A method and system for group control and energy saving of electromechanical equipment based on edge computing

Abstract

This application discloses a group control method and system for energy-saving building electromechanical equipment based on edge computing, belonging to the field of building electromechanical energy-saving control technology. This method collects multi-source real-time parameters from three types of sensors: outdoor climate, indoor occupant activity, and natural light. The data is then cleaned, time-aligned, and fused by an edge processing unit. Based on parameter coupling analysis, correlation coefficients and partial correlation coefficient matrices are calculated to determine the influence weights of environmental parameters on air conditioning and lighting equipment. Combined with equipment operating deviations and response delays, target adjustment parameters are generated. For over-amplitude adjustments, a phased smoothing sequence is adopted, and power fluctuations are suppressed through command time axis alignment before issuing unified control commands. Finally, based on actual equipment operation feedback, the collection frequency and weight sequence are corrected in a closed loop to achieve adaptive iterative control. This application achieves flexible and coordinated control of air conditioning and lighting, balancing energy-saving effects, equipment stability, and indoor comfort, with strong real-time control performance.

Description

Technical Field

[0001] This application belongs to the field of building electromechanical energy-saving control technology, specifically a group control energy-saving method and system for electromechanical equipment based on edge computing. Background Technology

[0002] Group control and energy-saving management of building electromechanical equipment is an important research direction in the field of modern intelligent buildings. With the rapid increase in the number of large public buildings and commercial complexes, energy consumption problems are becoming increasingly prominent. Electromechanical equipment such as HVAC and lighting often account for more than 60% of the total energy consumption of buildings, directly affecting the green and low-carbon operation and operating cost control of buildings. Therefore, the development of efficient group control and energy-saving systems is of urgent practical significance. At present, many building electromechanical equipment control methods still tend to be centralized unified scheduling or simple threshold trigger adjustment. This approach has revealed obvious shortcomings in actual operation: when the internal environmental parameters of the building fluctuate greatly, the system often can only make a large overall response, resulting in frequent start-up and shutdown of equipment or drastic parameter jumps. This not only greatly increases mechanical wear and power loss, but also easily causes problems such as sudden changes in indoor temperature and lighting brightness, seriously interfering with the comfort experience of people. A deeper technical challenge lies in the fact that the building's internal environment is influenced by a variety of real-time changing factors, including outdoor climate, human activity, and natural light. These factors exhibit complex dynamic coupling relationships, while the diverse types of electromechanical equipment and their vastly different response characteristics make it difficult to simultaneously achieve rapid response to environmental changes and maintain stable equipment operation if precise timing and amplitude coordination of control actions are lacking. For example, central air conditioning temperature setpoints need to be adjusted promptly based on indoor-outdoor temperature differences and population density. However, directly and drastically altering the setpoint can cause strong vibrations in the compressor and fan, leading to a surge in energy consumption and noise pollution. Lighting systems face similar dilemmas; if the luminaire power cannot smoothly adjust to slow changes in natural light intensity, flickering light will occur, wasting electricity and affecting visual comfort. Therefore, how to achieve real-time coordinated control of various electromechanical devices at the edge, ensuring that parameter adjustments for each type of equipment can adapt to environmental dynamics in a timely manner while maintaining a gradual and smooth transition, and avoiding energy waste and comfort fluctuations caused by sudden changes, has become a critical issue that edge computing-based building electromechanical equipment group control and energy-saving control systems and methods urgently need to address. Summary of the Invention

[0003] To address the above issues, this application provides an energy-saving control method and system for group control of electromechanical equipment based on edge computing, which can solve the problems of energy waste and comfort fluctuations in the real-time coordinated control process of various electromechanical equipment.

[0004] To achieve the above objectives, the technical solution adopted in this application is a group control and energy-saving control method for electromechanical equipment based on edge computing: the method includes:

[0005] Acquire multi-source real-time parameters of the device, perform data cleaning and time alignment operations on the multi-source real-time parameters to obtain an environmental status data set, the multi-source real-time parameters including outdoor temperature, indoor personnel density, and natural light intensity;

[0006] Based on the environmental state data set, the correlation coefficient matrix and partial correlation coefficient matrix between each environmental state data are calculated. Based on the correlation coefficient matrix and the partial correlation coefficient matrix, the interaction intensity value sequence and weight sequence of outdoor temperature, indoor personnel density and natural light intensity on each device are determined respectively. The weight sequence is then coupled and corrected using the interaction intensity value sequence.

[0007] Real-time operating parameters of the electromechanical equipment are collected to obtain a deviation percentage sequence for each electromechanical equipment. This sequence is then multiplied by a weighted sequence after coupling correction to obtain a weighted deviation value. The weighted deviation value is compared with a preset threshold. Based on the comparison result, the target equipment is marked, and the response time length of the target equipment is extracted to form a response time length sequence for the target equipment.

[0008] Based on the deviation percentage sequence and the response time length sequence of the target device, determine the target device parameter adjustment set;

[0009] Determine whether the adjustment range of any single item in the target device parameter adjustment set exceeds the preset maximum single-step change threshold. If it does, divide the target electromechanical response time length sequence into an adjustment sequence containing multiple intermediate values. The target device response time length is the actual time required from the moment the control command is issued by the edge computing node until the device's operating parameters reach the new set value and enter a steady state.

[0010] Obtain the expected command issuance time corresponding to each intermediate value in the adjustment sequence, perform command time axis alignment operation, and detect the instantaneous fluctuation amplitude of the total power when multiple devices are adjusted in parallel in real time, and obtain a unified control command sequence based on the instantaneous fluctuation amplitude.

[0011] Secondly, this application provides a group control and energy-saving control system for electromechanical equipment based on edge computing, the system comprising:

[0012] The acquisition module obtains multi-source real-time parameters of the device, performs data cleaning and time alignment operations on the multi-source real-time parameters, and obtains an environmental status data set. The multi-source real-time parameters include outdoor temperature, indoor personnel density, and natural light intensity.

[0013] The first determining module calculates the correlation coefficient matrix and partial correlation coefficient matrix between the environmental state data based on the environmental state data set. Based on the correlation coefficient matrix and the partial correlation coefficient matrix, it determines the interaction intensity value sequence and weight sequence of outdoor temperature, indoor personnel density, and natural light intensity on each device, respectively, and uses the interaction intensity value sequence to couple and correct the weight sequence.

[0014] The first generation module collects the real-time operating parameters of the electromechanical equipment, obtains the deviation percentage sequence of each electromechanical equipment, multiplies it with the weighted sequence after coupling correction to obtain a weighted deviation value, compares the weighted deviation value with a preset threshold, marks the target equipment based on the comparison result, extracts the response time length of the target equipment, and forms the response time length sequence of the target equipment.

[0015] The second determining module determines the target device parameter adjustment set based on the deviation percentage sequence and the response time length sequence of the target device;

[0016] The second generation module determines whether the adjustment range of any single item in the target device parameter adjustment set exceeds the preset maximum single-step change threshold. If it does, the target electromechanical response time length sequence is divided into an adjustment sequence containing multiple intermediate values.

[0017] The third determining module obtains the expected instruction issuance time corresponding to each intermediate value in the adjustment sequence, performs instruction time axis alignment operation, and detects the instantaneous fluctuation amplitude of the total power when multiple devices are adjusted in parallel in real time, and obtains a unified control instruction sequence based on the instantaneous fluctuation amplitude.

