Zero-carbon factory energy consumption and air pressure efficiency integrated monitoring platform

By using multi-source time-series alignment and dynamic carbon-efficiency coupling modeling, the data misalignment problem between compressed air system carbon emission monitoring and zero-carbon target management is solved, enabling real-time quantification and causal analysis of carbon emissions, and supporting precise decision-making by factory managers.

CN122797966BActive Publication Date: 2026-11-10GUANGDONG XINZHUAN ENERGY SAVING TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202611296037.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-08-25
Publication Date
2026-11-10
Estimated Expiration
2046-08-25

AI Technical Summary

Technical Problem

Existing compressed air system carbon emission monitoring and zero-carbon target management suffer from data time-series misalignment and a lack of real-time dynamic evaluation methods. This makes it impossible to accurately quantify the impact of factors such as air pressure fluctuations and pipeline leaks on carbon emissions, leading to carbon emission calculation errors and management difficulties.

Method used

A multi-source timing alignment module is used to synchronize high-frequency aerodynamic operating parameters and low-frequency energy consumption data. A dynamic carbon efficiency coupling modeling module maps air pressure efficiency to dynamic carbon emission equivalent. A pneumatic carbon contribution evaluation module is used to quantify the contribution or drag of the compressed air system to the zero-carbon goal.

Benefits of technology

It enables real-time, dynamic, traceable quantification of carbon emissions from compressed air systems, accurately identifies the specific contributions of factors such as air pressure fluctuations and pipeline leaks to carbon emissions, supports factory managers in quickly identifying and locating the main causes, and provides data support for zero-carbon target management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122797966B_ABST
    Figure CN122797966B_ABST
Patent Text Reader

Abstract

This invention provides an integrated monitoring platform for energy consumption and pressure efficiency in zero-carbon factories, relating to the fields of industrial energy conservation and intelligent manufacturing. It includes modules for multi-source time-series alignment, dynamic carbon-efficiency coupling modeling, and aerodynamic carbon contribution evaluation. The invention uses a multi-source time-series alignment module to accurately align high-frequency aerodynamic parameters with low-frequency energy consumption data using timestamps. A dynamic carbon-efficiency coupling modeling module converts multi-dimensional aerodynamic indicators such as pressure fluctuations, pipeline pressure drop, and leakage into dynamic carbon emission equivalents at each level. An aerodynamic carbon contribution evaluation module calculates an aerodynamic carbon contribution index, quantifying the system's drag on zero-carbon goals. A softmax function outputs a causal weight vector and matching investigation guidance labels, achieving integrated monitoring of compressed air system energy consumption and pressure efficiency under carbon constraints, and precise source tracing and diagnosis of carbon emission contributions.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of industrial energy conservation and intelligent manufacturing, specifically to an integrated monitoring platform for energy consumption and air pressure efficiency in zero-carbon factories. Background Technology

[0002] In the construction of zero-carbon factories, electricity consumption, the operating efficiency of compressed air systems, and the resulting carbon emissions have become core concerns for manufacturing enterprises' energy management. As a major electricity consumer within the factory, the supply efficiency, distribution status, and end-user air consumption behavior of compressed air systems directly affect the factory's overall energy consumption and carbon emissions performance.

[0003] Compressed air system management technology has evolved from single-unit monitoring to integrated energy and carbon assessment. Early systems primarily relied on sensors to collect equipment-level data and manual judgment based on experience. This was followed by station-level energy efficiency assessments, which effectively reduced equipment idle rates through flow monitoring and load analysis. In recent years, empowered by the Internet of Things (IoT) and artificial intelligence (AI), the system has entered the integrated energy and carbon monitoring stage, achieving real-time tracking of energy consumption and carbon emission indicators based on a cloud-edge-device architecture.

[0004] However, the following shortcomings still exist in the accurate monitoring and quantitative evaluation of carbon emissions from compressed air systems for zero-carbon factories: The air pressure and flow data of the compressed air system are high-frequency dynamic data (second-level / millisecond-level), while the factory's total electricity meter, carbon emission conversion data, green electricity ratio data, and carbon emission quota limit are all low-frequency management data. The sampling periods of the two types of data are significantly different. If the high-frequency data is simply extracted based on the timestamp of the low-frequency data, or if numerical matching is performed directly, it will lead to the accumulation of time sequence misalignment errors, resulting in significant deviations in subsequent carbon emission calculations.

