Public institution kitchen energy data management system based on cloud data

Abstract

The invention discloses a public institution kitchen energy data management system based on cloud data, particularly relates to the field of energy management, and comprises a multi-source heterogeneous data acquisition module, an edge intelligent preprocessing module, a cloud data lake storage module, a multi-dimensional analysis engine module, an intelligent decision center module and a visual interaction module. According to the method, space-time alignment of multi-source data is realized through protocol adaptive conversion and millisecond time synchronization technologies, a high-consistency analysis basis is constructed, real-time index calculation and adaptive model optimization are implemented by edge calculation and cloud dynamic modeling collaborative architecture, hidden energy loss is accurately identified, and the real-time performance of the system is improved. The intelligent decision center is combined with digital twinborn verification to generate a control strategy, a minute-level closed-loop response mechanism is formed, automatic adjustment and optimization of equipment parameters and intelligent distribution of maintenance work orders are achieved, manual intervention delay is eliminated, timeliness and execution accuracy of an energy strategy are guaranteed, and multi-system collaborative management efficiency is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of energy management, and more specifically, to a public institution kitchen energy data management system based on cloud data. Background Art

[0002] Traditional technical solutions adopt a discrete architecture, which collects equipment operating parameters by deploying independent sensors. After the data is initially processed by the local industrial computer, it is manually exported to the central server on a regular basis. The operation process includes four stages: data collection, local storage, offline analysis, and manual decision-making. The calculation cycle of key indicators usually exceeds 24 hours. Existing systems generally lack edge computing capabilities. The direct upload of raw data increases network bandwidth pressure, and differences in protocols of equipment from different brands make data integration difficult.

[0003] The existing technology has three significant defects: first, multi-source heterogeneous data lack a unified spatiotemporal benchmark, and the time deviation between thermal parameters and electrical signals often exceeds 500ms, affecting the calculation accuracy of composite indicators; second, the analysis model relies on static threshold judgment and cannot dynamically adapt to changes in operating conditions such as equipment aging and load fluctuations; finally, decision execution relies on manual intervention, and there is often a delay of several hours from anomaly identification to on-site disposal, making it difficult to achieve online optimization of carbon emissions and rapid response to equipment failures. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a public institution kitchen energy data management system based on cloud data. Through the following scheme, the traditional scheme proposed in the above background technology is solved, such as the multi-source data time deviation exceeds 500ms, the static analysis model cannot adapt to dynamic working conditions, and the manual decision-making is delayed for several hours, resulting in energy efficiency optimization lag and slow fault response.

[0005] To achieve the above objectives, the present invention provides the following technical solutions: a public institution kitchen energy data management system based on cloud data, comprising: Multi-source heterogeneous data acquisition module: responsible for real-time data acquisition of kitchen equipment and environmental parameters, covering three major categories of physical quantities: electrical, thermal, and mechanical, through an industrial-grade sensor network, and using adaptive protocol conversion technology to achieve multi-source data standardization; Edge intelligent preprocessing module: deployed on the on-site edge computing node, filters and cleans the raw data and calculates primary indicators, eliminates noise interference and compresses the data size; Cloud data lake storage module: Build a cloud-based tiered storage system and use a collaborative architecture of time series database and relational database to achieve data management; Multi-dimensional analysis engine module: Integrates dynamic modeling and machine learning algorithms to perform in-depth analysis of equipment energy efficiency, environmental thermodynamics, and equipment health; Intelligent Decision-making Central Module: Generates management strategies based on a preset rule library and digital twin simulation, realizing closed-loop management from data analysis to execution control; Visualization Interaction Module: Provides a 3D data dashboard and an interactive analysis interface, integrating real-time monitoring data and analysis decision results.

[0006] Technical Effects and Advantages of the Present Invention: The present invention innovatively constructs a multi-source data fusion system. Through edge layer protocol adaptive conversion and millisecond-level time synchronization technology, it realizes the spatio-temporal alignment of multi-dimensional data such as electrical, thermal, and mechanical data. An industrial Internet of Things gateway is used to integrate heterogeneous device interfaces, eliminating the data island problem caused by brand differences in traditional solutions, providing a high-consistency data basis for the calculation of composite energy efficiency indexes, and significantly improving the integrity of device state characterization; Through the collaborative computing architecture of deploying an edge intelligent preprocessing module and a cloud dynamic modeling engine, the present invention breaks through the limitations of traditional offline analysis. Edge nodes perform real-time calculations of primary indicators and implement data denoising. The cloud uses an adaptive weight optimization algorithm to dynamically correct the analysis model, enabling the calculation accuracy of the energy efficiency aggregation index and environmental entropy value to evolve autonomously with the device operation state, effectively identifying hidden energy efficiency losses that are difficult to capture by traditional methods; The deep integration of the intelligent decision-making center of the present invention with the device control network forms a complete perception-analysis-execution closed loop. The system automatically generates control strategies based on a preset rule library, and after verification through digital twin simulation, directly issues them to on-site devices, achieving minute-level response from anomaly detection, parameter optimization to maintenance execution. This machine-to-machine collaboration mechanism significantly reduces the frequency of manual intervention, ensuring the timeliness and execution consistency of energy management strategies. Description of the Drawings

[0007] Figure 1 It is a schematic diagram of the connection relationship of the present invention.

[0008] Figure 2 It is a schematic diagram of the system flow structure of the present invention. Detailed Embodiments

[0009] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without making creative efforts shall fall within the protection scope of the present invention.

