A plant energy-saving control method and system based on the Internet of Things

Abstract

This invention proposes an IoT-based energy-saving control method and system for factory buildings, comprising: S1, constructing a multi-layered coupled system architecture including an infrastructure layer, a control decision layer, an intelligent algorithm layer, and a feedback optimization layer; S2, collecting illumination, equipment status, and environmental parameters through various sensors deployed on-site in the factory, and generating real-time data comprehensive correction coefficients; S3, establishing an illumination demand prediction model based on production scheduling, personnel distribution, meteorological conditions, and holiday factors; S4, constructing a flexible grouping model based on the physical location of lighting fixtures, functional area attributes, and network addresses, and establishing a delayed shutdown control mechanism based on personnel presence and area; S5, realizing multi-protocol information interaction between the manufacturing execution system, programmable logic controller, and lighting terminals through an edge gateway; and S6, monitoring the energy consumption of the lighting system in real time and calculating energy-saving efficiency, and adaptively optimizing control parameters through gradient descent to achieve dynamic adjustment of system performance.

Description

Technical Field

[0001] This invention relates to the field of energy conservation, and in particular to an energy-saving control method and system for factory buildings based on the Internet of Things. Background Technology

[0002] Traditional manual switches require manual operation and cannot automatically adjust brightness or on / off status based on actual lighting needs and usage scenarios. For example, lights may be forgotten to be turned off during the day when natural light is abundant, resulting in energy waste; or lights may remain on even when no one is working in certain areas. They also cannot provide independent brightness adjustments for different areas. For instance, using the same lighting intensity in areas with low lighting requirements (such as passageways or adhesive storage areas) and areas with high requirements (such as production areas) leads to excessive energy consumption in areas with low lighting needs.

[0003] Because tire production workshops are large and have numerous lighting devices, staff need to manually turn lights on or off in each area, which is time-consuming and labor-intensive. Especially in emergencies requiring rapid lighting adjustments, manual switches are difficult to respond to promptly. In large workshops, multiple staff members may be responsible for lighting control, leading to inconsistencies in operation. For example, someone might turn on a light in one area and forget to turn it off, while others may not realize this, making effective unified management impossible. This necessitates a solution from someone skilled in the art to address these technical challenges. Summary of the Invention

[0004] This invention aims to at least solve the technical problems existing in the prior art, and in particular, it innovatively proposes an energy-saving control method and system for factories based on the Internet of Things.

[0005] To achieve the above-mentioned objectives of this invention, this invention provides an energy-saving control method for factory buildings based on the Internet of Things, comprising:

[0006] S1 constructs a multi-layered coupled system architecture that includes an infrastructure layer, a control and decision-making layer, an intelligent algorithm layer, and a feedback optimization layer, and quantifies the degree of collaborative integration of each layer through weighted configuration.

[0007] S2 collects light, equipment status and environmental parameters through various sensors deployed on the factory site, and generates a real-time data comprehensive correction coefficient through normalization processing and time-weighted fusion.

[0008] S3 establishes an illumination demand prediction model based on production scheduling, personnel distribution, weather conditions and holiday factors, and combines equipment operating status index to make multi-factor coupled decisions and output comprehensive control decision values.

[0009] S4 constructs a flexible grouping model based on the physical location of the lamps, functional area attributes, and network address. It dynamically adjusts the output brightness of the lamps according to the comprehensive control decision value and establishes a delayed shutdown control mechanism based on personnel residue and area.

[0010] S5 enables multi-protocol information exchange between the manufacturing execution system, programmable logic controller and lighting terminal through the edge gateway, and implements edge computing load balancing to optimize communication efficiency;

[0011] S6 monitors the energy consumption of the lighting system in real time and calculates the energy-saving efficiency. It adaptively optimizes the control parameters through the gradient descent method to achieve dynamic adjustment of system performance.

[0012] In the preferred embodiment of the above technical solution, the overall architecture of the multi-layer coupled system in S1 includes: an infrastructure layer containing sensor networks and execution terminals, a control decision layer containing logic processing units, an intelligent algorithm layer containing prediction models and optimization strategies, and a feedback optimization layer containing state monitoring and parameter self-calibration. Each layer achieves collaborative integration through weight coefficient configuration.

[0013] In the preferred embodiment of the above technical solution, the multiple types of sensors in S2 include light sensors, production line status acquisition terminals, and environmental parameter sensors; the normalization process adopts the minimum-maximum standardization method; and the time-dimensional weighted fusion adopts an exponential decay function to smooth historical data.

[0014] In the preferred embodiment of the above technical solution, the illuminance demand prediction model in S3 adopts a fuzzy logic reasoning system to dynamically determine the target illuminance value based on shift time, regional coverage, weather conditions, and holiday attributes; the equipment operation status index is calculated based on different equipment operation, shutdown, fault, and maintenance states and their corresponding illuminance demand coefficients.

