Integrated SIP (Session Initiation Protocol) chip for predictive maintenance of industrial equipment

Abstract

The utility model discloses an integrated SIP (Session Initiation Protocol) chip for predictive maintenance of industrial equipment, which comprises a packaging space formed by a substrate and a packaging shell, a vibration sensor, an analog-to-digital converter, a microprocessor, a Flash memory, an SRAM (Static Random Access Memory) memory, an AI (Artificial Intelligence) core and a wireless communication module which are integrated on the substrate are arranged in the packaging space, and the Flash memory, the SRAM memory, the AI core and the wireless communication module are electrically connected with the microprocessor. The vibration sensor is electrically connected with the input end of the AI core through the analog-to-digital converter; the vibration sensor monitors the running state of equipment and outputs a vibration analog signal, the analog-to-digital converter converts the vibration analog signal into a digital signal, the AI core performs feature extraction and analysis on the digital signal and gives out fault prediction, and the microprocessor coordinates the work of each module and processes according to the fault prediction. And the information is received and sent to the cloud through the wireless communication module. The system is small in size, reduces the installation complexity, and improves the integration level, reliability and operation efficiency of the system.

Description

Technical Field

[0001] This utility model relates to the field of predictive maintenance technology for industrial equipment, and in particular to an integrated SIP chip for predictive maintenance of industrial equipment. Background Technology

[0002] With the rapid development of Industry 4.0 and intelligent manufacturing, predictive maintenance (PdM) for industrial equipment has become a key technology for improving equipment operating efficiency and reducing maintenance costs. Predictive maintenance achieves precise maintenance by monitoring equipment status in real time (such as vibration and sound) and combining data analysis to predict equipment failures. However, existing technologies have the following problems:

[0003] Separation of sensors and processing modules: Traditional industrial equipment monitoring systems typically consist of independent sensors, data acquisition modules, and processing units, resulting in complex systems, low integration, high power consumption, and high installation and maintenance costs.

[0004] Low data processing efficiency: Traditional solutions typically rely on external servers or the cloud for data processing, which results in data transmission delays and network dependencies. This not only fails to meet the high real-time requirements of scenarios but also increases a lot of ineffective data transmission.

[0005] Limited functionality: Existing monitoring equipment has limited functionality and cannot simultaneously achieve integrated operation of data acquisition, processing and wireless transmission.

[0006] Large size: Modules that integrate the above-mentioned multiple functions will have a large size. For example, the acquisition system disclosed in Chinese invention patent announcement number CN112327784B is difficult to install on some small and medium-sized devices.

[0007] Low reliability: Since most application scenarios are in harsh environments, monitoring equipment must withstand long-term and continuous vibration. The more complex the equipment functions and the more chips contained in the module, the higher its failure rate.

[0008] Therefore, there is an urgent need for a highly integrated, low-power, multifunctional solution to meet the actual needs of predictive maintenance for industrial equipment. Utility Model Content

[0009] The purpose of this invention is to address the shortcomings of existing technologies by proposing an integrated SIP chip for predictive maintenance of industrial equipment. By integrating a vibration sensor, ADC, AI core, microprocessor, Flash, SRAM, and wireless communication module into a single chip, it achieves seamless integration of vibration data acquisition, processing, analysis, and wireless transmission, thereby reducing system size, installation complexity, and improving system integration, reliability, and operating efficiency.

[0010] The technical solution to achieve the purpose of this utility model is:

[0011] An integrated SIP chip for predictive maintenance of industrial equipment includes a packaged space formed by a substrate and a packaged housing covering the substrate. The packaged space houses a vibration sensor, an analog-to-digital converter (ADC), a microprocessor, and Flash memory, SRAM memory, an AI core, and a wireless communication module, all electrically connected to the microprocessor. The vibration sensor is electrically connected to the input terminal of the AI ​​core via the ADC. The vibration sensor monitors the operating status of the equipment and outputs a vibration analog signal. The ADC converts the vibration analog signal into a digital signal. The AI ​​core extracts and analyzes features from the digital signal and provides a fault prediction. The microprocessor coordinates the operation of each module, processes the fault prediction, and receives and sends information to the cloud via the wireless communication module.

[0012] Furthermore, it also includes a sound sensor, which is electrically connected to the analog-to-digital converter to collect the device's analog sound signals.

[0013] Furthermore, the sound sensor is a MEMS sound sensor with a sound hole on the top, and the package housing has an opening. The diaphragm of the sound sensor is flush with the top of the opening and exposed on the package surface through the opening.

[0014] Furthermore, the top of the opening is covered with a protective layer.

