Intelligent gateway and device monitoring method based on intelligent gateway, and electronic device

Abstract

The application discloses an intelligent gateway, a device monitoring method based on the intelligent gateway and an electronic device. The calculation module of the application processes data information collected by the collection module into input of a preset neural network algorithm, calculates an output result, and judges whether the device is normally operated according to the output result. When an abnormality is determined, the weight of the input of the preset neural network algorithm is adjusted. When the calculation module determines that the adjustment is invalid, a prediction alarm information is generated. When it is determined that the device is still in a normal state, the weight parameter of the preset neural network is updated to the adjusted weight parameter, the self-optimization of the neural network algorithm is realized, the problem that it is usually difficult and tedious for a person to design a perfect neural network algorithm model by experience, and when there is a deviation in parameter setting, the deviation also has a certain influence on the determination of the monitoring result, and even an analysis error is caused is solved, the intelligent degree of the application is further enhanced, and the application scenario of the application is greatly expanded.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of intelligent gateway devices, and specifically discloses an intelligent gateway and a device monitoring method based on the intelligent gateway and an electronic device. BACKGROUND

[0002] With the continuous development of various industries, more and more devices use machine learning and neural network algorithms to monitor the running state of the devices. With the gradual increase in the use of devices, the devices are prone to be affected by the fluctuations or aging of their own electronic components or the interference of external environmental factors, such as electromagnetic field fluctuations, fluctuating environmental temperature, environmental humidity and the like, which may have certain influence on the monitoring result analysis. The device products are various, and it is usually difficult and tedious for a person to design a perfect neural network algorithm model by experience, and when there is a deviation in the parameter setting, it will also have certain influence on the determination of the monitoring result, and even cause analysis errors. Considering the monitoring quality and algorithm self-optimization, the requirements for the intelligence degree and usability of the monitoring device are higher and higher, and there is an urgent need for an intelligent gateway with high intelligence and high usability and a device monitoring method based on the intelligent gateway to realize self-optimization of the neural network algorithm and realize high-intelligent and high-usable device monitoring.

[0003] Therefore, the prior art still needs to be further developed and improved. SUMMARY

[0004] In view of various deficiencies of the prior art and in order to solve the above problems, the present application provides a mold management method, a mold management system and an electronic device with high reliability and high intelligence, and the present application provides the following technical solutions:

[0005] According to a first aspect of the present application, a device monitoring method based on an intelligent gateway is provided, characterized in that,

[0006] The current signal and the voltage signal of the device collected by the collection module of the intelligent gateway are acquired according to a first detection period;

[0007] The signals are analyzed, and the current data and the voltage data obtained after analysis are processed according to a preset rule and then input into a preset neural network algorithm according to a first weight, a first factor value, a second weight and a second factor value to calculate an output result. When the output result is greater than a preset threshold, the current signal and the voltage signal of the device collected by the collection module are acquired according to a second detection period;

[0008] The current signal and the voltage signal are respectively acquired and analyzed for a first preset number of times, the current data and the voltage data obtained after the analysis are processed according to a preset rule, and then the processed current data and voltage data are respectively input into a preset neural network algorithm according to a first weight, a first factor value, a second weight and a second factor value, and an output result is calculated, when the number of times that the output result is greater than a preset threshold value exceeds a second preset number of times, the first weight and the second weight of a third preset number of times are synchronously adjusted according to a first preset proportion, and after the first weight and the second weight are adjusted each time, the current signal and the voltage signal collected by the acquisition module are acquired according to a third detection period;

[0009] The current signal and the voltage signal are respectively acquired and analyzed for a fourth preset number of times according to a third detection period, the current data and the voltage data obtained after the analysis are processed according to a preset rule, and then the processed current data and voltage data are respectively input into a neural network algorithm according to a first factor value, an adjusted first weight and a second factor value, an adjusted second weight, an output result is calculated, and the output result is compared and analyzed with a preset threshold value, if the data increment of the output result close to the preset threshold value is greater than or equal to the data increment away from the preset threshold value, the first weight and the second weight are continuously adjusted according to a first preset proportion, when the number of times that the first weight and the second weight are adjusted accumulatively is less than or equal to a third preset number of times and the output result is less than the preset threshold value, the calculation module determines that the equipment is still in a normal operating state at this time, and the first weight and the second weight of the preset neural network are updated to the first weight after the last adjustment and the adjusted second weight.

[0010] Further, the method further comprises:

[0011] If the data increment of the output result close to the preset threshold value is greater than or equal to the data increment away from the preset threshold value in the current third detection period, the first weight and the second weight are decreased according to a first preset proportion until the output result is less than the preset threshold value or the number of times that the first weight and the second weight are adjusted accumulatively is greater than or equal to a third preset number of times and the output result is still greater than the preset threshold value.

[0012] If the data increment of the output result close to the preset threshold value is less than the data increment away from the preset threshold value in the current third detection period, the first weight and the second weight are increased according to a second preset proportion until the output result is less than the preset threshold value or the number of times that the first weight and the second weight are adjusted accumulatively is greater than or equal to a third preset number of times and the output result is still greater than the preset threshold value.

[0013] Further, the method further comprises:

[0014] If the number of times that the first weight and the second weight are adjusted accumulatively is greater than or equal to a third preset number of times and the output result is still greater than the preset threshold value, the calculation module determines that the equipment is in an abnormal state at this time, the adjusted first weight and the second weight are restored to the initial values of this cycle, the calculation module generates a prediction alarm information and sends it to the server.

[0015] Further, the method further comprises:

[0016] The output result is calculated by summing the product of the first weight and the first factor value and the product of the second weight and the second factor value.

[0017] Further, the method further comprises:

[0018] The second detection period is less than the first detection period.

[0019] Further, the method further comprises:

[0020] The third detection period is less than the second detection period.

[0021] Further, the method further comprises:

[0022] The first weight is 20, the second weight is 10, the first preset proportion is 0.2%, the second preset proportion is 0.3%, the first preset number of times is 20, the second preset number of times is 15, the third preset number of times is 5, the fourth preset number of times is 10, the first detection period is 10 minutes, the second detection period is 5 minutes, and the third detection period is 5 seconds.

