Intelligent production equipment digital service method and system based on Internet of Things

Abstract

The invention provides an intelligent production equipment digital service method and system based on the Internet of Things, and relates to the technical field of industrial Internet of Things and intelligent manufacturing, and the method comprises the steps: collecting vibration, temperature, current and image data through a multi-mode sensor network; constructing an equipment digital twinborn model and calculating a comprehensive efficiency index CEI and a production quality fluctuation coefficient PQFC; generating a dynamic optimization strategy based on deep reinforcement learning; reverse control compensation is implemented through an edge computing node; and when the CEI is lower than a threshold value, triggering a three-dimensional visual maintenance scheme. The system comprises a sensor array, a processor, a memory, an execution mechanism controller and a user interface device, and all the components realize data interaction through an industrial Internet of Things protocol. According to the invention, through multi-modal data fusion and a two-parameter evaluation system, the problems of single monitoring dimension and rigid control strategy in the prior art are solved, and collaborative optimization of equipment efficiency and production quality is realized in combination with digital twin real-time mapping and a dynamic optimization algorithm.

Description

Technical Field

[0001] The present invention relates to the field of industrial Internet of Things and intelligent manufacturing technology, and in particular to a digital service method and system for intelligent production equipment based on the Internet of Things. Background Art

[0002] In the fields of Industrial Internet of Things and intelligent manufacturing, digital service technologies for production equipment have become a key means of improving manufacturing efficiency and quality. However, existing technologies still have significant shortcomings in practical applications, which restricts further advancements in intelligent equipment management. First, traditional equipment monitoring systems typically rely on a single type of sensor to collect data, such as vibration sensors or temperature sensors. This results in a single dimension for evaluating equipment status and fails to fully reflect the overall performance of equipment operation. This one-sided monitoring approach struggles to capture the complex degradation characteristics of equipment under the influence of multi-physical field coupling, and can easily lead to missed detection of key fault signs.

[0003] Secondly, the current mainstream equipment health assessment methods are mostly based on fixed thresholds or single performance indicators (such as vibration amplitude exceeding the limit) for abnormal judgment, and lack quantitative analysis of the correlation between equipment performance and production quality. This fragmented evaluation system leads to a disconnect between equipment control strategies and production quality requirements, and cannot achieve the coordinated optimization of production efficiency and product quality. For example, equipment may be in a sub-healthy state but still meet short-term production requirements, but long-term operation will lead to a continuous decline in yield rate, and existing technologies make it difficult to identify such hidden risks in a timely manner.

[0004] Furthermore, most existing control strategies employ static adjustment mechanisms based on preset parameters, failing to dynamically optimize based on the real-time status of equipment and changes in the production environment. When equipment performance degrades or load fluctuates, control systems with fixed parameters struggle to maintain optimal operation, leading to energy waste and reduced machining accuracy. Furthermore, traditional predictive maintenance solutions often rely on simple threshold triggering mechanisms and lack integrated analysis of multi-dimensional equipment degradation trends. This results in inaccurate judgments about maintenance timing, potentially leading to over-maintenance or delayed maintenance.

[0005] Finally, existing digital twin models suffer from data lags in their mapping and updating mechanisms, resulting in inaccurate synchronization between the virtual model and the real-time state of the physical device. Model parameter updates are infrequent and rely on offline calibration data, making it difficult to adapt to the rapid state changes during dynamic device operation. Consequently, the digital twin cannot effectively support real-time control decisions. These technical limitations severely limit the application of intelligent production equipment in complex industrial scenarios. A digital service solution is urgently needed that can achieve multi-dimensional state perception, dynamic strategy optimization, and precise maintenance decision-making. Summary of the Invention

[0006] In order to solve the technical problems in the existing technology such as insufficient monitoring dimensions of a single sensor, lack of correlation analysis between equipment efficiency and production quality, static and rigid control strategy, and low real-time synchronization accuracy of digital twin models, the present invention provides a digital service method and system for intelligent production equipment based on the Internet of Things.