[0010] Refer to Figure 1-2 A public institution kitchen energy data management system based on cloud data as shown, including: Multi-source heterogeneous data acquisition module: Responsible for real-time data acquisition of kitchen equipment and environmental parameters, covering three categories of physical quantities, namely electricity, heat, and mechanics, through an industrial-grade sensor network, and adopting adaptive protocol conversion technology to standardize multi-source data.

[0011] The input of the multi-source heterogeneous data acquisition module is the original signal generated by the sensor, and the output is a structured data stream with time stamps, which is directly transmitted to the edge intelligent preprocessing module, and provides a receiving interface for reverse control instructions for the intelligent decision-making center module.

[0012] The multi-source heterogeneous data acquisition module deploys a three-phase power quality analyzer in the main distribution cabinet of the kitchen equipment, and uses a CT sensor to collect the effective value of each phase current and the voltage harmonic distortion rate in real time, with a sampling frequency not less than 2 kHz; installs a turbine flowmeter with temperature compensation on the main gas pipeline, and adopts a dual-pulse output mode to ensure a measurement accuracy of ±0.5%; uses an infrared temperature measurement array on the surface of the equipment shell to collect temperature field distribution data at a frequency of 5 Hz, and uploads it to the edge computing gateway through the LoRaWAN transmission protocol.

[0013] The multi-source heterogeneous data acquisition module installs a distributed temperature and pressure composite sensor in the exhaust system, arranges a 5-point measurement array on the duct cross-section according to the ISO 3966 standard, and synchronously collects temperature gradient and dynamic pressure data; deploys an ultrasonic flowmeter at the fresh air inlet and integrates a temperature and humidity compensation algorithm; uses an explosion-proof pressure transmitter and a fast-response thermocouple for the refrigeration circuit, with a sampling interval ≤ 100 ms; configures a high-temperature resistant mass flowmeter for the steam recovery pipeline, with an internal steam dryness compensation function. All sensors achieve microsecond-level synchronization through a PROFINET ring network architecture.

[0014] The multi-source heterogeneous data acquisition module installs a three-axis MEMS accelerometer at the motor drive end to collect broadband vibration signals, with a frequency response range of 5 Hz - 15 kHz; integrates an on-line dielectric spectroscopy sensor in the lubricating oil circuit, and completes a frequency band scan of 0.1 - 10 MHz every 30 seconds; configures a broadband Rogowski coil on the power supply side to capture the current waveform at a sampling rate of 100 kHz; installs an acoustic emission sensor on the bearing seat, with a center frequency of 250 kHz, and cooperates with a 16-bit ADC to achieve a sampling rate of 40 MS / s. All high-frequency signals are transmitted to a dedicated acquisition station through optical fiber isolation, and the IEEE 1588 precise clock protocol is adopted, with the time base jitter controlled within ±5 ns.

[0015] Edge intelligent preprocessing module: Deployed at the on-site edge computing node, it filters, cleans, and calculates primary indicators for the original data, eliminates noise interference, and compresses the data scale.

[0016] The edge intelligent preprocessing module is uploaded to the cloud data lake storage module through an encrypted channel, and at the same time provides a real-time data subscription service for the multi-dimensional analysis engine module.

[0017] The edge intelligence preprocessing module performs 3σ outlier filtering and EMA exponential smoothing on the original sensor data to eliminate noise interference and fill data gaps, and then calculates primary metrics based on the cleaned standardized dataset.

[0018] The primary metrics include the device energy efficiency dynamic matrix, the environmental thermal cycle efficiency group, and the intelligent device health group; the device energy efficiency dynamic matrix includes the dynamic energy efficiency ratio, the harmonic heat loss coefficient, the instantaneous carbon emission intensity, and the thermal inertia index, the environmental thermal cycle efficiency group includes the thermosiphon effect value, the phase change hysteresis coefficient, the aerosol energy carrying rate, and the dynamic defrosting factor, and the intelligent device health group includes the friction pair health index, the electromagnetic degradation degree, the contact stress fluctuation factor, and the dynamic lubrication efficiency.

[0019] When calculating the total three-phase active power of the dynamic energy efficiency ratio, it is specifically expressed as: , DER represents the dynamic energy efficiency ratio, I k represents the current of the k-th phase, U k represents the voltage of the k-th phase, cosφ k the power factor of the k-th phase, H gas represents the calorific value of gas, Q gas represents the gas flow rate, ΔT represents the temperature difference between the surface temperature and the environment, α represents the heat dissipation coefficient, and the α value is calibrated according to the thermal conductivity and surface area of the device housing material through Fourier's law. Multiply the current of each phase by the corresponding phase voltage, then multiply by the power factor of that phase, sum them up and divide by 1000 to convert to kW unit. The gas energy consumption term is obtained by multiplying the calorific value of gas by the flow rate, and the heat dissipation term is calculated by multiplying the square of the surface temperature difference by the heat dissipation coefficient α. Finally, substitute the three terms into the formula to complete the calculation.

[0020] The harmonic heat loss coefficient is specifically expressed as: , HLC represents the harmonic heat loss coefficient, THD U represents the total harmonic distortion rate of voltage, THD I represents the total harmonic distortion rate of current, f base represents the fundamental frequency, f0 represents the nominal frequency deviation threshold, and take the product of the natural logarithm of the total harmonic distortion rates of voltage and current and the ratio of the fundamental frequency to the nominal frequency deviation threshold.