[0015] In the preferred embodiment of the above technical solution, the flexible grouping model in S4 performs logical AND operations based on the wireless network address, physical location coordinates, and functional area attributes to construct a set of lamps that meets dual constraints; the delayed shutdown control mechanism calculates the delayed shutdown duration based on the basic delay time, the number of remaining personnel, and the area.

[0016] In the preferred embodiment of the above technical solution, the multi-protocol information interaction in S5 includes message queue telemetry transmission protocol communication between the manufacturing execution system and the edge gateway, industrial Ethernet protocol communication between the edge gateway and the programmable logic controller, and low-power wireless protocol communication between the programmable logic controller and the lighting terminal.

[0017] In the preferred embodiment of the above technical solution, the real-time monitoring in S6 includes calculating energy consumption values ​​based on the rated power of the lamp, brightness percentage, running time, and power factor; the adaptive optimization approximates the gradient of the loss function using the numerical difference method and iteratively updates the decision weight coefficient and delay control coefficient.

[0018] The present invention also discloses a computer system, comprising:

[0019] processor;

[0020] Memory used to store processor-executable instructions;

[0021] The processor is configured to implement the energy-saving control method for factories based on the Internet of Things as described in any one of claims 1 to 7 when executing the executable instructions.

[0022] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0023] Energy consumption analysis of the lighting system revealed that adopting intelligent lighting control strategies can significantly reduce energy consumption. For example, preliminary estimates suggest that implementing the lighting optimization plan could save electricity compared to other branch offices each month, allowing for a faster recovery of investment costs.

[0024] In the long run, implementing lighting control solutions also reduces the frequency of lamp replacement and maintenance costs. This is because intelligent lighting systems can adjust brightness and operating time according to actual needs, extending the lifespan of lamps and reducing repair and replacement costs.

[0025] By rationally dividing and optimizing lighting, suitable illumination conditions can be provided for different production areas. This intelligent lighting control method can better adapt to the production rhythm of the workshop and improve overall production efficiency.

[0026] By constructing a four-layer weighted coupling architecture comprising an infrastructure layer, a control and decision-making layer, an intelligent algorithm layer, and a feedback optimization layer, the system achieves automatic sensing and dynamic adjustment of real-time lighting needs within industrial plants, overcoming the limitations of traditional lighting systems being separated from production systems. The weight coefficients of each layer are adjustable according to actual operating conditions, ensuring that the system architecture adapts to the needs of plants of different sizes.

[0027] Secondly, by establishing a dual-drive decision-making mechanism of equipment status association model and illuminance demand prediction model, the lighting system is deeply linked with production equipment, automatically matching the corresponding illuminance level according to different equipment states such as operation, shutdown, fault, and maintenance. The fuzzy logic rule base can be manually adjusted, ensuring both production safety and energy conservation.

[0028] Third, Zigbee wireless communication technology enables flexible grouping of lighting fixtures, eliminating the dependence on physical wiring in traditional wired control. Grouping is constructed based on logical constraints of location area and functional attributes. When the function of the production area changes on site, precise control can be achieved simply by reorganizing the backend logic, without changing the wiring, significantly reducing maintenance costs.

[0029] Fourth, by leveraging edge computing and a multi-protocol converged communication architecture, a combination of rapid local decision-making and centralized remote monitoring is achieved. The MQTT, PROFINET, and Zigbee three-layer protocol stack ensures reliable information transmission from the management system to the execution terminal. Edge computing load accounts for no less than 70%, allowing managers to view the status of lighting fixtures, energy consumption data, and fault information in various areas in real time via mobile terminals, enabling centralized management across regions and improving enterprise operational management.

[0030] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0031] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0032] Figure 1 This is a schematic diagram of the overall invention;

[0033] Figure 2 This is a schematic diagram of the control signal generated by the LCC and executed by the PLC according to the present invention;

[0034] Figure 3 This is a schematic diagram of the software used for lighting in the illumination area of ​​this invention;

[0035] Figure 4 This is a software illustration of the lighting area of ​​the present invention;

[0036] Figure 5 This is another software illustration of the lighting area of ​​the present invention;

[0037] Figure 6 This is a diagram illustrating the energy-saving control effect of the lighting system according to the present invention.

[0038] Figure 7 This is a comparative schematic diagram of the lighting energy-saving control of the present invention. Detailed Implementation

[0039] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0040] like Figure 1 As shown, this invention discloses a factory energy-saving control method and system based on the Internet of Things, including the following:

[0041] By adopting PROFINET industrial Ethernet communication technology, MQTT communication technology, Zigbee communication technology, edge gateway, and MES to realize the lighting control solution of the Internet of Things: the illuminance on site is collected by PLC, the production line status is obtained by the on-site LCC, the PLC returns the control logic to the IOT, the IOT sends the signal to the Zigbee coordinator, and then the signal is wirelessly transmitted to the terminal to drive the lighting fixtures.