[0007] The technical solutions provided by the present invention are as follows:

[0008] First aspect:

[0009] The present invention provides a digital service method for intelligent production equipment based on the Internet of Things, comprising:

[0010] S1. Collecting device operation data through a multimodal sensor network deployed on the device body, wherein the multimodal sensor network includes a vibration sensor, a temperature sensor, a current sensor, and an image acquisition unit;

[0011] S2. Build a digital twin model of the equipment, map the collected data into a virtual space to generate a holographic image of the equipment, and simultaneously calculate the equipment comprehensive efficiency index (CEI) and production quality fluctuation coefficient (PQFC);

[0012] S3. Establish a dynamic optimization strategy library based on deep reinforcement learning algorithm, and select the best control parameter combination from the strategy library according to the real-time changes of CEI and PQFC;

[0013] S4. Perform reverse control compensation on the equipment actuator through the edge computing node and synchronously update the digital twin model parameters;

[0014] S5. When the CEI is lower than the preset threshold, the predictive maintenance module is triggered to generate a three-dimensional visual maintenance plan and transmit it to the terminal device.

[0015] Furthermore, the S1 further includes:

[0016] The multimodal sensor network adopts a distributed data acquisition architecture. Each sensor node communicates with the edge computing node through the LoRaWAN protocol. The sampling frequency is dynamically adjusted according to the spindle speed of the device. The adjustment rule is:

[0017]

[0018] Where k is the sensor type correlation coefficient, and RPM is the real-time speed of the device.

[0019] Furthermore, the S2 further includes:

[0020] The equipment comprehensive efficiency index CEI is calculated by the following formula:

[0021]

[0022] Among them, E v is the characteristic energy entropy of the vibration signal, which is obtained by performing wavelet packet decomposition on the vibration signal. T is the dynamic time warping distance of the temperature time series data, σ I is the standard deviation of the current waveform, and C1, C2, and C3 are characteristic constants of the device type.

[0023] Furthermore, the S2 further includes:

[0024] The calculation method of the production quality fluctuation coefficient PQFC requires collecting the key quality characteristic value X of the product in N consecutive production cycles. i , calculated using the following formula:

[0025]

[0026] in, is the arithmetic mean of the quality characteristic values, N≥100.

[0027] Furthermore, the S3 further includes:

[0028] The deep reinforcement learning algorithm adopts a dual deep Q network DDQN architecture. The network input layer contains the temporal change gradient features of CEI and PQFC, and the output layer generates a device control parameter adjustment vector.

[0029] Furthermore, the S4 further includes:

[0030] The reverse control compensation adopts fuzzy proportional integral control algorithm, and the proportional coefficient K p Dynamic adjustment is performed based on the exponential moving average of CEI, and the integral time constant is negatively correlated with PQFC.

[0031] Furthermore, the S5 further includes:

[0032] The predictive maintenance module generates a three-dimensional visualization solution that includes spare parts replacement priorities and maintenance time estimates by analyzing the digital twin model parameter deviation matrix.

[0033] Furthermore, the device type characteristic constants C1, C2, and C3 are determined by offline optimization, and the optimization objective function is:

[0034]

[0035] Among them, λ is the regularization factor and T is the total amount of training data.

[0036] Furthermore, the S2 further includes:

[0037] The device digital twin model is constructed using a gated recurrent unit (GRU) network. The number of hidden layer nodes in the network is 2 according to the number of device sensor channels. n Automatic rule configuration

[0038] Second aspect:

[0039] The present invention provides a digital service system for intelligent production equipment based on the Internet of Things, comprising:

[0040] The sensor array, including vibration sensors, temperature sensors, current sensors, and image acquisition units, is physically deployed on the production equipment itself;

[0041] a processor, connected to the sensor array via a communication interface, and configured to execute digital twin model construction, efficiency parameter calculation, and dynamic optimization strategy generation;

[0042] Memory, which stores historical device operation data, optimization strategy library, and preset threshold parameters;

[0043] Actuator controller, connected to the processor via an industrial bus, used to convert the optimization strategy into equipment control instructions;

[0044] A user interface device receives the maintenance plan output by the processor and realizes three-dimensional visual presentation;

[0045] Among them, the processor generates a holographic image of the equipment by real-time analysis of sensor data. When it detects that the comprehensive performance index is lower than the preset threshold in the memory, it triggers the generation of a maintenance plan and transmits it to the user interface device, and at the same time implements reverse control compensation through the actuator controller.