[0021] The instantaneous carbon emission intensity is specifically expressed as: , ICE represents the instantaneous carbon emission intensity, C elec represents the electricity carbon emission factor, P elec represents the real-time electric power, C gasLet \(C_{g}\) denote the gas carbon emission factor, and \(\eta(t)\) denote the time-varying combustion efficiency function. \(\eta(t)\) is dynamically calculated through a regression model established based on the flue gas temperature and oxygen content. The model coefficients are updated every 30 days using the least squares method. The electricity carbon emission term is the product of the real-time electric power and a fixed carbon emission factor, and the gas carbon emission term is the product of the flow rate and the time-varying combustion efficiency.

[0022] The specific expression of the thermal inertia index is as follows: , where \(TII\) represents the thermal inertia index, \(m\) represents the mass of the equipment, \(c\) represents the specific heat capacity of the material, \(\Delta T(t)\) represents the real-time temperature difference function, \(t_2 - t_1\) represents the monitoring period. When performing time integration on the surface temperature difference, the trapezoidal rule is used for discretized calculation, and the integration period is consistent with the thermal relaxation time of the equipment. The equipment mass and specific heat capacity are extracted from the equipment technical documentation and need to be converted to the standard SI units during calculation.

[0023] The specific expression of the thermosiphon effect value is as follows: , where \(TEV\) represents the thermosiphon effect value, \(g\) represents the acceleration due to gravity, \(h\) represents the height of the vertical air duct, \(\Delta\rho\) represents the density difference between the exhaust air and air, and \(v\) new represents the fresh air flow rate. The density difference term is calculated through the ideal gas state equation, and the exhaust air temperature, ambient temperature, and relative humidity need to be input synchronously. The acceleration due to gravity takes the standard value of \(9.81m / s^2\). The square root operation in the formula represents the driving force of natural convection, and the denominator fresh air flow rate is processed using a moving average filter.

[0024] The specific expression of the phase change hysteresis coefficient is as follows: , where \(PLC\) represents the phase change hysteresis coefficient, \(Q\) steam represents the actual steam recovery amount, \(\Delta h\) represents the enthalpy difference of the refrigerant, \(Q\) ref represents the refrigerant flow rate, and \(L\) represents the latent heat of vaporization of water vapor. When calculating the theoretical steam recovery amount, the enthalpy difference of the refrigerant is obtained by referring to the refrigerant physical property table, and the flow rate is taken from the electromagnetic flowmeter. The time differential of the absolute difference is calculated using the five-point central difference method, and the differential step size is taken as 10 seconds. The latent heat value \(L\) is fixed at \(2260kJ / kg\) and is corrected every six months according to the atmospheric pressure in the target area.

[0025] The specific expression of the aerosol energy carrying rate is as follows: , where \(AER\) represents the aerosol energy carrying rate, \(\omega\) represents the moisture content of the exhaust air, \(\rho\) oil represents the concentration of oil fume particles, \(\Delta c\) p represents the specific heat capacity difference between air and oil mist, and \(e\) is a constant. In the natural exponential term, the moisture content of the exhaust air is calculated from the dew point temperature. The specific heat capacity difference takes the fixed difference between air and typical cooking oil mist. The formula correlates humidity and heat energy loss through an exponential decay function.

[0026] The aerosol energy carrying rate is used to quantify the energy loss carried by oil fume suspensions during the heat exchange process.

[0027] The specific expression of the dynamic defrosting factor is as follows: , where DDF represents the dynamic defrosting factor, and δ i represents the frost layer thickness of each evaporator, and λ i represents the thermal conductivity of the frost layer, and t d represents the defrosting cycle, dδ / dt represents the frosting rate. The thermal conductivity of the frost layer is determined by looking up the table according to the frost density, and the density is inversely calculated by the infrared thickness gauge. The square term of the defrosting cycle amplifies the risk of long-term operation. The frosting rate is calculated by the linear regression method, and the time window takes the data of the nearest 30 minutes.

[0028] The specific expression of the friction pair health index is as follows: , where FHI represents the friction pair health index, and E 3-5kHz represents the vibration energy of 3 - 5 kHz, and ε oil represents the dielectric change rate of the lubricating oil. The calculation of the vibration energy is limited to the frequency band of 3 - 5 kHz, and the power spectrum integral is calculated by using the Hanning window FFT. The dielectric change rate of the lubricating oil takes the linear fitting slope of the data of the nearest 5 minutes. The formula is designed in the reciprocal form, and it exponentially decreases when the growth rate of the vibration energy exceeds the deterioration rate of the dielectric performance.

[0029] The specific expression of the electromagnetic degradation degree is as follows: , where EMD represents the electromagnetic degradation degree, and S I represents the Shannon entropy of the current waveform, A1 represents the fundamental wave amplitude, and A3 represents the third harmonic amplitude. The entropy value of the current waveform is calculated by the Shannon entropy formula, the sampling rate ≥ 100 kHz, and the harmonic amplitude ratio takes the ratio of the effective values of the fundamental wave and the third harmonic. The natural logarithm operation strengthens the influence of the low-order harmonics.

[0030] The specific expression of the contact stress fluctuation factor is as follows: , where CSF represents the contact stress fluctuation factor, and N AE represents the acoustic emission counting rate, and K represents the kurtosis coefficient of the vibration envelope, and f mod represents the raceway frequency modulation index. The kurtosis coefficient of the envelope is calculated by the fourth-order moment statistical method, and the raceway frequency modulation index calculates the correlation coefficient between the theoretical fault frequency and the actual spectrum through the bearing geometric parameters. The numerator term of the formula represents the impact energy, and the denominator term amplifies the abnormality of the modulation effect.

