Mems technology-based black light factory mom collaborative platform
The MOM (Manufacturing Controller) collaborative platform for the dark factory, powered by MEMS technology, enables proactive perception of equipment status and optimization of business scheduling. It solves the problem of the lack of commercial risk control rules in the MOM system and improves the production efficiency and operational effectiveness of the dark factory.
Patent Information
- Application Number
- CN202611618507.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-20
- Publication Date
- 2026-11-17
AI Technical Summary
In a dark factory, the existing MOM system lacks scheduling upper and lower limit constraints based on the enterprise's overall business risk control rules. This makes it easy for aggressive rescheduling of a single plan to cause the entire factory's supply chain to break down, internal logistics distribution to become unbalanced, and inventory costs to accumulate abnormally, leading to economic losses and fluctuations in operating costs.
The MOM (Multi-Manager) collaborative platform for the "lights-out" factory, based on MEMS technology, collects equipment status signals through a data acquisition module, assesses equipment health status through a status evaluation module, calculates business scheduling reference benchmark values through a benchmark determination module, calculates global business strategy adjustment amounts through a correction calculation module, and performs boundary limiting processing through a decision execution module, thereby achieving proactive perception of equipment status degradation and optimization of business scheduling.
It achieves precise linkage between microscopic physical degradation of equipment and macroscopic management decisions, ensuring the control of commercial risk limits, avoiding supply chain shocks caused by changes in production plans, improving the stability of enterprise management decisions and equipment lifespan, and increasing asset utilization and operational efficiency.
Smart Images

Figure CN123117737A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of enterprise manufacturing operations management, and in particular to a MOM collaborative platform for lights-out factories based on MEMS technology. Background Technology
[0002] A "dark factory," as an advanced form of smart manufacturing, refers to a modern factory that achieves fully autonomous production through automated equipment, industrial robots, and intelligent control systems with minimal or no human intervention. Its core characteristic is production even in the dark, representing a fundamental leap in manufacturing from human-operated machines to systems making autonomous decisions.
[0003] In the architecture of a lights-out factory, the MOM (Manufacturing Operations Management) system plays a crucial bridging role. The MOM system receives production instructions from the ERP (Enterprise Resource Planning) layer above and connects with the execution layers such as MES (Manufacturing Execution System), EAM (Enterprise Asset Management), QMS (Quality Management System), and WMS (Warehouse Management System) below, coordinating and managing the five major functional areas of production planning, equipment scheduling, quality inspection, equipment maintenance, and material distribution. The quality of decisions made by the MOM system directly determines the production efficiency, product quality, and resource utilization rate of the lights-out factory.
[0004] Meanwhile, MEMS (Micro-Electro-Mechanical Systems) technology has made groundbreaking progress in recent years. MEMS sensors, with their advantages of small size, low power consumption, high integration, and strong anti-interference capabilities, have been widely used in the field of industrial equipment condition monitoring. By deploying MEMS accelerometers, temperature sensors, and acoustic emission sensors on key manufacturing equipment, vibration, temperature, and acoustic signals during equipment operation can be collected in real time, providing rich physical layer data support for equipment health status assessment.
[0005] Even if a few advanced systems attempt to incorporate the status data of underlying assets into MOM business decisions, their business correction mechanisms often suffer from a lack of management flexibility.
[0006] Due to the lack of upper and lower limits for scheduling based on the overall business risk control rules of the enterprise, aggressive rescheduling of a single plan can easily lead to the disconnection of the entire plant's supply chain, imbalance of internal logistics allocation, and abnormal accumulation of inventory costs, thereby causing serious economic losses and fluctuations in operating costs within the enterprise. Therefore, we propose a MEMS-based MOM (Multi-Manager Manufacturing) collaborative platform for lights-out factories to address the aforementioned issues. Summary of the Invention
[0007] This invention provides a MEMS-based MOM collaborative platform for lights-out factories, which enables manufacturing systems to proactively perceive equipment degradation in unmanned, lights-out factory scenarios.
[0008] The first aspect of this invention provides a MEMS-based MOM (Manufacturing Maker) collaborative platform for a lights-out factory. The platform includes: a data acquisition module for collecting operating status signals of manufacturing equipment to extract real-time status data and acquiring baseline status data of the manufacturing equipment under normal operating conditions; a status assessment module for comparing the real-time status data with the baseline status data to determine the health status indicators of the manufacturing equipment; and a baseline determination module for converting the health status indicators into comprehensive influence factors of each functional link based on the correlation weights between the manufacturing equipment and various functional links of the manufacturing operation management system, and combining this with a preset initial setting. The system defines parameters to determine the business scheduling reference benchmark value for each functional link; a correction calculation module is used to enable each functional link to calculate its own deviation signal based on its corresponding actual business plan and the business scheduling reference benchmark value, and to obtain the neighbor deviation signal of neighboring functional links; a global business strategy adjustment amount is calculated based on the self-deviation signal, the neighbor deviation signal, the actual business plan, and the business scheduling reference benchmark value; a decision execution module is used to perform boundary limiting processing on the global business strategy adjustment amount to obtain the final business scheduling parameters, and to update the original business plan of each functional link using the final business scheduling parameters to obtain the actual executed enterprise operation management plan.
[0009] Optionally, in a first implementation of the first aspect of the present invention, a parameter initialization and calibration module is further included: The manufacturing equipment is operated under preset benchmark conditions, the operating status signals are collected and statistical feature values are calculated to obtain the benchmark status data; Assess the dependence strength between the manufacturing equipment and each functional link, generate corresponding initial weight values, and establish an association weight table to provide the association weights; Each functional component is assigned a collaborative operation parameter, which includes at least the initial setting parameter. The actual business plan of each functional link is initialized to be consistent with the normal setting value in the initial setting parameter, and the initial value of the calculation item associated with the global business strategy adjustment amount is set to a preset zero value.
[0010] Optionally, in a second implementation of the first aspect of the present invention, the parameter initialization and calibration module establishes an associated weight table, specifically used for: Acquire historical abnormal event data of the manufacturing equipment, and extract the response time interval between the occurrence of an abnormality in the manufacturing equipment and the first change in the business plan of the corresponding functional link; The dependency strength is determined based on the response time interval, and the dependency strength decreases as the response time interval increases. The calculated dependency strength value is weighted and fused with the initial weight value obtained from the score to obtain a fused weight value, and an association weight table is constructed using the fused weight value.
[0011] Optionally, in a third implementation of the first aspect of the present invention, the data acquisition module collects the operating status signal of the manufacturing equipment to extract real-time status data, specifically for: The operating status signals are collected by a sensor array, and the operating status signals include vibration signals, temperature signals, and acoustic emission signals; The operating status signal is preprocessed to extract real-time status data including effective vibration value, average temperature value and acoustic emission pulse count rate.
[0012] Optionally, in a fourth implementation of the first aspect of the present invention, the status assessment module compares the real-time status data with the baseline status data to determine the health status indicators of the manufacturing equipment, specifically for: Acquire the baseline state data, which includes the effective value of vibration, the average temperature, and the average baseline values corresponding to the acoustic emission pulse count rate; The effective value of vibration, the average temperature, and the acoustic emission pulse count rate in the real-time status data are compared with the corresponding average values in the reference status data to obtain the corresponding deviation ratios. The physical deviation is obtained by assigning corresponding weighting coefficients to each deviation ratio and summing them. The health impact coefficient is determined based on the comparison between the physical deviation and the preset threshold, and the health impact coefficient is used as a health status indicator.
[0013] Optionally, in a fifth implementation of the first aspect of the present invention, before assigning corresponding weight coefficients to each of the deviation ratios and summing them, the state evaluation module is further configured to perform a consistency check on each of the deviation ratios: Calculate the absolute value of the difference between each pair of the deviation ratios mentioned in each item; If there exists one and only one deviation ratio whose absolute difference from all other deviation ratios exceeds a preset deviation threshold, then that deviation ratio is removed. The weighting coefficients that originally corresponded to the eliminated deviation ratios are redistributed proportionally to the remaining deviation ratios, and then summed.
[0014] Optionally, in a sixth implementation of the first aspect of the present invention, the benchmark determination module, based on the correlation weights of the manufacturing equipment and the various functional links of the manufacturing operation management system, transforms the health status indicators into comprehensive influence factors for each functional link, and, in conjunction with preset initial setting parameters, determines the business scheduling reference benchmark value for each functional link, specifically used for: Retrieve the association weight table containing the aforementioned association weights; The health status indicators are weighted, summed, and normalized using the correlation weights in the correlation weight table to obtain a comprehensive influence factor. Calculate the physical reference correction amount based on the comprehensive impact factor and the response sensitivity of each functional component; The normal setting values of each functional component are superimposed with the physical reference correction value to obtain the business scheduling reference baseline value; wherein, the response sensitivity and the normal setting value are the initial setting parameters.