[0046] The beneficial effects brought about by the technical solution provided by the present invention include at least:

[0047] (1) In the present invention, vibration, temperature, current, and image data are collected synchronously through a multimodal sensor network. Combined with wavelet packet decomposition, dynamic time warping algorithm, and standard deviation calculation, the multidimensional characteristics of equipment operation are extracted. Based on the dual-parameter fusion calculation of the equipment comprehensive efficiency index CEI and the production quality fluctuation coefficient PQFC, the correlation between the degradation of equipment mechanical performance and product quality is quantified. This method breaks through the limitations of traditional single-indicator monitoring, accurately identifies the hidden quality risks in the sub-health state of equipment, and provides comprehensive data support for dynamic optimization.

[0048] (2) In this invention, a deep reinforcement learning algorithm is used to construct a dynamic optimization strategy library, generating control parameter adjustment vectors in real time based on the temporal gradient of CEI and PQFC. Combined with a fuzzy proportional-integral control algorithm, the proportional coefficient is dynamically adjusted using the exponential moving average of CEI, and the integral time constant is optimized based on the negative correlation between PQFC and CEI. This method overcomes the response hysteresis problem of static control strategies, enabling adaptive adjustment of equipment operating parameters and ensuring an optimal balance between machining accuracy and energy consumption under complex working conditions.

[0049] (3) In this invention, a highly synchronized digital twin model is constructed through a gated recurrent unit network to map device status in real time and update the parameter deviation matrix. When the CEI is continuously below the threshold, a three-dimensional visual maintenance plan is generated based on multi-dimensional deviation analysis to clarify spare parts replacement priorities and repair path planning. This method avoids the blindness of traditional maintenance that relies on empirical thresholds, enables early warning of equipment degradation trends and accurate decision-making, and significantly reduces the risk of unplanned downtime. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0051] Figure 1 A flowchart of a digital service method for intelligent production equipment based on the Internet of Things provided by an embodiment of the present invention;

[0052] Figure 2 A schematic diagram of a process for preprocessing sensor data in a digital service method for intelligent production equipment based on the Internet of Things provided by an embodiment of the present invention;

[0053] Figure 3 A flowchart of a network training process in a digital service method for intelligent production equipment based on the Internet of Things provided by an embodiment of the present invention;

[0054] Figure 4 A schematic diagram of the workflow of a predictive maintenance module in a digital service method for intelligent production equipment based on the Internet of Things provided by an embodiment of the present invention;

[0055] Figure 5 A schematic diagram of the structure of a digital service system for intelligent production equipment based on the Internet of Things provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0057] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0058] In the embodiments of the present invention, the terms "image" and "picture" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same. The terms "of," "corresponding," and "corresponding" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same.

[0059] In the embodiments of the present invention, sometimes a subscript such as W1 may be mistakenly written as a non-subscript form such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0060] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0061] Reference Manual Figure 1 , which shows a flow chart of a digital service method for intelligent production equipment based on the Internet of Things provided by an embodiment of the present invention.

[0062] An embodiment of the present invention provides a digital service method for intelligent production equipment based on the Internet of Things. The method can be implemented by a digital service device for intelligent production equipment based on the Internet of Things. The digital service device for intelligent production equipment based on the Internet of Things can be a terminal or a server. The processing flow of the digital service method for intelligent production equipment based on the Internet of Things may include the following steps:

[0063] S1. Collect device operation data through a multimodal sensor network deployed on the device body. The multimodal sensor network includes a vibration sensor, a temperature sensor, a current sensor, and an image acquisition unit.