[0042] After the MES obtains these states, it sends different signals to the IOT according to the given rules. The signal is then sent to the PLC via the edge gateway. The PLC combines the signals and logic obtained from the field sensors and sends a signal back to the IOT. The IOT sends the signal to the Zigbee coordinator, which transmits it to the terminal via wireless signal to drive the lights.

[0043] Wireless control is achieved through Zigbee communication technology, reducing the need for control wiring. This technology also enables flexible grouping of individual lights and zones. When functions change in the production area, precise control can be achieved through grouping in the background without modifying the wiring.

[0044] The vulcanization inspection workshop has 30 skylights and 16 sets of ventilation units (smoke exhaust skylights) on its roof, providing natural light. The skylights are evenly distributed, covering 50% of the workshop area. During the day, natural light is the primary source of illumination. On sunny days at midday, the light intensity at the component site ranges from 600 to 3000 lux. On cloudy days or during the early morning or evening, the natural light intensity decreases significantly, with the light intensity at the vulcanization site ranging from 400 to 1100 lux. The brightness is even lower in the north and south areas and under the robotic arms of the equipment.

[0045] Preliminary data collection for a specific production area:

[0046] Production area Key areas Passage Area Key areas Rubber material transfer and storage area Key areas Type 1 adhesive thread yes Passage 1 no Rubber Material Transfer Area no Type 2 adhesive thread yes Passage 2 no Rubber storage area no Type 3 adhesive thread yes Passage 3 no Type 4 adhesive thread yes Passage 4 no thin film yes South passage no calendering yes North passage no Inner lining yes South passage Multi-tool yes North passage Cutting machine yes South passage Transplanter yes North passage

[0047] The equipment is densely distributed, including equipment in vulcanization ditches #1, #2, #3, #4, and #5, as well as the inspection area.

[0048] Production area Key areas Passage Area Key areas #1 Sulfurized Ditch yes Passage 1 no #2 Sulfurized Ditch yes Passage 2 no #3 Sulfurized Ditch yes Passage 3 no #4 Sulfurized Ditch yes Passage 4 no 5# Sulfurized Ditch yes Channel 5 no Inspection area equipment yes Channel Six no South passageway no North passageway no

[0049] S1, Construct a multi-layered coupled system architecture: Reflect the degree of synergistic integration of the infrastructure layer, control and decision-making layer, intelligent algorithm layer, and feedback optimization layer through a weighted summation method; Provide the raw input data foundation for subsequent data cleaning;

[0050] Establish the overall system architecture model:

[0051] ;

[0052] in, This represents a dimensionless index representing the overall evaluation index of the system architecture. This represents the infrastructure layer, which includes sensor networks, communication devices, and execution terminals. This represents the control decision layer, which includes PLC, edge gateway, and logic processing unit. This represents the intelligent algorithm layer, which includes data analysis, predictive models, and optimization strategies. This represents the feedback optimization layer, which includes state monitoring, effect evaluation, and parameter self-calibration. , , , These represent the weight coefficients for each layer, are dimensionless, and range from 0 to 1, while satisfying the constraints. The settings are typically based on the importance of each layer in the system. Take 0.25, Take 0.30, Take 0.25, Take 0.20. This index represents the overall system architecture comprehensive evaluation index. It is dimensionless and used during the system design phase to assess the integration and coupling degree of the four-layer architecture. Its value ranges from 0 to 1, with higher values ​​indicating better system integration. This index is determined by weight configuration before system deployment and does not change during real-time control.

[0053] S2 performs multi-source data acquisition and preprocessing, collecting raw data through light sensors, production line status acquisition terminals, and environmental parameter sensors deployed on-site in the factory. Since light data, equipment status data, and environmental parameter data have different physical dimensions, they must be normalized separately before weighted fusion.

[0054] ;

[0055] in, This represents a comprehensive index of the original collected data, and is dimensionless. This represents the illuminance value collected by the i-th light sensor, in lux, where i∈[1,n]; This represents the maximum reference illuminance, in lux, and is taken as the measured maximum natural illuminance during the daytime in the factory or 10000 lux. The weighting coefficient for illumination data is dimensionless and is set according to the sensor's position accuracy. The value ranges from 0.8 to 1.2. For sensors with high position accuracy and frequent calibration, the value is 1.2, while for sensors with low position accuracy and no calibration for a long time, the value is 0.8. Let j represent the state parameter of the j-th production equipment. It is dimensionless and takes the value of 1.0 when the equipment is in operation, 0.2 when it is stopped, 0.5 when it is in fault state, and 0.5 when it is under maintenance state, where j∈[1,m]. This represents the maximum reference value for the device status; it is dimensionless and has a value of 1.0. The weighting coefficient for equipment status data is dimensionless and is set according to the importance of the equipment. The value ranges from 0.5 to 1.5, with 1.5 for critical production equipment and 0.5 for auxiliary equipment. Let k represent the k-th environmental parameter, including temperature, humidity, and population density, where k∈[1,p]. These represent reference values ​​for the corresponding environmental parameters: temperature reference value is 40℃, humidity reference value is 100%, and personnel density reference value is 10 people / square meter. The weighting coefficient for environmental parameters is dimensionless and ranges from 0.3 to 0.8, set according to the correlation between the parameters and lighting requirements; n represents the total number of light sensors; m represents the total number of production equipment; and p represents the number of environmental parameter types.