[0023] Further, the method further comprises:

[0024] The first weight is 20, the second weight is 10, the first preset proportion is 0.2%, the second preset proportion is 0.3%, the first preset number of times is 20, the second preset number of times is 15, the third preset number of times is 5, the fourth preset number of times is 10, the first detection period is 10 minutes, the second detection period is 5 minutes, and the third detection period is 5 seconds.

[0025] According to a second aspect of the present application, an intelligent gateway is provided, characterized by comprising a collection module, a calculation module and a server;

[0026] The collection module is configured to collect data information of the device.

[0027] The calculation module is configured to process the voltage signal and the current signal collected by the collection module as the input of the preset neural network algorithm and calculate the output result, and determine whether the device is normally running according to the output result, adjust the weight of the input of the preset neural network algorithm when determining that the device is abnormal, and generate a prediction alarm information and send it to the server when the calculation module determines that the adjustment is invalid and the device is in an abnormal state.

[0028] A server is configured to uniformly manage control programs of all devices and issue configuration instructions.

[0029] According to a third aspect of the present application, an electronic device is provided, comprising:

[0030] A memory and a processor, wherein computer readable instructions are stored on the memory and are executed by the processor to implement the device monitoring method based on the intelligent gateway.

[0031] The present application has the following advantages:

[0032] 1. The present application can update the first factor value and the second factor value of the neural network algorithm in real time by processing the current and voltage data of the device collected by the collection module when the current and voltage data of the device collected by the collection module fluctuate, realize intelligent integration monitoring of the voltage, current and running state of the device, greatly improve the intelligence level of the intelligent gateway, greatly improve the monitoring quality, and greatly expand the application scenarios of the present application.

[0033] 2. The present application can prevent the device from being easily affected by normal fluctuations or aging of its own electronic components or interference from external environmental factors such as electromagnetic field fluctuations, fluctuating environmental temperature, environmental humidity and other factors that may affect the monitoring result analysis, greatly improving the anti-interference and stability of the system.

[0034] 3. The present application can realize intelligent monitoring and analysis of the running state of the device through algorithm only, without complex system and modeling operation, greatly simplifying the system architecture under the condition of ensuring analysis accuracy, greatly reducing the system production, installation and maintenance cost, and greatly improving the enterprise benefit.

[0035] 4. The calculation module of the present application processes the data information collected by the collection module into the input of the neural network algorithm and calculates the output result, and judges whether the device is running normally according to the output result, adjusts the weight of the input of the neural network algorithm when an abnormality is determined, generates a prediction warning information when the calculation module determines that the adjustment is invalid, updates the weight parameter of the preset neural network to the adjusted weight parameter when it is determined that the device is still in a normal state, realizes self-optimization of the neural network algorithm, solves the problem that it is usually difficult and tedious for a person to design a perfect neural network algorithm model by experience, and the parameter setting deviation will also affect the determination of the monitoring result, even cause analysis error, further enhances the intelligence level of the present application, and greatly expands the application scenarios of the present application.

[0036] 5、The first weight is 20, the second weight is 10, the first preset ratio is 0.2%, the second preset ratio is 0.3%, the first preset number is 20 times, the second preset number is 15 times, the third preset number is 5 times, the fourth preset number is 10 times, the first detection period is 10 minutes, the second detection period is 5 minutes, and the third detection period is 5 seconds, which are designed ingeniously. By comprehensively considering the possible conditions on site and the circuit implementation mode, the parameters of the circuit are optimally designed, the influence of various conditions on site on the judgment result of the system is avoided, the program architecture is optimized, the system analysis efficiency is greatly improved, the time and labor cost are greatly saved, and the ultra-low power consumption and high reliability operation of the intelligent gateway are realized. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 A device monitoring method flowchart realized based on an intelligent gateway in specific embodiments of the present application is shown in the figure.

[0038] Figure 2 A schematic diagram of a device monitoring method realized based on an intelligent gateway in another specific embodiment of the present application is shown in the figure.

[0039] Figure 3 A principle block diagram of an intelligent gateway in specific embodiments of the present application is shown in the figure. DETAILED DESCRIPTION

[0040] In order for those skilled in the art to better understand the technical solutions of the present application, the technical solutions of the present application will be described clearly and completely below in combination with the drawings of the present application. Based on the embodiments in the present application, other similar embodiments obtained by those skilled in the art without making creative efforts should all belong to the scope of protection of the present application. In addition, the direction words mentioned in the following embodiments, such as "up", "down", "left", "right", etc. are only reference directions of the drawings, therefore, the direction words used are used for illustration and not for limiting the present application.

[0041] The present application will be further described below in combination with the drawings and preferred embodiments.

[0042] Please refer to Figure 1 The present application provides a device monitoring method realized based on an intelligent gateway, which comprises:

[0043] S100, acquiring the current signal and voltage signal of the device collected by the acquisition module of the intelligent gateway according to a first detection period.

[0044] It should be noted here that the acquisition module is pre-set in the device of the present application, which is used to collect the current and voltage signals of the device in real time, i.e. the acquisition module can generate signals about the changes of current and voltage in real time.

[0045] Before step S100, the first detection period, the first weight, the second weight, the first preset proportion, the second preset proportion, the first preset number, the second preset number, the third preset number, the fourth preset number, the second detection period, and the third detection period are preset. Based on the preset first detection period, the calculation module periodically collects the detection signal generated by the collection module. Preferably, the first detection period is 10 minutes, the first weight is 20, the second weight is 10, the first preset proportion is 0.2%, the second preset proportion is 0.3%, the first preset number is 20 times, the second preset number is 15 times, the third preset number is 5 times, the fourth preset number is 10 times, the second detection period is 5 minutes, and the third detection period is 5 seconds.

[0046] In S200, the signal is analyzed, and the current data and voltage data obtained after analysis are processed according to a preset rule and input into a preset neural network algorithm according to the first weight, the first factor value, the second weight, and the second factor value to calculate an output result. When the output result is greater than a preset threshold, the current signal and the voltage signal of the collection device are collected according to the second detection period.