[0015] Furthermore, the protective layer is, for example, a mesh film or a hydrophobic coating.

[0016] Furthermore, the vibration sensor is a triaxial vibration sensor.

[0017] Furthermore, the sampling accuracy of the analog-to-digital converter is at least 16 bits, and the sampling frequency is not less than twice the response frequency of the vibration sensor.

[0018] Furthermore, the AI ​​core integrates a lightweight neural network and supports edge computing, and the model size does not exceed 1MB.

[0019] Furthermore, the wireless communication module is a wireless communication chip using a 2.4GHz WIFI, Bluetooth, LoRa or Zigbee wireless communication protocol. The wireless communication chip is covered with a metal shielding cover, and the substrate is provided with an independent grounding layer for the wireless communication chip.

[0020] By adopting the above technical solution, this utility model has the following beneficial effects:

[0021] (1) This utility model integrates a vibration sensor, analog-to-digital converter, AI core, microprocessor, Flash memory, SRAM memory and wireless communication module into a single packaged chip, replacing the traditional PCB-based functional modules with a single chip, realizing seamless connection of data acquisition, processing, analysis and wireless transmission of various types of data related to predictive maintenance. It has high integration and small size, which reduces installation complexity and improves system reliability. Furthermore, the application of AI core in this chip has edge computing capabilities, reduces network dependence, improves real-time performance, and thus improves overall operating efficiency.

[0022] (2) By adding a sound sensor, this utility model can work with a vibration sensor to determine the operating status of the equipment and improve the accuracy of data acquisition.

[0023] (3) By adding an opening and aligning the diaphragm of the sound sensor with the opening, this utility model enables the sound sensor to pick up sound signals better and further improves the accuracy of sound data acquisition.

[0024] (4) By setting a protective layer, this utility model meets the IPX level requirements, protects the sound sensor from environmental influences, and further improves the accuracy of sound data acquisition.

[0025] (5) This utility model uses a triaxial vibration sensor, which has more comprehensive monitoring capabilities, higher diagnostic accuracy and data analysis potential. At the same time, only one sensor is needed to complete the all-round monitoring task that originally might have required three independent sensors, reducing the installation workload and cost.

[0026] (6) This utility model avoids radio frequency signals from interfering with other parts by adding a metal shielding cover and a grounding layer. Attached Figure Description

[0027] To make the content of this utility model easier to understand, the present utility model will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein:

[0028] Figure 1 This is a structural block diagram of the present invention;

[0029] Figure 2 This is a simplified diagram of the hardware structure of this utility model.

[0030] The labels in the attached diagram are:

[0031] 1. Substrate; 2. Encapsulation housing; 3. Vibration sensor; 4. Sound sensor; 5. Analog-to-digital converter; 6. Microprocessor; 7. Flash memory; 8. SRAM memory; 9. AI core; 10. Wireless communication module. Detailed Implementation

[0032] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0033] (Example 1)

[0034] like Figures 1 to 2 The integrated SIP chip shown is for predictive maintenance of industrial equipment. It includes a packaged space formed by a substrate 1 and a package housing 2 covering the substrate 1. Within the packaged space are integrated components on the substrate 1: a vibration sensor 3, a sound sensor 4, an analog-to-digital converter 5, a microprocessor 6, a flash memory 7, an SRAM memory 8, an AI core 9, and a wireless communication module 10. The vibration sensor 3 and sound sensor 4 are electrically connected to the input of the analog-to-digital converter 5, respectively, to collect analog vibration and sound signals from the device under test. The output of the analog-to-digital converter 5 is sequentially connected to the AI ​​core 9, the microprocessor 6, and the wireless communication module 10. The flash memory 7 and SRAM memory 8 are both electrically connected to the microprocessor 6. The analog-to-digital converter 5 converts the analog signals into digital signals. The AI ​​core 9 extracts and analyzes the features of the digital signals and provides fault predictions. The microprocessor 6 coordinates the operation of each module, processes the fault predictions, and receives and sends information to the cloud via the wireless communication module 10.

[0035] Specifically, vibration sensor 3 is used to collect vibration signals of industrial equipment in real time and monitor the operating status of the equipment. It adopts a triaxial vibration sensor and selects a sensor with a frequency response range of not less than 2kHz. It has more comprehensive monitoring capabilities, higher diagnostic accuracy and data analysis potential. At the same time, only one sensor is needed to complete the all-round monitoring tasks that originally might have required three independent sensors, reducing the amount of installation work and cost.