[0056] Each component of the original data was cleaned and standardized separately, using the illumination data as an example. For example, its standardization process is as follows:

[0057] ;

[0058] in, This represents standardized illumination data after cleaning, and is dimensionless. This represents the minimum value of historical data, in lux, and its value is 0 or the lower limit of sensor detection. The maximum value in historical data is represented by lux, with a value of 10000 lux; α represents the scaling factor, dimensionless, which controls the normalization range of the data, usually α=100; β represents the offset factor, dimensionless, which adjusts the baseline value of the data, usually β=0; temperature, humidity, and personnel density data are processed using the same standardization method to obtain the corresponding standardized data. , , .

[0059] Further time-dimensional data fusion was performed, and the standardized data at time t was... Through time-weighted integration, a smooth fusion of historical and current data is achieved:

[0060] ;

[0061] in, This represents the standardized illumination data after fusion along the time dimension; it is dimensionless and is... The weighted average value within the fusion time window; The standardized illumination data at time t is dimensionless. Represents the time decay function. λ represents the attenuation coefficient, measured in minutes, which controls the degree of influence of historical data. Typically, λ = 0.1 is used, meaning that the weight of data from 10 minutes ago is attenuated to approximately 37% of the weight at the current moment. This indicates the start time of data fusion, typically taken as 10 minutes to 1 hour prior to the current time, depending on the frequency of data fluctuations. This indicates the end time of data fusion, i.e., the current time. Temperature, humidity, and population density data are processed using the same time fusion method to obtain... , , ;

[0062] in This indicates the result of time-fusion of temperature data. This represents the result of time-fusion of humidity data. This represents the result of time-fusion of population density data.

[0063] By weighted averaging the time-integrated data of light intensity, temperature, humidity, and population density, a normalized comprehensive correction coefficient for real-time data is obtained.

[0064] ;

[0065] in, This represents the normalized real-time data comprehensive correction coefficient, which is dimensionless and ranges from 0 to 1. , , , , representing the weighting coefficients of the fused data for illumination, temperature, humidity, and personnel density, respectively. These coefficients are dimensionless, ranging from 0.1 to 0.4, and satisfy the constraints. ,generally Take 0.4, Take 0.2, Take 0.2, Take 0.2.

[0066] S3 establishes an intelligent decision-making model based on production scheduling, personnel distribution, weather conditions, and holiday factors to create an illuminance demand prediction model. A Mamdani-type fuzzy inference system is used to achieve nonlinear mapping: by weighted summation and fusion of three decision factors—illuminance demand, equipment status, and real-time data—a comprehensive control decision value is output, providing the final decision basis for the lamp brightness calculation in S4. ;

[0067] ;

[0068] in, This indicates the required illuminance value for the target area, in lux. This indicates the shift time parameters, distinguishing between day shifts, night shifts, and shift handover periods. Day shifts are from 8:00 to 20:00, and night shifts are from 20:00 to 8:00 the next day. This represents the occupancy rate of a region, which is the current percentage of population density in the region. It is obtained through a population counting sensor and ranges from 0% to 100%. This represents weather condition parameters, characterizing the degree of influence of sunny, cloudy, and rainy days on natural light. The value is 1.0 for sunny days, 0.6 for cloudy days, and 0.3 for rainy days. The holiday factor distinguishes between production patterns on weekdays, weekends, and holidays, with a value of 1.0 for weekdays, 0.8 for weekends, and 0.5 for holidays; f() represents a fuzzy logic inference function. After fuzzification of the input variables, it infers and outputs precise illuminance values ​​based on a preset rule base. Specific rules include: if... For night shift and If it is high, then It is 500 lux; if For sunny days and If it is low, then 200 lux; if If it is a holiday, then Reduced by 30% from the base value.

[0069] Construct a correlation model between equipment status and illuminance requirements:

[0070] ;

[0071] in, This represents the overall condition index of the equipment, and is dimensionless. A dimensionless Boolean identifier representing the x-th device state when the device is in state x. =1, otherwise =0, and satisfy This ensures that only one state is true at any given time; x∈{1: running, 2: stopped, 3: fault, 4: under maintenance}; This represents the illuminance demand coefficient corresponding to state x, which is dimensionless. Specifically, =1.0 indicates the running status. =0.3 indicates a shutdown state. =0.5 indicates a fault condition. =0.5 indicates a maintenance status.