[0047] It needs to be explained here that the server analyzes the obtained current and voltage signals, and the current and voltage information data of the device at the current time obtained by the acquisition module are included in the analyzed data. The first factor value is obtained by rounding the current value, and the second factor value is obtained by rounding the voltage value. The calculation method of the output result is to sum the product of the first weight and the first factor value and the product of the second weight and the second factor value. The first weight is 20, and the second weight is 10. The first weight is set to 20 and the second weight is set to 10, which is obtained by the inventors through a large number of experiments. The current fluctuation of the device is usually the determining factor of causing damage or other abnormal conditions of the device. Therefore, in the neural network algorithm of the present application, the first weight is assigned a value of 20, and the first weight corresponds to the first factor value. The second weight is assigned a value of 10, and the second weight corresponds to the second factor value. This setting can greatly simplify the system architecture and system complexity, analyze the current and voltage state of the device, greatly reduce the production and maintenance cost of the system, greatly improve the usability of the present application, and greatly expand the application scenarios of the present application. This calculation method can realize the processing of the current and voltage information collected by the acquisition module into the input of the neural network algorithm and the calculation of the output result, and determine whether the device is running normally. When an abnormality is determined, the weight of the input of the neural network algorithm is adjusted. When the calculation module determines that the adjustment is invalid, a prediction warning information is generated. When it is determined that the device is still in a normal state, the weight parameters of the preset neural network are updated to the adjusted weight parameters, realizing the self-optimization of the neural network algorithm. The problem of difficulty and tediousness in designing and perfecting a neural network algorithm model by artificial experience, and the problem of deviation in parameter setting also affecting the determination of the monitoring result and even causing analysis errors are solved. The intelligent degree of the present application is further enhanced, and the application scenarios of the present application are greatly expanded. The preset threshold value is compared with the analyzed data, and the second detection period is 5 minutes. The second detection period is set to 5 minutes, which is obtained by the inventors through a large number of experiments. When the output result calculated periodically by the calculation module according to the first detection period is greater than the preset threshold value, it indicates that the current and / or voltage of the device during operation fluctuates. At this time, the calculation module needs to determine whether the current and / or voltage fluctuation at this time is due to the fluctuation or aging of the electronic components of the device or the interference of external environmental factors such as electromagnetic field fluctuation, fluctuating environmental temperature, and environmental humidity. The server needs to determine whether the device is still in a normal state. Shortening the second detection period to 5 minutes can ensure the reliability of the detection result while preventing system misjudgment caused by fluctuations of various factors, further enhancing the stability of the system, reducing the computational load of the calculation module during normal monitoring, realizing low-power operation of the calculation module, and greatly saving computational resources.When the output result calculated according to the first detection period is less than the preset threshold value, the calculation module judges that the equipment is still in a normal operation state, at this time, the system continues to periodically acquire the detection signal of the acquisition module according to the first detection period. By analyzing only the current and voltage detection signals of the equipment, the analysis of the quality of the injection workpiece can be completed, which greatly simplifies the analysis process, greatly improves the analysis efficiency, reduces the system complexity, greatly reduces the cost of each link of system production, installation and operation and maintenance, and greatly improves the enterprise benefit.

[0048] S300, respectively acquire the current signal and the voltage signal of the first preset number of times and analyze, input the current data and the voltage data obtained after analysis into a preset neural network algorithm according to a first weight, a first factor value, a second weight and a second factor value after processing according to a preset rule, when the number of times that the output result is greater than the preset threshold value exceeds the second preset number of times, adjust the first weight and the second weight of the third preset number of times according to the first preset proportion synchronously, and acquire the current signal and the voltage signal of the equipment collected by the acquisition module according to the third detection period after adjusting the first weight and the second weight each time.

[0049] Specifically, when the output result is compared with the preset threshold value, the calculation module periodically detects the output result in the first detection period, and then another workflow is started, that is, a preset second detection period is called to obtain the current and voltage detection signals of the equipment collected by the collection module. The second detection period is 5 minutes. The second detection period of 5 minutes is obtained by the technical personnel of the present application through a large number of experiments. When the output result calculated by the neural network algorithm of the calculation module is greater than the preset threshold value at a certain moment, an abnormal situation event must occur. At this time, the second detection period is set in advance and is less than the first detection period, which greatly improves the reliability of subsequent data analysis. By optimizing the program structure, only when the current and voltage of the equipment fluctuate, the second detection period is entered, and the ultra-low power consumption operation of the system of the present application is realized. Therefore, the second detection period of 5 minutes can improve the reliability of data analysis after a dangerous or emergency event occurs and realize the ultra-low power consumption operation of the system of the present application. The first preset number is 20, and the second preset number is 15. The first preset number is set to 20 and the second preset number is set to 15, which is obtained by the technical personnel of the present application through a large number of experiments. Because, when the output result is greater than the preset threshold value at a certain moment, an abnormal situation event must occur. At this time, the detection signal is obtained by the detection frequency as much as possible, and the analysis is performed. While ensuring the accuracy of the analysis result, unnecessary CPU resource waste is avoided. If it is a normal fluctuation or aging of the state of the electronic components, the device should still be in a normal operating state. If it is an interference of external environmental factors, such as electromagnetic field fluctuation, fluctuating environmental temperature, environmental humidity and other factors, the running state of the device should be restored to normal within 5 minutes, that is, 20 output results obtained by the neural network algorithm within 5 minutes, 15 or more output results are less than the preset threshold value. If the running current and voltage of the device cannot be restored to normal within 5 minutes, that is, only 15 or fewer output results are less than the preset threshold value among the 20 pressure data collected within 5 minutes, the server judges that the device running state cannot be restored automatically and needs to enter the weight adjustment stage, so as to avoid system misjudgment due to normal fluctuation or aging of the state of the electronic components. Because if 5 or more output results obtained by calculation within 5 minutes are still greater than the preset threshold value, it represents that the running state of the device fluctuates greatly at this time, and the weight adjustment stage is entered. On the premise of reliable decomposition result, further waste of various costs is avoided. This setting can better analyze whether the device voltage and current fluctuation at this time can be restored automatically, and the degree is reasonably designed, which greatly improves the accuracy of the analysis result of the present application and reduces the energy consumption of the system of the present application.

