AI server remote after-sales maintenance service system based on wireless communication

Abstract

The invention discloses an AI server remote after-sales maintenance service system based on wireless communication, and relates to the technical field of intelligent maintenance, and the system collects the operation data of an industrial robot through a sensor group, transmits the data to an AI server through a wireless network, carries out the preprocessing, obtains an operation health data group, and transmits the operation health data group to the AI server. The comprehensive health assessment index jpg is calculated, preliminary health assessment is carried out on the comprehensive health assessment index jpg and a preset industrial robot health threshold value Z, if operation is abnormal, operation parameters of the robot are remotely adjusted through an AI server, and if operation is normal, health prediction is carried out; a health prediction model is constructed through a convolutional neural network, a comprehensive health assessment index jpg (t + t) at a future moment is predicted, a potential fault risk is identified in advance, and a maintenance period is optimized. According to the system, through intelligent remote monitoring and diagnosis, the running state of the robot can be adjusted in real time, the maintenance cost is reduced, and the stability and the after-sales response speed of equipment are improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent maintenance technology, and specifically to an AI server remote after-sales maintenance service system based on wireless communication. Background Art

[0002] As the global manufacturing industry continues to develop towards intelligence and automation, the application of industrial robots on production lines is becoming more and more popular, especially in the automotive, electronics and consumer goods industries. These robots have greatly improved production efficiency, reduced labor costs, and are able to perform delicate and repetitive work. However, with the diversification of application scenarios and the increasing complexity of robots, after-sales maintenance issues have gradually become a challenge for companies. Traditional after-sales service models usually rely on manual inspections and periodic maintenance, which is not easy to monitor and predict potential failures of robots in real time, resulting in production interruptions and high maintenance costs. Therefore, modern enterprises are increasingly relying on intelligent after-sales maintenance services, especially remote monitoring and diagnosis systems based on AI and wireless communication technologies, to improve maintenance efficiency, reduce downtime, and reduce overall operating costs. In this context, developing more efficient and intelligent after-sales maintenance solutions has become a necessary means for the manufacturing industry to enhance its competitiveness and reduce long-term operating risks.

[0003] At present, many industrial robots rely on traditional maintenance mode, and are regularly overhauled and inspected. However, this traditional mode has many problems. First, the health monitoring of many robots relies on manual inspection, which lacks real-time and accuracy, and easily leads to hidden dangers being ignored or delayed. Secondly, the existing maintenance method lacks intelligence and is not easy to automatically adjust according to the real-time working status and load conditions. Therefore, when many robots face sudden failures, it is not easy to accurately locate and repair them in time, resulting in a long maintenance cycle and even affecting the normal production. In addition, the traditional maintenance method fails to effectively integrate data analysis and prediction technology, lacks effective prediction of potential risks in the future, is not easy to warn in advance, and easily causes unnecessary downtime and economic losses. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides an AI server remote after-sales maintenance service system based on wireless communication, which solves the problems in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a remote after-sales maintenance service system for an AI server based on wireless communication, comprising a health monitoring module, an AI server analysis module, a comprehensive evaluation module, an AI server tuning module and a risk prediction module; The health monitoring module is used to collect the operation data of the industrial robot in real time based on the sensor group installed on the industrial robot, and transmit it to the AI ​​server through the wireless network for preprocessing to obtain the operation health data group; The AI ​​server analysis module is used to perform summary calculations based on the acquired operation health data group to obtain a power health index dlj, a load health index fzj, and a vibration and mechanical stress health index zjz; The comprehensive evaluation module is used to perform summary calculation based on the acquired power health index dlj, load health index fzj and vibration and mechanical stress health index zjz, obtain a comprehensive health evaluation index jpg, and perform a preliminary health evaluation with a preset industrial robot health threshold Z; The AI ​​server training module is used to construct a training model through the AI ​​server when the initial health assessment shows that the industrial robot is operating abnormally, input the comprehensive health assessment index jpg into the training model, obtain the industrial robot training coefficient stj, and calculate the trained comprehensive health assessment index jpg by comparing it with the comprehensive health assessment index jpg. j (t), and conduct risk assessment after adjustment with the preset industrial robot health threshold Z; The risk prediction module is used to build a health prediction model when the operation risk assessment after the preliminary health assessment and adjustment shows that the industrial robot is operating normally, obtain the predicted comprehensive health assessment index jpg(t+∆t) through the health prediction model, and perform a predicted risk assessment with the preset industrial robot health threshold Z.

[0006] Preferably, the health monitoring module includes a health monitoring unit and a data processing unit; The health monitoring unit is used to collect the operation data of the industrial robot in real time based on the sensor group installed at the key position of the industrial robot, and mark the collection timestamp; The key positions include the various motion axes, drive systems and pneumatic control systems of the industrial robot; The sensor group includes an accelerometer, a rotary encoder, a torque sensor, a current sensor, a voltage sensor, a power sensor, a temperature sensor, a vibration sensor, and a frequency analyzer.