[0072] Comprehensive decision output:

[0073] ;

[0074] in, This represents the output value of the comprehensive decision, which is dimensionless and ranges from 0 to 1.2. This represents the maximum reference illuminance, in lux, with a value of 1000 lux. Normalized to a dimensionless ratio; This represents the overall condition index of the equipment, and is dimensionless. This represents the normalized real-time data comprehensive correction coefficient, dimensionless. This represents the illuminance demand weighting coefficient, which is dimensionless. ∈[0,1], is set according to the requirements of the national standard GB / T50034-2024, and is usually taken as 0.4 to 0.6; This represents the equipment status weighting coefficient, which is dimensionless. ∈[0,1], set according to production priority, usually taken as 0.3 to 0.5; This represents the weighting coefficients for real-time data, which are dimensionless. ∈[0,1], usually taking values ​​from 0.1 to 0.2; and satisfying the constraint conditions. .

[0075] S4 performs lighting fixture grouping and dynamic control, constructing a flexible grouping model based on Zigbee wireless network addresses, physical locations, and functional area attributes. A group is defined as a set of lighting fixture addresses that satisfy specific location and functional attribute constraints. Set operations characterize the flexible grouping logic of lighting fixtures based on the three attributes of location, functional area, and network address, providing a logical control unit for S5's multi-protocol communication. This will enable unified regional management.

[0076] ;

[0077] in, This represents a set of lighting fixture groups, that is, a set of lighting fixture network addresses for a specific logical control area. This represents the Zigbee network address of the q-th lamp, which can be a 64-bit extended address or a 16-bit short network address. This represents the physical location attribute of the q-th lamp, expressed using a region number or coordinate range. For the target physical region; This represents the functional area attributes of the q-th lamp, including the component area, vulcanization area, molding area, and inspection area. For the target functional area; The AND operation represents the logical operation, indicating that the luminaire must simultaneously satisfy both the location area and functional attribute constraints; Q represents the number of luminaires that fall into this group after screening, i.e., the set. The total number of elements in the middle is calculated by satisfying the constraints. The number is determined.

[0078] Based on the ratio of the comprehensive decision output value to the national standard illuminance, calculate the percentage of the actual output luminance of the luminaire, and consider luminous efficacy attenuation. Convert into executable lighting control commands .

[0079] Calculate the output brightness of the luminaire based on the comprehensive decision output value:

[0080] ;

[0081] in, This indicates the output brightness value of the luminaire, expressed as a percentage or lumen. Indicates the maximum brightness value of the lamp; The output value represents the comprehensive decision, which is dimensionless and includes the weighted normalized result of illuminance requirement, equipment status, and real-time data; η represents the luminous efficacy conversion coefficient, which is dimensionless and takes into account factors such as lamp aging and dust obstruction. η∈[0.7,1.0], with 1.0 for new lamps and 0.8 for lamps that have been used for 3 years; min() represents the minimum value function to ensure that the output does not exceed 100%.

[0082] Establish a delay-off time calculation model based on three factors: basic delay, personnel residue, and area, to provide delay control parameters for automatic lamp shutdown. This achieves energy-saving control by turning off lights when people leave, and includes a delayed shutdown control mechanism.

[0083] ;

[0084] in, Indicates the delayed shutdown time, in minutes; This indicates the base delay time, in minutes, with a default value of 5 minutes. This represents the personnel residual coefficient, expressed in minutes per person. It is set based on the population flow in the area, and its value ranges from 0.3 to 1.0 minutes per person. This represents an estimated number of remaining personnel, expressed in head count, obtained through personnel detection sensors. This represents the area coefficient, expressed in minutes per square meter, with a value ranging from 0.0005 to 0.002 minutes per square meter. This indicates the area, expressed in square meters.

[0085] S5 integrates multi-protocol communication and edge computing, enabling information exchange between the MES system, PLC, and Zigbee coordinator via an edge gateway. The information exchange process is defined as follows: the MES system sends production plans and equipment status data to the edge gateway via the MQTT protocol; the edge gateway exchanges control commands with the PLC via the PROFINET protocol; and the PLC sends brightness adjustment and on / off commands to the lighting terminals via the Zigbee protocol. Arrow symbols with protocol annotations represent the multi-level information flow paths between the MES system, edge gateway, PLC, and lighting terminals, clearly defining the system communication link structure and providing a data acquisition channel for S6's energy consumption monitoring.

[0086] To optimize communication efficiency, a comprehensive evaluation method combining communication latency and protocol reliability is used.