[0031] The specific expression of the dynamic lubrication efficiency is as follows: , where DLE represents the dynamic lubrication efficiency, β represents the dielectric response coefficient, dε / dt represents the gradient of the dielectric constant change, and Δf represents the main frequency offset, and f1 represents the rated frequency. The hyperbolic tangent function maps the gradient of the dielectric constant change to the interval [-1, 1]. β is determined by the accelerated aging test. The frequency offset rate is calculated as (measured main frequency - rated frequency) / rated frequency. The product relationship of the formula reflects the coupling effect of lubrication and mechanical vibration.

[0032] Cloud Data Lake Storage Module: Build a hierarchical storage system in the cloud, and adopt a collaborative architecture of time-series database and relational database to achieve data management.

[0033] The cloud data lake storage module receives the structured data from the edge preprocessing module and the calculation results of the multi-dimensional analysis engine as input, outputs a dataset stored in a normalized manner, provides a data retrieval interface for the analysis engine module, and opens a historical data access channel to the visualization interaction module.

[0034] The cloud data lake storage module establishes a hot and cold data hierarchical storage strategy. The hot data layer stores the monitoring data with high-frequency access in the past 30 days, and the cold data layer archives historical records and audit logs. The output end provides a real-time data query interface for the analysis engine module, and at the same time opens a standard ODBC data access channel to the visualization interaction module, supporting time range retrieval and device label filtering. The columnar storage optimization and adaptive compression algorithm are adopted in the data flow process to ensure the balance between storage efficiency and query performance.

[0035] Multi-dimensional Analysis Engine Module: Integrate dynamic modeling and machine learning algorithms to perform in-depth analysis of equipment energy efficiency, environmental thermodynamics, and equipment health.

[0036] The multi-dimensional analysis engine module takes the preprocessed data from the cloud data lake storage module as input, and the outputs include the equipment energy efficiency aggregation index, environmental thermodynamics entropy value, and equipment health risk value. The analysis results are pushed to the intelligent decision-making center module in real time, and at the same time, a data visualization interface is provided for the visualization interaction module.

[0037] The specific expression of the equipment energy efficiency aggregation index is: , where EEAI represents the equipment energy efficiency aggregation index, and α1 represents the normalization coefficient.

[0038] By introducing the exponential scaling effect of energy transfer, the equipment energy efficiency aggregation index first amplifies the characteristics of the high-efficiency area of the dynamic energy efficiency ratio DER by the 1.5th power, and at the same time weakens the linear influence of harmonic loss by the 0.3rd root of HLC. The carbon emission term is designed as a saturation function (1 - ICE / ICE_0), where ICE_0 takes the local environmental protection standard limit value to achieve an over-standard penalty mechanism. The thermal inertia term uses a logarithmic function to process the relationship between TII and the minimum threshold TII_min to avoid insufficient sensitivity in the low-temperature section. α is calibrated through typical working conditions, and α = 2.718 is taken to normalize EEAI to 100 in the reference state. The constant value is verified and determined from 10 groups of historical data through the 3σ principle.

[0039] The specific expression of the environmental thermodynamics entropy value is: , where ETHE represents the environmental thermodynamics entropy value, γ represents the thermodynamic coupling coefficient, and δ represents the weight coefficient.

[0040] The construction of the environmental thermodynamic entropy value conducts a two-channel coupling analysis. The main term processes the synergistic effect of the thermosiphon effect TEV and the phase change hysteresis PLC through dimensionless treatment, eliminates the dimension by dividing the geometric mean by the square root of the aerosol energy carrying rate AER, and optimizes the sensitivity curve with a 0.7th power. The secondary term uses the tangent function to describe the asymptotic saturation characteristics of the defrosting factor DDF and the maximum value DDF_max. δ is used as the weight coefficient, γ takes the reciprocal of the Carnot efficiency of the refrigeration system, δ is determined to be 0.618 through the heat exchanger efficiency experiment, DDF_max takes the maximum value in the recent 30 days using the moving window method, and the π / 2 coefficient in the formula ensures the continuity of the function within the domain and avoids divergence.

[0041] The device health risk value is specifically expressed as: , where EHRV represents the device health risk value, η represents the electromechanical coupling risk coefficient, θ represents the lubrication deterioration sensitivity coefficient, and DLE opt represents the reference value of the dielectric property of the lubricating oil.

[0042] The device health risk value adopts a dual-mode risk superposition architecture. The electromagnetic and mechanical composite risk term characterizes the electromechanical coupling effect through the product of EMD and CSF, weakens the basic influence of the health index with the 0.2nd root of FHI, and strengthens the abnormal state identification with a 1.8th power. The lubrication deviation term constructs a non-linear penalty function using the absolute value and a 2.5th power. DLE_opt takes the median of the optimal value interval recommended by the manufacturer, η is determined by Weibull distribution fitting, and θ is inversely calculated based on the lubricating oil replacement cycle.

[0043] Intelligent decision-making central module: Generates management strategies based on a preset rule base and digital twin simulation, and realizes closed-loop management from data analysis to execution control.

[0044] The intelligent decision-making central module receives the index calculation results of the multi-dimensional analysis engine module as input, outputs device control instructions and maintenance work orders, issues control parameter adjustment commands to the multi-source heterogeneous data acquisition module, and synchronizes the decision-making log to the visual interaction module.