[0050] S400, the fourth preset number of current signals and voltage signals are acquired respectively according to a third detection period and are analyzed, the current data and voltage data obtained after analysis are processed according to a preset rule, and then the processed current data and voltage data are input into a neural network algorithm according to a first factor value, an adjusted first weight and a second factor value, an adjusted second weight, and an output result is calculated, and the output result is compared and analyzed with a preset threshold value, if the data increment of the output result close to the preset threshold value is greater than or equal to the data increment far from the preset threshold value, the first weight and the second weight are continued to be adjusted according to a first preset proportion, when the number of cumulative adjustment of the first weight and the second weight is less than or equal to a third preset number and the output result is less than the preset threshold value, the calculation module determines that the equipment is still in a normal operation state at this time, and the first weight and the second weight of the preset neural network are updated to the last adjusted first weight and the adjusted second weight.

[0051] Specifically, if the data increment of the output result close to the preset threshold value is greater than or equal to the data increment far from the preset threshold value in the current third detection period, the first weight and the second weight are decreased according to the first preset proportion until the output result is less than the preset threshold value or the output result is still greater than the preset threshold value after the first weight and the second weight are cumulatively adjusted by the third preset number respectively.

[0052] If the data increment of the output result close to the preset threshold value is less than the data increment far from the preset threshold value in the current third detection period, the first weight and the second weight are increased according to the second preset proportion until the output result is less than the preset threshold value or the output result is still greater than the preset threshold value after the first weight and the second weight are cumulatively adjusted by the third preset number respectively.

[0053] Specifically, the first preset proportion is 0.2%, the second preset proportion is 0.3%, the third detection period is 5 seconds, the third preset number is 5 times, and the fourth detection number is 10 times. The first preset proportion of 0.2% and the second preset proportion of 0.3% are obtained by a large number of experiments of the present technical person, and the setting can better complete the fine adjustment of the first weight and the second weight, can better realize the traversal adjustment of the weight parameter, greatly simplifies the program system complexity, improves the parameter adjustment quality and efficiency, greatly reduces the cost required for production and installation of the present application, and greatly expands the application scenarios of the present application.

[0054] Generally, when the device is disturbed by the normal fluctuation or aging of its own electronic components or external environmental factors, such as electromagnetic field fluctuation, fluctuating ambient temperature, ambient humidity and other factors, the current will increase, and the current increase is usually the direct cause of abnormal conditions or even dangerous conditions, so the application preferentially reduces the first weight and the second weight according to the first preset ratio, and when the first weight and the second weight cannot make the output result close to the preset threshold, the application increases the first weight and the second weight according to the second preset ratio, which can greatly improve the efficiency and reliability of the weight adjustment process. The above setting can reduce the computational complexity of the system program on the basis of ensuring the adjustment accuracy, and the program design is ingenious and reasonable, which solves the problem that the existing technology is very complex and tedious through manual adjustment of the neural network model, greatly improves the weight adjustment density and weight adjustment efficiency, and greatly improves the enterprise benefit.

[0055] The third preset number of times is set to 5 times, the third detection period is set to 5 seconds, and the fourth preset number of times is set to 10 times, which is obtained by the present inventors through a large number of experiments. This setting can better analyze the changes of the current and voltage after each adjustment of the weight, save analysis time, improve the weight adjustment efficiency, and greatly expand the application scenarios of the device.

[0056] Specifically, the calculation module draws the 10 output results obtained in the third detection period in sequence, when the change trend of the adjacent two output results approaches the preset threshold, the change amount of the adjacent two output results is called the data sub-increment approaching the preset threshold, when the change trend of the adjacent two output results deviates from the preset threshold, the change amount of the adjacent two output results is called the data sub-increment deviating from the preset threshold, the server sums up each data sub-increment approaching the preset threshold as the data increment approaching the preset threshold in the current third detection period, sums up each data sub-increment deviating from the preset threshold as the data increment deviating from the preset threshold in the current third detection period, if the data increment approaching the preset threshold in the current third detection period is greater than the data increment deviating from the preset threshold, the calculation module judges that the output result in the current third period approaches the preset threshold, this weight adjustment is effective, the calculation module continues to adjust the first weight and the second weight according to the first preset proportion, until the output result is less than the preset threshold or the output result is still greater than the preset threshold after the first weight and the second weight are adjusted cumulatively for the third preset number of times. When the first weight and the second weight are adjusted cumulatively for the third preset number of times, the output result obtained by the calculation module is still greater than the preset threshold when detected in the next third period, at this time, the calculation module judges that the weight adjustment is invalid, at this time, the current current and voltage fluctuation of the device is an abnormal situation, all adjustments of the first weight and the second weight in this cycle are cancelled, a prediction alarm information is generated and sent to the server. The present application initiatively completes the analysis of the running current and voltage change of the device after the weight parameter adjustment only by analyzing the 10 output results obtained according to the third detection period, without complex modeling and algorithm operation, greatly improving the parameter adjustment efficiency, greatly simplifying the analysis algorithm, greatly reducing the cost required for system production, installation and maintenance, and greatly improving the enterprise benefit.

[0057] Please refer to Figure 2 The present application provides a device monitoring method based on an intelligent gateway, which comprises the following steps:

[0058] P0: Start.

[0059] P1: Obtain the current signal and voltage signal of the device collected by the collection module of the intelligent gateway according to the first detection period.

[0060] It should be noted that the device of the present application is provided with a collection module for collecting the current and voltage signals of the device in real time, that is, the collection module can generate signals about the change of current and voltage in real time.

[0061] Before step P1, the first detection period, the first weight, the second weight, the first preset ratio, the second preset ratio, the first preset number, the second preset number, the third preset number, the fourth preset number, the second detection period and the third detection period are preset. Based on the preset first detection period, the calculation module periodically collects the detection signal generated by the collection module. Preferably, the first detection period is 10 minutes, the first weight is 20, the second weight is 10, the first preset ratio is 0.2%, the second preset ratio is 0.3%, the first preset number is 20 times, the second preset number is 15 times, the third preset number is 5 times, the fourth preset number is 10 times, the second detection period is 5 minutes, and the third detection period is 5 seconds.

[0062] P2, the analyzed signal and the current data and voltage data obtained after analysis are processed according to the preset steps, and then input into a preset neural network algorithm according to the first weight, the first factor value and the second weight, the second factor value to calculate an output result.