[0015] Optionally, in a seventh implementation of the first aspect of the present invention, the correction amount calculation module calculates the global business strategy adjustment amount based on the self-deviation signal, the neighbor deviation signal, the actual business plan, and the business scheduling reference benchmark value, specifically for: Obtain the timestamp corresponding to the neighbor deviation signal, determine the information age based on the difference between the timestamp and the current time, and calculate the aging decay coefficient based on the information age; Calculate the difference between the comprehensive influence factor of the current functional link in each functional link and the comprehensive influence factor of the corresponding neighboring functional link, and dynamically adjust the basic collaborative coupling weight based on the difference to obtain the collaborative coupling weight; The cooperative correction component is calculated by combining the self-bias signal, the neighbor bias signal, the aging attenuation coefficient, and the cooperative coupling weight; Calculate the self-driven correction component by combining the actual business plan with the business scheduling reference benchmark value; The collaborative correction component and the self-driven correction component are superimposed to obtain the global business strategy adjustment amount.
[0016] Optionally, in an eighth implementation of the first aspect of the present invention, the correction calculation module dynamically adjusts the basic cooperative coupling weight based on the difference to obtain the cooperative coupling weight, specifically for: Obtain the absolute value of the difference to get the absolute value of the physical difference; If the absolute value of the physical difference is greater than the physical difference threshold, then calculate the reduction amount that increases as the absolute value of the physical difference increases; The reduction amount is subtracted from the basic cooperative coupling weight, and the result of the subtraction is not less than the lower limit of the coupling weight, and the result of the subtraction is taken as the cooperative coupling weight.
[0017] Optionally, in a ninth implementation of the first aspect of the present invention, the aging degradation coefficient is calculated using the following formula. : ; in, Information age, This refers to the effective period of information.
[0018] The mechanism of this invention is as follows: After being constrained by the business security boundary, the consistency-limited amplitude is superimposed on the original scheduling decision model, so that the MOM system can achieve progressive, self-organized collaborative scheduling across business links before triggering serious downtime and overdue risks. This transforms the changes in the efficiency of underlying resources into a continuous adaptive response to the upper-level business operation decisions, breaking the rigid open-loop scheduling mode of the traditional MOM system. Beneficial effects: It achieves precise binding between micro-physical degradation of equipment and macro-management decision parameters, enabling the MOM system to keenly perceive subtle changes in asset performance and automatically transform them into enterprise-level business scheduling plans; Commercial risk control limits ensure that no business flow rescheduling action will exceed the bottom line of enterprise operating costs and supply chain capacity, solving the internal management chaos and business shock caused by excessive adjustment of scheduling indicators in traditional systems, and achieving smooth downgrade and allocation in the process of fulfilling commercial orders. The system has a built-in risk control and limit mechanism based on the enterprise's business operation rules. While ensuring the business process responds responsively, it avoids business disruptions in the entire factory's logistics and supply chain caused by excessive changes in production plans, and ensures a smooth transition in enterprise management decisions and order fulfillment. The proactive intervention capability allows the system to proactively reduce production takt time, schedule maintenance in advance, and adjust the frequency of quality inspections, so as to adaptively absorb the impact of equipment performance degradation without interrupting production. This not only significantly reduces unplanned downtime but also extends the service life of equipment, truly demonstrating the self-healing capability of a lights-out factory under unattended operation. The MOM system is endowed with proactive business risk mitigation capabilities. By actively optimizing business processes (fine-tuning production scheduling, planning maintenance work orders in advance, and dynamically adjusting warehousing and logistics frequency, etc.), it can adaptively absorb the management risks brought about by capacity degradation without interrupting the company's operations, effectively improving the overall asset utilization rate and operational efficiency of the "lights-out" factory. Attached Figure Description
[0019] Figure 1This is a schematic diagram of an embodiment of the MOM (Multi-Manager) collaborative platform for a lights-out factory based on MEMS technology in this invention. Figure 2 This is a schematic diagram of another embodiment of the MOM collaborative platform for a lights-out factory based on MEMS technology in this invention. Figure 3 A schematic diagram of real-time data acquisition using MEMS sensors (accelerometer, temperature, sound pressure) mounted on a critical device (PECVD); Figure 4 This is a schematic diagram of an embodiment of a MOM collaborative device for a lights-out factory based on MEMS technology, as described in this invention. Detailed Implementation
[0020] This invention provides a MEMS-based MOM (Multi-Manager Manufacturing) collaborative platform for lights-out factories, enabling proactive sensing of equipment degradation in unmanned, lights-out factory scenarios. The terms first, second, third, fourth, etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms include or have, and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device comprising a series of steps or units is not necessarily limited to those explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.
[0021] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the MOM (Multi-Manager) collaborative platform for lights-out factories based on MEMS technology in this invention includes: The first data acquisition module 101 is used to install a MEMS sensor group on key manufacturing equipment to collect vibration signals, temperature signals and acoustic emission signals during equipment operation. The collected signals are filtered, de-DC and feature extracted by sliding window at the edge of the equipment to obtain real-time physical state data, and the baseline physical state data of the equipment under normal operating conditions is determined in advance.
[0022] It is understood that the executing entity of this invention can be a MEMS-based MOM (Multi-Manager Manufacturing) collaborative device for a lights-out factory, or it can be a terminal or a server; the specific implementation is not limited here. This embodiment of the invention will be illustrated using a server as the executing entity.
[0023] It should be noted that the spindle of a 5-axis CNC milling machine in a dark factory was selected as the key manufacturing equipment.
[0024] A set of industrial-grade MEMS sensors is rigidly mounted on the surface of the bearing housing at the front end of the CNC milling machine spindle. This sensor set integrates a 3-axis MEMS accelerometer, a MEMS temperature sensor, and a MEMS acoustic emission sensor.
[0025] When the CNC milling machine starts and performs cutting operations, the sensor group simultaneously collects three kinds of physical signals: the accelerometer collects the spindle vibration signal at a sampling rate of 10,000 Hz; the temperature sensor collects the temperature change signal on the spindle surface at a sampling rate of 10 Hz; and the acoustic emission sensor collects the high-frequency acoustic emission signal generated by internal bearing friction and cutting stress release at a sampling rate of 1,000,000 Hz.
[0026] Before the equipment was officially connected to the MOM system, baseline data under normal operating conditions was measured. The milling machine was set to run continuously for 12 hours under standard load (spindle speed 10000 r / min, constant feed rate cutting aluminum alloy blank). The edge-side computing unit statistically analyzed the data during this period to determine the baseline physical state data of the spindle in a healthy state: the root mean square value of the vibration signal was 0.8 mm / s, the baseline value of the stable operating temperature of the equipment was 45℃, and the baseline value of the peak energy of the acoustic emission signal was 120 mV. This baseline data was stored in the edge gateway.
[0027] In daily production, the real-time raw waveform data collected by the sensors is directly fed into the edge computing gateway attached to the device.
[0028] Filtering and DC removal: The gateway applies a bandpass filtering algorithm of 20Hz~2000Hz to the vibration signal to eliminate low-frequency machine tool base shaking and high-frequency electromagnetic interference. At the same time, it subtracts the arithmetic mean of the sampling sequence to eliminate DC zero-point drift during signal acquisition.
[0029] Sliding window feature extraction was performed: the gateway was set with a time window of 1 second in length and a sliding step of 0.5 seconds. Statistical features were extracted for 10,000 vibration data points, 10 temperature data points, and 1,000,000 acoustic emission data points within each window.
[0030] When the process reaches a certain cutting stage, the edge gateway calculates the following using the current sliding window: the root mean square value of vibration is 1.3 mm / s, the average temperature is 51℃, and the maximum peak value of acoustic emission is 180 mV. This set of data, after dimensionality reduction and feature extraction, constitutes the real-time physical state data.
[0031] The first state assessment module 102 is used to compare real-time physical state data with baseline physical state data to obtain physical deviation. The physical deviation is then subjected to unilateral saturation reduction to obtain a health impact coefficient. The health impact coefficient ranges from 0.6 to 1.0, where 1.0 corresponds to normal equipment operation and 0.6 corresponds to severe equipment degradation.
[0032] It should be noted that the system receives real-time physical state data extracted from the first data acquisition module 101, namely, a root mean square vibration value of 1.3 mm / s, an average temperature of 51℃, and a maximum peak acoustic emission of 180 mV. Simultaneously, the system retrieves pre-measured baseline physical state data, namely, a vibration of 0.8 mm / s, a temperature of 45℃, and an acoustic emission of 120 mV.