[0045] The intelligent decision-making central module establishes a hierarchical response management system according to the analysis results of the multi-dimensional analysis engine module, including energy efficiency aggregation index management, thermodynamic entropy value management, and health risk value management.

[0046] When the energy efficiency aggregation index management has EEAI < 85, it immediately adjusts the combustion efficiency parameter η(t) of the gas equipment. According to the equation η_new = η_old + k1ln(EEAI / 85), it starts the harmonic governance device, aiming to reduce HLC to: HLC_target = HLC_current×(85 / EEAI)^(2 / 3). It optimizes the production scheduling plan, and switches the equipment with thermal inertia TII > 2.5 to operate during the valley electricity period. η_old represents the current combustion efficiency coefficient, η_new is the target efficiency coefficient after regulation, k1 = 0.2 is the combustion adjustment coefficient, HLC_target represents the optimized target value of the harmonic loss system, and HLC_current represents the currently measured harmonic loss coefficient. When EEAI > 115, it activates the energy storage system to participate in peak shaving, and the adjustment power is: P_ESS = (EEAI - 115) / 10×P_base, where P_ESS represents the dynamic adjustment power of the energy storage system, and P_base represents the equipment reference power.

[0047] When the thermodynamic entropy value management has 45 < ETHE < 55, it dynamically adjusts the exhaust fan speed to make the TEV value meet: TEV_opt = (0.6ETHE + 18) ± 0.1v_new. It optimizes the defrosting cycle of the refrigeration system: t_d' = t_d×[1 - 0.05(ETHE - 50)], triggers the steam recovery enhancement mode, aiming to reduce the PLC value by 10 - 15%. TEV_opt represents the optimal value of the thermosiphon effect, v_new represents the exhaust system speed adjustment amount, t_d' represents the optimized defrosting cycle, and t_d represents the standard defrosting cycle parameter originally set by the equipment. When ETHE > 60, it starts the emergency ventilation protocol, and the fresh air flow is increased to: Q_new = min(2Q_base, Q_max×tanh(ETHE / 70)), and it forcibly implements the cleaning of the heat exchanger, aiming to reduce AER to: AER_target = AER_current×e^(-0.03(ETHE - 60)). Q_new represents the fresh air flow for environmental control, Q_base represents the reference fresh air flow parameter under the design condition, Q_max represents the engineering limit value of the maximum allowable fresh air flow in the ventilation system, AER_target represents the aerosol energy-carrying rate control target, and AER_current represents the currently measured aerosol energy-carrying rate parameter.

[0048] When the health risk value management has an EHRV > 7.5, it automatically generates a maintenance work order and gives priority to processing the subsystem corresponding to max{EMD×0.4, CSF×0.6, (1 - DLE)×9.0}, adjusts the equipment load to: P_safe = P_rated×[1 - 0.15(EHRV - 7)], activates the high-frequency monitoring mode, and increases the vibration sampling rate to 20 kHz; when EHRV > 9.0 continuously, it triggers the interlock shutdown protection, satisfying: shutdown threshold = 9.0 + 0.3∑(EHRV_daily_Δ), enables the digital twin system to conduct fault simulation, with input parameters: {FHI_current, CSF_history, DLE_trend}. P_safe represents the equipment safe operating power threshold, P_rated represents the rated power technical parameter marked on the equipment nameplate, EHRV_daily_Δ represents the daily change amount of the equipment health risk index, FHI_current represents the real-time measurement value of the equipment health index, CSF_history represents the historical data sequence of the current waveform entropy value, and DLE_trend represents the lubrication dielectric value change trend function.

[0049] Visualization interaction module: Provides a three-dimensional data dashboard and an interactive analysis interface, integrating real-time monitoring data and analysis and decision-making results.

[0050] The visualization interaction module inputs and obtains the historical data of the cloud data lake storage module and the real-time calculation results of the multi-dimensional analysis engine module, outputs visualization charts and warning information, provides a decision support interface for management personnel, and exports a structured report to the external audit system.

[0051] The present invention is based on the collaborative work of three-layer architectures of the application layer, edge layer, and cloud platform. The specific process is as follows: In the edge layer, the multi-source heterogeneous data acquisition module collects electrical, thermal, and mechanical parameters of kitchen equipment in real time through an industrial sensor network, including three-phase power quality, gas flow rate, temperature field distribution, vibration signals, and environmental parameters. The adaptive protocol conversion technology is used to standardize the original signals into time-stamped structured data streams; the edge intelligent preprocessing module is deployed on the field computing node to perform 3σ outlier filtering and EMA exponential smoothing processing on the original data, calculate primary indicators such as dynamic energy efficiency ratio, harmonic heat loss coefficient, and thermosiphon effect value, and upload them to the cloud data lake storage module in the cloud platform layer through an encrypted channel; the cloud platform adopts a collaborative architecture of time series database and relational database to implement hierarchical storage of hot and cold data. The hot data layer stores high-frequency access data within 30 days, and the cold data layer archives historical records. The multi-dimensional analysis engine module calls the stored data to perform in-depth analysis, and generates decision-making basis through the composite calculation of dynamic energy efficiency aggregation index, thermodynamic entropy value, and equipment health risk value; the intelligent decision-making center module generates control strategies based on the preset rule library and digital twin simulation, issues device parameter adjustment instructions to the edge layer, and generates maintenance work orders; the visualization interaction module in the application layer integrates a three-dimensional data dashboard and an interaction interface, and real-time displays monitoring data, analysis results, and warning information, realizes historical data traceability and structured report export through the standard ODBC interface, and forms a full closed-loop management from data acquisition, edge processing, cloud analysis to decision feedback.