[0063] It should be noted that the server analyzes the obtained current and voltage signals, and the analysis data includes the current and voltage information data of the equipment at the current time obtained by the collection module. The first factor value is obtained by rounding the current value, and the second factor value is obtained by rounding the voltage value. The calculation method of the output result is to sum the product of the first weight and the first factor value and the product of the second weight and the second factor value. The first weight is 20, and the second weight is 10. The first weight is set to 20 and the second weight is set to 10, which is obtained by the inventors through a large number of experiments. The current fluctuation of the equipment is usually the determining factor of equipment damage or other abnormal conditions. Therefore, in the neural network algorithm of the present application, the first weight is assigned a value of 20, and the first weight corresponds to the first factor value. The second weight is assigned a value of 10, and the second weight corresponds to the second factor value. This setting can greatly simplify the system architecture and system complexity, analyze the current and voltage state of the equipment, greatly reduce the production and maintenance cost of the system, greatly improve the usability of the present application, and greatly expand the application scenarios of the present application. This calculation method can realize the processing of the current and voltage information collected by the collection module into the input of the neural network algorithm and the calculation of the output result, and determine whether the equipment is running normally according to the output result. When an abnormality is determined, the weight of the input of the neural network algorithm is adjusted. When the calculation module determines that the adjustment is invalid, a prediction warning information is generated. The neural network algorithm is self-optimized, which solves the problems that it is usually difficult and tedious for a person to design and perfect a neural network algorithm model by experience, and when the parameter setting has a deviation, it will also have a certain impact on the determination of the monitoring result, and even cause analysis errors. The intelligence level of the present application is further enhanced, and the application scenarios of the present application are greatly expanded.

[0064] P3, compare whether the output result is greater than a preset threshold value? If yes, execute step P4, if not, return to step P1.

[0065] It should be noted that the preset threshold value is compared with the parsed data, the first detection number is 20 times, the first detection number is 20 times, and the second detection period is set to 5 minutes. The first detection number is 20 times, and the second detection period is set to 5 minutes. The technical personnel of the present application obtains through a large number of experiments. When the output result calculated by the calculation module periodically according to the first detection period is greater than the preset threshold value, it is indicated that the current and / or voltage of the running device fluctuates. At this time, the calculation module needs to judge whether the current and / or voltage fluctuation at this time is due to the fluctuation or aging of the electronic components of the calculation module or the interference of external environmental factors, such as electromagnetic field fluctuation, fluctuating environment temperature, environment humidity and the like. The server needs to judge whether the device is still in a normal state at this time. Shortening the second detection period to 5 minutes can ensure the reliability of the detection result while preventing system misjudgment caused by various factor fluctuations, further enhancing the stability of the system, and reducing the calculation amount of the calculation module during normal monitoring, realizing low-power consumption operation of the calculation module, greatly saving the calculation resources. When the output result calculated according to the first detection period is less than the preset threshold value, the calculation module judges that the device is still in a normal running state, and the system continues to periodically acquire the detection signal of the acquisition module according to the first detection period. Only by analyzing the current and voltage detection signals of the device, the quality of the injection molded workpiece can be analyzed, which greatly simplifies the analysis process, greatly improves the analysis efficiency, reduces the system complexity, greatly reduces the system production, installation, operation and maintenance cost, and greatly improves the enterprise benefit.

[0066] P4, acquire the current signal and voltage signal of the device collected by the acquisition module for the first preset number of times according to the second detection period respectively and parse, input the parsed current data and voltage data according to the first factor value, the first weight and the second factor value, the second weight into the preset neural network algorithm to calculate the output result.

[0067] P5, the calculation module compares whether the number of times that the output result is greater than the preset threshold value exceeds the second preset number of times? If yes, execute step P6; if not, return to step P1.

[0068] When the output result is compared with the preset threshold value, the calculation module periodically detects the output result in the first detection period, and then another workflow is started, that is, a preset second detection period is called to obtain the current and voltage detection signals of the equipment collected by the collection module. The second detection period is 5 minutes. The second detection period of 5 minutes is obtained by the technical personnel of the application through a large number of experiments. When the output result calculated by the calculation module through the neural network algorithm is greater than the preset threshold value at a certain moment, an abnormal situation event must occur. At this time, the second detection period is set in advance and is less than the first detection period, which greatly improves the reliability of subsequent data analysis. By optimizing the program structure, only when the current and voltage of the equipment fluctuate, the second detection period is entered, and the ultra-low power consumption operation of the system of the application is realized. The first preset number is 20, and the second preset number is 15. The first preset number is set to 20, and the second preset number is set to 15. This is obtained by the technical personnel of the application through a large number of experiments. Because, when the output result is greater than the preset threshold value at a certain moment, an abnormal situation event must occur. At this time, the detection signal is obtained by the detection frequency as much as possible, and the analysis is performed. While ensuring the accuracy of the analysis result, unnecessary CPU resource waste is avoided. If it is a normal fluctuation or aging of the state of the electronic components, the device should still be in a normal operating state. If it is an interference of external environmental factors, such as electromagnetic field fluctuation, fluctuating environmental temperature, environmental humidity and other factors, the running state of the device should be restored to normal within 5 minutes, that is, 20 output results obtained by the neural network algorithm within 5 minutes, 15 or more output results are less than the preset threshold value. If the running current and voltage of the device cannot be restored to normal within 5 minutes, that is, only 15 or fewer output results are less than the preset threshold value among the 20 pressure data collected within 5 minutes, the server judges that the device running state cannot be restored automatically, and needs to enter the weight adjustment stage, so as to avoid system misjudgment due to normal fluctuation or aging of the state of the electronic components. Because if 5 or more output results obtained by calculation within 5 minutes are still greater than the preset threshold value, it represents that the running state of the device fluctuates greatly at this time, and the weight adjustment stage is entered. On the premise of reliable decomposition result, further waste of various costs is avoided. This setting can better analyze whether the device voltage and current fluctuation at this time can be restored automatically, and the degree design is reasonable, which greatly improves the accuracy of the analysis result of the application and reduces the energy consumption of the system of the application.