[0007] Preferably, the data processing unit is used to establish a communication connection between the sensor group and the AI ​​server through a wireless network according to the AI ​​server used by the manufacturer, and transmit the operating data collected by the sensor group to the AI ​​server in real time for preprocessing to obtain the operating health data group; The preprocessing includes denoising, filling missing values, filtering, dimensionless processing and electrical processing. The electrical processing is used to perform dimensionless processing on the operation data, and then calculate the power consumption gx, current fluctuation ∆I and voltage fluctuation ∆V of the robot through the root mean square fluctuation method based on the current I and voltage V obtained by the current sensor and the voltage sensor. The specific algorithm formula is: power consumption gx=current I×voltage V, , , where T represents the time window, I(t) and V(t) represent the instantaneous current and instantaneous voltage at time t, respectively. and They represent the average current and average voltage in the time period T, respectively, and dt represents the integral sign in the integral function; The operational health data group includes a power data group, a load data group and a stress data group; The power data set includes speed w, torque , acceleration a, current fluctuation ∆I and voltage fluctuation ∆V; The load data set includes the robot internal temperature jw, the ambient temperature hw, the robot current load df, the robot power consumption gx and the heat conduction flow rc; The stress data set includes vibration frequency zp, vibration amplitude zf, vibration period zz, vibration acceleration av and solid acceleration response jx.

[0008] Preferably, the AI ​​server analysis module analyzes the acquired operation health data group through the AI ​​server to obtain the operation health status of the industrial robot; The AI ​​server analysis module includes a power analysis unit, a load analysis unit, and a vibration and mechanical stress analysis unit; The power analysis unit is used to perform summary calculations based on the acquired power data group, obtain the power health index dlj, and analyze the operating stability of the drive system of the industrial robot; The power health index dlj is calculated and obtained by the following formula: ; In the formula, i represents the x, y and z axes, w i Indicates the rotation speed of each axis of the robot. Represents the torque of each axis of the robot, a i Indicates the acceleration of each axis of the robot; The load analysis unit is used to perform summary calculations based on the acquired load data group, obtain the load health index fzj, and analyze the thermal state of the industrial robot under load operation; The load health index fzj is calculated by the following formula: ; In the formula, ln represents the natural logarithm function, d represents the thermal diffusion coefficient of the robot material, which is set according to the properties of the robot material; The vibration and mechanical stress analysis unit is used to perform summary calculations based on the acquired stress data group, obtain the vibration and mechanical stress health index zjz, and analyze the potential damage of the industrial robot caused by vibration during operation; The vibration and mechanical stress health index zjz is calculated and obtained by the following formula: .

[0009] Preferably, the comprehensive assessment module includes a comprehensive health analysis unit and a health assessment unit; The comprehensive health analysis unit is used to perform summary calculation based on the acquired power health index dlj, load health index fzj and vibration and mechanical stress health index zjz, obtain a comprehensive health assessment index jpg, and analyze the operation health status of the industrial robot under the interaction between different factors; The comprehensive health assessment index jpg is calculated by the following formula: ; Wherein, ln represents the natural logarithmic function, and e represents the exponential function.

[0010] Preferably, the health assessment unit presets the health threshold Z of the industrial robot based on the industrial robot industry specifications and standards, and performs a preliminary health assessment with the obtained comprehensive health assessment index jpg to comprehensively assess the operating health of the industrial robot. The specific assessment scheme is as follows; When the comprehensive health assessment index jpg≤the preset industrial robot health threshold Z, it means that the industrial robot is operating normally, and the fault risk prediction instruction is executed at this time; When the comprehensive health assessment index jpg> the preset industrial robot health threshold Z, it means that the industrial robot is operating abnormally. At this time, the industrial robot is systematically adjusted through the AI ​​server.

[0011] Preferably, the AI ​​server tuning module includes a system tuning unit and a tuning evaluation unit; The system tuning unit is used to perform remote industrial robot operation system tuning on the industrial robot through the AI ​​server when the initial health assessment shows that the industrial robot is operating abnormally, and to build a tuning model, and then input the obtained comprehensive health assessment index jpg into the tuning model to obtain the industrial robot tuning coefficient stj; The industrial robot training coefficient stj is calculated and obtained by the following formula: ; Where stj(t) represents the industrial robot training coefficient at time t, jpg(t) represents the predicted comprehensive health assessment index, bz represents the target ideal health index, b represents the proportional coefficient, which controls the response intensity of health deviation, and ɑ represents the adjustment factor of the change rate of the comprehensive health assessment index. represents the differential operator, which differentiates time t, It represents the rate of change of the comprehensive health assessment index; The comprehensive health assessment index jpg after adjustment is obtained by calculating the obtained industrial robot adjustment coefficient stj(t) and the comprehensive health assessment index jpg(t). j (t), specifically:jpg j (t)=stj(t)+jpg(t), j represents the modification mark after the comprehensive health assessment index is adjusted.

[0012] Preferably, the training evaluation unit is used to perform a post-training operation risk evaluation based on the acquired post-training comprehensive health evaluation index jpgj(t) and a preset industrial robot health threshold Z. The specific evaluation scheme is as follows: When the comprehensive health assessment index after training j When (t)≤the preset health threshold Z of the industrial robot, it means that the industrial robot is normal after adjustment, and the fault risk prediction is performed at this time; When the comprehensive health assessment index after training j (t)>the preset industrial robot health threshold Z, it means that the industrial robot still operates abnormally after adjustment. At this time, iterative adjustment is performed. When the adjustment times are greater than three times and the problem is still not corrected, it is judged as a mechanical failure. At this time, fault information is generated and transmitted to the after-sales maintenance department through the wireless network.