[0005] Traditional energy efficiency assessment methods can only reflect the energy consumption level of equipment under specific operating conditions, and cannot dynamically convert real-time aerodynamic / thermodynamic indicators such as air pressure fluctuations, pipeline pressure drop, leakage rate, and equipment start-up and shutdown frequency into corresponding carbon emission equivalents. Zero-carbon factory carbon emission accounting requires real-time, dynamic, and traceable carbon data, but existing methods can only provide static energy efficiency labels. Furthermore, existing monitoring systems lack the ability to estimate transmission-side factors such as pipeline pressure drop and leakage rate online, and cannot quantify the additional carbon emissions caused by decreased pipeline transmission efficiency.

[0006] While existing monitoring systems can output energy efficiency ratings for air compressor units, these ratings only assess the energy consumption of the equipment itself under standard operating conditions. They cannot dynamically characterize the contribution or drag on the plant's overall zero-carbon goals based on the current operating status. When total carbon emissions approach the limit, managers struggle to quantify the impact of the compressed air system on carbon emissions exceeding the limit based on existing monitoring data. They also cannot distinguish whether the main factors causing carbon emissions to deviate from the target are due to decreased air compressor unit efficiency, increased pipeline transport losses, or a mismatch between operating strategies and load demand.

[0007] The accumulation of these three issues has resulted in a long-term disconnect between carbon emission monitoring of compressed air systems and zero-carbon target management. Summary of the Invention

[0008] The purpose of this invention is to provide an integrated monitoring platform for energy consumption and air pressure efficiency in zero-carbon factories, aiming to solve the technical problems that the existing compressed air system operation evaluation mainly focuses on energy efficiency, is difficult to couple with the factory's zero-carbon goals in real time, and lacks a means of quantifying the contribution to the zero-carbon goals.

[0009] To achieve the above objectives, the present invention adopts the following technical solution: The zero-carbon factory energy consumption and pressure efficiency integrated monitoring platform includes a multi-source time-series alignment module, a dynamic carbon efficiency coupling modeling module, and an aerodynamic carbon contribution evaluation module that are interconnected. Among them: The multi-source time-series alignment module is configured to synchronously collect high-frequency aerodynamic operating parameters of the compressed air system and low-frequency energy consumption data and green electricity ratio data of the factory power grid, and obtain the current time window carbon emission quota upper limit issued by the factory energy management system through the data interface. It performs timestamp alignment on the high-frequency aerodynamic operating parameters and low-frequency energy consumption data, and associates and matches the green electricity ratio data and the carbon emission quota upper limit with the corresponding time window to output a time-series synchronized fusion feature dataset. The dynamic carbon efficiency coupling modeling module is configured to receive the fused feature dataset, construct a pressure efficiency-carbon emission mapping model, which is used to convert the aerodynamic operating state into a dynamic carbon emission equivalent and output the converted dynamic carbon emission equivalent. The pneumatic carbon contribution evaluation module is configured to receive the dynamic carbon emission equivalent, calculate the pneumatic carbon contribution index based on the benchmark carbon emission equivalent, determine the degree of contribution or drag on the zero-carbon target of the current operating state of the compressed air system according to the numerical range of the pneumatic carbon contribution index, and output early warning information.

[0010] As a preferred embodiment of this application, the multi-source time-series alignment module includes an edge gateway deployed on the air compressor station side, a factory-level energy consumption acquisition module, a carbon emission data interface module, and a time-series database; The edge gateway connects to the local controller of the air compressor via an industrial bus or Modbus protocol to collect high-frequency pneumatic operating parameters, including air compressor outlet pressure, main pipe pressure, terminal pressure, instantaneous flow rate, motor current, active power, and exhaust temperature data. The factory-level energy consumption acquisition module connects to the factory's main electricity meter via a power gateway to collect low-frequency energy consumption data. The carbon emission data interface module connects to the power grid carbon factor data platform via an API interface to obtain green electricity ratio data and power structure prediction data, and connects to the factory energy management system via an API interface to obtain the carbon emission quota limit allocated to the current time window in real time. The time-series database is used to store the fused feature dataset after time-series synchronization.

[0011] As a preferred embodiment of this application, the multi-source time-series alignment module is further configured to: perform time window merging on the collected high-frequency data stream with a preset window width, perform steady-state identification and feature extraction on the data subset within each window, and calculate the feature vector used to characterize the operating state of the window; and map the feature vector of each window onto the time coordinate axis of the low-frequency data using the timestamp of the low-frequency energy consumption data as the interpolation node, thereby forming a timestamp-aligned same-source time window dataset.