[0069] P6, reducing the first weight and the second weight according to the first preset ratio for a third preset number of times and obtaining the current signal and the voltage signal of the device collected by the collection module according to a third detection period after each adjustment of the first weight and the second weight.

[0070] It should be noted that the first preset ratio is 0.2%, the second preset ratio is 0.3%, the third detection period is 5 seconds, the third preset number of times is 5 times, and the fourth preset number of times is 10 times. The first preset ratio of 0.2% and the second preset ratio of 0.3% are obtained by the inventors through a large number of experiments. This setting can better fine-tune the first weight and the second weight, can better realize the traversal adjustment of the weight parameters, greatly simplifies the complexity of the program system, improves the parameter adjustment quality and efficiency, greatly reduces the cost required for production and installation of the application, and greatly expands the application scenarios of the application.

[0071] Generally, when the device is subjected to normal fluctuations or aging of its own electronic components or interference from external environmental factors, such as electromagnetic field fluctuations, fluctuating environmental temperature, environmental humidity, and other factors, the current will increase, and the increase in current is usually the direct cause of abnormal situations or even dangerous situations. Therefore, the application preferentially reduces the first weight and the second weight according to the first preset ratio. When the first weight and the second weight cannot be reduced to make the output result close to the preset threshold, the application further increases the first weight and the second weight according to the second preset ratio. This setting can greatly improve the efficiency and reliability of the weight adjustment process. The above setting can reduce the computational complexity of the system program on the basis of ensuring the adjustment accuracy, and the program design is ingenious and reasonable, solving the problem of complex and tedious manual adjustment of the neural network model in the prior art, greatly improving the weight adjustment density and weight adjustment efficiency, and greatly improving the enterprise efficiency.

[0072] Setting the third preset number of times to 5 times, the third detection period to 5 seconds, and the fourth preset number of times to 10 times is obtained by the inventors through a large number of experiments. This setting can better analyze the changes in current and voltage after each adjustment of the weight, save analysis time, improve weight adjustment efficiency, and greatly expand the application scenarios of the device.

[0073] P7, obtaining the current signal and the voltage signal for a fourth preset number of times according to a third detection period and analyzing them, inputting the analyzed current data and voltage data according to a first factor value, an adjusted first weight, a second factor value, and an adjusted second weight into a preset neural network algorithm to calculate an output result, and comparing and analyzing the output result with a preset threshold.

[0074] It should be noted that the third preset number of times is set to 5 times, the third detection period is set to 5 seconds, and the fourth preset number of times is set to 10 times, which is obtained by the technical personnel of the present application through a large number of experiments. This setting can analyze the changes of current and voltage after adjusting the weight each time, save analysis time, improve weight adjustment efficiency, and greatly expand the application scenarios of the device.

[0075] P8, is the data increment close to the preset threshold greater than or equal to the data increment away from the preset threshold? If yes, step P9 is executed; if no, step P12 is executed.

[0076] It should be noted that the calculation module sequentially connects and draws the 10 output results obtained by the calculation module according to the third detection period. When the change trend of the adjacent two output results is close to the preset threshold, the change amount of the adjacent two output results is called the data sub-increment close to the preset threshold. When the change trend of the adjacent two output results is away from the preset threshold, the change amount of the adjacent two output results is called the data sub-increment away from the preset threshold. The server collects each data sub-increment close to the preset threshold as the data increment close to the preset threshold in the current third detection period, and collects each data sub-increment away from the preset threshold as the data increment away from the preset threshold in the current third detection period. If the data increment close to the preset threshold in the current third detection period is greater than the data increment away from the preset threshold, the calculation module determines that the output result in the current third period is close to the preset threshold, this weight adjustment is effective, and the calculation module continues to adjust the first weight and the second weight according to the first preset proportion until the output result is less than the preset threshold or the first weight and the second weight are adjusted a total of the third preset number of times and the output result is still greater than the preset threshold. When the first weight and the second weight are adjusted a total of the third preset number of times, the output result calculated by the calculation module is still greater than the preset threshold in the next third period detection. At this time, the calculation module determines that the weight adjustment is ineffective, the current current and voltage fluctuation of the device is an abnormal situation, all adjustments of the first weight and the second weight in this cycle are cancelled, a prediction warning information is generated and sent to the server. The present application initiatively analyzes the running current and voltage change of the device after the weight parameter adjustment only by analyzing the 10 output results obtained according to the third detection period, without complex modeling and algorithm operation, greatly improves the parameter adjustment efficiency, greatly simplifies the analysis algorithm, greatly reduces the cost required for system production, installation and maintenance, and greatly improves the enterprise efficiency.

[0077] P9, is the output result greater than the preset threshold? If yes, step P10 is executed; if no, step P1 is returned.

[0078] P10, is the cumulative adjustment number greater than the third preset number of times? If yes, step P14 is executed; if no, step P11 is executed.

[0079] It should be noted that the adjustment of the first weight and the second weight is a synchronous adjustment, and the synchronous adjustment of the first weight and the second weight is recorded as one adjustment in the present application.

[0080] P11, continue to reduce the first weight and the second weight according to the first preset proportion, and then execute step P7.

[0081] It should be noted that if the data increment close to the preset threshold is greater than the data increment away from the preset threshold in the current third detection period, the calculation module determines that the output result in the current third period is close to the preset threshold, this weight adjustment is effective, the calculation module continues to adjust the first weight and the second weight according to the first preset proportion until the output result is less than the preset threshold or the output result is still greater than the preset threshold after the third preset number of adjustments of the first weight and the second weight respectively.

[0082] P12, increase the first weight and the second weight according to the second preset proportion.

[0083] It should be noted that generally, when the device is disturbed by normal fluctuations or aging of its own electronic components or external environmental factors such as electromagnetic field fluctuations, fluctuating ambient temperature, ambient humidity and the like, the current will increase, and generally the increase in current is the direct cause of abnormal conditions or even dangerous conditions, so the present application preferentially reduces the first weight and the second weight according to the first preset proportion, and when the first weight and the second weight cannot be reduced to make the output result close to the preset threshold, the present application increases the first weight and the second weight according to the second preset proportion, which can greatly improve the efficiency and reliability of the weight adjustment process. The above setting can reduce the computational complexity of the system program on the basis of ensuring the adjustment accuracy, and the program design is ingenious and reasonable, which solves the problem of complex and tedious manual adjustment of the neural network model in the prior art, greatly improves the weight adjustment density and efficiency, and greatly improves the enterprise efficiency.