[0013] Preferably, the risk prediction module includes a risk prediction unit and a prediction evaluation unit; The risk prediction unit is used to, when the operation risk assessment after the preliminary health assessment and adjustment shows that the industrial robot is operating normally, the AI ​​server constructs a health prediction model through a convolutional neural network, and inputs the health prediction model for model training based on the historical operation health data group of the industrial robot; After the model training is completed, the comprehensive health assessment index jpg obtained in real time is input into the health prediction model, and the predicted comprehensive health assessment index jpg(t+∆t) is obtained through the health prediction model. The specific algorithm formula is as follows; ; In the formula, jpg(t) represents the health assessment index at the current time t, m represents the growth and decline rate of the health index, and controls the amplitude of the change, t 0represents the time reference point, s represents the time constant of health change, and controls the speed of health change.

[0014] Preferably, the prediction and evaluation unit is used to perform a prediction risk evaluation based on the acquired prediction comprehensive health evaluation index jpg(t+∆t) and a preset industrial robot health threshold Z to evaluate the predicted health status of the industrial robot. The specific evaluation scheme is as follows; When the predicted comprehensive health assessment index jpg(t+∆t)>the preset industrial robot health threshold Z, it means that the industrial robot has hidden dangers. At this time, the industrial robot is systematically tuned through the AI ​​server tuning module; When the predicted comprehensive health assessment index jpg(t+∆t)≤the preset industrial robot health threshold Z, it indicates that the industrial robot is operating normally and the normal maintenance cycle is maintained for maintenance.

[0015] The present invention provides an AI server remote after-sales maintenance service system based on wireless communication. It has the following beneficial effects: (1) The health monitoring module of the system collects the operation data of the industrial robot in real time through the sensor group installed in the key parts of the industrial robot, and marks the collection timestamp. The collected data is transmitted to the AI ​​server through the wireless network for preprocessing to obtain the operation health data group. The processing process includes denoising, filling missing values, filtering, dimensionless processing and electrical processing. In the electrical processing stage, the power consumption gx, current fluctuation ∆I and voltage fluctuation ∆V of the robot are calculated through the current I and voltage V, providing accurate data support for the subsequent health index calculation.

[0016] (2) The AI ​​server analysis module of the system performs summary calculations based on the operation health data group transmitted by the health monitoring module to obtain the power health index dlj, load health index fzj and vibration and mechanical stress health index zjz. The power health index dlj reflects the operation stability of the industrial robot drive system, the load health index fzj analyzes the load working thermal state of the robot, and the vibration and mechanical stress health index zjz evaluates the potential damage caused by vibration during the operation of the robot. The comprehensive evaluation module further summarizes and calculates the comprehensive health evaluation index jpg, and performs a preliminary health evaluation with the preset industrial robot health threshold Z. When the comprehensive health evaluation index jpg is lower than or equal to the preset industrial robot health threshold Z, the system will determine that the industrial robot is operating normally and execute the fault risk prediction instruction. When the comprehensive health evaluation index jpg exceeds the preset industrial robot health threshold Z, the system will determine that the robot is operating abnormally and start the training module to remotely train the robot.

[0017] (3) When the robot is abnormal, the AI ​​server training module of the system constructs a training model, inputs the training model according to the comprehensive health assessment index jpg, obtains the robot training coefficient stj, and when the health assessment index jpg after training is j (t) is lower than or equal to the preset industrial robot health threshold Z, the system determines that the industrial robot is normal after adjustment, and then performs fault risk prediction. If the adjusted health assessment index jpg j (t) still exceeds the preset industrial robot health threshold Z, the system will iteratively adjust. When the number of adjustments is more than three times and still not corrected, it is judged as a mechanical fault and fault information is generated. The fault information is transmitted to the after-sales maintenance department through the wireless network for processing. The risk prediction module constructs a health prediction model through a convolutional neural network to predict the future t+∆ prediction comprehensive health assessment index jpg(t+∆t). According to the comparison between the prediction result and the preset industrial robot health threshold Z, it is judged whether the robot has hidden dangers. If there are hidden dangers, the system will automatically start the adjustment module for adjustment. If it runs normally, the maintenance cycle is maintained for routine maintenance. This system can realize real-time monitoring and accurate evaluation of the robot's operating health status. Through efficient adjustment and prediction mechanisms, potential faults can be identified and processed in advance, which not only improves the intelligent level of industrial robot fault diagnosis and maintenance, but also effectively reduces the frequency and cost of manual intervention, while extending the service life of robot equipment and ensuring the continuity and reliability of production. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 The present invention is a flowchart of an AI server remote after-sales maintenance service system based on wireless communication. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Example 1 See also Figure 1 The present invention provides an AI server remote after-sales maintenance service system based on wireless communication. To achieve the above purpose, the present invention is implemented through the following technical solutions: including a health monitoring module, an AI server analysis module, a comprehensive evaluation module, an AI server tuning module and a risk prediction module; The health monitoring module is used to collect the operation data of the industrial robot in real time based on the sensor group installed on the industrial robot, and transmit it to the AI ​​server through the wireless network for preprocessing to obtain the operation health data group; The AI ​​server analysis module is used to perform summary calculations based on the acquired operation health data group to obtain a power health index dlj, a load health index fzj, and a vibration and mechanical stress health index zjz; The comprehensive evaluation module is used to perform summary calculation based on the acquired power health index dlj, load health index fzj and vibration and mechanical stress health index zjz, obtain a comprehensive health evaluation index jpg, and perform a preliminary health evaluation with a preset industrial robot health threshold Z; The AI ​​server training module is used to construct a training model through the AI ​​server when the initial health assessment shows that the industrial robot is operating abnormally, input the comprehensive health assessment index jpg into the training model, obtain the industrial robot training coefficient stj, and calculate the trained comprehensive health assessment index jpg by comparing it with the comprehensive health assessment index jpg. j (t), and conduct risk assessment after adjustment with the preset industrial robot health threshold Z; The risk prediction module is used to build a health prediction model when the operation risk assessment after the preliminary health assessment and adjustment shows that the industrial robot is operating normally, obtain the predicted comprehensive health assessment index jpg(t+∆t) through the health prediction model, and perform a predicted risk assessment with the preset industrial robot health threshold Z.