[0033] The percentage change of each real-time data point from the baseline data was calculated to obtain the individual physical deviation. The vibration deviation was 62.5%, the temperature deviation was approximately 13.3%, and the acoustic emission deviation was 50%.
[0034] To comprehensively assess the equipment status, the system calculates the overall physical deviation according to preset empirical weights. The weights for vibration characteristics are set at 50%; temperature characteristics at 20%; and acoustic emission characteristics at 30%. Each deviation is multiplied by its corresponding weight and summed to obtain an overall physical deviation of approximately 48.9% for the spindle at that moment.
[0035] After obtaining the comprehensive physical deviation, the system performs a one-sided saturation reduction calculation.
[0036] The one-sided characteristic here means that when the real-time state of the device is better than or equal to the baseline state, that is, when the comprehensive physical deviation is ≤0, the system does not apply a positive premium to the state and directly outputs the health impact coefficient as the upper limit of 1.0.
[0037] The saturation reduction characteristic refers to the system's preset maximum tolerance threshold for comprehensive physical deviation being 100%. The system maps the deviation range of 0% to 100% to a health impact coefficient range of 1.0 to 0.6. When the deviation exceeds 100%, the result saturates, and the health impact coefficient is forcibly capped at the lower limit of 0.6, no longer decreasing.
[0038] The current calculated overall physical deviation is 48.9%. The maximum attenuation span is 0.4. The system multiplies the overall physical deviation by this span to obtain a reduction of approximately 0.2. Subtracting this reduction from the upper limit of 1.0, the current spindle health impact coefficient is calculated to be 0.8.
[0039] The first benchmark determination module 103 is used to preset the correlation weight table between key manufacturing equipment and each functional link of the MOM system. Each functional link includes production planning, equipment scheduling, quality inspection, equipment maintenance, and material distribution. The module weights and normalizes the health impact coefficients according to the correlation weight table to obtain the comprehensive physical impact factor for each functional link. Based on the preset normal setpoints and response sensitivities of each functional link, and combined with the comprehensive physical impact factor, the module determines the physical business scheduling reference benchmark value for each functional link. Specifically, for positive indicators where an increase in value indicates enhanced production or quality capabilities, the physical business scheduling reference benchmark value decreases as the comprehensive physical impact factor decreases; for negative indicators where an increase in value indicates a conservative level of safety, the physical business scheduling reference benchmark value increases as the comprehensive physical impact factor decreases. The physical business scheduling reference benchmark value is then stored in the industrial communication server.
[0040] It should be noted that the spindle health impact coefficient of 0.8 generated by the first state assessment module 102 indicates a 20% performance degradation in the spindle. The system retrieves a pre-defined correlation weight table, which defines the degree of influence of the spindle status on each link of the MOM system: equipment maintenance has a weight of 1.0; equipment scheduling has a weight of 0.9; production planning has a weight of 0.8; quality inspection has a weight of 0.7; and material delivery has a weight of 0.4.
[0041] The 20% performance degradation was multiplied by the weight of each stage and then deducted from the maximum health score of 1.0. The calculations showed that the comprehensive physical impact factor for equipment maintenance was 0.8, for equipment scheduling it was 0.82, for production planning it was 0.84, and for material distribution it remained at 0.92.
[0042] The correction baseline value is calculated based on the preset indicator type, normal setting value, and response sensitivity of each stage.
[0043] Taking the production planning stage as an example, its control indicator is the standard number of cut pieces per hour, which is a positive indicator. The normal set value is 50 pcs / h, and the response sensitivity is 0.5. The comprehensive physical impact factor for this stage is 0.84, representing a 16% decrease from the relatively healthy state. Multiplying the 16% decrease by the sensitivity of 0.5 yields a result indicating an 8% reduction in expected capacity. Therefore, the physical business scheduling reference benchmark decreases as the factor decreases, changing to 46 pcs / h.
[0044] Taking equipment maintenance as an example, the control indicator is the automatic lubrication frequency of the spindle bearing, which is a reverse indicator. The normal set value is 10 times / 100,000 times, and the response sensitivity is 1.0. The impact factor for this step is 0.8, decreasing by 20%. This 20% attenuation is fully compensated for in reverse to safety measures. The physical business scheduling reference benchmark increases as the factor decreases, changing to 12 times / 100,000 times.
[0045] After the calculation is completed, the system writes the correction benchmark values of 46pcs / h in the production planning stage and 12times / 100000r in the equipment maintenance stage into a specific address in the industrial communication server's memory database in the workshop.
[0046] The first correction calculation module 104 is used for each functional link to read the physical service scheduling reference benchmark value from the industrial communication server according to its own business evaluation cycle, and to read the current actual collaborative service plan of the current link; to divide the difference between the current actual collaborative service plan and the physical service scheduling reference benchmark value by the normal setting value to obtain the normalized deviation signal of the current link; to write the normalized deviation signal of the current link along with the timestamp into the industrial communication server; to read the normalized deviation signal and timestamp of the neighboring functional links from the industrial communication server, wherein the neighboring functional links are other functional links that are pre-configured and have a relationship with the current link; and to calculate the information age based on the difference between the current time and the timestamp of the neighboring functional links. The aging attenuation coefficient is used to correct the difference between the normalized deviation signal of the neighboring functional link and the normalized deviation signal of this link, resulting in the aging attenuation deviation difference. Based on the degree of difference in the comprehensive physical influence factors between this link and the neighboring functional links, the collaborative coupling weight between them is adjusted. The aging attenuation deviation difference is multiplied by the collaborative coupling weight and boundedly compressed to obtain the neighbor coordination correction component. The difference between the current actual collaborative business plan and the physical business scheduling reference benchmark is subjected to first-order inertial tracking to obtain the physical drive correction component. The neighbor coordination correction component and the physical drive correction component are superimposed to obtain the global business strategy adjustment amount.
[0047] It should be noted that during the current business evaluation cycle, the production planning stage reads 46 pcs of physical business scheduling reference baseline values generated in the first baseline determination module 103 from the industrial communication server, and reads 48 pcs of the currently executing actual collaborative business plan. The difference between the current business plan and the baseline value is calculated to be 2 pcs. This difference is divided by the normal setting value of 50 pcs in the production planning stage, and the normalized deviation signal of this stage is calculated to be 0.04. Subsequently, this signal, along with the current timestamp (denoted as baseline 0min), is written to the industrial communication server.
[0048] The production planning stage retrieves pre-configured data from neighboring equipment maintenance stages from the server. The normalized deviation signal for the equipment maintenance stage is read as -0.10, and its timestamp is 2 minutes earlier than the current time.
[0049] The information aging and decay coefficient is calculated based on a 2-minute time difference. The system rule is that the information validity decays by 0.05 for every 1-minute delay, with an initial value of 1.0. The currently calculated aging and decay coefficient is 0.9.
[0050] The system subtracts the deviation signal of this stage (0.04) from the deviation signal of the neighbor (-0.10) to obtain a difference of -0.14. This difference is then multiplied and corrected using an attenuation coefficient of 0.9, resulting in an aging-attenuated deviation difference of -0.126. This process reduces asynchronous time difference interference in MOM service flow.
[0051] Extract the comprehensive physical impact factors of the two stages (planning 0.84, maintenance 0.8). Based on the difference of 0.04, dynamically increase the preset initial collaborative coupling weight of 0.4 to 0.5 to enhance the pull-down protection of the maintenance stage with poor health status.
[0052] Multiplying the deviation difference of -0.126 after aging decay by the weight of 0.5 yields -0.063. The system performs a bounded compression check on this value, which does not exceed the compression boundary of ±0.1. The neighbor coordination correction component is confirmed to be -0.063.
[0053] The difference between the actual business plan and the baseline value of 2 pcs in this step is subjected to first-order inertial tracking. The tracking gain coefficient is set to -0.4, and the two are multiplied to obtain the physical drive correction component of -0.8.
[0054] In the production planning stage, the horizontal neighbor coordination correction component of -0.063 and the vertical physical drive correction component of -0.8 are superimposed to obtain a global business strategy adjustment amount of -0.863.
[0055] The first decision execution module 105 is used to constrain the global business strategy adjustment amount by a safety boundary. The safety boundary consists of the minimum and maximum correction limits preset by each functional link. The constrained global business strategy adjustment amount is written as an external correction amount to the industrial communication server. In the next business evaluation cycle, each functional link adds the external correction amount to its original MOM business plan to obtain the revised enterprise operation management plan. The revised enterprise operation management plan updates the operating parameters of each functional link and controls the corresponding execution layer equipment actions. This allows each functional link to adapt to the equipment status degradation with a consistent and limited adjustment range before the equipment alarms itself, realizing the distributed dynamic collaboration of the MOM system in the dark factory.