[0084] P13, compare whether the data increment of the output result close to the preset threshold is greater than or equal to the data increment away from the preset threshold? If yes, execute step P15, if no, execute step P14.

[0085] P14, cancel all weight adjustments in this cycle, generate a prediction warning information and send it to the server, and then execute step P17.

[0086] P15, whether the output result is greater than the preset threshold? If yes, execute step P16; if no, return to step P1.

[0087] P16, whether the cumulative adjustment number is greater than the third preset number? If yes, execute step P14; if no, return to step P12.

[0088] P17, end.

[0089] It should be noted that when the first weight and the second weight are adjusted for the third preset number of times, the output result calculated by the calculation module is still greater than the preset threshold value in the next third detection period, at which time the calculation module judges that the weight adjustment is invalid, the current of the device and the voltage fluctuation are abnormal, all adjustments to the first weight and the second weight in the current cycle are cancelled, a prediction alarm information is generated and sent to the server. The application initiatively analyzes the running current and voltage change of the device after the weight parameter adjustment by analyzing 10 output results obtained according to the third detection period, without complex modeling and algorithm operation, greatly improving the parameter adjustment efficiency, greatly simplifying the analysis algorithm, greatly reducing the cost required for system production, installation and maintenance, and greatly improving the enterprise benefit.

[0090] Please refer to Figure 3 The application provides another embodiment, which provides an intelligent gateway, the intelligent gateway comprises:

[0091] The intelligent gateway comprises a collection module 1, a calculation module 2 and a server 100.

[0092] The collection module 1 is used for collecting the current and voltage data of the device in real time.

[0093] The calculation module is used for processing the current information and the voltage information collected by the collection module 1 into inputs of a neural network algorithm according to a preset rule, calculating an output result, judging whether the device is normally running according to the output result, adjusting the weight of the input of the neural network algorithm when an abnormality is determined, generating a prediction alarm information when the calculation module judges that the adjustment is invalid, and updating the weight parameter of the preset neural network to the adjusted weight parameter when it is determined that the device is still in a normal state, so as to realize self-optimization of the neural network algorithm. The application solves the problems that it is usually difficult and tedious for a person to design a perfect neural network algorithm model by experience, and when there is a deviation in the parameter setting, the deviation will also have a certain influence on the determination of the monitoring result, and even cause analysis errors, further enhances the intelligent degree of the application, and greatly expands the application scenarios of the application.

[0094] The collection module 1 is used for collecting real-time data information of various devices, each type of device has a respective protocol, the collection module 1 is compatible with protocols of multiple devices, including serial port protocols and network protocols, and the device types include Internet of Things terminals, PLC controllers, numerical control machining devices, welding robots, sensor devices and various instrument devices. The collection module 1 is used for collecting three-phase voltage and current data of the motor monitoring terminal, and the data includes original waveform data.

[0095] The computing module 2 inputs the real-time data collected after being processed according to preset rules into a neural network algorithm and calculates an output result to determine whether the device is working in a normal state, and generates a prediction warning information and sends it to the server 100 when an abnormality is determined. The computing module 2 can extract the current effective value, voltage effective value, power factor, active power, reactive power, voltage imbalance, current imbalance, load rate, frequency, total harmonic rate and other parameter indexes in the collected original waveform. The fault feature recognition and the nature and severity of the fault can be determined by performing FFT and Hilbert transform on the time domain waveform.

[0096] During the monitoring of the device, if the power is interrupted, the computing module 2 will recover from the last breakpoint and continue to learn.

[0097] The specific implementation process of a hardware scheme of the present application is as follows:

[0098] The core control module of the present application adopts an ARM architecture high energy efficiency processor i.MX 8 as a main control chip, 4GB LPDDR4 (64 Bit) RAM storage and 16GB Flash storage.

[0099] The computing module supports industrial router function, realizes 2-way local area network access and video data collection and uploading, supports WAN / LAN, 4G, WIFI and other functions to meet the needs of remote communication in different environments.

[0100] The computing module adopts Linux as an operating system, integrates Python development environment and C language development environment, provides standard API interface and development guidance, provides a stable and fast platform for secondary application development of users, users can easily call various interfaces and resources of the system, and the usability of the present application is greatly improved.

[0101] The present application can analyze the health status of the motor by analyzing the current information and voltage information of the device, to analyze the current working state of the device in real time, and can realize real-time response, fast connection, intelligent application, security and privacy protection and other business processing to provide computing resources, which can speed up the processing and transmission speed of the collected data and reduce the delay.

[0102] In the preferred embodiment, the present application also provides an electronic device, which comprises:

[0103] The computer device can be a server, a terminal, or any other electronic device with necessary computing and / or processing capabilities in a broad sense. In an embodiment, the computer device can include a processor, a memory, a network interface, a communication interface, and the like connected by a system bus. The processor of the computer device can be configured to provide necessary computing, processing, and / or control capabilities. The memory of the computer device can include a non-volatile storage medium and an internal memory. The non-volatile storage medium can store an operating system, a computer program, and the like therein or thereon. The internal memory can provide an environment for running the operating system and the computer program in the non-volatile storage medium. The network interface and the communication interface of the computer device can be configured to connect and communicate with external devices through a network. The computer program, when executed by the processor, performs the steps of the method of the present application.

[0104] The present application can be implemented as a computer readable storage medium having a computer program stored thereon, which, when executed by a processor, causes the steps of the method of the embodiments of the present application to be performed. In an embodiment, the computer program is distributed over a plurality of computer devices or processors coupled by a network, such that the computer program is stored, accessed, and executed by one or more computer devices or processors in a distributed manner. A single method step / operation, or two or more method steps / operations, can be performed by a single computer device or processor, or by two or more computer devices or processors. One or more method steps / operations can be performed by one or more computer devices or processors, and one or more other method steps / operations can be performed by one or more other computer devices or processors. One or more computer devices or processors can perform a single method step / operation, or perform two or more method steps / operations.