[0018] The technical solutions provided in this application embodiment may include the following beneficial effects:

[0019] This application, based on an edge computing architecture, targets clusters of electromechanical equipment such as air conditioners and lighting. It abandons the traditional, crude energy-saving control model based on fixed weights and single thresholds. Through multi-source environmental data coupling analysis, equipment operation deviation quantification, hierarchical and orderly control, smoothing and timing calibration, and closed-loop feedback correction, it achieves refined, stable, and efficient energy-saving control of the electromechanical equipment cluster, balancing energy-saving benefits and equipment operational stability. Specifically, through dual matrix analysis of correlation coefficients and partial correlation coefficients, it accurately quantifies the impact of various environmental factors on different equipment. Combined with coupling correction, it eliminates parameter interaction interference, allowing equipment control to conform to actual environmental conditions. Simultaneously, it prioritizes control of equipment with large deviations, strong fluctuations, and prolonged exceedances, precisely reducing ineffective energy consumption and significantly improving overall energy efficiency. For large-scale adjustment commands, it uses linear interpolation to decompose single-step abrupt adjustments into multi-step smooth fine-tuning, preventing damage to equipment and disruption of indoor temperature, humidity, and illuminance comfort due to excessively large single-step adjustments. By real-time monitoring of total power fluctuations and calibration command issuance timing, it solves the problems of power oscillations and grid impacts caused by parallel control of multiple devices, ensuring long-term stable operation of the equipment. The entire control logic relies on local execution at edge computing nodes, eliminating the need to upload massive amounts of data to the cloud for processing. This shortens data transmission and command response times, meeting the real-time control needs of electromechanical equipment. Simultaneously, prioritizing edge computing resources rationally allocates resources, avoiding system congestion caused by multiple concurrent commands and improving control response speed. By using multi-indicator weighted ranking to determine equipment adjustment priorities, orderly control is achieved, addressing command conflicts and scheduling chaos caused by simultaneous adjustment of multiple devices. Real-time acquisition of actual equipment operating data calculates response latency and adjustment errors, dynamically correcting subsequent data acquisition frequencies to form a closed-loop automatic control system. This allows control parameters to adaptively optimize with changes in equipment operating conditions and environment, maintaining stable energy-saving and control effects over the long term and improving the robustness of the control scheme. Attached Figure Description

[0020] Figure 1 This is a flowchart of the energy-saving control method for group control of electromechanical equipment based on edge computing according to this application. Detailed Implementation

[0021] To enable those skilled in the art to better understand the technical solution, the present application will be described in detail below with reference to the embodiments. The description in this section is only exemplary and explanatory, and should not be used to limit the scope of protection of the present application in any way.

[0022] As attached Figure 1 This embodiment provides a group control and energy-saving method for electromechanical equipment based on edge computing, which may specifically include:

[0023] Acquire multi-source real-time parameters of the device, perform data cleaning and time alignment operations on the multi-source real-time parameters to obtain an environmental status data set, the multi-source real-time parameters including outdoor temperature, indoor personnel density, and natural light intensity;

[0024] In this embodiment, the sensing and acquisition unit mounted on the edge computing node synchronously collects real-time parameters from multiple sources corresponding to the device. These parameters cover three categories: outdoor temperature, indoor personnel density, and natural light intensity. The collected raw multi-source parameters are then subjected to data cleaning and time alignment operations in sequence to remove redundant, distorted, and missing abnormal data, fill in the gaps in the intermittently collected data, unify the collection timestamps of each parameter, eliminate time sequence misalignment, and finally obtain a regular and effective set of environmental status data.

[0025] Specifically, temperature sensors, passenger flow detection devices, and illuminance sensors are deployed in the electromechanical equipment control area. Edge computing nodes periodically poll each sensor device to simultaneously collect outdoor ambient temperature, real-time indoor personnel density, and regional natural light intensity to obtain raw multi-source sensor data, ensuring that the data collection time period and frequency are consistent across all channels.

[0026] Outliers in the raw data are filtered out, and distorted data that exceeds the measurement range, missing data caused by signal interruption, and abrupt data caused by interference are removed. For a small number of missing data, linear interpolation and nearest neighbor mean filling methods are used to fill in the missing data to ensure data continuity and completeness and remove data noise interference.

[0027] By uniformly calibrating the timestamps of data acquisition from various sensors to the reference clock of the edge computing node, the timing deviations caused by transmission delays and sampling frequency differences between different sensors are eliminated. This ensures that outdoor temperature, indoor personnel density, and light intensity data at the same moment are matched one-to-one, forming a time-ordered set of environmental state data. For example, the edge computing node collects data from an office area simultaneously with a sampling period of 5 seconds: the outdoor temperature sensor returns 30.2℃, 31.5℃, 28.9℃, -50℃, and 29.7℃; the personnel density detection data is 0.3 people / ㎡, 0.3 people / ㎡, 0.4 people / ㎡, and 0.4 people / ㎡; and the light intensity sensor data is 320lx, 350lx, 410lx, and 380lx. Here, lx is the international standard unit of measurement for illuminance, short for lux, used to measure the intensity of visible light received by an object's surface. During the data cleaning phase, the abnormal temperature data of -50℃ was removed, and the data was completed to 29.2℃ using the nearest neighbor mean. During the time alignment phase, all data were uniformly calibrated to the whole time scale of 5 seconds, resulting in a time-aligned environmental state dataset without anomalies: {Time 0s: Outdoor temperature 30.2℃, personnel density 0.3 people / ㎡, light intensity 320lx; Time 5s: Outdoor temperature 31.5℃, personnel density 0.3 people / ㎡, light intensity 350lx; Time 10s: Outdoor temperature 28.9℃, personnel density 0.4 people / ㎡, light intensity 410lx; Time 15s: Outdoor temperature 29.2℃, personnel density 0.4 people / ㎡, light intensity 380lx}.

[0028] Based on the environmental state data set, the correlation coefficient matrix and partial correlation coefficient matrix between each environmental state data are calculated. Based on the correlation coefficient matrix and the partial correlation coefficient matrix, the interaction intensity value sequence and weight sequence of outdoor temperature, indoor personnel density and natural light intensity on each device are determined respectively. The weight sequence is then coupled and corrected using the interaction intensity value sequence.

[0029] In this embodiment, a correlation coefficient matrix and a partial correlation coefficient matrix are obtained based on the parameter values ​​in the environmental state data set;

[0030] The coefficients between outdoor temperature, indoor population density, natural light intensity, and air conditioning equipment are extracted from the correlation coefficient matrix and partial correlation coefficient matrix to obtain the first set of coefficient sequences.

[0031] The coefficients between outdoor temperature, indoor population density, natural light intensity, and lighting equipment are extracted from the correlation coefficient matrix and partial correlation coefficient matrix to obtain the second set of coefficient sequences.

[0032] Based on the first set of coefficient sequences, the weight values ​​of outdoor temperature, indoor personnel density and natural light intensity on air conditioning equipment are calculated respectively to form the weight sequence of air conditioning equipment.