[0052] Secondly, in the attached drawings of the disclosed embodiments of the present invention, only the structures related to the disclosed embodiments are involved. For other structures, reference can be made to the general design. Without conflict, the same embodiment and different embodiments of the present invention can be combined with each other; Finally, the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A public institution kitchen energy data management system based on cloud data, characterized in that, Including: Multi-source heterogeneous data acquisition module: Responsible for real-time data acquisition of kitchen equipment and environmental parameters, covering three major physical quantities of electricity, heat, and machinery through an industrial-grade sensor network, and adopting adaptive protocol conversion technology to standardize multi-source data; Edge intelligent preprocessing module: Deployed on-site edge computing nodes, filtering, cleaning, and calculating primary indicators for the original data to eliminate noise interference and compress the data scale; The primary indicators include the equipment energy efficiency dynamic matrix, the environmental heat cycle efficiency group, and the intelligent device health group; the equipment energy efficiency dynamic matrix includes the dynamic energy efficiency ratio, the harmonic heat loss coefficient, the instantaneous carbon emission intensity, and the thermal inertia index, the environmental heat cycle efficiency group includes the thermosiphon effect value, the phase change hysteresis coefficient, the aerosol energy-carrying rate, and the dynamic defrosting factor, and the intelligent device health group includes the friction pair health index, the electromagnetic degradation degree, the contact stress fluctuation factor, and the dynamic lubrication efficiency; Cloud data lake storage module: Construct a cloud-based hierarchical storage system, and adopt a collaborative architecture of time-series database and relational database to achieve data management; Multi-dimensional analysis engine module: Integrate dynamic modeling and machine learning algorithms to perform in-depth analysis of equipment energy efficiency, environmental thermodynamics, and equipment health; The multi-dimensional analysis engine module inputs the preprocessed data from the cloud data lake storage module, and the outputs include the equipment energy efficiency aggregation index, the environmental thermodynamics entropy value, and the equipment health risk value, and pushes the analysis results to the intelligent decision-making center module in real time, and at the same time provides a data visualization interface for the visualization interaction module; Intelligent decision-making center module: Generate management strategies based on a preset rule library and digital twin simulation to achieve closed-loop management from data analysis to execution control; Visualization interaction module: Provide a three-dimensional data dashboard and an interactive analysis interface to integrate real-time monitoring data and analysis and decision-making results.

2. The public institution kitchen energy data management system based on cloud data according to claim 1, characterized in that: The input of the multi-source heterogeneous data acquisition module is the original signal generated by the sensor, and the output is a structured data stream with time stamps, which is directly transmitted to the edge intelligent preprocessing module, and provides a receiving interface for the reverse control instruction of the intelligent decision-making center module; The edge intelligent preprocessing module is uploaded to the cloud data lake storage module through an encrypted channel, and at the same time provides a real-time data subscription service to the multi-dimensional analysis engine module; The edge intelligent preprocessing module performs 3σ outlier filtering and EMA exponential smoothing processing on the original sensor data to eliminate noise interference and fill in data missing, and then calculates primary indicators based on the cleaned standardized data set; The cloud data lake storage module inputs and receives the structured data from the edge preprocessing module and the calculation results of the multi-dimensional analysis engine, and the output is a dataset stored in a standardized manner, provides a data retrieval interface for the analysis engine module, and opens a historical data access channel to the visualization interaction module; The intelligent decision-making center module receives the indicator calculation results of the multi-dimensional analysis engine module, outputs equipment control instructions and maintenance work orders, sends control parameter adjustment commands to the multi-source heterogeneous data acquisition module, and synchronizes the decision log to the visual interaction module; The intelligent decision-making center module establishes a hierarchical response management system according to the analysis results of the multi-dimensional analysis engine module, including energy efficiency aggregation index management, thermodynamic entropy value management and health risk value management; The visualization interaction module inputs and obtains the historical data of the cloud data lake storage module and the real-time calculation results of the multi-dimensional analysis engine module, outputs visualization charts and warning information, provides a decision support interface for managers, and exports structured reports to the external audit system.