[0036] The sound sensor 4 is used to collect sound signals from industrial equipment in real time. Combined with data from the vibration sensor, it helps determine the equipment's operating status. It employs a MEMS sound sensor with a sound aperture on the top. The housing 2 has an opening, and the diaphragm of the sound sensor is flush with the top of the opening and exposed on the housing surface, allowing for better sound signal pickup. To protect the sound sensor 4 from environmental influences, the top of the opening is covered with a protective layer, such as a mesh film or a hydrophobic coating, meeting IPX rating requirements and further improving the accuracy of sound data acquisition.

[0037] The analog-to-digital converter 5 has a sampling accuracy of at least 16 bits and a sampling frequency of no less than twice the response frequency of the vibration sensor 3. It is responsible for converting the analog signals collected by various sensors into digital signals to support subsequent data processing.

[0038] AI Core 9 is an edge-based AI core that integrates lightweight neural networks, an AI algorithm library, and a pre-trained fault model. It supports edge computing, with a model size not exceeding 1MB. It is used for real-time analysis and fault prediction of various sensor data, reducing reliance on cloud computing. For general predictive maintenance applications, AI Core 9's processing power should be no less than 100 GOPS. AI Core 9 extracts and analyzes features from digital signals, and combined with the pre-trained fault prediction model, provides fault prediction conclusions.

[0039] The microprocessor 6 can be selected from various options, such as a microprocessor based on ARM or RISC V architecture as the main control unit, responsible for coordinating the work of various modules, performing data processing, algorithm scheduling and communication tasks, receiving the inference results from the AI ​​core, and uploading relevant information to the cloud by calling the wireless communication module according to a preset strategy. This embodiment uses a Cortex-M4 core with a main frequency of 120MHz and integrates an FPU to support floating-point operations.

[0040] The Flash memory 7 has a capacity of 1MB and is used to store firmware, AI models, and historical data, supporting offline data storage.

[0041] The SRAM memory has a capacity of 256KB, providing high-speed data caching and supporting real-time data processing.

[0042] The wireless communication module 10 is used to transmit processed data or alarm information to a remote server or terminal device. It employs a 2.4GHz wireless communication chip using WIFI, Bluetooth, LoRa, or Zigbee protocols. The wireless communication chip is covered with a metal shielding cover, and the substrate 1 has an independent grounding layer for the wireless communication chip to prevent radio frequency signals from interfering with other components. The wireless communication circuit typically includes components such as radio frequency circuits, antennas, matching circuits, and crystal oscillators. In a SiP (System-in-Package), these components can be integrated into the same package, reducing external connections and optimizing space. Due to the use of the 2.4GHz high-frequency communication protocol, the antenna size is relatively small, and the antenna can also be integrated inside the SiP. This internal antenna uses an on-chip antenna or an inverted-F antenna. Considering the signal attenuation caused by the SiP packaging material, a low-dielectric-constant polymer should be used for the SiP packaging material, preferably a liquid crystal polymer (LCP).

[0043] Among the aforementioned functional modules, vibration sensor 3, analog-to-digital converter 5, AI core 9, microprocessor 6, flash memory 7, SRAM memory 8, and wireless communication module 10 all use bare dies to reduce the overall size of the SIP.

[0044] Taking a rotating machine (such as an electric motor) in a factory as an example, the vibration sensor 3 is a Bosch Sensortec BMA456, which is a high-performance triaxial accelerometer with a frequency response range of up to 2kHz, a range of ±2g / ±4g / ±8g selectable, a resolution of 16bit, low power consumption, and is suitable for industrial vibration monitoring.

[0045] The sound sensor 4, model Infineon IM69D130, is a high-performance MEMS microphone with a signal-to-noise ratio of 69dB, a frequency response range of 20Hz to 20kHz, and a sound hole on top, making it suitable for sound signal acquisition in industrial equipment.

[0046] The Texas Instruments ADS1256 analog-to-digital converter (ADC) is a 24-bit high-precision delta-sigma ADC with a maximum sampling rate of 30 kSPS, meeting the high-frequency sampling requirements of vibration sensors (more than twice the 2 kHz response frequency) and sound sensors. It supports multi-channel input and can process both vibration and analog sound signals simultaneously.

[0047] The microprocessor 6 is model STMicroelectronics STM32F429IG, based on the ARM Cortex-M4 core, with a main frequency of 180MHz. It integrates a floating-point unit (FPU), 1MB of Flash and 256KB of SRAM, making it suitable for real-time data processing and module coordination. It is packaged as a bare die and supports AI core inference result processing and wireless communication task scheduling.

[0048] The Flash memory 7, model number Micron MT25QL01GBBB, is a 1MB serial NOR Flash memory that supports high-speed read and write, has low power consumption, and is used to store firmware, AI models, and historical data, meeting the needs of offline data caching.