[0056] It should be noted that each functional component reads the global business strategy adjustment amount generated in the first correction amount calculation module 104 and makes a limit determination based on its own process safety specifications.
[0057] For the production planning stage, the global business strategy adjustment calculated in the previous stage was -0.863 pcs. To prevent excessively large single capacity adjustments from causing disruptions in upstream and downstream logistics in the workshop, the system preset a single-cycle safety boundary for this stage: a minimum correction limit of -0.5 pcs and a maximum correction limit of +0.5 pcs. Since -0.863 exceeded the lower limit, the system clamped it down to -0.5.
[0058] For the equipment maintenance phase, based on the same distributed computing logic, the calculated global business policy adjustment is +1.500 times. The preset safety boundary for this phase is -1.0 times to +1.0 times. After thresholding, the adjustment amount for this phase is truncated to +1.0.
[0059] The system then uses these two clamped data points as external correction values and writes them to the industrial communication server.
[0060] Upon entering the next business evaluation cycle, each functional unit reads its own external adjustment amount from the server and directly adds it to the original MOM business plan of that unit. Specific data changes are shown in Table 1 below: Table 1 Functional components Global business strategy adjustment amount Single-cycle safety limit External correction amount Existing business plan Revised Enterprise Operation and Management Plan Production planning stage -0.863pcs [-0.5, +0.5]pcs -0.5pcs 48.0 pcs / h 47.5 pcs / h Equipment maintenance +1.500 times [-1.0,+1.0]times +1.0times 11.0 times / 100000r 12.0 times / 100000r The system uses the revised enterprise operation management plan in Table 1 to overwrite the operation control parameters of the underlying equipment.
[0061] After receiving a new target of 47.5 pcs / h in the production planning stage, the CNC control system automatically fine-tunes the feed rate and spindle speed of the machining program, so that the machining pace of the milling machine is moderately slowed down. At the same time, after receiving a new instruction of 12 times / 100000r in the equipment maintenance stage, the parameters of the automatic lubrication controller are updated accordingly, increasing the frequency of injecting lubricating oil into the spindle bearing.
[0062] Through these actions, the system autonomously reduces equipment operating load and enhances maintenance efforts before spindle wear triggers a shutdown alarm. Adjustments are strictly limited to a single-cycle safety margin, allowing all manufacturing processes in the lights-out factory to smoothly adapt to equipment degradation and achieving distributed dynamic collaboration.
[0063] Please see Figure 2 and Figure 3 Another embodiment of the MOM (Multi-Manager) collaborative platform for a lights-out factory based on MEMS technology in this invention includes: The second data acquisition module 201 is used to install a MEMS sensor group on the key manufacturing equipment to collect vibration signals, temperature signals and acoustic emission signals during equipment operation. The acquired signals are filtered, de-DC and feature extracted by sliding window at the edge of the equipment to obtain real-time physical state data, and the baseline physical state data of the equipment under normal operating conditions is determined in advance.
[0064] Specifically, the MEMS sensor array includes a MEMS triaxial accelerometer, a MEMS temperature sensor, and a MEMS acoustic pressure sensor. The MEMS triaxial accelerometer is installed on the bottom flange of the equipment's process chamber, with a range of ±16g and a bandwidth of 2kHz, acquiring vibration signals at a sampling rate of 2000Hz. The MEMS temperature sensor is installed on the outer wall of the equipment's cooling pipes, acquiring temperature signals at a sampling rate of 1Hz. The MEMS acoustic pressure sensor is installed on the outer wall of the equipment's chamber, with a frequency response range of 20Hz to 10kHz, acquiring acoustic emission signals at a sampling rate of 10000Hz. The acquired signals are pre-processed at the equipment's edge. During processing, the DC component of the vibration signal is removed, and the effective vibration value is calculated with a window length of 256 sampling points and a sliding step size of 128 sampling points. The temperature signal is filtered by moving average to obtain the average temperature value. The number of pulses exceeding the preset amplitude threshold in the acoustic emission signal is counted to obtain the acoustic emission pulse count rate. The effective vibration value, average temperature value, and acoustic emission pulse count rate are used as real-time physical state data. After the equipment passes acceptance, the equipment is run continuously and normally for 24 hours. The baseline average values of the effective vibration value, average temperature value, and average acoustic emission pulse count rate during these 24 hours are obtained using the same preprocessing method and used as the baseline physical state data.
[0065] It should be noted that a PECVD machine in the darkroom of the solar cell factory was selected as the key manufacturing equipment, and collaborative calculations were performed in conjunction with the screen printing equipment and its corresponding MOM business function.
[0066] A MEMS 3-axis accelerometer with a range of ±16g and a bandwidth of 2kHz is rigidly installed at the bottom flange of the process chamber of the PECVD equipment to acquire vibration signals at a sampling rate of 2000Hz. A MEMS temperature sensor is installed on the outer wall of the cooling pipes of the equipment to acquire temperature signals at a sampling rate of 1Hz. A MEMS acoustic pressure sensor with a response range of 20Hz to 10kHz is installed on the outer wall of the chamber to acquire acoustic emission signals at a sampling rate of 10000Hz.
[0067] Sensor signals are transmitted to the edge computing gateway. The gateway removes the DC component of the vibration signal, calculates the effective vibration value using 256 sampling points as a window and 128 points as a step size; it performs a moving average filter on the temperature signal to obtain the average temperature; and it obtains the pulse count rate from the acoustic emission waveforms with amplitude exceeding the limit. These three data points together constitute the real-time physical state data.
[0068] After the equipment passes acceptance testing, it operates continuously for 24 hours under standard process load. The edge gateway collects data from these 24 hours using the same preprocessing method and archives the baseline physical state data of the PECVD equipment: the baseline mean value of vibration RMS is 0.15g, the baseline mean value of temperature is 40℃, and the baseline mean value of acoustic emission pulse count rate is 20 times / s. This baseline data is stored in the edge-side storage area.
[0069] The second state assessment module 202 is used to compare real-time physical state data with baseline physical state data to obtain physical deviation. The physical deviation is then subjected to unilateral saturation reduction to obtain the health impact coefficient. The health impact coefficient ranges from 0.6 to 1.0, where 1.0 corresponds to normal equipment operation and 0.6 corresponds to severe equipment degradation.
[0070] Specifically, the effective vibration value, average temperature value, and acoustic emission pulse count rate in the real-time physical state data are compared with the corresponding baseline average values of the effective vibration value, average temperature value, and average acoustic emission pulse count rate in the baseline physical state data, respectively, to obtain the vibration deviation ratio, temperature deviation ratio, and acoustic emission deviation ratio. A first weighting coefficient is assigned to the vibration deviation ratio, a second weighting coefficient to the temperature deviation ratio, and a third weighting coefficient to the acoustic emission deviation ratio. The three deviation ratios after weighting are added together to obtain the physical deviation degree; the first weighting coefficient is 1, and the second weighting coefficient is 0. The third weighting coefficient is 0.3; the health impact coefficient is determined based on the physical deviation: when the physical deviation is less than or equal to the normal threshold, the health impact coefficient takes the normal state value; when the physical deviation is greater than the normal threshold but less than the degradation threshold, the health impact coefficient decreases linearly from the normal state value to the lower limit protection value as the physical deviation increases; when the physical deviation is greater than or equal to the degradation threshold, the health impact coefficient takes the lower limit protection value; the normal threshold is 1.0, the degradation threshold is 1.5, the normal state value is 1.0, and the lower limit protection value is 0.6; the health impact coefficient is used as the input for weighted normalization.
[0071] Furthermore, before weighted summing the vibration deviation ratio, temperature deviation ratio, and acoustic emission deviation ratio, a consistency check is performed on the three deviation ratios. The consistency check includes: comparing the absolute values of the differences between each pair of the three deviation ratios with preset deviation thresholds; if there exists one and only one deviation ratio whose absolute value of the difference with the other two deviation ratios exceeds the preset deviation threshold, then the sensor data corresponding to that deviation ratio is deemed abnormal, and that deviation ratio is discarded; if the absolute value of the difference between any two of the three deviation ratios does not exceed the preset deviation threshold, then all three deviation ratios are deemed valid, and all deviation ratios are retained; if two or three deviation ratios satisfy the condition that the absolute value of the difference with their respective corresponding other deviation ratios exceeds the preset deviation threshold, then all three deviation ratios are deemed invalid. The system retains all deviation ratios and generates a physical deviation using a conservative approach corresponding to a preset deviation threshold. When a deviation ratio is removed, its original weight coefficient is redistributed to the remaining two deviation ratios according to the proportions of the weight coefficients corresponding to the remaining two deviation ratios, resulting in adjusted weight coefficients. The remaining two deviation ratios are multiplied by the adjusted weight coefficients, summed, and then divided by the sum of the adjusted weight coefficients to obtain the corrected physical deviation, which is used as the physical deviation. When all deviation ratios are retained, a weighted sum is performed according to the first, second, and third weight coefficients to obtain the physical deviation. The physical deviation is used to compare with the normal threshold and the degradation threshold to determine the health impact coefficient, which serves as the input for weighted normalization.