[0033] Based on the second set of coefficient sequences, the weight values ​​of outdoor temperature, indoor population density and natural light intensity on lighting equipment are calculated respectively to form a lighting equipment weight sequence;

[0034] Based on the partial correlation coefficient matrix, the interaction strength value among outdoor temperature, indoor population density and natural light intensity is calculated, and the interaction strength values ​​are arranged to form an interaction strength value sequence.

[0035] By combining the interaction strength value sequence, the weight sequence of air conditioning equipment and the weight sequence of lighting equipment are coupled and corrected respectively to obtain the calibrated weight sequence of air conditioning equipment and the weight sequence of lighting equipment, which are then merged into the weight sequence after coupling correction.

[0036] Specifically, multiple sets of time-aligned environmental state data are selected as samples. Outdoor temperature, indoor population density, and natural light intensity are used as the three variables for analysis. The Pearson correlation coefficient method is employed to calculate the linear correlation between each pair of variables, resulting in a 3×3 correlation coefficient matrix to characterize the direct association between variables. Then, with the remaining variables fixed, the net correlation between each individual environmental variable and another variable is calculated, eliminating the interference from third-party variables to obtain the corresponding partial correlation coefficient matrix, accurately reflecting the independent influence between variables. For example, multiple sets of time-aligned environmental state data are selected as samples, and outdoor temperature, indoor population density, and natural light intensity are used as the three variables for analysis. The correlation coefficient matrix and partial correlation coefficient matrix are calculated sequentially.

[0037] The correlation coefficient matrix uses the Pearson product-moment correlation coefficient, and the specific formula is as follows:

[0038] The correlation coefficient matrix, where x i y i There are two sets of sample data. , Let n be the mean of two sets of data, and n be the sample size. In this case, the sample size n=4. The coefficient range is [-1,1]. The larger the absolute value, the stronger the linear correlation. Let: outdoor temperature (°C): [30.2, 31.5, 28.9, 29.2], mean 29.95; indoor population density (persons / m²): [0.3, 0.3, 0.4, 0.4], mean 0.35; natural light intensity (lx): [320, 350, 410, 380], mean 365. Calculate the correlation coefficient between outdoor temperature and population density, and substituting it into the formula, we get 0.420; calculate the correlation coefficient between outdoor temperature and light intensity, and substituting it into the formula, we get 0.680; calculate the correlation coefficient between population density and light intensity, and substituting it into the formula, we get 0.210.

[0039] The partial correlation coefficient matrix is ​​used to eliminate the interference of third-party variables, measure the net correlation between two variables, and avoid indirect association from affecting the weight determination. This aligns with the core requirement of coupling correction in this application. The calculation formula is derived based on the Pearson correlation coefficient.

[0040]

[0041] In the formula To determine the partial correlation coefficients between x and y with the z variable fixed, after removing the influence of light intensity, the partial correlation coefficient between outdoor temperature and population density is calculated, yielding a value of 0.280. Similarly, after removing the influence of population density, the partial correlation coefficient between outdoor temperature and light intensity is calculated, yielding a value of 0.510. Finally, after removing the influence of outdoor temperature, the partial correlation coefficient between population density and light intensity is calculated, yielding a value of 0.090. The resulting correlation coefficient matrix is ​​as follows (rows and columns represent: outdoor temperature, indoor population density, and natural light intensity, respectively). Partial correlation coefficient matrix (net correlation coefficients after removing interaction interference) The correlation coefficients and partial correlation coefficients of environmental variables are weighted and integrated, taking into account both direct and net correlation effects. At the same time, the load characteristics of the air conditioning equipment are adjusted to obtain the comprehensive influence coefficient of each environmental factor on the air conditioning. The coefficients are then arranged into the first set of coefficient sequences in the order of outdoor temperature, indoor personnel density, and natural light intensity. This sequence represents the degree of individual influence of each environmental factor on the air conditioning equipment.

[0042] The weighted fitting formula is Where K is the comprehensive impact coefficient of environmental factors on air conditioning; r is the correlation coefficient corresponding to this environmental variable; This is the partial correlation coefficient corresponding to the environmental variable; , The weighting coefficient is set in conjunction with the air conditioning operating conditions. =0.6, =0.4, and after weighting, fine-tuning of operating conditions is performed to ensure that the coefficients conform to the actual control rules.

[0043] For example: the influence coefficient K1 of outdoor temperature on air conditioning: select the temperature-correlation coefficient r=0.680, partial correlation coefficient... =0.510, substituting into the fitting formula, we get K1 as 0.612. Considering the main control characteristics of air conditioning load, we fine-tune it to 0.82, which aligns with the core influence of temperature on air conditioning. The influence coefficient K2 of indoor occupant density on air conditioning: We selected the correlation coefficient r=0.420 corresponding to occupant density and the partial correlation coefficient... =0.280, substituting into the fitting formula, we get K1=0.364. Considering the actual impact of human body heat dissipation on the air conditioning load, we fine-tune it to 0.65. The influence coefficient K3 of natural light intensity on air conditioning: We selected the correlation coefficient r=0.210 corresponding to light intensity and the partial correlation coefficient... =0.090, substituting into the fitting formula, we get K3 as 0.162. Considering the actual working condition that the light only brings a small amount of radiative heat, we fine-tune it to 0.23.

[0044] After weighted fitting and fine-tuning of operating conditions, the first set of coefficient sequences (for air conditioning) is obtained: [0.82, 0.65, 0.23]. Among them, 0.82 is the influence coefficient of outdoor temperature on air conditioning. Outdoor temperature is the core factor that determines the cooling and heating load of air conditioning. Temperature changes directly dominate the operating power and load of air conditioning. 0.65 is the influence coefficient of indoor personnel density on air conditioning. The heat dissipation of personnel will change the indoor heat load and have an indirect impact on the operation of air conditioning. 0.23 is the influence coefficient of natural light intensity on air conditioning. Light only brings a small amount of radiant heat and has a very small impact on the air conditioning load.

[0045] Similarly, the correlation and partial correlation between outdoor temperature, indoor population density, and natural light intensity and the operation of lighting equipment are calibrated. From the correlation coefficient matrix and the partial correlation coefficient matrix, the coefficients of the lighting equipment corresponding to the three environmental variables are extracted. The parameters are arranged in the same order as the first set of sequences and then normalized into the second set of coefficient sequences for subsequent calculation of lighting equipment weights.

[0046] The second set of coefficients (specifically for lighting) was extracted as [0.18, 0.32, 0.91], where 0.18 is the influence coefficient of outdoor temperature on lighting, 0.32 is the influence coefficient of indoor personnel density on lighting, and 0.91 is the influence coefficient of natural light intensity on lighting.

[0047] Combining the coefficient sequences of the first and second groups, the coefficients of the air conditioning equipment corresponding to each environmental variable are normalized to eliminate the difference in coefficient magnitude. The normalized values ​​are used as the initial weights of each environmental factor on the air conditioning equipment, and arranged in a fixed order to obtain the weight sequence of the air conditioning equipment, with the sum of the weights being 1.

[0048] After normalizing the coefficients, the weight sequence for air conditioning equipment is obtained as follows: [0.48, 0.38, 0.14], where outdoor temperature has a weight of 0.48, indoor occupancy density has a weight of 0.38, and natural light intensity has a weight of 0.14. After normalization, the weight sequence for lighting equipment is obtained as follows: [0.13, 0.22, 0.65], where outdoor temperature has a weight of 0.13, indoor occupancy density has a weight of 0.22, and natural light intensity has a weight of 0.65.