[0012] As a preferred embodiment of this application, the dynamic carbon efficiency coupling modeling module includes a nonlinear performance mapping sub-model for air compressors, a real-time estimation sub-model for pipeline pressure drop and leakage, a power consumption increment conversion sub-model, and a dynamic carbon emission equivalent conversion model. The nonlinear performance mapping sub-model of the air compressor is configured to receive the speed, exhaust pressure and load rate data within the current time window and output the specific power characteristic value under the current working condition. The real-time estimation sub-model for pipeline pressure drop and leakage is configured to receive the fields of total pipeline pressure, terminal pressure and flow rate change rate in the current time window from the fused feature dataset, estimate the pipeline pressure drop in real time, and estimate the real-time leakage rate based on the pipeline pressure decay data collected during non-production periods. The power consumption increment calculation sub-model is configured to calculate the additional power consumption caused by the inefficient operation of the compressed air system based on the specific power characteristic value, the real-time leakage rate, the pipeline pressure drop value, and the standard deviation of pressure fluctuation. The dynamic carbon emission equivalent conversion sub-model is configured to obtain the dynamic grid carbon factor corresponding to the current time window from the time-synchronized fusion feature dataset, convert the additional power consumption into a dynamic carbon emission equivalent based on the dynamic grid carbon factor, and output the converted dynamic carbon emission equivalent.

[0013] As a preferred embodiment of this application, the nonlinear performance mapping sub-model of the air compressor outputs the specific power characteristic value S under the current operating condition. t The calculation formula is as follows: S t =f(n t , P t , L t ) Where, n t P represents the rotational speed within the current time window t. t L represents the exhaust pressure within the current time window t. t f( represents the load rate within the current time window t). ) is a nonlinear function characterizing the specific power as a function of speed, exhaust pressure and load rate.

[0014] As a preferred embodiment of this application, the additional power consumption ΔE t The calculation formula is:

[0015] Where w1, w2, w3, and w4 are weighting coefficients. This represents the standard deviation of pressure fluctuations within the current time window. This represents the pipeline pressure drop within the current time window. This represents the specific power characteristic value within the current time window. This represents the real-time leakage rate within the current time window. is the overall efficiency parameter of the air compressor under the current operating conditions, and b is the fixed loss term under the reference operating conditions.

[0016] As a preferred embodiment of this application, the dynamic carbon emission equivalent conversion sub-model converts the additional power consumption into a dynamic carbon emission equivalent ΔC. t The calculation formula is as follows:

[0017] Among them, EF t This represents the dynamic grid carbon factor corresponding to the current time window t.

[0018] As a preferred embodiment of this application, the dynamic grid carbon factor EF t The method of obtaining it is: Priority should be given to using carbon intensity data released in real time by the power grid dispatch center; When the grid does not release real-time data, a weighted estimate is made based on the historical carbon factor data of the regional power grid and the predicted value of the power structure for the day.

[0019] As a preferred embodiment of this application, the dynamic carbon-efficiency coupling modeling module is also equipped with a calibration mechanism: When the deviation between the output of the nonlinear performance mapping sub-model of the air compressor and the measured specific power exceeds the preset calibration deviation threshold, the calibration process is triggered, and the parameters of the nonlinear performance mapping sub-model of the air compressor are corrected by recursive least squares method or Kalman filtering method.

[0020] As a preferred embodiment of this application, in the aerodynamic carbon contribution evaluation module, the actual carbon emission equivalent... benchmark carbon emission equivalent With the aforementioned dynamic additional carbon emission equivalent ΔC t sum.

[0021] Furthermore, the aerodynamic carbon contribution evaluation module is configured with a dual-mode evaluation mechanism based on the load boundary: (1) Relative contribution evaluation model: when ≥L min At that time, the carbon contribution index PPCI t The calculation formula is: Furthermore, the aerodynamic carbon contribution index PPCI t The calculation formula is:

[0022] in, This represents the actual carbon emission equivalent within the current time window t. This represents the theoretical baseline carbon emission equivalent under the same production load and gas supply constraints. Indicates the production load factor. This represents a correction term to prevent the denominator from being zero, where θ is a preset minimum operating baseline load compensation coefficient used to prevent... Approaching zero leads to evaluation failure, so it is important to avoid misjudging obvious inefficiencies as having no impact on the zero-carbon target.