[0021] In this embodiment, the health monitoring module collects the operation data of the industrial robot in real time based on the sensor group installed on the industrial robot, and transmits it to the AI ​​server through a wireless network for preprocessing to obtain the operation health data group, which provides accurate real-time data support for subsequent data processing. Unlike the traditional method of relying on manual inspection and regular maintenance, this sensor-based real-time monitoring can continuously track the operation status of the robot, quickly discover potential faults, and thus provide early warnings to avoid downtime losses caused by equipment failures. The AI ​​server analysis module relies on the operation health data group obtained from the health monitoring module to calculate the power health index dlj, load health index fzj and vibration and mechanical stress health index zjz to comprehensively evaluate the operation status of the robot in all aspects. The comprehensive evaluation module integrates these health indexes to calculate the comprehensive health evaluation index jpg, and compares it with the preset industrial robot health threshold Z to quickly determine whether the robot is in an abnormal state, reducing the time and resource consumption of traditional manual inspections. This data-driven evaluation method is more scientific and accurate than traditional technical means, and can understand the health status of the robot in real time, avoiding the risk of missing the fault opportunity due to the long inspection cycle in traditional means. The AI ​​server tuning module and risk prediction module further enhance the intelligence level of the system. When an abnormality is found in the preliminary health assessment, a tuning model is constructed through the AI ​​server, and the robot's parameters are adjusted to restore its normal operating state. By comparing with the preset industrial robot health threshold Z, the robot's working state is adjusted in time to reduce the probability of failure. The risk prediction module predicts the comprehensive health assessment index jpg(t+∆t) based on the convolutional neural network, which can predict the health status of the robot in the future and further optimize the maintenance plan. This predictive and proactive maintenance method can more effectively prevent the occurrence of major failures than the traditional method that relies on simple fault detection and regular inspections, reduce maintenance costs, and improve production continuity and equipment service life. In general, compared with the prior art, the technical solution of the present invention breaks through the traditional maintenance model and brings about a comprehensive improvement in intelligence, predictability and efficiency.

[0022] Example 2 This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the health monitoring module includes a health monitoring unit and a data processing unit; The health monitoring unit is used to collect the operation data of the industrial robot in real time based on the sensor group installed at the key position of the industrial robot, and mark the collection timestamp; The key positions include the various motion axes, drive systems and pneumatic control systems of the industrial robot; The sensor group includes an accelerometer, a rotary encoder, a torque sensor, a current sensor, a voltage sensor, a power sensor, a temperature sensor, a vibration sensor, and a frequency analyzer.

[0023] The data processing unit is used to establish a communication connection between the sensor group and the AI ​​server through a wireless network according to the AI ​​server used by the manufacturer, and transmit the operating data collected by the sensor group to the AI ​​server in real time for preprocessing to obtain the operating health data group; The preprocessing includes denoising, filling missing values, filtering, dimensionless processing and electrical processing. The electrical processing is used to perform dimensionless processing on the operation data, and then calculate the power consumption gx, current fluctuation ∆I and voltage fluctuation ∆V of the robot through the root mean square fluctuation method based on the current I and voltage V obtained by the current sensor and the voltage sensor. The specific algorithm formula is: power consumption gx=current I×voltage V, , , where T represents the time window, I(t) and V(t) represent the instantaneous current and instantaneous voltage at time t, respectively. and They represent the average current and average voltage in the time period T, respectively, and dt represents the integral sign in the integral function; The operational health data group includes a power data group, a load data group and a stress data group; The power data set includes speed w, torque , acceleration a, current fluctuation ∆I and voltage fluctuation ∆V; The load data set includes the robot internal temperature jw, the ambient temperature hw, the robot current load df, the robot power consumption gx and the heat conduction flow rc; The stress data set includes vibration frequency zp, vibration amplitude zf, vibration period zz, vibration acceleration av and solid acceleration response jx.