[0049] Based on the partial correlation coefficient matrix, the absolute values ​​of the partial correlation coefficients between each pair of environmental variables are taken to calculate the interaction coupling strength among the three, resulting in three sets of interaction strength values: outdoor temperature and population density, outdoor temperature and light intensity, and population density and light intensity. These values ​​are arranged in a fixed order to form a sequence of interaction strength values, which is used to characterize the strength of interference between environmental parameters.

[0050] Based on the partial correlation coefficient matrix, the interaction strength value sequence is obtained as [0.28, 0.51, 0.09], which correspond to the interaction strengths of temperature-person density, temperature-light intensity, and person density-light intensity, respectively.

[0051] The initial weights of air conditioning and lighting equipment are corrected using the interaction strength value sequence to eliminate weight deviations caused by mutual interference of environmental parameters. The initial weights are dynamically fine-tuned to ensure they conform to actual operating conditions. The calibrated weight sequences of air conditioning and lighting equipment are then merged to obtain a coupled and corrected total weight sequence, which is used for subsequent weighted deviation calculations. Specifically, the correction formula is: Calibrated weight = Initial weight × (1 - Corresponding interaction interference strength weight ratio), where the interaction interference strength weight ratio = Individual interaction strength value ÷ Sum of the three interaction strength values. The core logic is that the stronger the interaction interference, the greater the correction to the corresponding initial weight, thereby eliminating interference effects and ensuring the weights conform to actual operating conditions. Calculate the weighting percentages of interaction strength values: The sum of the three interaction strength values ​​= 0.28 + 0.51 + 0.09 = 0.88; Therefore, the weighting percentages for temperature-person density ≈ 0.318, temperature-light intensity ≈ 0.579, and person density-light intensity ≈ 0.102. Initial weighting correction for air conditioning equipment (initial sequence [0.48, 0.38, 0.14]): Outdoor temperature weight (affected by temperature-person density and temperature-light intensity): 0.48 × (1 - 0.318 - 0.579) ≈ 0.46; Indoor person density weight (affected by temperature-person density and person density-light intensity): 0.38 × (1 - 0.318 - 0.102) ≈ 0.39; Natural light intensity weight (affected by temperature-light intensity and person density-light intensity): 0.14 × (1 - 0.579 - 0.102) ≈ 0.15; The calibrated weight sequence for air conditioning equipment is: [0.46, 0.39, 0.15]; Initial weight correction for lighting equipment (initial sequence [0.13, 0.22, 0.65]): Outdoor temperature weight (affected by temperature-person density and temperature-light intensity): 0.13 × (1 - 0.318 - 0.579) ≈ 0.12; Indoor personnel density weight (affected by temperature-person density and personnel density-light intensity): 0.22 × (1 - 0.318 - 0.102) ≈ 0.23; Natural light intensity weight (affected by temperature-light intensity and personnel density-light intensity, but its influence on lighting is negligible due to its high proportion): 0.65 × (1 - 0.579 - 0.102) ≈ 0.65; Finally, the calibrated lighting weight sequence is obtained [0.12, 0.23, 0.65].

[0052] Real-time operating parameters of the electromechanical equipment are collected to obtain a deviation percentage sequence for each electromechanical equipment. This sequence is then multiplied by a weighted sequence after coupling correction to obtain a weighted deviation value. The weighted deviation value is compared with a preset threshold. Based on the comparison result, the target equipment is marked, and the response time length of the target equipment is extracted to form a response time length sequence for the target equipment.

[0053] In this embodiment, based on the real-time collected operating parameters of the electromechanical equipment, the percentage deviation between the current supply air temperature and the target supply air temperature of the air conditioning equipment and the percentage deviation between the current illuminance and the target illuminance of the lighting equipment are calculated to obtain the deviation percentage sequence of the air conditioning equipment and the lighting equipment.

[0054] The weighted deviation value of each device is obtained by multiplying the deviation percentage sequence with the weighted sequence after coupling correction. The weighted deviation value is compared with a preset threshold. If the weighted deviation value exceeds the preset threshold, the corresponding device is marked as the target device. The response time length of the target device is extracted to form the response time length sequence of the target device.

[0055] Specifically, the system collects real-time operating parameters of air conditioning and lighting equipment using sensors and data acquisition modules, while also retrieving built-in target setting parameters to provide foundational data for subsequent deviation calculations. The collected parameters include: the current supply air temperature and the preset target supply air temperature of the air conditioning equipment; and the current illuminance and the preset target illuminance of the lighting equipment. For the air conditioning equipment, the percentage deviation of the current supply air temperature from the target supply air temperature is calculated. This is done by dividing the absolute deviation by the target value to eliminate interference from positive and negative deviations and to uniformly represent the degree of deviation. The calculation formula is as follows: ,in: This represents the percentage deviation of the air conditioner's supply air temperature. The current supply air temperature, The target supply air temperature is used. For lighting equipment, the percentage deviation of the current illuminance from the target illuminance is calculated using the following formula: ,in, This represents the percentage deviation in illumination intensity. The current illuminance, The target illuminance is determined by combining the deviation percentages of the two types of equipment according to a fixed equipment sorting rule (air conditioning first, lighting second) to form a one-dimensional deviation percentage sequence, ensuring the consistency of the subsequent weight matching and calculation order.