[0023] (2) Absolute carbon drag evaluation model: when <L min When the system is determined to be in a dormant or shutdown period, the calculation of the relative index PCCI will be automatically stopped. t Instead, they calculated the absolute carbon drag equivalent. :

[0024] At this point, the aerodynamic carbon contribution evaluation module directly bases its evaluation on the absolute carbon drag equivalent. The compressed air system is compared with the preset carbon emission tolerance threshold during the shutdown period to determine whether it is hindering the zero-carbon target. As a preferred embodiment of this application, the theoretical benchmark carbon emission equivalent The upper limit of carbon emission equivalents allowed to be produced by the compressed air system within the current time window is determined by the following two constraints: first, based on the current gas production rate. Specific power characteristic value under current operating conditions and dynamic grid carbon factor The first is the calculated theoretical minimum carbon emission equivalent; the second is the upper limit of carbon emission allowances allocated to this time window based on the factory's zero-carbon target baseline and the current production plan. The actual setting value is the smaller of the two values ​​mentioned above, that is... .

[0025] As a preferred embodiment of this application, the production load factor The calculation formula is:

[0026] in, This represents the actual gas production within the current time window. This refers to the rated gas production under design conditions. The number of gas-consuming equipment in operation. This represents the total number of gas-using devices. and For preset weighting coefficients and .

[0027] Furthermore, the values ​​of the weighting coefficients α and β are determined based on the distribution characteristics of the factory's gas-consuming equipment: When a factory has a large number of small gas-consuming devices (many devices, small gas consumption per unit, and frequent start-up and shutdown), the influence weight of β should be increased, and it is recommended that α:β = 5:5 to 4:6. When a factory has a small number of large gas-consuming devices (few devices, large gas consumption per unit, and continuous operation), the proportion of gas production is more indicative, and it is recommended that α:β = 7:3 to 6:4. As a preferred embodiment of this application, the pneumatic carbon contribution evaluation module is further configured with a threshold elastic scaling mechanism: the threshold elastic scaling mechanism stores a baseline threshold, calculates an urgency factor in real time based on the carbon quota consumption progress of the current accounting cycle of the factory, and dynamically reconstructs the baseline threshold based on the urgency factor to generate an effective threshold for the current judgment.

[0028] Furthermore, the aerodynamic carbon contribution evaluation module performs a graded evaluation based on the comparison results between the numerical range of the aerodynamic carbon contribution index and the effective threshold after dynamic reconstruction: when When the compressed air system is deemed to have made a positive contribution to the zero-carbon goal, it is determined that the compressed air system contributes positively to the goal. when At that time, it is determined to be an acceptable operating area; when When the compressed air system is deemed to be hindering the zero-carbon target, a warning signal is triggered.

[0029] As a preferred embodiment of this application, the drag-off causal weights include the contribution of air compressor body efficiency deviation, pipeline transportation efficiency deviation, and operating strategy deviation; the pneumatic carbon contribution evaluation module is specifically configured as follows when calculating the contribution of each deviation: Based on the relative deviation between the specific power characteristic value under current operating conditions and the theoretical optimal specific power, the contribution of the air compressor body efficiency deviation is calculated: Based on the relative deviation between the actual pipeline pressure drop and the design pressure drop, and by integrating the real-time leakage rate after unifying the dimensions with conversion factors, the contribution of the pipeline transportation efficiency deviation is calculated. The deviation contribution of the operation strategy is calculated based on the relative deviation between the actual loading rate, start and stop frequency and the target value corresponding to the optimal operation strategy. The pneumatic carbon contribution evaluation module uses the softmax function to perform probability distribution mapping on the above three deviation contributions, and outputs a normalized causal weight vector with the sum of the weights of each item being 1.

[0030] The causal weight vector is used to trigger the generation of the warning information, which includes one or more of the following: The numerical value and classification evaluation results of the aerodynamic carbon contribution index within the current time window; A status indicator indicating whether a contribution or a hindrance is made; Normalized causal weight vector; Trend prediction results are obtained by using the first-order exponential smoothing method or the linear regression extrapolation method based on historical PCCI data. Suggested investigation directions are automatically generated based on the component with the largest proportion in the causal weight vector.

[0031] Compared with the prior art, the present invention has the following beneficial effects: This invention deploys an edge gateway within a multi-source time-series alignment module and employs an adaptive time window merging strategy. Under steady-state conditions, it maintains a window width consistent with low-frequency energy consumption data to reduce computational load. Under non-steady-state conditions, it automatically compresses the window to preserve high-frequency variation characteristics. Combined with steady-state identification and latency compensation mechanisms, it precisely aligns second- / millisecond-level high-frequency aerodynamic parameters with minute- / hour-level low-frequency energy consumption data and green electricity ratio data along the time axis, generating a time-synchronized fused feature dataset. This mechanism eliminates the accumulated time-series errors caused by timestamp misalignment in inter-frequency data, ensuring that subsequent carbon emission equivalent calculations are based on accurate and traceable data. It solves the technical problem of existing monitoring systems failing to perform correlation analysis due to data layer misalignment.