[0049] The SRAM memory 8 is model Cypress CY62167EV30, a 256KB low-power SRAM with an access time of 45ns, suitable for high-speed data caching and supporting real-time data processing.

[0050] AI Core 9, model Eta Compute ECM3532, is an ultra-low-power AI processor that integrates a lightweight neural network with a model size of less than 1MB and a computing power of 100 GOPS, supporting edge computing. It features an embedded pre-trained fault prediction model suitable for feature extraction and fault diagnosis of vibration and sound data.

[0051] The wireless communication module 10, model number Espressif ESP32-S3, is a wireless communication chip that supports 2.4GHz WiFi and Bluetooth 5.0. It integrates RF circuits, matching circuits, and an on-chip inverted F antenna, and has low power consumption, making it suitable for industrial IoT applications.

[0052] The SIP chip is mounted on the surface of the device, and the specific workflow is as follows:

[0053] Vibration sensor 3 and sound sensor 4 collect vibration and sound signals of the equipment in real time, and analog-to-digital converter 5 converts analog signals into digital signals.

[0054] AI Core 9 performs frequency domain analysis on the data, extracts the characteristics of the vibration signal (such as peak value and frequency components), and compares them with the pre-trained model to determine whether there are any abnormalities (such as bearing wear).

[0055] The microprocessor 6 generates an alarm signal based on the analysis results of the AI ​​core 9, and sends the results to the factory's monitoring center via the wireless communication module 10.

[0056] If the network is interrupted, the Flash memory 7 can cache the data and transmit it after the network is restored.

[0057] This invention integrates a vibration sensor, analog-to-digital converter, AI core, microprocessor, Flash memory, SRAM memory, and wireless communication module into a single packaged chip. It replaces the traditional PCB-based functional modules with a single chip, achieving seamless integration of data acquisition, processing, analysis, and wireless transmission related to predictive maintenance. The high integration and small size reduce installation complexity while improving system reliability. Furthermore, the application of the AI ​​core in this chip provides edge computing capabilities, reducing network dependence, improving real-time performance, and thus enhancing overall operating efficiency.

[0058] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this utility model. It should be understood that the above descriptions are merely specific embodiments of this utility model and are not intended to limit this utility model. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this utility model should be included within the protection scope of this utility model.

Claims

1. An integrated SIP chip for predictive maintenance of industrial equipment, characterized in that: The device includes an encapsulation space formed by a substrate and an encapsulation shell covering the substrate. Within the encapsulation space are integrated a vibration sensor, an analog-to-digital converter (ADC), a microprocessor, and Flash memory, SRAM memory, an AI core, and a wireless communication module, all electrically connected to the microprocessor. The vibration sensor is electrically connected to the input of the AI ​​core via the ADC. The vibration sensor monitors the device's operating status and outputs a vibration analog signal. The ADC converts the vibration analog signal into a digital signal. The AI ​​core extracts and analyzes features from the digital signal and provides fault prediction. The microprocessor coordinates the work of each module, processes the fault prediction, and receives and sends information to the cloud via the wireless communication module.

2. The integrated SIP chip for predictive maintenance of industrial equipment according to claim 1, characterized in that: It also includes a sound sensor, which is electrically connected to the analog-to-digital converter to collect the device's analog sound signals.

3. An integrated SIP chip for predictive maintenance of industrial equipment according to claim 2, characterized in that: The sound sensor is a MEMS sound sensor with a sound hole on the top. The package housing has an opening, and the diaphragm of the sound sensor is flush with the top of the opening and exposed on the package surface through the opening.

4. An integrated SIP chip for predictive maintenance of industrial equipment according to claim 3, characterized in that: The top of the opening is covered with a protective layer.

5. An integrated SIP chip for predictive maintenance of industrial equipment according to claim 1, characterized in that: The vibration sensor is a triaxial vibration sensor.

6. An integrated SIP chip for predictive maintenance of industrial equipment according to claim 1, characterized in that: The analog-to-digital converter has a sampling accuracy of at least 16 bits and a sampling frequency of not less than twice the response frequency of the vibration sensor.

7. An integrated SIP chip for predictive maintenance of industrial equipment according to claim 1, characterized in that: The AI ​​core integrates a lightweight neural network and supports edge computing, with a model size not exceeding 1MB.

8. An integrated SIP chip for predictive maintenance of industrial equipment according to claim 1, characterized in that: The wireless communication module is a wireless communication chip that uses a 2.4GHz WIFI, Bluetooth, LoRa or Zigbee wireless communication protocol. The wireless communication chip is covered with a metal shielding cover, and the substrate has an independent grounding layer for the wireless communication chip.

Images