[0056] The deviation percentage sequence is multiplied digit by digit of the weight sequence to obtain the weighted deviation values ​​for air conditioning and lighting equipment, respectively. These weighted values ​​are then incorporated into the weighting priority to improve the rationality of deviation determination. The weighted deviation value of each device is compared with its corresponding preset weighted deviation threshold to determine if the device is in an abnormal, pending control state. If the weighted deviation value of a device is greater than the corresponding preset threshold, the device is marked as a target device; if the weighted deviation value is less than or equal to the preset threshold, it is considered a normally operating device and is not marked. The preset response time length (i.e., the time required for the device to execute control commands) of each target device is retrieved. Following the device order consistent with the deviation sequence and weight sequence, the response times of the target devices are arranged sequentially to form a response time length sequence for the target devices; if there are no target devices, this sequence is empty. For example, two air conditioners and two lighting fixtures in an office area are selected as the measurement objects. The air conditioners are calculated using supply air temperature as the indicator, with a target supply air temperature of 26℃, resulting in an air conditioner deviation percentage sequence of [12.5%, 8.2%]. The lighting fixtures are calculated using illuminance as the indicator, with a target illuminance of 350 lx, resulting in a lighting deviation percentage sequence of [22.3%, 4.1%]. The weighted deviation value is calculated by multiplying the deviation percentage of a single device by the weighted sum of the corresponding device's weighted sequence after coupling correction. The core purpose is to incorporate weight priority (i.e., the degree of influence of each environmental factor on the device), avoiding the coarseness of judging a single deviation percentage and making the deviation judgment more consistent with actual working conditions. Because this example only calculates a single core operating parameter (air conditioner supply air temperature, lighting illuminance), the deviation percentage is a single value. During calculation, it is directly multiplied by the weighted sum of the corresponding device's weight sequence (the normalized weighted sum is always 1.0). If multiple parameters are calculated, the deviation percentage of each parameter needs to be multiplied by its corresponding weight before summing. The weighted sequence after coupling correction for air conditioning is [0.46, 0.39, 0.15], with a weight sum of 0.46 + 0.39 + 0.15 = 1.0; the weighted sequence after coupling correction for lighting is [0.12, 0.23, 0.65], with a weight sum of 0.12 + 0.23 + 0.65 = 1.0; the percentage deviation sequence for air conditioning is [12.5%, 8.2%] (for air conditioners 1 and 2); the percentage deviation sequence for lighting is [22.3%, 4.1%] (for lighting fixtures 1 and 2). The weighted deviation value of air conditioner No. 1 is 12.5 × 1.0 = 12.5, which becomes 11.82 after fine-tuning based on weight priority adaptation (fitting the air conditioner's operating conditions and correcting minor deviations); The weighted deviation value of air conditioner No. 2 is 8.2 × 1.0 = 8.2, which becomes 7.75 after fine-tuning; The weighted deviation value of lighting No. 1 is 22.3 × 1.0 = 22.3, which becomes 20.17 after fine-tuning; The weighted deviation value of lighting No. 2 is 4.1 × 1.0 = 4.1, which becomes 3.89 after fine-tuning.The weighted deviation value is used to accurately determine whether the equipment is in an abnormal, uncontrollable state. Its core significance lies in incorporating weighted priorities to make the deviation judgment more reasonable. Target equipment identification: A preset deviation threshold of 10 is set. After comparison, the weighted deviation values ​​of air conditioner No. 1 and lighting No. 1 both exceed the threshold, and these two devices are marked as target equipment. A response time length sequence is generated, and the response time of air conditioner No. 1 is extracted as the response time of lighting No. 1, ultimately obtaining the response time length sequence of the target equipment.

[0057] Based on the deviation percentage sequence and the response time length sequence of the target device, determine the target device parameter adjustment set;

[0058] In this embodiment, target device response fluctuation data is obtained based on the deviation percentage sequence and the target device response time length sequence to determine the target device performance evaluation result. The evaluation result is compared with a preset evaluation threshold. If the comparison result exceeds the preset evaluation threshold, the target device response time fluctuation range is obtained. Based on the target device's current configuration state and the target device's response time fluctuation range, a preliminary parameter optimization strategy is determined. Based on the preliminary parameter optimization strategy, the target device parameter adjustments are prioritized by combining the weighted deviation values ​​of each target device, the target device response time fluctuation amplitude, and the deviation exceeding the standard duration. For the sorted device parameters, a support vector machine algorithm is used to classify the sorted device parameters, determine the influence weight of each type of parameter on the device's operating efficiency, and determine the target device parameter adjustment set.

[0059] Specifically, the deviation percentage sequence and response time length sequence of the target equipment within several consecutive control cycles are retrieved, abnormal discrete values ​​in the sequence are removed, and valid time series data are retained to obtain the normalized deviation time series subset and response time time series subset.

[0060] Based on the normalized deviation time series subset, the mean of the deviation percentage, the standard deviation of the deviation percentage, and the maximum range of the deviation are calculated respectively. The standard deviation of the deviation percentage is used to characterize the dispersion of deviation fluctuation, and the maximum range of the deviation is the difference between the maximum and minimum deviation values ​​in the series. All the statistics together constitute the deviation fluctuation characteristics.

[0061] Based on the normalized response time time series subset, the response time mean, response time standard deviation, response time fluctuation range, and response time fluctuation interval are calculated respectively. The response time standard deviation is used to characterize the stability of response delay, and the fluctuation interval is the range between the minimum and maximum response times within the time series subset. All statistics together constitute the response time fluctuation characteristics.

[0062] By integrating deviation fluctuation characteristics and response time fluctuation characteristics, eliminating redundant features, and summarizing the filtered quantitative features, a complete response fluctuation dataset for the target device is formed. This dataset covers the steady-state and fluctuation characteristics of deviation and response delay.

[0063] By pre-setting weighting coefficients for each fluctuation characteristic quantity, a comprehensive performance evaluation formula is constructed. The weighted summation of each characteristic quantity in the response fluctuation dataset is performed to obtain the quantitative performance evaluation value of the target equipment. The evaluation value directly reflects the equipment's operational stability and control adaptability. The higher the evaluation value, the higher the degree of equipment deviation exceeding the standard and the worse the response stability.

[0064] The calculated quantitative performance evaluation value is compared with the preset performance evaluation threshold. Based on the comparison result, the operating performance level of the target equipment is determined, and a standardized equipment performance evaluation result is generated.

[0065] The generated equipment performance evaluation results are compared with preset evaluation thresholds. If the evaluation value does not exceed the preset threshold, the equipment is considered to be operating stably, and only routine minor parameter adjustments are performed. If the evaluation value exceeds the preset threshold, the equipment is considered to have insufficient operational stability, and fine-tuned parameter adjustments are required. The response time fluctuation range of the target equipment is then retrieved and locked. The current operating configuration status of the target equipment is collected, including the equipment's operating level, set parameters, output power, and load conditions. Combined with the response time fluctuation range, the abnormal type of equipment response—whether it is lagging, overly fast, or fluctuating—is determined. A preliminary parameter optimization strategy is then established, clarifying the direction of parameter adjustment and the basic adjustment step size, adapting to the current operating conditions and fluctuation characteristics of the equipment, and avoiding excessively large or small parameter adjustments that could lead to control failure.

[0066] Based on the established preliminary parameter optimization strategy, a priority scoring model is constructed by combining three indicators: the weighted deviation value of each target device, the response time fluctuation range, and the duration of deviation exceeding the standard. Each indicator is assigned a corresponding weighted score, and the adjustment priority score of each device parameter is calculated by weighting. The adjustment items of the target device parameters are sorted in descending order of the scores to obtain the parameter adjustment priority sequence. The higher the score, the stronger the urgency of parameter adjustment and the higher the control priority.

[0067] Historical equipment operation and control data are retrieved as training samples to construct a support vector machine (SVM) classification model. For example, historical data from 12 consecutive months of operation of three identical air conditioners and three lighting circuits within the same building are retrieved, sampled at 5-second intervals, resulting in approximately 6.3 million raw data records. After cleaning and removing outliers, 5.8 million valid samples are retained and divided into training and validation sets in a 7:3 ratio. The sorted equipment parameters are used as input features and fed into the trained SVM model. The model distinguishes the degree of influence of various parameters on equipment operating efficiency and energy-saving effects, classifying them into high-impact, medium-impact, and low-impact parameters. High-impact parameters and some medium-impact parameters are filtered out, while low-impact parameters are removed. The filtered valid parameters are then aggregated to form the final target equipment parameter adjustment set.

[0068] Determine whether the adjustment range of any single item in the target device parameter adjustment set exceeds the preset maximum single-step change threshold. If it does, divide the target electromechanical response time length sequence into an adjustment sequence containing multiple intermediate values.

[0069] In this embodiment, it is determined whether the adjustment range of each individual parameter in the target device parameter adjustment set exceeds the preset single-step maximum change threshold. If the adjustment range of any individual parameter exceeds the threshold, the response time length sequence of the corresponding target device is retrieved.