[0064] In one possible implementation, a multimodal sensor network is deployed within the device itself, comprising vibration sensors, temperature sensors, current sensors, and an image acquisition unit. The vibration sensors utilize piezoelectric accelerometers, positioned axially and radially along the device's main shaft to capture vibration signals during operation. The temperature sensors utilize platinum resistance temperature sensors, located in key heat-generating areas of the device, including the bearing seat and motor windings. The current sensors utilize Hall-effect sensors, monitoring the motor's three-phase current waveform in real time. The image acquisition unit utilizes an industrial-grade, high-resolution camera, deployed at key production line locations to capture images of product surfaces for quality inspection.

[0065] The multimodal sensor network uses a distributed data acquisition architecture, with each sensor node communicating with the edge computing node via the Long Range Wide Area Network (LoRaWAN) protocol. The sampling frequency is dynamically adjusted based on the device's spindle speed. The specific adjustment rules are: when the device's real-time RPM is less than 2000, the sampling frequency is set to the product of the sensor type correlation coefficient k and the RPM; when the RPM is greater than or equal to 2000, the sampling frequency is adjusted to twice the product of k and the RPM. The sensor type correlation coefficient k ranges from 0.8 to 1.2, and the specific value is determined by the sensor type and installation location. For example, the k value for a vibration sensor is set to 0.9, and the k value for a temperature sensor is set to 1.1.

[0066] S2. Build a digital twin model of the equipment, map the collected data to the virtual space to generate a holographic image of the equipment, and simultaneously calculate the equipment comprehensive efficiency index CEI and production quality fluctuation coefficient PQFC.

[0067] like Figure 2 As shown in the figure, the collected multimodal data is mapped to the virtual space to build a digital twin model of the device. First, the sensor data is preprocessed:

[0068] S201. Vibration signal processing: Decompose the vibration signal to the fourth layer through wavelet packet decomposition, extract the energy distribution of each frequency band, and calculate the characteristic energy entropy E v Characteristic energy entropy E v The calculation formula is:

[0069]

[0070] Among them, p i It represents the energy proportion of the i-th frequency band. The number of wavelet packet decomposition layers is set to 4 according to the vibration characteristics of the equipment, and a total of 16 frequency bands are obtained.

[0071] S202. Temperature data processing: Dynamic Time Warping (DTW) is used to calculate the dynamic time warping distance (DTWT) between the temperature time series data and a reference temperature curve. The reference temperature curve is generated by averaging 24 consecutive hours of temperature data under normal equipment operation.

[0072] S203, current data processing: perform fast Fourier transform on the current waveform, filter out power frequency interference and calculate the current waveform standard deviation σ I , used to characterize the fluctuation degree of the current signal.

[0073] Based on the preprocessed data, the Comprehensive Efficiency Index (CEI) and the Production Quality Fluctuation Coefficient (PQFC) are calculated:

[0074] The equipment comprehensive effectiveness index CEI is calculated using the following formula:

[0075]

[0076] Among them, C1, C2, and C3 are characteristic constants of the equipment type, which are determined through offline optimization. The offline optimization process includes the following steps:

[0077] Collect historical operation data of equipment and extract E v DTW T , σ I The true value of

[0078] Establish an objective function to minimize the mean square error between the predicted CEI and the measured CEI, and add an L1 regularization term to control the amplitude of the constant change:

[0079]

[0080] Among them, λ is the regularization factor, the default value is 0.01; T is the total amount of training data.

[0081] Collect the key quality characteristic value X of the product in N consecutive production cycles i (N≥100), calculate the arithmetic mean of the quality characteristic values The PQFC is then calculated using the following formula:

[0082]

[0083] For example, for part size detection, X i represents the diameter measurement value of the i-th part, For nominal size, N was set to 100 to ensure statistical significance.