[0087] ;

[0088] in, It represents a comprehensive indicator of communication efficiency and is dimensionless. This represents the average transmission rate, measured in kbps. It is obtained by averaging the actual transmission rates of the three protocols, PROFINET, MQTT, and Zigbee, over a specific period of time. This represents the average bandwidth utilization rate, is dimensionless, and ranges from 0 to 1. This represents the average reliability coefficient, which is dimensionless and ranges from 0.95 to 0.99, calculated based on the packet loss rate. This represents the average transmission delay, measured in milliseconds. The optimization objective is to make... Maximizing efficiency is achieved by adjusting packet priority and caching strategies. The difference between protocol performance weighting and transmission latency penalty is used to quantify a comprehensive communication efficiency index, providing an evaluation metric for communication protocol selection and parameter optimization. .

[0089] Implement edge computing load balancing:

[0090] ;

[0091] in, This indicates the percentage of edge computing load. This indicates the amount of data processed locally by the edge gateway, in MB or GB. This indicates the amount of data uploaded to the cloud for processing, in MB or GB; the optimization goal is to make... ≥ 70%, this threshold is set based on the balance between the edge gateway's processing capacity and network bandwidth. When the response rate is below 70%, local preprocessing logic is added to reduce the amount of data uploaded to the cloud, thereby improving response speed.

[0092] Calculate the percentage of data processed locally by the edge gateway relative to the total data processed, quantify the edge computing load ratio, and provide a basis for edge computing resource configuration for S6's adaptive optimization.

[0093] S6 performs energy consumption monitoring and adaptive optimization. By summing the products of power, brightness percentage, running time, and power factor, it calculates the real-time energy consumption of the lighting system, providing a data foundation for energy efficiency assessment and adaptive optimization. Calculate real-time energy consumption:

[0094] ;

[0095] in, This represents real-time energy consumption, expressed in kWh. This indicates the rated power of the g-th group of lamps, in watts (W). This represents the current brightness percentage of the g-th group of lamps, in percentage form, with a value ranging from 0 to 100. When calculating, divide by 100 to convert it into a proportionality factor. This represents the operating time of the g-th group of lights, in hours (h). The power factor of the g-th group of luminaires is dimensionless; G represents the total number of luminaire groups.

[0096] Calculate the percentage energy saving of this system compared to traditional control methods, quantify the energy saving effect of the system, and provide a reference objective function for adaptive optimization.

[0097] Assess energy efficiency:

[0098] ;

[0099] in, Expresses energy efficiency as a percentage, with the unit being percentage; The benchmark energy consumption is expressed in kWh and is defined as the measured monthly energy consumption or industry statistical value under the same production plan and regional layout, using traditional manual control or simple time-controlled lighting methods. This represents the actual energy consumption of the system, expressed in kWh.

[0100] Continuously optimize system performance through parameter self-learning:

[0101] ;

[0102] in, This indicates the updated parameter value; This represents the current parameter value, specifically referring to the weighting coefficient in S3. , Or the delay control coefficient in S4 , μ represents the learning rate, which is dimensionless and controls the step size of the parameter update, μ∈[0.001,0.1]. The gradient of the loss function with respect to the parameters is approximated using the numerical difference method: , where Δθ is a small increment with a value of 0.01; J() represents the loss function, defined as the sum of squares of the deviations between the predicted illuminance and the actual demand, i.e. Gradient descent is used to update system control parameters, minimizing the loss function and achieving adaptive optimization of system parameters, thereby continuously improving control accuracy.

[0103] Example 1: Lighting control in the component area and vulcanization area

[0104] In the component and vulcanization areas, where skylights are installed on the roof of the factory building, natural light conditions vary significantly over time. The system uses light sensors to collect real-time data on natural light intensity. During the day, when the sunlight is bright, the illuminance can reach 8000 lux, while in the evening when the light is reduced, the illuminance is about 1200 lux.

[0105] The system calculates the required target illuminance based on fuzzy logic rules. During the day shift For day shift, weather For sunny days, use 1.0, and the area coverage rate is [missing information]. When the level is 60%, or medium, reasoning is based on the rule base: For day shift and The illuminance requirement is moderate, with a basic illuminance requirement of 400 lux. On a sunny day with ample natural light, the need for artificial lighting is reduced by 30%; calculations show... = 400 × (1 - 0.3) = 280 lux. Normalization process: = 280 / 1000 = 0.28.

[0106] The system obtains the production line status from the MES system via an edge gateway. When a production worker logs into their LCC account, the equipment status is displayed. The running state (x=1) is activated. =1, the rest , satisfying ∑ =1. Based on the equipment status association model. .

