighting working law of the lighting device. Since the lighting time period of the lighting device may change, the lighting start moment of the lighting device is fuzzified, and finally several regular working time periods of the lighting device are determined. In this way, not only can the data generated by the lighting device be reasonably processed, but also the regular working time periods of the lighting device are determined, improving the efficiency and quality of subsequent prediction and maintenance.

[0169] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. A predictive maintenance method for a lighting device, characterized in that, Including: Step 1: According to a number of real-time parameter information of the lighting device, construct the lighting working rule of the lighting device, and determine a number of regular working time periods of the lighting device; Step 2: Use the autoregressive model to predict the historical working data corresponding to each of the regular working time periods respectively, and obtain the remaining regular working times of the lighting device; Step 3: Deduce the predicted remaining life of the lighting device, construct the real-time working damage of the lighting device by using the real-time parameter information, and determine the cumulative damage characteristics corresponding to each lighting device in the lighting device; Step 4: Mark the device complete loss moment corresponding to each lighting device in the predicted remaining life according to the cumulative damage characteristics, and determine the device maintenance moment corresponding to each lighting device.

2. The predictive maintenance method for a lighting device according to claim 1, characterized in that, The said Step 1 includes: Step 11: Transmit the real-time parameter information of the lighting device through the communication network, perform data encoding on the real-time parameter information in the application layer of the communication network to obtain the real-time application data of the lighting device, and identify the first byte data of the real-time application data. Construct the real-time data characteristics of the lighting device according to the data content included in the first byte data; Step 12: Identify the second byte data of the real-time application data, construct the key byte value of the real-time application data according to the first byte data and the second byte data, use AI technology to perform type matching on the key byte value to determine the device attribute of the lighting device, and perform data calibration on the real-time application data based on the legal data packet type of the device attribute to obtain the real-time valid data of the lighting device; Step 13: Perform normalization processing on the real-time valid data corresponding to different moments, sort the normalized data according to the time sequence, perform linear analysis on the sorting result to obtain the data linear relationship of the lighting device, and identify a number of lighting segments of the lighting device in the data linear relationship to construct the lighting working rule of the lighting device; Step 14: Map the lighting working rule into the time space to obtain a number of historical lighting time periods of the lighting device, perform fuzzification processing on the lighting start moment corresponding to each historical lighting time period respectively to obtain the fuzzy lighting start point corresponding to each historical lighting time period, and determine a number of regular working time periods of the lighting device.

3. The predictive maintenance method for a lighting device according to claim 2, characterized in that, It also includes: Determine the lighting duration of the corresponding historical lighting time period according to each lighting start moment, and determine the historical average lighting duration of the lighting device; Determine the adjustable duration corresponding to each historical lighting time period based on the historical average lighting duration; Perform fuzzification processing on the corresponding lighting start point within the adjustable duration range to obtain the fuzzy lighting start point corresponding to the corresponding historical lighting time period.

4. The predictive maintenance method for a lighting device according to claim 1, wherein, The said Step 2 includes: Step 21: Use the autoregressive model to perform autocorrelation calculation on each historical working data respectively, determine the current order threshold of the autoregressive model, and adjust the initial window length according to the current order threshold to obtain the effective window function of the lighting device; Step 22: Use the window function to perform windowing processing on each piece of the historical working data to obtain several optimized lighting time periods of the lighting device, and feed the optimized lighting time periods back into the autoregressive model for model fitting to obtain the autoregressive coefficients of the lighting device; Step 23: Perform regression prediction on the historical working data according to the autoregressive coefficients to obtain several predicted lighting moments of the lighting device, determine the failure lighting time periods of the lighting device according to the predicted lighting effect characteristics corresponding to each predicted lighting moment, and determine the remaining normal working times of the lighting device.

5. The predictive maintenance method for a lighting device according to claim 4, wherein It further includes: After obtaining the autoregressive coefficients, perform residual identification on the autoregressive model to obtain the residuals of the autoregressive model; When there is an autocorrelation relationship between the residuals and the historical working data, use the residuals to correct the current order of the autoregressive model to obtain an updated autoregressive model; Use the updated autoregressive model to perform model fitting on the optimized lighting time periods, and update the autoregressive coefficients according to the fitting results.