[0084] S3. A dynamic optimization strategy library is established based on the deep reinforcement learning algorithm. The optimal control parameter combination is selected from the strategy library according to the real-time changes of CEI and PQFC.

[0085] Specifically, a dynamic optimization strategy library was constructed using the Dual Deep Q-Network (DDQN) architecture from deep reinforcement learning algorithms. The network's input layer contains the temporal gradient characteristics of the CEI and PQFC, specifically the first-order derivative of the CEI and the sliding window variance of the PQFC over the past 10 minutes. The output layer generates a vector for adjusting the device control parameters, including the spindle speed adjustment ΔRPM, the feed rate correction value, and the coolant flow rate adjustment ratio.

[0086] like Figure 3 As shown in Figure 2, the network training process includes the following steps:

[0087] S301, Experience Replay: Store historical state-action-reward tuples into the experience replay pool, with the pool capacity set to 1000 sets of data;

[0088] S302, target network update: after every 100 training cycles, copy the current network parameters to the target network;

[0089] S303, Strategy Optimization: Update the network weights through the mean square error loss function, which is defined as:

[0090]

[0091] Among them, Q is the action value function of the current network, θ Q is the network parameter, y j is the expected value calculated for the target network, and M is the number of batch training samples.

[0092] S4. Perform reverse control compensation on the equipment actuator through the edge computing node and synchronously update the digital twin model parameters.

[0093] Specifically, the edge computing node receives the optimal control parameter combination generated by the optimization strategy library and performs reverse control compensation on the device actuator through the fuzzy proportional-integral-derivative (PID) control algorithm. p Dynamic adjustment based on the Exponential Moving Average (EMA) of CEI. The specific formula is:

[0094] K p =0.5×EMA(CEI)+1.2

[0095] Integration time constant T i It is negatively correlated with PQFC, and the calculation formula is:

[0096]

[0097] Control commands are transmitted to the equipment actuators via the industrial bus, synchronously updating the bearing wear coefficient, motor efficiency parameters, and thermal conductivity in the digital twin model. Model parameter updates are set to once per minute to ensure the virtual model is synchronized with the physical equipment status.

[0098] S5. When the CEI is lower than the preset threshold, the predictive maintenance module is triggered to generate a three-dimensional visual maintenance plan and transmit it to the terminal device.

[0099] Specifically, if Figure 4 As shown in FIG, when the equipment comprehensive effectiveness index CEI is lower than the preset threshold (for example, 0.75) for three consecutive times, the predictive maintenance module is triggered. The module performs the following operations:

[0100] S501, Deviation Matrix Analysis: Extract the deviation values of key parameters in the digital twin model, including bearing wear deviation, motor efficiency deviation, and temperature distribution deviation, and construct a 5×5 deviation matrix;

[0101] S502, Maintenance Plan Generation: Based on the deviation matrix calculation results, determine the priority of spare parts replacement and maintenance time estimation. The priority rule is: when the bearing wear deviation is greater than 15%, replace the bearing first; when the motor efficiency deviation is greater than 8%, repair the motor winding;

[0102] S503, 3D visualization presentation: The maintenance plan is converted into a 3D model, and superimposed onto the real-life equipment image through augmented reality (AR) technology, with the disassembly path, tool list, and safety precautions marked.

[0103] Reference Manual Figure 5 , which shows a structural diagram of a digital service system for intelligent production equipment based on the Internet of Things provided by an embodiment of the present invention.