3. The cloud-based public institution kitchen energy data management system according to claim 2, characterized in that: When calculating the total three-phase active power of the dynamic energy efficiency ratio, it is specifically expressed as: , where DER represents the dynamic energy efficiency ratio, I k represents the current of the k-th phase, U k represents the voltage of the k-th phase, cosφ k is the power factor of the k-th phase, H gas represents the calorific value of the gas, Q gas represents the gas flow rate, ΔT represents the temperature difference between the surface temperature and the environment temperature, α represents the heat dissipation coefficient, and the value of α is calibrated according to the thermal conductivity and surface area of the equipment housing material through Fourier's law. Multiply the current of each phase by the corresponding phase voltage, then multiply by the power factor of that phase, sum them up and divide by 1000 to convert to the unit of kW. The gas energy consumption term is obtained by multiplying the calorific value of the gas by the flow rate, and the heat dissipation term is calculated by multiplying the square of the surface temperature difference by the heat dissipation coefficient α. Finally, substitute the three terms into the formula to complete the calculation; The harmonic heat loss coefficient is specifically expressed as: , where HLC represents the harmonic heat loss coefficient, and THD U represents the total harmonic distortion rate of voltage, and THD I represents the total harmonic distortion rate of current, f base represents the fundamental frequency, and f0 represents the nominal frequency deviation threshold, which is obtained by multiplying the natural logarithm of the ratio of the total harmonic distortion rates of voltage and current to the deviation threshold of the fundamental frequency relative to the nominal frequency; The instantaneous carbon emission intensity is specifically expressed as: , where ICE represents the instantaneous carbon emission intensity, C elec represents the carbon emission factor of electricity, P elec represents the real-time electric power, C gas represents the carbon emission factor of gas, η(t) represents the time-varying combustion efficiency function, η(t) is dynamically calculated through a regression model established by the flue gas temperature and oxygen content, and the model coefficients are updated every 30 days using the least squares method. The electricity carbon emission term is the real-time electric power multiplied by the fixed carbon emission factor, and the gas carbon emission term is the flow rate multiplied by the time-varying combustion efficiency; The specific expression of the thermal inertia index is as follows: , where TII represents the thermal inertia index, m represents the mass of the device, c represents the specific heat capacity of the material, ΔT(t) represents the real-time temperature difference function, t2 - t1 represents the monitoring period. When performing time integration on the surface temperature difference, the trapezoidal rule is used for discretized calculation, and the integration period is consistent with the thermal relaxation time of the device. The device mass and specific heat capacity are extracted from the device technical documentation and need to be converted to standard SI units during calculation.

4. The cloud-based public institution kitchen energy data management system according to claim 2, characterized in that: The value of the thermosiphon effect is specifically expressed as: , where TEV represents the value of the thermosiphon effect, g represents the acceleration due to gravity, h represents the height of the vertical air duct, Δρ represents the density difference between the exhaust air and the air, and v new represents the fresh air flow rate. The density difference term is calculated by the ideal gas state equation, and the exhaust air temperature, ambient temperature, and relative humidity need to be input synchronously. The acceleration due to gravity takes the standard value of 9.81 m / s². The square root operation in the formula represents the driving force of natural convection, and the fresh air flow rate in the denominator is processed by moving average filtering; The phase change hysteresis coefficient is specifically expressed as: , where PLC represents the phase change hysteresis coefficient, Q steam represents the actual steam recovery amount, Δh represents the refrigerant enthalpy difference, Q ref represents the refrigerant flow rate, L represents the latent heat of water vapor. When calculating the theoretical steam recovery amount, the refrigerant enthalpy difference is obtained by referring to the refrigerant property table, the flow rate is taken from the electromagnetic flowmeter, the time differential of the absolute difference is calculated using the five-point central difference method, the differential step size is taken as 10 seconds, and the latent heat value L is fixed at 2260 kJ / kg, which is corrected every six months according to the atmospheric pressure in the target area; The specific expression of the energy-carrying rate of the aerosol is as follows: , where AER represents the energy-carrying rate of the aerosol, ω represents the moisture content of the exhaust air, and ρ oil represents the concentration of oil fume particles, and Δc p represents the specific heat capacity difference between air and oil mist. e is a constant. The moisture content of the exhaust air in the natural exponential term is calculated from the dew point temperature. The specific heat capacity difference takes a fixed difference between air and typical edible oil mist. The formula correlates humidity and heat energy loss through an exponential decay function; The specific expression of the dynamic defrosting factor is as follows: , where DDF represents the dynamic defrosting factor, and δ i represents the frost layer thickness of each evaporator, and λ i represents the frost layer thermal conductivity, and t d represents the defrosting cycle, dδ / dt represents the frosting rate. The frost layer thermal conductivity is determined by looking up the table according to the frost density, and the density is inversely calculated by an infrared thickness gauge. The square term of the defrosting cycle amplifies the risk of long-term operation. The frosting rate is calculated by the linear regression method, and the time window takes the data of the most recent 30 minutes.

5. The cloud-based public institution kitchen energy data management system according to claim 2, characterized in that: The friction pair health index is specifically expressed as: , where FHI represents the friction pair health index, and E 3-5kHz represents the vibration energy at 3 - 5 kHz, and ε oil represents the lubricating oil dielectric change rate. The vibration energy calculation is limited to the 3 - 5 kHz frequency band, and the power spectrum integral is calculated using the Hanning window FFT. The lubricating oil dielectric change rate takes the linear fitting slope of the data in the most recent 5 minutes. The formula is designed in a reciprocal form, and the index decreases when the vibration energy growth rate exceeds the dielectric performance degradation rate. The specific expression of the electromagnetic degradation degree is as follows: , where EMD represents the electromagnetic degradation degree, S I represents the Shannon entropy of the current waveform, A1 represents the fundamental wave amplitude, A3 represents the third harmonic amplitude. The entropy value of the current waveform is calculated by the Shannon entropy formula, the sampling rate ≥ 100 kHz, the harmonic amplitude ratio takes the ratio of the effective values of the fundamental wave and the third harmonic, and the natural logarithm operation strengthens the influence of the low-order harmonics; The contact stress fluctuation factor is specifically expressed as: , where CSF represents the contact stress fluctuation factor, N AE represents the acoustic emission count rate, K represents the kurtosis coefficient of the vibration envelope, f mod represents the raceway frequency modulation index. The kurtosis coefficient of the envelope is calculated using the fourth-order moment statistical method. The raceway frequency modulation index calculates the correlation coefficient between the theoretical fault frequency and the actual spectrum through the bearing geometric parameters. The numerator term of the formula characterizes the impact energy, and the denominator term amplifies the abnormality of the modulation effect; The specific expression of the dynamic lubrication efficiency is as follows: , where DLE represents the dynamic lubrication efficiency, β represents the dielectric response coefficient, dε / dt represents the gradient of dielectric constant change, Δf represents the main frequency offset, f1 represents the rated frequency. The hyperbolic tangent function maps the gradient of dielectric constant change to the interval [-1, 1]. β is determined through an accelerated aging test. The frequency offset rate is calculated as (measured main frequency - rated frequency) / rated frequency. The product relationship in the formula reflects the coupling effect between lubrication and mechanical vibration.