[0107] Assuming the real-time data correction coefficient after time fusion The value is 0.5. The system calculates according to the comprehensive decision formula: [The value is missing here, likely indicating a value of 0.5]. =0.5 (Illuminance requirements take priority), =0.4 (equipment status is secondary) =0.1 (real-time data fine-tuning), satisfies ;but = 0.5×0.28 + 0.4×1.0 + 0.1×0.5 = 0.14 +0.4 + 0.05 = 0.59.

[0108] Further calculations are performed based on the brightness adjustment formula: assuming =100%, η=0.9, the luminous efficacy of the lamps has slightly decreased after 2 years of use. =0.59, = 100% × 0.59 × 0.9 = 53.1%. After rounding, the lamps are turned on at 53% brightness, which satisfies the illuminance requirement of 280 lux while avoiding over-illumination.

[0109] When an employee finishes production and logs out of their LCC account, the equipment status changes to "stopped" (x=2). =1, , = 0.3. At this time... The corresponding delay is reduced, and the system calculates the delay time according to the delay shutdown formula. Let the base delay be... =5 minutes, personnel residual coefficient =0.5 minutes / person, remaining personnel =2 people, area coefficient k_2=0.001 minutes / square meter, area =500 square meters, then = 5 + 0.5×2 + 0.001×500 = 5 + 1 + 0.5 = 6.5 minutes. The system will automatically turn off the lights after a 6.5-minute delay to allow employees to complete the machine cleaning activities.

[0110] Example 2: Lighting Control in the Molding Area

[0111] The forming area is a production zone without skylights and operates 24 / 7, resulting in significantly higher lighting energy consumption compared to other areas. The system optimizes control logic for this area: when the production machine is running normally, the MES system sends equipment operation signals to the edge gateway via the MQTT protocol. The edge gateway then transmits the commands to the PLC via the PROFINET protocol. The PLC, combining data from field sensors and performing calculations using a decision model, sends control commands to the corresponding lighting fixtures in the G_group via a Zigbee coordinator.

[0112] Assuming this area The requirement is 500 lux, which is normalized to 0.5. =1.0 indicates the running status. =0.6, weight taken =0.4, =0.5, =0.1, then = 0.4×0.5 + 0.5×1.0+ 0.1×0.6 = 0.2 + 0.5 + 0.06 = 0.76. Take η=0.85, = 100% × 0.76 × 0.85 = 64.6%, meaning the lights are turned on at approximately 65% ​​brightness to achieve full illumination coverage.

[0113] When a piece of equipment has a scheduled maintenance day, the operator marks the equipment as being under maintenance via the MES system. At this time, the equipment status... Mid-maintenance status activated. =1, =0.5, = 0.5. Recalculate = 0.4×0.5+0.5×0.5+0.1×0.6= 0.2+0.25+0.06= 0.51; = 100%×0.51×0.85=43.4%. The PLC only sends a dimming signal to the group to which the machine belongs during maintenance, and only turns on some of the lights in the maintenance area. For example, turning on the lights intermittently can achieve an illumination of about 50%, avoiding unnecessary energy consumption, while the machine area in normal production maintains an illumination of 65%.

[0114] like Figure 2 As shown, the component area and vulcanization area: With the design of the top skylight, the natural light conditions are good. The light sensor detects that the natural light intensity is sufficient, and the illuminance can reach 15,000 lux, for example, bright sunlight during the day; when the light is reduced in the evening, the illuminance is around 1,200 lux. According to the program, when the current illuminance is less than or equal to the set illuminance value, the PLC logs into the LCC terminal account and starts the lighting based on the machine production status. The PLC obtains the equipment production status through the IoT and, based on the lighting conditions, immediately controls the lighting of the corresponding area to ensure that the equipment area in the workshop has sufficient lighting to ensure production safety. At the same time, after the employee finishes production and logs out of the LCC terminal account, the system automatically delays for five minutes to turn off the lighting, allowing the employee to complete activities such as cleaning the machine.

[0115] Molding Area: As an area without skylights and operating 24 / 7, the molding area consumes significantly more electricity for lighting than the parts area and vulcanization area. Therefore, the control logic is optimized: information from the LCC is obtained through the IoT. Production machines send a full-on signal to the PLC, while maintenance machines send a half-on signal. Different signals control different switches. When there is a maintenance plan for the equipment on the same day, the PLC only turns on the appropriate area lighting for the maintenance machine, turning on only half of the lights in the maintenance area to avoid unnecessary energy consumption, while machines in normal production are not affected.

[0116] like Figure 3 , 4 Figure 5 shows a schematic diagram of the working status of lighting equipment in a production area displayed through a software interface. On sunny days at noon, the ambient light intensity at the component site ranges from 500 to 2000 lux, with varying brightness depending on the installation location of the equipment. On cloudy days or during the early morning or evening, the natural light intensity decreases significantly, with the ambient light intensity at the component site ranging from 300 to 1000 lux, and even lower in the north and south areas and below the production line.