[0070] The total number of staged adjustment steps is calculated based on the response time length sequence of the target device. Based on the calculated total number of steps, multiple intermediate values ​​are generated between the current operating value of the target device and the target adjustment value using a linear interpolation method, resulting in an adjustment sequence containing multiple intermediate values.

[0071] Specifically, the system iterates through all the parameters to be adjusted in the target equipment parameter adjustment set, calculates the difference between the current operating value and the target adjustment value of each individual parameter, obtains the theoretical adjustment range of each individual parameter, and compares each theoretical adjustment range with the preset single-step maximum change threshold to determine whether there are any parameter items that exceed the threshold for adjustment.

[0072] If the theoretical adjustment range of all individual parameters does not exceed the preset maximum change threshold for a single step, the single-step parameter adjustment is performed directly according to the target adjustment value; if the theoretical adjustment range of any individual parameter exceeds the preset threshold, single-step direct adjustment is prohibited, and the target equipment response time length sequence corresponding to the parameter that exceeds the standard is retrieved as the timing basis for phased adjustment.

[0073] Based on the retrieved response time length sequence, the average response time of the sequence is calculated. Combined with the total adjustment time limit, the total number of steps for phased adjustment is calculated to ensure that the duration of each adjustment step matches the device response delay, thus ensuring that the device can stably respond to each adjustment command.

[0074] Based on the total number of phased adjustment steps calculated, a linear interpolation algorithm is used to generate multiple intermediate transition values ​​evenly between the current operating value of the target equipment and the target adjustment value. The intermediate values ​​are arranged sequentially, and abnormal interpolation results are eliminated to form a phased adjustment sequence containing multiple levels of smooth transition values, ensuring that the parameter adjustment is smooth and without shock.

[0075] Taking air conditioner supply air temperature adjustment as an example: the preset maximum single-step change threshold is ±2℃, the target air conditioner's current supply air temperature is 22℃, the target adjustment value is 28℃, and the theoretical adjustment range is 6℃, which far exceeds the maximum single-step change threshold, so staged interpolation adjustment is required.

[0076] Retrieve the air conditioner response time length sequence. If the response time length sequence is [1.2s, 1.3s, 1.4s, 1.5s, 1.6s], calculate the average response time as 1.4s, set the total adjustment time to not exceed 7s, and calculate the total number of steps for phased adjustment as 5 steps.

[0077] Using linear interpolation, intermediate adjustment values ​​are uniformly generated between 22℃ and 28℃, resulting in a phased adjustment sequence: 22℃→23.2℃→24.4℃→25.6℃→26.8℃→28℃. This step-by-step adjustment avoids excessive single-step temperature adjustment, which could lead to equipment shock and sudden changes in energy consumption.

[0078] Obtain the expected command issuance time corresponding to each intermediate value in the adjustment sequence, perform command time axis alignment operation, and detect the instantaneous fluctuation amplitude of the total power when multiple devices are adjusted in parallel in real time, and obtain a unified control command sequence based on the instantaneous fluctuation amplitude.

[0079] In this embodiment, an expected delivery time sequence is generated based on each intermediate value in the adjustment sequence, and an instruction time axis alignment operation is performed to obtain a preliminarily aligned delivery time sequence.

[0080] Based on the initial aligned release time sequence, the parallel adjustment time of multiple devices is obtained. The total power of the devices at each adjustment time is collected and calculated in real time to obtain the instantaneous fluctuation amplitude of the total power between adjacent times.

[0081] The instantaneous fluctuation amplitude is compared with a preset power fluctuation threshold. If it exceeds the preset power fluctuation threshold, the device corresponding to the instantaneous fluctuation amplitude is marked as a late-developing device.

[0082] For the subsequent sending device, the sending time is corrected to obtain a calibrated sending time sequence;

[0083] The control commands of each device are rearranged according to the calibrated issuance time sequence to obtain a unified control command sequence.

[0084] Specifically, based on each intermediate adjustment value in the phased adjustment sequence and combined with the device's single-step response delay, the expected instruction issuance time corresponding to each intermediate value is calculated sequentially, and each time is sorted in time sequence to generate an expected issuance time sequence. Using the unified clock of the edge nodes as a reference, a time axis alignment operation is performed on the expected issuance time of each target device to eliminate time sequence offset and obtain a preliminarily aligned issuance time sequence.

[0085] Based on the initial aligned time sequence, time nodes for synchronous parallel adjustment of multiple devices are defined. The operating power of each device is collected in real time at each adjustment time, and the total power of all devices at that time is calculated. The absolute value of the difference in total power between adjacent adjustment times is calculated in turn to obtain the instantaneous fluctuation amplitude of the total power.

[0086] The calculated instantaneous fluctuation amplitude is compared with the preset power fluctuation threshold. If the instantaneous fluctuation amplitude does not exceed the threshold, the original sending time is retained. If the instantaneous fluctuation amplitude exceeds the threshold, it is determined that the control at that time caused a power surge, and the corresponding device that triggered the excessive fluctuation is marked as the subsequent sending device.

[0087] For the marked late-issued equipment, the time of its instruction issuance is postponed to avoid the centralized control period, thus eliminating the power superposition impact. The issuance time of each period exceeding the standard is calibrated to obtain a calibrated issuance time sequence with regular timing and no power impact.

[0088] Based on the calibrated issuance time sequence, the adjustment commands for each stage of each target device are reordered and timed, conflicting commands are eliminated, and integrated into a unified control command sequence that is suitable for coordinated control of multiple devices and is stable and without fluctuations.

[0089] Taking the parallel group control of air conditioning units and lighting circuits as an example: the preset power fluctuation threshold is 5kW. After initial alignment, the command issuance times for air conditioning and lighting equipment are 0s, 1.4s, and 2.8s, respectively. The two devices adjust synchronously, and the instantaneous power fluctuation reaches 7.2kW, exceeding the preset threshold.

[0090] The lighting circuit was marked as a follower device, and the timing of its instruction issuance was delayed. The lighting adjustment timing was staggered to 0.7s, 2.1s, and 3.5s, and arranged alternately with the air conditioning adjustment timing. The instantaneous fluctuation of the total power at adjacent times was recalculated to be 3.1kW, which meets the threshold requirement.

[0091] After calibration, the step-by-step adjustment commands for air conditioning and lighting are arranged according to the staggered issuance sequence to form a complete and unified control command sequence. This ensures that the equipment is adjusted in place and avoids sudden increases or decreases in power caused by the simultaneous start-up, shutdown, and adjustment of multiple devices, thus ensuring the stable operation of the electromechanical equipment group control and energy saving.

[0092] This application, based on an edge computing architecture, enables refined energy-saving group control of clusters of electromechanical equipment such as air conditioners and lighting. First, it collects multi-source environmental data including outdoor temperature, indoor occupancy density, and natural light intensity. After preprocessing, it uses correlation and partial correlation coefficient matrices to eliminate interactive interference between environmental factors and completes weighted coupling correction. Next, it calculates the percentage deviation based on equipment operating parameters, combines this with the corrected weights to obtain a weighted deviation value, and filters out target equipment deviating from optimal operating conditions. Then, based on equipment deviation and response characteristics, it determines parameter adjustment strategies and prioritizes them, performing step-by-step smoothing for large adjustments, while calibrating the timing of multi-device commands and suppressing total power fluctuations. Finally, it issues control commands and uses actual equipment operating data to close the loop and correct subsequent data collection frequencies, achieving automated group control that balances energy saving and stability.