[0072] It should be noted that the system reads the current real-time physical state data of the PECVD equipment: the effective vibration value is 0.315g, the average temperature is 84℃, and the acoustic emission pulse count rate is 22 times / s. The system extracts the benchmark data measured by the second data acquisition module 201 and calculates the deviation ratios: the vibration deviation ratio is 1.1, the temperature deviation ratio is 1.1, and the acoustic emission deviation ratio is 0.1.
[0073] Consistency checks were performed on the three deviation ratios, calculating the absolute values of the pairwise differences. The difference between vibration and temperature was 0; the difference between vibration and acoustic emission was 1.0; and the difference between temperature and acoustic emission was 1.0. The preset deviation threshold was 0.3. Comparison revealed that only the acoustic emission item's difference (all 1.0) exceeded the limit. The system determined the acoustic emission sensor data to be abnormal and removed that deviation ratio.
[0074] The first weighting factor (vibration) is set to 1, the second (temperature) to 0.2, and the third (acoustic emission) to 0.3. Since the acoustic emission item is removed, its 0.3 weight is redistributed to the remaining two items at a ratio of 1:0.2. After adjustment, the vibration weight becomes 1.25, the temperature weight becomes 0.25, and the sum is 1.5.
[0075] The system multiplies the retained deviation ratio by the new weights: 1.1 × 1.25 + 1.1 × 0.25 = 1.65. Dividing by the weights and 1.5, the corrected physical deviation is calculated to be 1.1.
[0076] The normal threshold is 1.0, the degradation threshold is 1.5, the normal state value is 1.0, and the lower limit protection value is 0.6. The current deviation of 1.1 is within the range. A span of 0.5 corresponds to a decrease of 0.4, and the current exceedance of 0.1 corresponds to a decrease of 0.08. Subtracting 0.08 from the normal state value of 1.0 outputs a health impact coefficient of 0.92 for this PECVD equipment.
[0077] The second benchmark determination module 203 is used to preset the correlation weight table between key manufacturing equipment and each functional link of the MOM system. Each functional link includes production planning, equipment scheduling, quality inspection, equipment maintenance, and material distribution. The module weights and normalizes the health impact coefficients according to the correlation weight table to obtain the comprehensive physical impact factor for each functional link. Based on the preset normal setpoints and response sensitivities of each functional link, and combined with the comprehensive physical impact factor, the module determines the physical business scheduling reference benchmark value for each functional link. Specifically, for positive indicators where an increase in value indicates enhanced production or quality capabilities, the physical business scheduling reference benchmark value decreases as the comprehensive physical impact factor decreases; for negative indicators where an increase in value indicates a conservative level of safety, the physical business scheduling reference benchmark value increases as the comprehensive physical impact factor decreases. The physical business scheduling reference benchmark value is then stored in the industrial communication server.
[0078] Specifically, a correlation weight table is pre-established between key manufacturing equipment and each functional link of the MOM system. Each key manufacturing equipment in the correlation weight table is assigned a weight value for each functional link. The weight value represents the degree to which the health status of the key manufacturing equipment affects the decision-making objective of the functional link, and the weight value ranges from 0 to 1. Each functional link includes production planning, equipment scheduling, quality inspection, equipment maintenance, and material distribution. For each functional link, the sum of all weight values corresponding to that functional link in the correlation weight table is used as the normalized denominator. The health impact coefficients of all key manufacturing equipment associated with that functional link are multiplied by their respective weight values and summed. The result is divided by the normalized denominator to obtain the comprehensive physical impact factor of that functional link, which ranges from 0.6 to 1.0. The comprehensive physical impact factor of all functional links serves as the input for subsequent adjustments to the collaborative coupling weights between functional links. Each functional stage has a pre-defined normal setting value and response sensitivity. The normal setting value is the decision target value adopted by the corresponding functional stage when all equipment is in normal condition. The response sensitivity is the adjustment range of the decision target value of the corresponding functional stage as the comprehensive physical influence factor changes. For each functional stage, the difference between the comprehensive physical influence factor of the stage and the normal state value of 1.0 is multiplied by the response sensitivity of the stage to obtain the physical reference correction amount for that stage. When the decision target value of the functional stage is a positive indicator, the normal setting value and the physical reference correction amount are added to obtain the physical business scheduling reference benchmark value for that functional stage. When the decision target value of the functional stage is a negative indicator, the normal setting value and the physical reference correction amount are subtracted to obtain the physical business scheduling reference benchmark value for that functional stage. A positive indicator is one indicative of increased production or quality capability, while a negative indicator is one indicative of increased conservative safety level. The physical business scheduling reference benchmark values of each functional stage are stored in an industrial communication server. Each functional stage reads the corresponding physical business scheduling reference benchmark value from the industrial communication server according to its own business evaluation cycle.
[0079] Furthermore, when the sum of all weight values corresponding to a certain functional link in the association weight table is zero, it is determined that the functional link is not constrained by the health status of any key manufacturing equipment. The comprehensive physical influence factor of the functional link is directly set to the normal state value, and the physical business scheduling reference benchmark value of the functional link is set to be equal to its normal setting value. The comprehensive physical influence factor serves as the basis for adjusting the collaborative coupling weight between the functional link and other functional links, and the physical business scheduling reference benchmark value serves as the benchmark for calculating the normalized deviation signal of the functional link.
[0080] It should be noted that currently operating screen printing equipment is included (health impact coefficient 1.0). From the correlation weight table, the weight of the production planning stage for PECVD is 0.8, and the weight for screen printing is 0.2 (normalized denominator 1.0). The calculated comprehensive physical impact factor for the production planning stage is 0.936.
[0081] The weight of equipment maintenance for PECVD is 0.9, and the weight for screen printing is 0.1. The overall physical influence factor is calculated to be 0.928.
[0082] The decision target in the production planning stage is the number of silicon wafers coated per hour, which is a positive indicator. The normal setpoint is 5000 pcs / h, and the response sensitivity is strictly set to 1.5 due to the legal range constraints. The difference between the comprehensive factor of 0.936 and 1.0 in this stage is -0.064. Multiplying this by the sensitivity of 1.5 yields a physical reference correction of -0.096. Adding -0.096 to the normal setpoint of 5000 gives a correction baseline value of 4999.904 pcs / h.
[0083] The target for equipment maintenance is the monthly maintenance frequency of the vacuum pump, which is a reverse indicator. The normal setpoint is 4 times / month, and the sensitivity is strictly set to 1.0. The difference between the comprehensive factor 0.928 and 1.0 is -0.072. Multiplying this by the sensitivity of 1.0 gives a correction of -0.072. Subtracting -0.072 from the normal setpoint of 4 gives a corrected baseline value of 4.072 times / month.
[0084] The system stores baseline values such as 4999.904 pcs / h in the production planning stage and 4.072 times / month in the equipment maintenance stage into the industrial communication server for backup.
[0085] The second correction calculation module 204 is used for each functional link to read the physical service scheduling reference benchmark value from the industrial communication server according to its own business evaluation cycle, and to read the current actual collaborative service plan of the current link; to divide the difference between the current actual collaborative service plan and the physical service scheduling reference benchmark value by the normal setting value to obtain the normalized deviation signal of the current link; to write the normalized deviation signal of the current link along with the timestamp into the industrial communication server; to read the normalized deviation signal and timestamp of the neighboring functional links from the industrial communication server, wherein the neighboring functional links are other functional links that are pre-configured and have a relationship with the current link; and to calculate the information age based on the difference between the current time and the timestamp of the neighboring functional links. The aging attenuation coefficient is used to correct the difference between the normalized deviation signal of the neighboring functional link and the normalized deviation signal of this link, resulting in the aging attenuation deviation difference. Based on the degree of difference in the comprehensive physical influence factors between this link and the neighboring functional links, the collaborative coupling weight between them is adjusted. The aging attenuation deviation difference is multiplied by the collaborative coupling weight and boundedly compressed to obtain the neighbor coordination correction component. The difference between the current actual collaborative business plan and the physical business scheduling reference benchmark is subjected to first-order inertial tracking to obtain the physical drive correction component. The neighbor coordination correction component and the physical drive correction component are superimposed to obtain the global business strategy adjustment amount.