[0024] In this embodiment, the health monitoring module collects the operating data of the industrial robot in real time through the sensor group installed at the key position of the industrial robot, which can comprehensively and accurately reflect the operating status of the robot. The data processing unit transmits the sensor data to the AI ​​server through a wireless network, and performs denoising, filling missing values, filtering, dimensionless processing and electrical processing to obtain the operating health data group, providing high-quality data input for subsequent analysis. Through electrical processing, the current and voltage data are converted into indicators such as power consumption gx, current fluctuation ∆I and voltage fluctuation ∆V, further revealing the operating efficiency and energy consumption of the robot. This comprehensive data collection and efficient data processing solution not only improves the accuracy of robot fault diagnosis, but also provides strong data support for subsequent health assessment and prediction, realizes real-time health monitoring and intelligent maintenance of the robot, greatly improves the stability of robot operation and the efficiency of life cycle management, and reduces production interruptions and maintenance costs caused by failures.

[0025] Example 3 This embodiment is explained in Example 2. Please refer to Figure 1 , specifically: the AI ​​server analysis module analyzes the acquired operation health data group through the AI ​​server to obtain the operation health status of the industrial robot; The AI ​​server analysis module includes a power analysis unit, a load analysis unit, and a vibration and mechanical stress analysis unit; The power analysis unit is used to perform summary calculations based on the acquired power data group, obtain the power health index dlj, and analyze the operating stability of the drive system of the industrial robot; The power health index dlj is calculated and obtained by the following formula: ; In the formula, i represents the x, y and z axes, w i Indicates the rotation speed of each axis of the robot. Represents the torque of each axis of the robot, a i Indicates the acceleration of each axis of the robot; The load analysis unit is used to perform summary calculations based on the acquired load data group, obtain the load health index fzj, and analyze the thermal state of the industrial robot under load operation; The load health index fzj is calculated by the following formula: ; In the formula, ln represents the natural logarithm function, d represents the thermal diffusion coefficient of the robot material, which is set according to the properties of the robot material; The vibration and mechanical stress analysis unit is used to perform summary calculations based on the acquired stress data group, obtain the vibration and mechanical stress health index zjz, and analyze the potential damage of the industrial robot caused by vibration during operation; The vibration and mechanical stress health index zjz is calculated and obtained by the following formula: .

[0026] In this embodiment, the AI ​​server analysis module provides a comprehensive assessment of the health status of the industrial robot through a multi-dimensional analysis of the running health data, greatly improving the accuracy and response speed of fault diagnosis. The power analysis unit calculates the power health index dlj through the power data group, so that it can monitor the operating stability of the robot drive system in real time and ensure the stability of the power performance of the robot during work. The load analysis unit calculates the load health index fzj based on the load data group to help analyze the thermal state of the robot under different load conditions, timely discover the thermal stress caused by high-load work, and avoid failures caused by overheating. The vibration and mechanical stress analysis unit calculates the vibration and mechanical stress health index zjz through the stress data group, which can identify the potential damage caused by vibration or mechanical stress during the operation of the robot in advance, thereby effectively reducing the occurrence of mechanical failures. This multi-angle and all-round health monitoring and analysis method not only greatly improves the reliability of the robot's operation, but also provides a scientific basis for subsequent maintenance decisions, greatly optimizes the allocation and utilization efficiency of maintenance resources, and extends the service life of robot equipment.

[0027] Example 4 This embodiment is explained in Example 3, please refer to Figure 1 ,Specifically: the comprehensive assessment module includes a comprehensive health analysis unit and a health assessment unit; The comprehensive health analysis unit is used to perform summary calculation based on the acquired power health index dlj, load health index fzj and vibration and mechanical stress health index zjz, obtain a comprehensive health assessment index jpg, and analyze the operation health status of the industrial robot under the interaction between different factors; The comprehensive health assessment index jpg is calculated by the following formula: ; Wherein, ln represents the natural logarithmic function, and e represents the exponential function.

[0028] The health assessment unit presets the health threshold Z of the industrial robot based on the industrial robot industry specifications and standards, and performs a preliminary health assessment with the obtained comprehensive health assessment index jpg to comprehensively assess the operating health of the industrial robot. The specific assessment scheme is as follows; When the comprehensive health assessment index jpg≤the preset industrial robot health threshold Z, it means that the industrial robot is operating normally, and the fault risk prediction instruction is executed at this time; When the comprehensive health assessment index jpg> the preset industrial robot health threshold Z, it means that the industrial robot is operating abnormally. At this time, the industrial robot is systematically adjusted through the AI ​​server.

[0029] In this embodiment, the comprehensive health analysis unit calculates the comprehensive health assessment index jpg by summarizing the power health index dlj, the load health index fzj and the vibration and mechanical stress health index zjz, and deeply analyzes the interaction of multiple factors, so as to accurately reflect the overall operation health of the robot. The health assessment unit presets the industrial robot health threshold Z based on industry specifications and standards, and performs a preliminary health assessment with the obtained comprehensive health assessment index jpg, thereby realizing intelligent judgment of the robot's health status. When the comprehensive health assessment index jpg is lower than and equal to the industrial robot health threshold Z, it indicates that the robot is in a normal state, and the system will automatically start the fault risk prediction to maintain the stability of the robot's operation; and when the comprehensive health assessment index jpg exceeds the industrial robot health threshold Z, it indicates that the robot has potential abnormalities. At this time, the system will automatically start the tuning mechanism and remotely tune the robot through the AI ​​server. This intelligent and automated evaluation and tuning method significantly improves the safety and efficiency of the robot during operation, avoids errors and delays in human judgment, and at the same time, through accurate fault prediction and tuning, greatly reduces equipment downtime and maintenance costs, extends the service life of the robot, and greatly improves the stability and economic benefits of the production line.