[0093] Abandoning the traditional fixed-weight model, this system precisely quantifies environmental impact through a dual-coefficient matrix and accurately identifies inefficiently operating equipment by combining weighted deviation. Targeted regulation reduces ineffective energy consumption, significantly improving overall energy efficiency. Step-by-step smoothing of over-amplitude adjustments, coupled with command timing calibration, avoids equipment shocks and power oscillations caused by sudden changes in single-step parameters and parallel control of multiple devices, reducing equipment losses and ensuring stable equipment operation. Edge computing enables localized data processing and command generation, eliminating the need for cloud uploads, shortening data transmission and response times, and rationally allocating computing power to meet the real-time control needs of electromechanical equipment. Weighted multi-indicator prioritization of adjustments achieves orderly control, addressing command conflicts and system congestion caused by simultaneous adjustment of multiple devices. By collecting actual equipment operating data, calculating response latency and adjustment errors, and dynamically correcting subsequent data collection frequencies, the control scheme adaptively optimizes with operating conditions, maintaining stable control performance over the long term.

[0094] Secondly, this embodiment provides a group control and energy-saving control system for electromechanical equipment based on edge computing, the system comprising:

[0095] The acquisition module obtains multi-source real-time parameters of the device, performs data cleaning and time alignment operations on the multi-source real-time parameters, and obtains an environmental status data set. The multi-source real-time parameters include outdoor temperature, indoor personnel density, and natural light intensity.

[0096] The first determining module calculates the correlation coefficient matrix and partial correlation coefficient matrix between the environmental state data based on the environmental state data set. Based on the correlation coefficient matrix and the partial correlation coefficient matrix, it determines the interaction intensity value sequence and weight sequence of outdoor temperature, indoor personnel density, and natural light intensity on each device, respectively, and uses the interaction intensity value sequence to couple and correct the weight sequence.

[0097] The first generation module collects the real-time operating parameters of the electromechanical equipment, obtains the deviation percentage sequence of each electromechanical equipment, multiplies it with the weighted sequence after coupling correction to obtain a weighted deviation value, compares the weighted deviation value with a preset threshold, marks the target equipment based on the comparison result, extracts the response time length of the target equipment, and forms the response time length sequence of the target equipment.

[0098] The second determining module determines the target device parameter adjustment set based on the deviation percentage sequence and the response time length sequence of the target device;

[0099] The second generation module determines whether the adjustment range of any single item in the target device parameter adjustment set exceeds the preset maximum single-step change threshold. If it does, the target electromechanical response time length sequence is divided into an adjustment sequence containing multiple intermediate values.

[0100] The third determining module obtains the expected instruction issuance time corresponding to each intermediate value in the adjustment sequence, performs instruction time axis alignment operation, and detects the instantaneous fluctuation amplitude of the total power when multiple devices are adjusted in parallel in real time, and obtains a unified control instruction sequence based on the instantaneous fluctuation amplitude.

[0101] It should be noted that, in this document, the terms "comprising," "including," and any other variations are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Specific examples have been used in this document to illustrate the principles and implementation methods of the technical solutions of this application. The above examples are only for the purpose of helping to understand the methods and core ideas of this application. The above descriptions are merely preferred embodiments of this application. It should be pointed out that, due to the limitations of written expression and the objective existence of infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this application, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes, or combinations, or the direct application of the concept and technical solutions of this application to other situations without modification, should all be considered within the scope of protection of this application.

Claims

1. A group control and energy-saving method for electromechanical equipment based on edge computing, characterized in that, The method includes: Acquire multi-source real-time parameters of the device, perform data cleaning and time alignment operations on the multi-source real-time parameters to obtain an environmental status data set, the multi-source real-time parameters including outdoor temperature, indoor personnel density, and natural light intensity; Based on the environmental state data set, the correlation coefficient matrix and partial correlation coefficient matrix between each environmental state data are calculated. Based on the correlation coefficient matrix and the partial correlation coefficient matrix, the interaction intensity value sequence and weight sequence of outdoor temperature, indoor personnel density and natural light intensity on each device are determined respectively. The weight sequence is then coupled and corrected using the interaction intensity value sequence. Real-time operating parameters of the electromechanical equipment are collected to obtain a deviation percentage sequence for each electromechanical equipment. This sequence is then multiplied by a weighted sequence after coupling correction to obtain a weighted deviation value. The weighted deviation value is compared with a preset threshold. Based on the comparison result, the target equipment is marked, and the response time length of the target equipment is extracted to form a response time length sequence for the target equipment. Based on the deviation percentage sequence and the response time length sequence of the target device, determine the target device parameter adjustment set; Determine whether the adjustment range of any single item in the target equipment parameter adjustment set exceeds the preset maximum single-step change threshold. If it does, divide the target electromechanical response time length sequence into an adjustment sequence containing multiple intermediate values. Obtain the expected command issuance time corresponding to each intermediate value in the adjustment sequence, perform command time axis alignment operation, and detect the instantaneous fluctuation amplitude of the total power when multiple devices are adjusted in parallel in real time, and obtain a unified control command sequence based on the instantaneous fluctuation amplitude.

2. The energy-saving control method for group control of electromechanical equipment based on edge computing according to claim 1, characterized in that, After obtaining the unified control instruction sequence, the method further includes: The unified control command sequence is sent to each device, and the real-time actual values ​​of the air supply temperature, air volume, and illuminance of the device are collected to obtain an updated set of device operating status information. Based on the updated set of device operating status information, the time difference between the instruction issuance time and the actual device startup response time is extracted, as well as the error correction amount between the target adjustment value and the actual operating value. Through the feedback processing module, the data acquisition frequency for the next cycle is corrected.

3. The energy-saving control method for group control of electromechanical equipment based on edge computing according to claim 1, characterized in that, Based on the environmental state data set, the correlation coefficient matrix and partial correlation coefficient matrix between each environmental state data are calculated. Based on the correlation coefficient matrix and the partial correlation coefficient matrix, the interaction intensity value sequence and weight sequence of outdoor temperature, indoor personnel density, and natural light intensity on each device are determined respectively. The weight sequence is then coupled and corrected using the interaction intensity value sequence, including: Based on the parameter values ​​in the environmental state data set, the correlation coefficient matrix and the partial correlation coefficient matrix are obtained; The coefficients between outdoor temperature, indoor population density, and natural light intensity and air conditioning equipment are extracted from the correlation coefficient matrix and partial correlation coefficient matrix to obtain the first set of coefficient sequences. The coefficients between outdoor temperature, indoor population density, natural light intensity, and lighting equipment are extracted from the correlation coefficient matrix and partial correlation coefficient matrix to obtain the second set of coefficient sequences. Based on the first set of coefficient sequences, the weight values ​​of outdoor temperature, indoor personnel density and natural light intensity on air conditioning equipment are calculated respectively to form the weight sequence of air conditioning equipment. Based on the second set of coefficient sequences, the weight values ​​of outdoor temperature, indoor population density and natural light intensity on lighting equipment are calculated respectively to form a lighting equipment weight sequence; Based on the partial correlation coefficient matrix, the interaction strength value among outdoor temperature, indoor population density and natural light intensity is calculated, and the interaction strength values ​​are arranged to form an interaction strength value sequence. By combining the interaction strength value sequence, the weight sequence of air conditioning equipment and the weight sequence of lighting equipment are coupled and corrected respectively to obtain the calibrated weight sequence of air conditioning equipment and the weight sequence of lighting equipment, which are then merged into the weight sequence after coupling correction.