[0086] Specifically, each functional link reads its physical business scheduling reference value from the industrial communication server according to its own business evaluation cycle, and reads its current actual collaborative business plan. The difference between the current actual collaborative business plan and the physical business scheduling reference value is divided by the normalized set value to obtain the normalized deviation signal. The normalized deviation signal is a signed value; a positive deviation indicates that the current actual collaborative business plan is biased towards the risk side relative to the physical business scheduling reference value, and a negative deviation indicates that it is biased towards the conservative side. The normalized deviation signal of this link, along with its timestamp, is written to the industrial communication server. The normalized deviation signal and its timestamp of the neighboring functional link are read from the industrial communication server, and the information age is determined based on the difference between the current time and the timestamp of the neighboring functional link. A preset information validity period is set. When the information age is greater than the information validity period, the information of the neighboring functional link is judged to be expired, and the corresponding aging attenuation coefficient is set to zero. When the information age is less than or equal to the information validity period, the aging attenuation coefficient is calculated according to the preset attenuation rate, so that the aging attenuation coefficient decreases as the information age increases. The aging attenuation coefficient is multiplied by the difference between the normalized deviation signal of the neighboring functional link and the normalized deviation signal of the current link to obtain the deviation difference after aging attenuation. A basic cooperative coupling weight and a lower limit for the coupling weight are preset. The reduction amount is calculated based on the absolute value of the difference between the comprehensive physical influence factor of the current link and the comprehensive physical influence factor of the neighboring functional links; the reduction amount increases with the increase of the absolute value. The reduction amount is subtracted from the basic cooperative coupling weight to obtain the dynamically adjusted cooperative coupling weight, which is then limited to not less than the lower limit of the coupling weight. The deviation difference after aging attenuation is multiplied by the dynamically adjusted cooperative coupling weight to obtain the weighted deviation difference. The weighted deviation difference is then compressed using hyperbolic tangent compression. The system maps the weighted deviation difference of any value to an output value between -1 and 1, and uses this output value as the neighbor coordination correction component. It presets an inertia coefficient and a tracking ratio coefficient, and adjusts the physical drive correction component in the current cycle relative to the previous cycle towards the tracking input value using the tracking ratio coefficient, while simultaneously suppressing the adjustment rate with the inertia coefficient. The tracking input value is the difference between the current actual collaborative business plan and the physical business scheduling reference value in this stage, thus obtaining the physical drive correction component. The physical drive correction component gradually approaches the tracking input value. The neighbor coordination correction component and the physical drive correction component are superimposed to obtain the global business strategy adjustment amount. The global business strategy adjustment amount serves as the input for security boundary clamping.
[0087] Furthermore, the basic collaborative coupling weight, coupling weight lower limit, physical difference threshold, and maximum reduction ratio are preset. The specific process for calculating the reduction amount is as follows: First, calculate the absolute value of the difference between the comprehensive physical influence factor of this stage and the comprehensive physical influence factor of neighboring functional stages to obtain the absolute value of the physical difference; compare the absolute value of the physical difference with the physical difference threshold; if the absolute value of the physical difference is less than or equal to the physical difference threshold, the reduction amount is set to zero; if the absolute value of the physical difference is greater than the physical difference threshold, the reduction amount is made to increase linearly with the increase of the difference between the absolute value of the physical difference and the physical difference threshold, and the reduction amount does not exceed the maximum reduction ratio; after obtaining the reduction amount, subtract the reduction amount from the basic collaborative coupling weight to obtain the adjusted weight. The adjusted candidate values for the collaborative coupling weight are compared with the lower limit of the coupling weight. If the candidate value is less than the lower limit, the lower limit is used as the dynamically adjusted collaborative coupling weight. If the candidate value is greater than or equal to the lower limit, the candidate value is used as the dynamically adjusted collaborative coupling weight. The dynamically adjusted collaborative coupling weight is multiplied by the deviation difference after aging and decay to obtain the weighted deviation difference. The weighted deviation difference is compressed, and the compressed output value is used as the neighbor coordination correction component. The neighbor coordination correction component is superimposed with the physical drive correction component to obtain the global business policy adjustment amount. The global business policy adjustment amount is used as the input for security boundary clamping.
[0088] It should be noted that the production planning stage reads the correction baseline value of 4999.904 pcs / h for this stage, and simultaneously reads the current actual collaborative business plan of 4980 pcs / h for this stage. The difference is calculated to be -19.904 pcs / h, which, divided by the normal setting value of 5000 pcs / h, yields a normalized deviation signal of -0.00398 for this stage. This signal, along with a timestamp, is written to the server.
[0089] The extracted normalized deviation signal from the neighboring equipment maintenance process is -0.08, and the time age is... The effective period of the information is 60 seconds. The time is 180 seconds. To ensure smooth decay, the information aging decay coefficient is calculated using the following formula. : ; The calculated attenuation coefficient is 0.866. Subtracting the deviation of this stage from the neighbor deviation of -0.08 gives a difference of -0.07602. Multiplying by the coefficient 0.866, the deviation difference after aging attenuation is -0.0658.
[0090] The absolute value of the difference between the two influencing factors is 0.008. The preset threshold is 0.010. Since 0.008 ≤ 0.010, the reduction is set to 0. The base weight of 0.6 is reduced by 0 and compared with the lower limit of 0.2, resulting in a weight of 0.6. The weighted deviation difference is -0.0658 × 0.6 = -0.0395.
[0091] Hyperbolic tangent compression is applied to this value, resulting in an output of -0.0395. The system directly assigns this value to operational dimensions, with the neighbor coordination correction component being -0.0395 pcs / h, to achieve sub-unit level flexible and smooth fine-tuning.
[0092] First-order inertial tracking was performed on the actual difference of -19.904. The inertial coefficient was set to 0.1, and the tracking proportional coefficient was set to -0.2. The physical drive correction component was calculated as: 0.1×0+(-0.2)×(-19.904)=3.9808pcs / h.
[0093] The neighbor coordination correction component (-0.0395 pcs / h) and the physical drive correction component (3.9808 pcs / h) are superimposed to obtain a global business policy adjustment amount of 3.9413 pcs / h.
[0094] The second decision execution module 205 is used to constrain the global business strategy adjustment amount by a safety boundary. The safety boundary consists of the minimum and maximum correction limits preset by each functional link. The constrained global business strategy adjustment amount is written as an external correction amount to the industrial communication server. In the next business evaluation cycle, each functional link adds the external correction amount to its original MOM business plan to obtain the revised enterprise operation management plan. The revised enterprise operation management plan updates the operating parameters of each functional link and controls the corresponding execution layer equipment actions. This allows each functional link to adapt to equipment status degradation with a consistent and limited adjustment range before the equipment alarms itself, realizing the distributed dynamic collaboration of the MOM system in the dark factory.
[0095] Specifically, minimum and maximum correction limits are preset for each functional link. The minimum correction limit is a negative value, and the maximum correction limit is a positive value. The minimum and maximum correction limits together define the allowable adjustment range of the external correction amount for each functional link. The global business strategy adjustment amount is compared with the minimum and maximum correction limits: if the global business strategy adjustment amount is greater than the maximum correction limit, the maximum correction limit is taken as the external correction amount; if the global business strategy adjustment amount is less than the minimum correction limit, the minimum correction limit is taken as the external correction amount; if the global business strategy adjustment amount is greater than or equal to the minimum correction limit and less than or equal to the maximum correction limit, the global business strategy adjustment amount remains unchanged and is directly used as the external correction amount; the external correction amount, along with the identifier of this link and the generation timestamp, is written to the industrial communication server. At the start of the next business evaluation cycle, each functional segment reads the corresponding external correction amount and corresponding generation timestamp from the industrial communication server according to its segment identifier; calculates the time difference between the current time and the generation timestamp; if the time difference is greater than the preset allowed storage time, the external correction amount is determined to be expired data, discarded, and the enterprise operation management plan of the previous business evaluation cycle of this segment is maintained as the corrected enterprise operation management plan for the current cycle; if the time difference is less than or equal to the allowed storage time, it is further determined whether the value of the external correction amount is within the closed interval formed by the minimum correction limit and the maximum correction limit; if it is within the closed interval, the external correction amount is superimposed on the original MOM business plan of this segment to obtain the corrected enterprise operation management plan. If the external correction amount is not within the closed interval, it is determined to be abnormal data, discarded, and the enterprise operation management plan of the previous business evaluation cycle is maintained as the corrected enterprise operation management plan for the current cycle. If the external correction amount cannot be read due to communication failure, the most recently successfully read external correction amount is read as the replacement value, and it is determined whether the replacement value is still within the allowed storage time. If it is, the replacement value is added to the original MOM business plan of this cycle to obtain the corrected enterprise operation management plan for the current cycle. If it is not, zero value is used as the replacement value, and the replacement value is added to the original MOM business plan of this cycle to obtain the corrected enterprise operation management plan for the current cycle. The revised enterprise operation management plan is written into the parameter configuration storage area of the MOM system, replacing the corresponding decision items in the original operating parameters of this stage, forming updated operating parameters. Control commands are generated based on the updated operating parameters and sent to the corresponding execution layer devices, driving the execution layer devices to operate according to the updated operating parameters. Among them, the execution layer device corresponding to the production planning stage is the planning management host, the execution layer device corresponding to the equipment scheduling stage is the process equipment controller, the execution layer device corresponding to the quality inspection stage is the online detection device, the execution layer device corresponding to the equipment maintenance stage is the maintenance alarm device, and the execution layer device corresponding to the material distribution stage is the AGV scheduling system. Each functional stage continues to execute in the next business evaluation cycle, so that each functional stage adapts to the equipment status degradation with a consistent and limited adjustment range before the equipment alarms itself, and keeps the MOM system operating parameters within the safety boundary, realizing the distributed dynamic collaboration of the MOM system in the lights-out factory.