[0030] Example 5 This embodiment is explained in Example 4. Please refer to Figure 1 ,Specifically: the AI ​​server tuning module includes a system tuning unit and a tuning evaluation unit; The system tuning unit is used to perform remote industrial robot operation system tuning on the industrial robot through the AI ​​server when the initial health assessment shows that the industrial robot is operating abnormally, and to build a tuning model, and then input the obtained comprehensive health assessment index jpg into the tuning model to obtain the industrial robot tuning coefficient stj; The industrial robot training coefficient stj is calculated and obtained by the following formula: ; Where stj(t) represents the industrial robot training coefficient at time t, jpg(t) represents the predicted comprehensive health assessment index, bz represents the target ideal health index, b represents the proportional coefficient, which controls the response intensity of health deviation, and ɑ represents the adjustment factor of the change rate of the comprehensive health assessment index. represents the differential operator, which differentiates time t, It represents the rate of change of the comprehensive health assessment index; The comprehensive health assessment index jpg after adjustment is obtained by calculating the obtained industrial robot adjustment coefficient stj(t) and the comprehensive health assessment index jpg(t). j (t), specifically:jpg j (t)=stj(t)+jpg(t), j represents the modification mark after the comprehensive health assessment index is adjusted.

[0031] The training and evaluation unit is used to obtain the comprehensive health evaluation index after training. j (t) After adjusting with the preset industrial robot health threshold Z, the operation risk assessment is performed. The specific assessment scheme is as follows; When the comprehensive health assessment index after training j When (t)≤the preset health threshold Z of the industrial robot, it means that the industrial robot is normal after adjustment, and the fault risk prediction is performed at this time; When the comprehensive health assessment index after training j (t)>the preset industrial robot health threshold Z, it means that the industrial robot still operates abnormally after adjustment. At this time, iterative adjustment is performed. When the adjustment times are greater than three times and the problem is still not corrected, it is judged as a mechanical failure. At this time, fault information is generated and transmitted to the after-sales maintenance department through the wireless network.

[0032] In this embodiment, when the system training unit finds that the industrial robot is abnormal, it inputs the training model based on the comprehensive health assessment index jpg and calculates the training coefficient stj, so as to accurately adjust the operating parameters of the robot and optimize its health status. j (t) is compared with the preset industrial robot health threshold Z to further evaluate the robot's operating status. When the adjusted comprehensive health assessment index is j When (t) is lower than and equal to the preset industrial robot health threshold Z, the system recognizes that the industrial robot is normal after adjustment and performs fault risk prediction; when the robot has not returned to normal, it is continuously optimized through iterative adjustment, and the number of adjustments is set. When the adjustment is invalid, the system identifies the mechanical failure and generates fault information, which is promptly transmitted to the after-sales maintenance department. The advantage of this adjustment and evaluation mechanism is that it not only reduces the need for human intervention, but also can intelligently judge and adjust the robot's operating status, prevent failures in advance, and restore the robot's healthy status in time through remote adjustment. Compared with the traditional reliance on manual diagnosis and maintenance methods, the present invention greatly improves maintenance efficiency and reduces downtime, while providing strong support for improving the service life and operational stability of the robot.

[0033] Example 6 This embodiment is explained in Example 4. Please refer to Figure 1 ,Specifically: the risk prediction module includes a risk prediction unit and a prediction evaluation unit; The risk prediction unit is used to, when the operation risk assessment after the preliminary health assessment and adjustment shows that the industrial robot is operating normally, the AI ​​server constructs a health prediction model through a convolutional neural network, and inputs the health prediction model for model training based on the historical operation health data group of the industrial robot; After the model training is completed, the comprehensive health assessment index jpg obtained in real time is input into the health prediction model, and the predicted comprehensive health assessment index jpg(t+∆t) is obtained through the health prediction model. The specific algorithm formula is as follows; ; In the formula, jpg(t) represents the health assessment index at the current time t, m represents the growth and decline rate of the health index, and controls the amplitude of the change, t 0 represents the time reference point, s represents the time constant of health change, and controls the speed of health change.

[0034] The prediction and evaluation unit is used to perform a prediction risk evaluation based on the obtained prediction comprehensive health evaluation index jpg(t+∆t) and the preset industrial robot health threshold Z to evaluate the predicted health status of the industrial robot. The specific evaluation scheme is as follows; When the predicted comprehensive health assessment index jpg(t+∆t)>the preset industrial robot health threshold Z, it means that the industrial robot has hidden dangers. At this time, the industrial robot is systematically tuned through the AI ​​server tuning module; When the predicted comprehensive health assessment index jpg(t+∆t)≤the preset industrial robot health threshold Z, it indicates that the industrial robot is operating normally and the normal maintenance cycle is maintained for maintenance.