[0104] The present invention also provides a digital service system for intelligent production equipment based on the Internet of Things, which is applied to a digital service method for intelligent production equipment based on the Internet of Things, including:

[0105] The sensor array, including vibration sensors, temperature sensors, current sensors, and image acquisition units, is physically deployed on the production equipment itself;

[0106] a processor, connected to the sensor array via a communication interface, and configured to execute digital twin model construction, performance parameter calculation, and dynamic optimization strategy generation;

[0107] Memory, which stores historical device operation data, optimization strategy library, and preset threshold parameters;

[0108] Actuator controller, connected to the processor via an industrial bus, used to convert the optimization strategy into equipment control instructions;

[0109] A user interface device receives the maintenance plan output by the processor and realizes three-dimensional visual presentation;

[0110] Among them, the processor generates a holographic image of the equipment by real-time analysis of sensor data. When it detects that the comprehensive performance index is lower than the preset threshold in the memory, it triggers the generation of a maintenance plan and transmits it to the user interface device. At the same time, reverse control compensation is implemented through the actuator controller.

[0111] The present invention provides a digital service system for intelligent production equipment based on the Internet of Things that can execute a digital service method for intelligent production equipment based on the Internet of Things and achieve the same or similar technical effects. To avoid repetition, the present invention will not go into details.

[0112] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0113] (1) In the present invention, vibration, temperature, current, and image data are collected synchronously through a multimodal sensor network. Combined with wavelet packet decomposition, dynamic time warping algorithm, and standard deviation calculation, the multidimensional characteristics of equipment operation are extracted. Based on the dual-parameter fusion calculation of the equipment comprehensive efficiency index CEI and the production quality fluctuation coefficient PQFC, the correlation between the degradation of equipment mechanical performance and product quality is quantified. This method breaks through the limitations of traditional single-indicator monitoring, accurately identifies the hidden quality risks in the sub-health state of equipment, and provides comprehensive data support for dynamic optimization.

[0114] (2) In this invention, a deep reinforcement learning algorithm is used to construct a dynamic optimization strategy library, generating control parameter adjustment vectors in real time based on the temporal gradient of CEI and PQFC. Combined with a fuzzy proportional-integral control algorithm, the proportional coefficient is dynamically adjusted using the exponential moving average of CEI, and the integral time constant is optimized based on the negative correlation between PQFC and CEI. This method overcomes the response hysteresis problem of static control strategies, enabling adaptive adjustment of equipment operating parameters and ensuring an optimal balance between machining accuracy and energy consumption under complex working conditions.

[0115] (3) In this invention, a highly synchronized digital twin model is constructed through a gated recurrent unit network to map device status in real time and update the parameter deviation matrix. When the CEI is continuously below the threshold, a three-dimensional visual maintenance plan is generated based on multi-dimensional deviation analysis to clarify spare parts replacement priorities and repair path planning. This method avoids the blindness of traditional maintenance that relies on empirical thresholds, enables early warning of equipment degradation trends and accurate decision-making, and significantly reduces the risk of unplanned downtime.

[0116] The above content is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

[0117] There are a few points to note:

[0118] (1) The drawings of the embodiments of the present invention only relate to the structures related to the embodiments of the present invention. Other structures may refer to conventional designs.

[0119] (2) For the sake of clarity, the thickness of layers or regions in the drawings used to describe the embodiments of the present invention are exaggerated or reduced, that is, these drawings are not drawn to scale. It is understood that when an element such as a layer, film, region, or substrate is referred to as being "on" or "under" another element, the element may be "directly" "on" or "under" the other element or intervening elements may be present.

[0120] (3) In the absence of conflict, the embodiments of the present invention and the features therein may be combined with each other to form new embodiments.

[0121] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A digital service method for intelligent production equipment based on the Internet of Things, characterized in that: include: S1. Collecting device operation data through a multimodal sensor network deployed on the device body, wherein the multimodal sensor network includes a vibration sensor, a temperature sensor, a current sensor, and an image acquisition unit; S2. Build a digital twin model of the equipment, map the collected data into a virtual space to generate a holographic image of the equipment, and simultaneously calculate the equipment comprehensive efficiency index (CEI) and production quality fluctuation coefficient (PQFC); S3. Establish a dynamic optimization strategy library based on deep reinforcement learning algorithm, and select the best control parameter combination from the strategy library according to the real-time changes of CEI and PQFC; S4. Perform reverse control compensation on the equipment actuator through the edge computing node and synchronously update the digital twin model parameters; S5. When the CEI is lower than the preset threshold, the predictive maintenance module is triggered to generate a three-dimensional visual maintenance plan and transmit it to the terminal device.