[0096] It should be noted that each functional stage imposes safety boundary constraints on the adjustment amount of the global business strategy. The production planning stage has a preset minimum correction limit of -50 pcs / h and a maximum correction limit of +50 pcs / h per cycle. The global business strategy adjustment amount calculated by this stage, 3.9413 pcs / h, falls within this allowable adjustment range, so the system keeps the value unchanged and directly stores it as the external correction amount on the server. Similarly, the global business strategy adjustment amount for the equipment maintenance stage is calculated to be +0.8 times / month, within the limit of [-1.0, +1.0], therefore the external correction amount is +0.8 times / month.
[0097] At the start of the next business evaluation cycle, the functional team reads the external correction amount and timestamp. The time difference is calculated to confirm it hasn't expired and that the value is within a closed interval. It is then overlaid onto the original MOM business plan. The data changes are shown in Table 2 below: Table 2 Functional components Global business strategy adjustment amount Single-cycle safety limit External correction amount Existing MOM business plan Revised Enterprise Operation and Management Plan Production planning stage +3.9413pcs / h [-50,+50]pcs / h +3.9413pcs / h 4980.0000pcs / h 4983.9413pcs / h Equipment maintenance +0.8000 times / month [-1.0, +1.0] times / month +0.8000 times / month 5,000 times / month 5,800 times / month The system replaces the original operating parameters with the revised enterprise operation management plan and issues control commands. The planning management host controls the AGV and film feeding cycle based on the target of 4983.9413 pcs / h; the maintenance alarm device adjusts its monitoring logic based on the target of 5.8 times / month. Consistent and restricted dynamic collaboration is achieved among all links during the equipment degradation period.
[0098] The parameter initialization and calibration module 206 is used for the parameter initialization and calibration process. The parameter initialization and calibration process includes: performing baseline physical state calibration for each key manufacturing equipment: After the equipment passes acceptance testing, it is continuously operated with standard process loads for no less than 24 hours. The output signals of the MEMS sensor array during this period are collected. After edge-side preprocessing similar to that used when collecting real-time physical state data, the statistical average values of the vibration effective value, temperature average, and acoustic emission pulse count rate are calculated respectively. These three statistical average values are used as the baseline physical state data for the equipment and stored in the edge-side storage area for comparison with real-time physical state data. Status data is retrieved and used when comparing; an association weight table is established: a review group composed of process engineers, equipment engineers and production planners scores the dependence of each key manufacturing equipment on the production planning, equipment scheduling, quality inspection, equipment maintenance and material distribution links item by item. The score range is 0 to 1, where 0 indicates that the health status of the equipment has no impact on the functional link and 1 indicates that the health status of the equipment has a decisive impact on the functional link. All score results are filled into the association weight table to form the initial weight value, which is stored in the industrial communication server for retrieval when weighting and normalizing the health impact coefficient according to the association weight table; Set the collaborative operation parameters for each functional link: For each functional link, set the normal setpoint, response sensitivity, minimum correction limit, maximum correction limit, tracking ratio coefficient, inertia coefficient, information validity period, and allowed storage duration. The normal setpoint is the historical average of the decision target value of this functional link when all devices are in normal condition. The initial value for response sensitivity is 1.0 to 1.5. The minimum and maximum correction limits are set according to the maximum fluctuation range of the business plan under historical abnormal operating conditions, with the maximum correction limit being positive and the minimum correction limit being negative. The initial value for tracking ratio coefficient is 0.1 to 0.5; the initial value for inertia coefficient is 0.1 to 0.3; the initial value for information validity period is 30 seconds to 300 seconds; and the initial value for allowed storage duration is 3 to 10 business evaluation cycles. Configure a list of neighboring functional links for each functional link. This list is used to identify neighboring functional links that are related to this functional link. All the parameters set above are stored in the industrial communication server for retrieval when determining the physical service scheduling reference value, calculating the normalized deviation signal and physical drive correction component, and identifying expired and abnormal data. The initial state settings of each functional link are completed: the current actual collaborative service plan of each functional link is initialized to be consistent with its normal setting value; the initial value of the physical drive correction component of each functional link is set to zero; the initial value of the global service strategy adjustment amount of each functional link in the previous service evaluation cycle is set to zero; all normalized deviation signals, timestamps and external correction amounts stored in the industrial communication server are cleared, the initialization is completed, and then the MEMS technology-based black-light factory MOM collaborative platform is executed.
[0099] Furthermore, the process of establishing the correlation weight table specifically includes: collecting abnormal event data recorded by each key manufacturing equipment during historical operation. The abnormal event data includes the time of occurrence of the abnormal event and the records of changes in the business plans of each functional link before and after the occurrence of the abnormal event. Extracting the response time interval between the time of occurrence of the abnormality of the equipment and the time of the first change in the business plan of the corresponding functional link from each abnormal event data. The shorter the response time interval, the higher the dependence strength between the equipment and the functional link. Converting the response time interval into a dependency strength calculation value according to a preset decreasing correspondence relationship. The decreasing correspondence relationship is set so that the dependency strength calculation value decreases as the response time interval increases. The dependency strength calculation value is normalized to the range of 0 to 1 to obtain the normalized dependency strength calculation value. The normalized dependency strength calculation value and the initial weight value obtained by the review group for item-by-item scoring are weighted and merged according to a preset ratio to obtain the merged weight value. The preset ratio is initially set to 50% each, and is adjusted by the review group after review based on the actual fault response during equipment operation. The review group reviews the merged weight value and corrects the merged weight value that is obviously inconsistent with the process understanding. The corrected merged weight value is filled into the correlation weight table as the correlation weight value and stored in the industrial communication server for use when weighting and normalizing the health impact coefficient according to the correlation weight table to generate the comprehensive physical impact factor of each functional link.
[0100] It should be noted that after the PECVD equipment passed acceptance testing, it was run continuously for 24 hours under process load, and the collected signals were preprocessed. Baseline physical state data were calculated: effective vibration value 0.15g, average temperature 40℃, and acoustic emission pulse count rate 20 times / s. This data was then stored in the edge-side storage area for later use.
[0101] The response time intervals from historical anomaly data are extracted and normalized to obtain the dependency strength. These are then scored by human experts from the review panel. The normalized dependency strength and the expert scores are weighted at a 50% ratio and merged, then reviewed. The merged dependency weight value is 0.9 for equipment maintenance and 0.8 for production planning. These determined weights are entered into a dependency weight table and stored on the server.
[0102] To ensure the algorithm logic and variable range are legal and consistent, the initial parameters for each step are strictly set according to the defined range, as shown in Table 3 below: Table 3 Functional components Normal setting value Response sensitivity Minimum correction limit Maximum correction limit Tracking ratio coefficient coefficient of inertia Information validity period Allowed storage duration Production planning stage 5000pcs / h 1.5 -50pcs / h +50pcs / h 0.2 0.1 180s 3 cycles Equipment maintenance 4 times / month 1.0 -1.0 times / month +1.0 times / month 0.3 0.2 180s 3 cycles Material distribution process 20times / shift 1.2 -2.0 times / shift +2.0 times / shift 0.1 0.1 180s 3 cycles After configuring the neighbor function list, save all the above parameters to the industrial communication server.
[0103] The system initializes the current actual collaborative business plan for each stage to be consistent with the normal settings. It clears the physical drive correction components of all stages and the initial values of the global business strategy adjustment amounts from the previous cycle, and removes all normalized deviation signals, timestamps, and external correction amounts from the server.
[0104] Figure 4 This is a schematic diagram of a MEMS-based MOM (Multi-Manager Manufacturing) collaborative device for a lights-out factory, provided in an embodiment of the present invention. The MEMS-based MOM collaborative device 300 may include a processor 301 and a memory 302. The memory 302 is used to store program instructions and / or data, and the processor 301 is used to execute the program instructions stored in the memory 302, thereby implementing the method described in the above embodiment.