[0035] In this embodiment, when the operation risk assessment after the preliminary health assessment and adjustment is that the industrial robot is operating normally, the health prediction model accurately predicts the comprehensive health assessment index jpg(t+∆t) based on the comprehensive health assessment index jpg of the industrial robot when it is operating normally and abnormally in history, and judges the health risk of the robot according to the predicted value. If the predicted comprehensive health assessment index jpg(t+∆t) exceeds the preset industrial robot health threshold Z, the system will immediately trigger the AI ​​server adjustment module to adjust and adjust to avoid potential failures; if the future comprehensive health assessment index jpg(t+∆t) is lower than and equal to the preset industrial robot health threshold Z, the robot maintains a normal maintenance cycle for maintenance. Compared with traditional maintenance methods, this system is not only highly forward-looking, but also reduces the unexpectedness and suddenness of failures through deep learning and intelligent prediction, and can intervene in advance before the health of the equipment deteriorates, significantly reducing maintenance costs and improving the stability and operation reliability of production equipment. This predictive and proactive adjustment and risk management gives robots more intelligent adaptive capabilities and provides more efficient and reliable maintenance support for modern industrial production.

[0036] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A remote after-sales maintenance service system for AI servers based on wireless communication, characterized in that: It includes health monitoring module, AI server analysis module, comprehensive evaluation module, AI server tuning module and risk prediction module; The health monitoring module is used to collect the operation data of the industrial robot in real time based on the sensor group installed on the industrial robot, and transmit it to the AI ​​server through the wireless network for preprocessing to obtain the operation health data group; The AI ​​server analysis module is used to perform summary calculations based on the acquired operation health data group to obtain a power health index dlj, a load health index fzj, and a vibration and mechanical stress health index zjz; The comprehensive evaluation module is used to perform summary calculation based on the acquired power health index dlj, load health index fzj and vibration and mechanical stress health index zjz, obtain a comprehensive health evaluation index jpg, and perform a preliminary health evaluation with a preset industrial robot health threshold Z; The AI ​​server training module is used to construct a training model through the AI ​​server when the initial health assessment shows that the industrial robot is operating abnormally, input the comprehensive health assessment index jpg into the training model, obtain the industrial robot training coefficient stj, and calculate the trained comprehensive health assessment index jpg by comparing it with the comprehensive health assessment index jpg. j (t), and conduct risk assessment after adjustment with the preset industrial robot health threshold Z; The risk prediction module is used to build a health prediction model when the operation risk assessment after the preliminary health assessment and adjustment shows that the industrial robot is operating normally, obtain the predicted comprehensive health assessment index jpg(t+∆t) through the health prediction model, and perform a predicted risk assessment with the preset industrial robot health threshold Z.

2. According to claim 1, a wireless communication-based AI server remote after-sales maintenance service system is characterized by: The health monitoring module includes a health monitoring unit and a data processing unit; The health monitoring unit is used to collect the operation data of the industrial robot in real time based on the sensor group installed at the key position of the industrial robot, and mark the collection timestamp; The key positions include the various motion axes, drive systems and pneumatic control systems of the industrial robot; The sensor group includes an accelerometer, a rotary encoder, a torque sensor, a current sensor, a voltage sensor, a power sensor, a temperature sensor, a vibration sensor, and a frequency analyzer.

3. According to claim 2, a wireless communication-based AI server remote after-sales maintenance service system is characterized by: The data processing unit is used to establish a communication connection between the sensor group and the AI ​​server through a wireless network according to the AI ​​server used by the manufacturer, and transmit the operating data collected by the sensor group to the AI ​​server in real time for preprocessing to obtain the operating health data group; The preprocessing includes denoising, filling missing values, filtering, dimensionless processing and electrical processing; The electrical processing is used to process the operation data dimensionlessly, and then calculate the power consumption gx, current fluctuation ∆I and voltage fluctuation ∆V of the robot through the root mean square fluctuation method according to the current I and voltage V obtained by the current sensor and the voltage sensor. The specific algorithm formula is: power consumption gx=I×V, , , where T represents the time window, I(t) and V(t) represent the instantaneous current and instantaneous voltage at time t, respectively. and They represent the average current and average voltage in the time period T, respectively, and dt represents the integral sign in the integral function; The operational health data group includes a power data group, a load data group and a stress data group; The power data set includes speed w, torque , acceleration a, current fluctuation ∆I and voltage fluctuation ∆V; The load data set includes the robot internal temperature jw, the ambient temperature hw, the robot current load df, the robot power consumption gx and the heat conduction flow rc; The stress data set includes vibration frequency zp, vibration amplitude zf, vibration period zz, vibration acceleration av and solid acceleration response jx.

4. According to claim 3, a wireless communication-based AI server remote after-sales maintenance service system is characterized by: The AI ​​server analysis module analyzes the acquired operation health data group through the AI ​​server to obtain the operation health status of the industrial robot; The AI ​​server analysis module includes a power analysis unit, a load analysis unit, and a vibration and mechanical stress analysis unit; The power analysis unit is used to perform summary calculations based on the acquired power data group, obtain the power health index dlj, and analyze the operating stability of the drive system of the industrial robot; The power health index dlj is calculated and obtained by the following formula: ； In the formula, i represents the axis in each direction, i∈{x-axis, y-axis, z-axis}, w i Indicates the rotation speed of each axis of the robot. Represents the torque of each axis of the robot, a i Indicates the acceleration of each axis of the robot; The load analysis unit is used to perform summary calculations based on the acquired load data group, obtain the load health index fzj, and analyze the thermal state of the industrial robot under load operation; The load health index fzj is calculated by the following formula: ； In the formula, ln represents the natural logarithm function, d represents the thermal diffusion coefficient of the robot material, which is set according to the properties of the robot material; The vibration and mechanical stress analysis unit is used to perform summary calculations based on the acquired stress data group, obtain the vibration and mechanical stress health index zjz, and analyze the potential damage of the industrial robot caused by vibration during operation; The vibration and mechanical stress health index zjz is calculated and obtained by the following formula: 。 5. According to claim 4, a wireless communication-based AI server remote after-sales maintenance service system is characterized by: The comprehensive assessment module includes a comprehensive health analysis unit and a health assessment unit; The comprehensive health analysis unit is used to perform summary calculation based on the acquired power health index dlj, load health index fzj and vibration and mechanical stress health index zjz, obtain a comprehensive health assessment index jpg, and analyze the operation health status of the industrial robot under the interaction between different factors; The comprehensive health assessment index jpg is calculated by the following formula: ； Wherein, ln represents the natural logarithmic function, and e represents the exponential function.