6. The predictive maintenance method for a lighting device according to claim 1, wherein, The said Step 3 includes: Step 31: Map the remaining normal working times into the lighting working pattern, expand the lighting working pattern according to the mapping positions corresponding to each remaining normal working time, and determine the predicted remaining life of the lighting device according to the expansion length; Step 32: Restore the real-time lighting state of the lighting device according to the real-time parameter information, and construct the cumulative lighting state of the lighting device. Deduce the historical cumulative damage of the lighting device according to the cumulative lighting state, and determine the increased damage of the historical cumulative damage according to the real-time lighting state; Step 33: According to the damage attributes of the increased damage, identify and count the number of damages with the same attributes included in the historical cumulative damage, and use the number of damages with the same attributes to enhance the damage attributes to determine the real-time working damage of the increased damage to the lighting device; Step 34: Use the real-time working damage to perform damage simulation on the lighting device, determine the damage positions corresponding to each lighting moment and the damage values corresponding to each damage position, and determine the cumulative damage characteristics corresponding to each lighting device in the lighting device.

7. The predictive maintenance method for a lighting device according to claim 1, wherein The said Step 4 includes: Step 41: Construct a damage lighting model of the lighting device according to the initial device parameters corresponding to each lighting device in combination with the cumulative damage characteristics, and run the damage lighting model to determine the device operation parameters corresponding to each lighting device; Step 42: Deduce the total damage moments of the corresponding lighting devices using the device operation parameters, mark the total damage moments in the predicted remaining life, and determine the total damage consequences corresponding to each lighting device based on the device attributes corresponding to each lighting device; Step 43: Set a first device sorting for the lighting devices according to the order of the total damage consequences from high to low, set a second device sorting for the lighting devices according to the order of the total damage moments from near to far, and re-sort the lighting devices according to the first device sorting and the second device sorting; Step 44: Feed back the sorting result to the predicted remaining life to determine the device maintenance time corresponding to each lighting device, generate a corresponding maintenance list and display it.

8. A predictive maintenance method for a lighting device according to claim 7, characterized in that, It further includes: Obtain the real-time device status corresponding to each lighting device respectively; Screen the target lighting devices with abnormal real-time device status, define the sorting of the target lighting devices as the first, generate an emergency maintenance warning and give corresponding alarms.

9. A predictive maintenance system for a lighting device, characterized in that, It includes: A rule analysis module for constructing the lighting working rule of the lighting device according to several real-time parameter information of the lighting device and determining several regular working time periods of the lighting device; A prediction execution module for using an autoregressive model to predict the historical working data corresponding to each regular working time period respectively to obtain the remaining regular working times of the lighting device; A damage derivation module for deriving the predicted remaining life of the lighting device, constructing the real-time working damage of the lighting device by using the real-time parameter information, and determining the cumulative damage characteristics corresponding to each lighting device in the lighting device; A maintenance assistance module for marking the device complete loss time corresponding to each lighting device in the predicted remaining life according to the cumulative damage characteristics and determining the device maintenance time corresponding to each lighting device.

10. A predictive maintenance system for a lighting device according to claim 9, characterized in that, The rule analysis module includes: A first analysis unit for transmitting the real-time parameter information of the lighting device through a communication network, encoding the real-time parameter information in the application layer of the communication network to obtain the real-time application data of the lighting device, identifying the first byte data of the real-time application data, and constructing the real-time data characteristics of the lighting device according to the data content included in the first byte data; A second analysis unit for identifying the second byte data of the real-time application data, constructing the key byte value of the real-time application data according to the first byte data and the second byte data, using AI technology to perform type matching on the key byte value to determine the device attribute of the lighting device, and calibrating the real-time application data based on the legal data packet type of the device attribute to obtain the real-time valid data of the lighting device; A third analysis unit for normalizing the real-time valid data corresponding to different times, sorting the normalized data according to the time sequence, performing linear analysis on the sorting result to obtain the data linear relationship of the lighting device, and identifying several lighting segments of the lighting device in the data linear relationship to construct the lighting working rule of the lighting device; A fourth analysis unit for mapping the lighting working rule into the time space to obtain several historical lighting time periods of the lighting device, respectively performing fuzzification processing on the lighting start time corresponding to each historical lighting time period to obtain the fuzzy lighting start point corresponding to each historical lighting time period, and determining several regular working time periods of the lighting device.

Images