6. The public institution kitchen energy data management system based on cloud data according to claim 1, characterized in that: The cloud data lake storage module establishes a tiered storage strategy for hot and cold data. The hot data layer stores the monitoring data with high frequency access in the past 30 days, and the cold data layer archives historical records and audit logs. The output end provides a real-time data query interface for the analysis engine module, and opens a standard ODBC data access channel to the visualization interaction module, supporting time range retrieval and device tag filtering. The data flow process adopts column storage optimization and adaptive compression algorithm to ensure a balance between storage efficiency and query performance.

7. The cloud-based public institution kitchen energy data management system according to claim 2, characterized in that: The device energy efficiency aggregation index is specifically expressed as: , where EEAI represents the device energy efficiency aggregation index and α1 represents the normalization coefficient; The environmental thermodynamic entropy value is specifically expressed as: , where ETHE represents the environmental thermodynamic entropy value, γ represents the thermodynamic coupling coefficient, and δ represents the weight coefficient; The specific expression of the device health risk value is as follows: , where EHRV represents the device health risk value, η represents the electromechanical coupling risk coefficient, θ represents the lubrication deterioration sensitivity coefficient, and DLE opt represents the reference value of the dielectric property of the lubricating oil.

8. The public institution kitchen energy data management system based on cloud data according to claim 2, characterized in that: The energy efficiency aggregation index management immediately adjusts the combustion efficiency parameter η(t) of the gas equipment when EEAI<85, and starts the harmonic control device according to the equation η_new=η_old+k1ln(EEAI / 85), with the goal of reducing the HLC to: HLC_target=HLC_current×(85 / EEAI)^(2 / 3) to optimize the production schedule, and convert the equipment with thermal inertia TII>2.5 to operate during the valley power period, η_old represents the current combustion efficiency coefficient, η_new is the target efficiency coefficient after regulation, k1=0.2 is the combustion regulation coefficient, HLC_target represents the harmonic loss system optimization target value, and HLC_current represents the harmonic loss coefficient actually measured at present; when EEAI>115, the energy storage system is activated to participate in the peak regulation, and the regulated power is: P_ESS=(EEAI-115) / 10×P_base, P_ESS represents the dynamic regulation power of the energy storage system, and P_base represents the equipment benchmark power.

9. The public institution kitchen energy data management system based on cloud data according to claim 2, characterized in that: The thermodynamic entropy value management dynamically adjusts the speed of the exhaust fan when 45 < ETHE < 55, so that the TEV value satisfies: TEV_opt = (0.6ETHE + 18) ± 0.1v_new, optimizes the defrosting cycle of the refrigeration system: t_d' = t_d × [1 - 0.05(ETHE - 50)], triggers the enhanced steam recovery mode, and the target PLC value drops by 10 - 15%. TEV_opt represents the optimal value of the thermosiphon effect, v_new represents the adjustment amount of the exhaust system speed, t_d' represents the optimized defrosting cycle, and t_d represents the standard defrosting cycle parameter originally set by the equipment; when ETHE > 60, start the emergency ventilation protocol, and the fresh air flow rate is increased to: Q_new = min(2Q_base, Q_max × tanh(ETHE / 70)), and the heat exchanger cleaning is forced to be carried out to reduce the AER to: AER_target = AER_current × e^(-0.03(ETHE - 60)). Q_new represents the fresh air flow rate for environmental control, Q_base represents the reference fresh air flow rate parameter under the design condition, Q_max represents the engineering limit value of the maximum allowable fresh air flow rate of the ventilation system, AER_target represents the aerosol energy-carrying rate control target, and AER_current represents the currently measured aerosol energy-carrying rate parameter.

10. A public institution kitchen energy data management system based on cloud data according to claim 2, characterized in that: The health risk value management automatically generates a maintenance work order when EHRV > 7.5, and preferentially processes the subsystem corresponding to max{EMD×0.4, CSF×0.6, (1 - DLE)×9.0}. EMD represents the electromagnetic degradation degree, adjusts the equipment load to: P_safe = P_rated × [1 - 0.15(EHRV - 7)], activates the high-frequency monitoring mode, and the vibration sampling rate is increased to 20 kHz; when EHRV > 9.0 continuously, triggers the interlock shutdown protection, satisfying: shutdown threshold = 9.0 + 0.3∑(EHRV_daily_Δ), enables the digital twin system to perform fault simulation, and inputs the parameters: {FHI_current, CSF_history, DLE_trend}. P_safe represents the equipment safe operating power threshold, P_rated represents the rated power technical parameter marked on the equipment nameplate, EHRV_daily_Δ represents the daily change amount of the equipment health risk index, FHI_current represents the measured value of the real-time equipment health index, CSF_history represents the historical data sequence of the current waveform entropy value, and DLE_trend represents the lubricating dielectric value change trend function.

Images