4. The energy-saving control method for group control of electromechanical equipment based on edge computing according to claim 1, characterized in that, The process involves collecting real-time operating parameters of the electromechanical equipment to obtain a deviation percentage sequence for each piece of equipment. This sequence is then multiplied by a weighted sequence after coupling correction to obtain a weighted deviation value. The weighted deviation value is compared with a preset threshold. Based on the comparison result, a target device is marked, and the response time length of the target device is extracted to form a response time length sequence for the target device, including: Based on the real-time collected operating parameters of the electromechanical equipment, calculate the percentage deviation between the current supply air temperature and the target supply air temperature of the air conditioning equipment and the percentage deviation between the current illuminance and the target illuminance of the lighting equipment, and obtain the deviation percentage sequence of the air conditioning equipment and the lighting equipment. The weighted deviation value of each device is obtained by multiplying the deviation percentage sequence with the weighted sequence after coupling correction. The weighted deviation value is compared with a preset threshold. If the weighted deviation value exceeds the preset threshold, the corresponding device is marked as the target device. The response time length of the target device is extracted to form the response time length sequence of the target device.

5. The energy-saving control method for group control of electromechanical equipment based on edge computing according to claim 1, characterized in that, The step of determining the target device parameter adjustment set based on the deviation percentage sequence and the response time length sequence of the target device includes: Based on the deviation percentage sequence and the response time length sequence of the target device, the response fluctuation data of the target device is obtained, and the evaluation result of the target device performance is determined. The evaluation result is compared with a preset evaluation threshold. If the comparison result exceeds the preset evaluation threshold, the response time fluctuation range of the target device is obtained. Based on the current configuration status of the target device and the response time fluctuation range of the target device, a preliminary parameter optimization strategy is determined; Based on the preliminary parameter optimization strategy, the target equipment parameter adjustments are prioritized by combining the weighted deviation values ​​of each target device, the fluctuation range of the target device response time, and the duration of deviation exceeding the standard. For the sorted equipment parameters, the support vector machine algorithm is used to classify the sorted equipment parameters, determine the influence weight of each type of parameter on the equipment operating efficiency, and determine the target equipment parameter adjustment set.

6. The energy-saving control method for group control of electromechanical equipment based on edge computing according to claim 1, characterized in that, The step involves determining whether the adjustment magnitude of any single item in the target equipment parameter adjustment set exceeds a preset maximum single-step change threshold. If it does, the target electromechanical response time length sequence is divided into an adjustment sequence containing multiple intermediate values, including: Determine whether the adjustment range of each individual parameter in the target device parameter adjustment set exceeds the preset single-step maximum change threshold. If the adjustment range of any individual parameter exceeds the threshold, retrieve the response time length sequence of the corresponding target device. The total number of staged adjustment steps is calculated based on the response time length sequence of the target device. Based on the calculated total number of steps, multiple intermediate values ​​are generated between the current operating value of the target device and the target adjustment value using a linear interpolation method, resulting in an adjustment sequence containing multiple intermediate values.

7. The energy-saving control method for group control of electromechanical equipment based on edge computing according to claim 1, characterized in that, The process involves obtaining the expected command issuance time corresponding to each intermediate value in the adjustment sequence, performing command time axis alignment, and real-time detection of the instantaneous fluctuation amplitude of the total power during parallel adjustment of multiple devices. Based on the instantaneous fluctuation amplitude, a unified control command sequence is obtained, including: Based on the intermediate values ​​in the adjustment sequence, a predicted delivery time sequence is generated, and an instruction time axis alignment operation is performed to obtain a pre-aligned delivery time sequence. Based on the initial aligned release time sequence, the parallel adjustment time of multiple devices is obtained. The total power of the devices at each adjustment time is collected and calculated in real time to obtain the instantaneous fluctuation amplitude of the total power between adjacent times. The instantaneous fluctuation amplitude is compared with a preset power fluctuation threshold. If it exceeds the preset power fluctuation threshold, the device corresponding to the instantaneous fluctuation amplitude is marked as a late-developing device. For the subsequent sending device, the sending time is corrected to obtain a calibrated sending time sequence; The control commands of each device are rearranged according to the calibrated issuance time sequence to obtain a unified control command sequence.

8. The energy-saving control method for group control of electromechanical equipment based on edge computing according to claim 2, characterized in that, Based on the updated set of device operating status information, the time difference between the instruction issuance time and the actual device startup response time is extracted, and the error correction amount between the target adjustment value and the actual operating value is also extracted. The feedback processing module adjusts the data acquisition frequency for the next cycle, including: The issuance time of the unified control command sequence is retrieved and compared with the corresponding actual response start time of the device in the updated device operating status information set to obtain the actual response lag of the device. By iterating through all the adjustment parameters of each device, calculating the difference between the target adjustment value and the measured operating value, the cumulative adjustment error correction amount is obtained; The actual response lag of the device and the cumulative adjustment error correction are input into the feedback processing to correct the data acquisition frequency for the next cycle.

9. A group control and energy-saving control system for electromechanical equipment based on edge computing, characterized in that, The system includes: The acquisition module obtains multi-source real-time parameters of the device, performs data cleaning and time alignment operations on the multi-source real-time parameters, and obtains an environmental status data set. The multi-source real-time parameters include outdoor temperature, indoor personnel density, and natural light intensity. The first determining module calculates the correlation coefficient matrix and partial correlation coefficient matrix between the environmental state data based on the environmental state data set. Based on the correlation coefficient matrix and the partial correlation coefficient matrix, it determines the interaction intensity value sequence and weight sequence of outdoor temperature, indoor personnel density, and natural light intensity on each device, respectively, and uses the interaction intensity value sequence to couple and correct the weight sequence. The first generation module collects the real-time operating parameters of the electromechanical equipment, obtains the deviation percentage sequence of each electromechanical equipment, multiplies it with the weighted sequence after coupling correction to obtain a weighted deviation value, compares the weighted deviation value with a preset threshold, marks the target equipment based on the comparison result, extracts the response time length of the target equipment, and forms the response time length sequence of the target equipment. The second determining module determines the target device parameter adjustment set based on the deviation percentage sequence and the response time length sequence of the target device; The second generation module determines whether the adjustment range of any single item in the target device parameter adjustment set exceeds the preset maximum single-step change threshold. If it does, the target electromechanical response time length sequence is divided into an adjustment sequence containing multiple intermediate values. The third determining module obtains the expected instruction issuance time corresponding to each intermediate value in the adjustment sequence, performs instruction time axis alignment operation, and detects the instantaneous fluctuation amplitude of the total power when multiple devices are adjusted in parallel in real time, and obtains a unified control instruction sequence based on the instantaneous fluctuation amplitude.

Images