[0105] Optionally, the memory 302 and the processor 301 are coupled. The coupling is an indirect coupling or communication connection between devices, units, or modules, and can be electrical, mechanical, or other forms, for information interaction between devices, units, or modules.
[0106] Optionally, the MEMS-based MOM (Multi-Manager Manufacturing) collaborative device 300 for a lights-out factory may further include a communication interface 303. The communication interface 303 is used to communicate with other devices via a transmission medium, for example, transmitting received signals from other communication devices to the processor 301, or transmitting signals from the processor 301 to other communication devices. The communication interface 303 can be a transceiver or an interface circuit, such as a transceiver circuit or transceiver chip.
[0107] This application embodiment does not limit the specific connection medium between the processor 301, memory 302, and communication interface 303. This application embodiment... Figure 4 The processor 301, memory 302, and communication interface 303 are connected via a bus 304. Figure 4 The connections between other components are shown in bold and are for illustrative purposes only, not as limiting information. The bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.
[0108] The present invention also provides a MEMS-based MOM (Multi-Manager Manufacturing) collaborative device for a lights-out factory. The MEMS-based MOM collaborative device includes a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor performs the steps of the MEMS-based MOM collaborative platform in the above embodiments.
[0109] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the MEMS-based light-out factory MOM collaborative platform.
[0110] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0111] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0112] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A MOM (Multi-Manager) collaborative platform for lights-out factories based on MEMS technology, characterized in that: include: The data acquisition module is used to collect the operating status signals of the manufacturing equipment to extract real-time status data, and to acquire the baseline status data of the manufacturing equipment under normal operating conditions. The status assessment module is used to compare the real-time status data with the baseline status data to determine the health status indicators of the manufacturing equipment. The benchmark determination module is used to convert the health status indicators into comprehensive influence factors of each functional link based on the correlation weights between the manufacturing equipment and each functional link of the manufacturing operation management system, and to determine the business scheduling reference benchmark value of each functional link in combination with preset initial setting parameters. The adjustment calculation module is used to enable each functional link to calculate its own deviation signal based on its corresponding actual business plan and the business scheduling reference benchmark value, and to obtain the neighbor deviation signal of the neighboring functional link; and to calculate the global business strategy adjustment amount based on the self-deviation signal, the neighbor deviation signal, the actual business plan and the business scheduling reference benchmark value. The decision execution module is used to perform boundary limiting processing on the global business strategy adjustment amount to obtain the final business scheduling parameters, and use the final business scheduling parameters to update the original business plans of each functional link to obtain the actual execution enterprise operation management plan.
2. The MEMS-based MOM (Multi-Manager Manufacturing) collaborative platform for a lights-out factory as described in claim 1, characterized in that, It also includes a parameter initialization and calibration module: The manufacturing equipment is operated under preset benchmark conditions, the operating status signals are collected and statistical feature values are calculated to obtain the benchmark status data; Assess the dependence strength between the manufacturing equipment and each functional link, generate corresponding initial weight values, and establish an association weight table to provide the association weights; Each functional component is assigned a collaborative operation parameter, which includes at least the initial setting parameter. The actual business plan of each functional link is initialized to be consistent with the normal setting value in the initial setting parameter, and the initial value of the calculation item associated with the global business strategy adjustment amount is set to a preset zero value.
3. The MEMS-based MOM (Multi-Manager Manufacturing) collaborative platform for a lights-out factory as described in claim 2, characterized in that, The parameter initialization and calibration module establishes an associated weight table, specifically used for: Acquire historical abnormal event data of the manufacturing equipment, and extract the response time interval between the occurrence of an abnormality in the manufacturing equipment and the first change in the business plan of the corresponding functional link; The dependency strength is determined based on the response time interval, and the dependency strength decreases as the response time interval increases. The calculated dependency strength value is weighted and fused with the initial weight value obtained from the score to obtain a fused weight value, and an association weight table is constructed using the fused weight value.
4. The MEMS-based MOM (Multi-Manager Manufacturing) collaborative platform for a lights-out factory according to claim 1, characterized in that, The data acquisition module collects the operating status signals of the manufacturing equipment to extract real-time status data, specifically for: The operating status signals are collected by a sensor array, and the operating status signals include vibration signals, temperature signals, and acoustic emission signals; The operating status signal is preprocessed to extract real-time status data including effective vibration value, average temperature value and acoustic emission pulse count rate.
5. The MEMS-based MOM (Multi-Manager Manufacturing) collaborative platform for a lights-out factory according to claim 4, characterized in that, The status assessment module compares the real-time status data with the baseline status data to determine the health status indicators of the manufacturing equipment, specifically for: Acquire the baseline state data, which includes the effective value of vibration, the average temperature, and the average baseline values corresponding to the acoustic emission pulse count rate; The effective value of vibration, the average temperature, and the acoustic emission pulse count rate in the real-time status data are compared with the corresponding average values in the reference status data to obtain the corresponding deviation ratios. The physical deviation is obtained by assigning corresponding weighting coefficients to each deviation ratio and summing them. The health impact coefficient is determined based on the comparison between the physical deviation and the preset threshold, and the health impact coefficient is used as a health status indicator.
6. The MEMS-based MOM (Multi-Manager Manufacturing) collaborative platform for a lights-out factory according to claim 5, characterized in that, Before assigning corresponding weight coefficients to each of the deviation ratios and summing them, the state assessment module also performs a consistency check on each of the deviation ratios: Calculate the absolute value of the difference between each pair of the deviation ratios mentioned in each item; If there exists one and only one deviation ratio whose absolute difference from all other deviation ratios exceeds a preset deviation threshold, then that deviation ratio is removed. The weighting coefficients that originally corresponded to the eliminated deviation ratios are redistributed proportionally to the remaining deviation ratios, and then summed.
7. The MEMS-based MOM (Multi-Manager Manufacturing) collaborative platform for a lights-out factory according to claim 1, characterized in that, The benchmark determination module, based on the correlation weights between the manufacturing equipment and each functional link of the manufacturing operation management system, transforms the health status indicators into comprehensive influence factors for each functional link, and, in conjunction with preset initial setting parameters, determines the business scheduling reference benchmark value for each functional link, specifically used for: Retrieve the association weight table containing the aforementioned association weights; The health status indicators are weighted, summed, and normalized using the correlation weights in the correlation weight table to obtain a comprehensive influence factor. Calculate the physical reference correction amount based on the comprehensive impact factor and the response sensitivity of each functional component; The normal setting values of each functional component are superimposed with the physical reference correction value to obtain the business scheduling reference baseline value; wherein, the response sensitivity and the normal setting value are the initial setting parameters.
8. The MEMS-based MOM (Multi-Manager Manufacturing) collaborative platform for a lights-out factory according to claim 1, characterized in that, The correction calculation module calculates the global business strategy adjustment amount based on the self-deviation signal, the neighbor deviation signal, the actual business plan, and the business scheduling reference benchmark value, specifically for: Obtain the timestamp corresponding to the neighbor deviation signal, determine the information age based on the difference between the timestamp and the current time, and calculate the aging decay coefficient based on the information age; Calculate the difference between the comprehensive influence factor of the current functional link in each functional link and the comprehensive influence factor of the corresponding neighboring functional link, and dynamically adjust the basic collaborative coupling weight based on the difference to obtain the collaborative coupling weight; The cooperative correction component is calculated by combining the self-bias signal, the neighbor bias signal, the aging attenuation coefficient, and the cooperative coupling weight; Calculate the self-driven correction component by combining the actual business plan with the business scheduling reference benchmark value; The collaborative correction component and the self-driven correction component are superimposed to obtain the global business strategy adjustment amount.
9. The MEMS-based MOM (Multi-Manager Manufacturing) collaborative platform for a lights-out factory according to claim 8, characterized in that, The correction calculation module dynamically adjusts the basic cooperative coupling weight based on this difference to obtain the cooperative coupling weight, which is specifically used for: Obtain the absolute value of the difference to get the absolute value of the physical difference; If the absolute value of the physical difference is greater than the physical difference threshold, then calculate the reduction amount that increases as the absolute value of the physical difference increases; The reduction amount is subtracted from the basic cooperative coupling weight, and the result of the subtraction is not less than the lower limit of the coupling weight, and the result of the subtraction is taken as the cooperative coupling weight.
10. The MEMS-based MOM (Multi-Manager Manufacturing) collaborative platform for a lights-out factory according to claim 8, characterized in that, The aging degradation coefficient is calculated using the following formula. : ; in, Information age, This refers to the effective period of information.