[0117] According to the GB50034-2013 Building Design Standard, the illumination standard for the production areas of the rubber industry's calendering, molding and cutting sections and vulcanization sections shall not be less than 300 LUX. Through energy-saving control methods, the workshop lighting control shall be based on the provisions of GB50034-2013 Building Design Standard.

[0118] Figure 6 and 7 The lighting unit consumption refers to the electrical energy consumed per unit of lighting. After the completion of the first phase of construction, the actual power consumption resulted in excessive power usage. After setting up a primary manual control adjustment, the power consumption was reduced. Finally, the intelligent control energy-saving method of this application was used to effectively reduce power loss and provide efficient lighting for the production environment.

[0119] The technical solution presented in this application provides intelligent control of factory lighting. By leveraging remote monitoring and management functions, it greatly facilitates lighting management in industrial plants. Regardless of their location, managers can simply log into the system monitoring interface via computer or mobile device to view the real-time operating status of the lighting system and remotely monitor and operate it. The remote monitoring and management functions also facilitate centralized management of lighting systems in multiple plants and areas, breaking down geographical limitations, reducing management costs, and improving the overall operational management level of the enterprise.

[0120] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A factory energy-saving control method based on the Internet of Things, characterized in that, include: S1 constructs a multi-layered coupled system architecture that includes an infrastructure layer, a control and decision-making layer, an intelligent algorithm layer, and a feedback optimization layer, and quantifies the degree of collaborative integration of each layer through weighted configuration. S2 collects light, equipment status and environmental parameters through various sensors deployed on the factory site, and generates a real-time data comprehensive correction coefficient through normalization processing and time-weighted fusion. S3 establishes an illumination demand prediction model based on production scheduling, personnel distribution, weather conditions and holiday factors, and combines equipment operating status index to make multi-factor coupled decisions and output comprehensive control decision values. S4 constructs a flexible grouping model based on the physical location of the lamps, functional area attributes, and network address. It dynamically adjusts the output brightness of the lamps according to the comprehensive control decision value and establishes a delayed shutdown control mechanism based on personnel residue and area. S5 enables multi-protocol information exchange between the manufacturing execution system, programmable logic controller and lighting terminal through the edge gateway, and implements edge computing load balancing to optimize communication efficiency; S6 monitors the energy consumption of the lighting system in real time and calculates the energy-saving efficiency. It adaptively optimizes the control parameters through the gradient descent method to achieve dynamic adjustment of system performance.

2. The IoT-based energy-saving control method for factory buildings according to claim 1, characterized in that, The overall architecture of the multi-layer coupled system in S1 includes: an infrastructure layer containing sensor networks and execution terminals, a control and decision layer containing logic processing units, an intelligent algorithm layer containing prediction models and optimization strategies, and a feedback optimization layer containing state monitoring and parameter self-calibration. Each layer achieves collaborative integration through weight coefficient configuration.

3. The IoT-based energy-saving control method for factory buildings according to claim 1, characterized in that, The various sensors in S2 include light sensors, production line status acquisition terminals, and environmental parameter sensors. The normalization process uses the minimum-maximum standardization method, and the time-dimensional weighted fusion uses an exponential decay function to smooth historical data.

4. The IoT-based energy-saving control method for factory buildings according to claim 1, characterized in that, The illuminance demand prediction model in S3 uses a fuzzy logic reasoning system to dynamically determine the target illuminance value based on shift time, regional coverage, weather conditions, and holiday attributes; the equipment operation status index is calculated based on different equipment operation, shutdown, fault, and maintenance states and their corresponding illuminance demand coefficients.

5. The energy-saving control method for factory buildings based on the Internet of Things according to claim 1, characterized in that, The flexible grouping model in S4 performs logical AND operations based on wireless network addresses, physical location coordinates, and functional area attributes to construct a set of lamps that meets dual constraints. The delayed shutdown control mechanism calculates the delayed shutdown duration based on a weighted average of the base delay time, the number of remaining personnel, and the area.

6. The energy-saving control method for factory buildings based on the Internet of Things according to claim 1, characterized in that, The multi-protocol information interaction in S5 includes message queue telemetry transmission protocol communication between the manufacturing execution system and the edge gateway, industrial Ethernet protocol communication between the edge gateway and the programmable logic controller, and low-power wireless protocol communication between the programmable logic controller and the lighting terminal.

7. The IoT-based energy-saving control method for factory buildings according to claim 1, characterized in that, The real-time monitoring in S6 includes calculating energy consumption values ​​based on the lamp's rated power, brightness percentage, running time, and power factor; the adaptive optimization approximates the gradient of the loss function using the numerical difference method and iteratively updates the decision weight coefficients and delay control coefficients.

8. A computer system, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to implement the energy-saving control method for factories based on the Internet of Things as described in any one of claims 1 to 7 when executing the executable instructions.

Images