[0105] It will be appreciated by those skilled in the art that the method steps of the present application can be instructed by a computer program to relevant hardware such as a computer device or a processor, which can be stored in a non-transitory computer readable storage medium, and which, when executed, causes the steps of the present application to be performed. Depending on the circumstances, any reference herein to a memory, storage, database, or other medium can include non-volatile and / or volatile memory. Examples of non-volatile memory include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, magnetic tapes, floppy disks, magneto-optical data storage devices, optical data storage devices, hard disks, solid-state disks, and the like. Examples of volatile memory include random access memory (RAM), external cache memory, and the like.

[0106] The technical features described above can be combined arbitrarily. Although all possible combinations of the technical features are not described, any combination of the technical features should be considered to be covered by the present specification, as long as there is no contradiction in such a combination.

[0107] Furthermore, it should be understood that, although the present specification is described in terms of embodiments, not every implementation embodies only one independent technical solution, and the present specification is described in this way only for the sake of clarity, and those skilled in the art should consider the present specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that those skilled in the art can understand.

Claims

1. A device monitoring method based on a smart gateway, characterized in that, The current and voltage signals of the device are acquired by the acquisition module of the smart gateway according to the first detection cycle. The signal is analyzed and the obtained current and voltage data are processed according to preset rules. The results are then input into a preset neural network algorithm according to the first weight, the first factor value, the second weight, and the second factor value to calculate the output result. When the output result is greater than the preset threshold, the current and voltage signals of the acquisition module are acquired according to the second detection cycle. The current signal and voltage signal are acquired and analyzed for a first preset number of times. The obtained current and voltage data are processed according to preset rules and then input into a preset neural network algorithm according to the first weight, the first factor value, the second weight, and the second factor value to calculate the output result. When the number of times the output result is greater than the preset threshold exceeds the second preset number of times, the first weight and the second weight are adjusted synchronously according to the first preset ratio for the third preset number of times. After each adjustment of the first weight and the second weight, the current signal and voltage signal of the device collected by the acquisition module are acquired according to the third detection cycle. The current and voltage signals are acquired and analyzed for the fourth preset number of times during the third detection cycle. The analyzed current and voltage data are processed according to preset rules and then input into a neural network algorithm according to the first factor value, adjusted first weight, second factor value, and adjusted second weight to calculate the output result. The output result is compared and analyzed with a preset threshold. When the change trend of two adjacent output results is close to the preset threshold, the change amount between these two adjacent output results is called the data sub-increment close to the preset threshold. When the change trend of two adjacent output results is far from the preset threshold, the change amount between these two adjacent output results is called the data sub-increment far from the preset threshold. The server will then process each data sub-increment close to the preset threshold. The data sub-increments are summarized as the data increments that are close to the preset threshold in the current third detection cycle, and the data sub-increments that are far from the preset threshold are summarized as the data increments that are far from the preset threshold in the current third detection cycle. If the data increments that are close to the preset threshold are greater than or equal to the data increments that are far from the preset threshold, then the first weight and the second weight are adjusted according to the first preset ratio. When the cumulative number of adjustments to the first weight and the second weight is less than or equal to the third preset number and the output result is less than the preset threshold, then the calculation module determines that the device is still in normal operation. The calculation module updates the first weight and the second weight of the preset neural network to the first weight and the second weight after the last adjustment. The method further includes: If the cumulative number of times the first and second weights are adjusted is greater than or equal to the third preset number and the output result is still greater than the preset threshold, the calculation module determines that the device is in an abnormal state. The calculation module restores the adjusted first and second weights to their initial values ​​for this cycle, and the calculation module generates a prediction alarm message and sends it to the server. The method further includes: The second detection cycle is shorter than the first detection cycle.

2. The device monitoring method based on a smart gateway according to claim 1, characterized in that, The method further includes: If, within the current third detection cycle, the data increment of the output result that is close to the preset threshold is greater than or equal to the data increment that is far from the preset threshold, then the first weight and the second weight are reduced according to the first preset ratio until the output result is less than the preset threshold or the cumulative number of times the first weight and the second weight are adjusted is greater than or equal to the third preset number and the output result is still greater than the preset threshold. If, during the current third detection cycle, the data increment of the output result that is close to the preset threshold is less than the data increment that is far from the preset threshold, then the first weight and the second weight are increased according to the second preset ratio until the output result is less than the preset threshold or the cumulative number of times the first weight and the second weight are adjusted is greater than or equal to the third preset number and the output result is still greater than the preset threshold.

3. The device monitoring method based on a smart gateway according to claim 1, characterized in that, The method further includes: The output result is calculated by summing the product of the first weight and the first factor value with the product of the second weight and the second factor value.

4. The device monitoring method based on a smart gateway according to claim 3, characterized in that, The method further includes: The first factor value is obtained by rounding up the current value, and the second factor value is obtained by rounding up the voltage value.

5. The device monitoring method based on a smart gateway according to claim 2, characterized in that, The method further includes: The third detection cycle is shorter than the second detection cycle.

6. The device monitoring method based on a smart gateway according to claim 2, characterized in that, The method further includes: The first weight is 20, the second weight is 10, the first preset ratio is 0.2%, the second preset ratio is 0.3%, the first preset number of times is 20, the second preset number of times is 15, the third preset number of times is 5, the fourth preset number of times is 10, the first detection period is 10 minutes, the second detection period is 5 minutes, and the third detection period is 5 seconds.

7. A smart gateway, characterized in that, The device monitoring method based on a smart gateway as described in any one of claims 1-6 includes a data acquisition module, a computing module, and a server. The acquisition module is used to acquire data information from the device; The calculation module is used to process the voltage and current signals collected by the acquisition module into the input of a preset neural network algorithm and calculate the output result. Based on this, it determines whether the device is operating normally. When an abnormality is detected, the weight of the input of the preset neural network algorithm is adjusted. When the calculation module determines that the adjustment is ineffective and the machine is in an abnormal state, a predictive alarm message is generated and sent to the server. The server is used to centrally manage the control programs of all devices and issue configuration commands.

8. An electronic device, characterized in that, include: Memory; The memory stores computer-readable instructions that, when executed by the processor, implement the device monitoring method based on a smart gateway according to any one of claims 1 to 6.

Images