2. The method for digital service of intelligent production equipment based on the Internet of Things according to claim 1, characterized in that: Said S1 further comprises: The multimodal sensor network adopts a distributed data acquisition architecture. Each sensor node communicates with the edge computing node through the LoRaWAN protocol. The sampling frequency is dynamically adjusted according to the spindle speed of the device. The adjustment rule is: Where k is the sensor type correlation coefficient, and RPM is the real-time speed of the device.

3. The method for digital service of intelligent production equipment based on the Internet of Things according to claim 2, characterized in that: Said S2 further comprises: The equipment comprehensive efficiency index CEI is calculated by the following formula: Among them, E v is the characteristic energy entropy of the vibration signal, which is obtained by performing wavelet packet decomposition on the vibration signal. T is the dynamic time warping distance of the temperature time series data, σ I is the standard deviation of the current waveform, and C1, C2, and C3 are characteristic constants of the device type.

4. The method for digital service of intelligent production equipment based on the Internet of Things according to claim 3, characterized in that: Said S2 further comprises: The calculation method of the production quality fluctuation coefficient PQFC requires collecting the key quality characteristic value X of the product in N consecutive production cycles. i , calculated using the following formula: in, is the arithmetic mean of the quality characteristic values, N≥100.

5. The method for digital service of intelligent production equipment based on the Internet of Things according to claim 4, characterized in that: Said S3 further comprises: The deep reinforcement learning algorithm adopts a dual deep Q network DDQN architecture. The network input layer contains the temporal change gradient features of CEI and PQFC, and the output layer generates a device control parameter adjustment vector.

6. The method for digital service of intelligent production equipment based on the Internet of Things according to claim 5, characterized in that: Said S4 further comprises: The reverse control compensation adopts fuzzy proportional integral control algorithm, and the proportional coefficient K p Dynamic adjustment is performed based on the exponential moving average of CEI, and the integral time constant is negatively correlated with PQFC.

7. The method for digital service of intelligent production equipment based on the Internet of Things according to claim 6, characterized in that: Said S5 further comprises: The predictive maintenance module generates a three-dimensional visualization solution that includes spare parts replacement priorities and maintenance time estimates by analyzing the digital twin model parameter deviation matrix.

8. The method for digital service of intelligent production equipment based on the Internet of Things according to claim 3, characterized in that: include: The equipment type characteristic constants C1, C2, and C3 are determined by offline optimization, and the optimization objective function is: Among them, λ is the regularization factor and T is the total amount of training data.

9. The method for digital service of intelligent production equipment based on the Internet of Things according to claim 1, characterized in that: Said S2 further comprises: The device digital twin model is constructed using a gated recurrent unit (GRU) network. The number of hidden layer nodes in the network is 2 according to the number of device sensor channels. n Rules are automatically configured.

10. A digital service system for intelligent production equipment based on the Internet of Things, characterized in that: include: The sensor array, including vibration sensors, temperature sensors, current sensors, and image acquisition units, is physically deployed on the production equipment itself; a processor, connected to the sensor array via a communication interface, and configured to execute digital twin model construction, efficiency parameter calculation, and dynamic optimization strategy generation; Memory, which stores historical device operation data, optimization strategy library, and preset threshold parameters; Actuator controller, connected to the processor via an industrial bus, used to convert the optimization strategy into equipment control instructions; A user interface device receives the maintenance plan output by the processor and realizes three-dimensional visual presentation; Among them, the processor generates a holographic image of the equipment by real-time analysis of sensor data. When it detects that the comprehensive performance index is lower than the preset threshold in the memory, it triggers the generation of a maintenance plan and transmits it to the user interface device, and at the same time implements reverse control compensation through the actuator controller.

Images