6. According to claim 5, a wireless communication-based AI server remote after-sales maintenance service system is characterized by: The health assessment unit presets the health threshold Z of the industrial robot based on the industrial robot industry specifications and standards, and performs a preliminary health assessment with the obtained comprehensive health assessment index jpg to comprehensively assess the operating health of the industrial robot. The specific assessment scheme is as follows; When the comprehensive health assessment index jpg≤the preset industrial robot health threshold Z, it means that the industrial robot is operating normally, and the fault risk prediction instruction is executed at this time; When the comprehensive health assessment index jpg> the preset industrial robot health threshold Z, it means that the industrial robot is operating abnormally. At this time, the industrial robot is systematically adjusted through the AI ​​server.

7. According to claim 6, a wireless communication-based AI server remote after-sales maintenance service system is characterized by: The AI ​​server tuning module includes a system tuning unit and a tuning evaluation unit; The system tuning unit is used to perform remote industrial robot operation system tuning on the industrial robot through the AI ​​server when the initial health assessment shows that the industrial robot is operating abnormally, and to build a tuning model, and then input the obtained comprehensive health assessment index jpg into the tuning model to obtain the industrial robot tuning coefficient stj; The industrial robot training coefficient stj is calculated and obtained by the following formula: ； Where stj(t) represents the industrial robot training coefficient at time t, jpg(t) represents the predicted comprehensive health assessment index, bz represents the target ideal health index, b represents the proportional coefficient, which controls the response intensity of health deviation, and ɑ represents the adjustment factor of the change rate of the comprehensive health assessment index. represents the differential operator, which differentiates time t, It indicates the rate of change of the comprehensive health assessment index; The comprehensive health assessment index jpg after adjustment is obtained by calculating the obtained industrial robot adjustment coefficient stj(t) and the comprehensive health assessment index jpg(t). j (t), specifically:jpg j (t)=stj(t)+jpg(t), j represents the modification mark after the comprehensive health assessment index is adjusted.

8. The AI ​​server remote after-sales maintenance service system based on wireless communication according to claim 7 is characterized by: The training and evaluation unit is used to obtain the comprehensive health evaluation index after training. j (t) After adjusting with the preset industrial robot health threshold Z, the operation risk assessment is performed. The specific assessment scheme is as follows; When the comprehensive health assessment index after training j When (t)≤the preset health threshold Z of the industrial robot, it means that the industrial robot is normal after adjustment, and the fault risk prediction is performed at this time; When the comprehensive health assessment index after training j (t)>the preset industrial robot health threshold Z, it means that the industrial robot still operates abnormally after adjustment. At this time, iterative adjustment is performed. When the adjustment times are greater than three times and the problem is still not corrected, it is judged as a mechanical failure. At this time, fault information is generated and transmitted to the after-sales maintenance department through the wireless network.

9. The AI ​​server remote after-sales maintenance service system based on wireless communication according to claim 6 is characterized by: The risk prediction module includes a risk prediction unit and a prediction evaluation unit; The risk prediction unit is used to, when the operation risk assessment after the preliminary health assessment and adjustment shows that the industrial robot is operating normally, the AI ​​server constructs a health prediction model through a convolutional neural network, and inputs the health prediction model for model training based on the historical operation health data group of the industrial robot; After the model training is completed, the comprehensive health assessment index jpg obtained in real time is input into the health prediction model, and the predicted comprehensive health assessment index jpg(t+∆t) is obtained through the health prediction model. The specific algorithm formula is as follows; ； Where jpg(t) represents the health assessment index at the current time t, m represents the growth and decline rate of the health index, which controls the amplitude of the change, t0 represents the time reference point, and s represents the time constant of health change, which controls the speed of health change.

10. The AI ​​server remote after-sales maintenance service system based on wireless communication according to claim 9, characterized in that: The prediction and evaluation unit is used to perform a prediction risk evaluation based on the obtained prediction comprehensive health evaluation index jpg(t+∆t) and the preset industrial robot health threshold Z to evaluate the predicted health status of the industrial robot. The specific evaluation scheme is as follows; When the predicted comprehensive health assessment index jpg(t+∆t)>the preset industrial robot health threshold Z, it means that the industrial robot has hidden dangers. At this time, the industrial robot is systematically tuned through the AI ​​server tuning module; When the predicted comprehensive health assessment index jpg(t+∆t)≤the preset industrial robot health threshold Z, it indicates that the industrial robot is operating normally and the normal maintenance cycle is maintained for maintenance.

Images