[0032] This invention utilizes a dynamic carbon-efficiency coupling modeling module to map multi-dimensional aerodynamic / thermodynamic indicators, such as pressure fluctuation standard deviation, pipeline pressure drop, leakage rate, and specific power characteristic value, to dynamic carbon emission equivalents through a series calculation of an additional power consumption conversion sub-model and a dynamic carbon emission equivalent conversion sub-model. This forms a complete quantitative chain from air pressure parameters to power consumption increments to carbon emission equivalents. This not only transforms the carbon emission impact of compressed air systems from empirical estimations into calculable and traceable quantitative results but also helps managers accurately identify the specific contributions of different factors, such as pressure fluctuations, pipeline leaks, and equipment efficiency degradation, to carbon emissions.

[0033] This invention defines a Pneumatic Carbon Contribution Index (PCCI) through a pneumatic carbon contribution evaluation module. This index uses the baseline carbon emission equivalent under zero-carbon targets as a reference, quantifying the actual carbon emission deviation of the compressed air system into a normalized index, and further decomposing it into three causal weights: the contribution of air compressor efficiency deviation, the contribution of pipeline transportation efficiency deviation, and the contribution of operational strategy deviation. Simultaneously, the module's built-in threshold elastic scaling mechanism dynamically maps the plant-wide carbon emission quota consumption progress to the real-time tightening or loosening of tiered evaluation thresholds, linking micro-level energy efficiency evaluation with macro-level carbon constraint risks. When carbon quotas approach depletion, it automatically increases early warning sensitivity, providing managers with intervention buffer time; and a lower limit protection mechanism prevents alarm overload caused by excessive threshold compression. This evaluation mechanism enables plant managers to quickly determine the current contribution or drag on the zero-carbon target of the compressed air system, pinpoint the main causes, and provide clear data support for carbon management decisions.

[0034] This invention connects the multi-source time-series alignment module, the dynamic carbon-efficiency coupling modeling module, and the aerodynamic carbon contribution evaluation module into an organic whole. The three modules support each other in function and transmit data step by step, realizing a complete information link from raw data to carbon emission equivalent conversion and then to zero-carbon target contribution evaluation. This provides a systematic technical tool for carbon footprint monitoring and zero-carbon target management of factory compressed air systems. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of the overall architecture of the present invention; Figure 2 This is a schematic diagram of the data processing flow in the multi-source timing alignment module of the present invention; Figure 3 This is a flowchart of the dynamic carbon-efficiency coupling modeling module of the present invention; Figure 4 This is a flowchart of the threshold elastic scaling mechanism in the pneumatic carbon contribution evaluation module of the present invention; Figure 5 This is a flowchart of the pneumatic carbon contribution evaluation module of the present invention. Detailed Implementation

[0036] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Those skilled in the art should understand that the following embodiments are only for explaining the present invention and are not intended to limit the scope of protection of the present invention.

[0037] Figure 1 This is a schematic diagram of the overall architecture of the present invention. (See diagram below.) Figure 1 As shown, the present invention provides an integrated monitoring platform for energy consumption and pressure efficiency in a zero-carbon factory, comprising a multi-source time-series alignment module, a dynamic carbon-efficiency coupling modeling module, and an aerodynamic carbon contribution evaluation module that are interconnected.

[0038] The multi-source time-series alignment module is configured to synchronously collect high-frequency aerodynamic operating parameters of the compressed air system and low-frequency energy consumption data and green electricity ratio data of the factory power grid. It also obtains the current time window carbon emission quota limit issued by the factory energy management system through a data interface, timestamps the high-frequency aerodynamic operating parameters and low-frequency energy consumption data, and associates and matches the green electricity ratio data and the carbon emission quota limit with the corresponding time window to output a time-series synchronized fused feature dataset.

[0039] Figure 2 This is a schematic diagram of the data processing flow in the multi-source timing alignment module of the present invention. For example... Figure 2 As shown, this module is used to solve the timestamp misalignment problem between the high-frequency aerodynamic parameters (second-level / millisecond-level) of the compressed air system and the low-frequency energy consumption data (minute-level / hour-level) of the factory power grid, providing a time-synchronized fusion feature dataset for subsequent dynamic carbon efficiency coupling modeling.

[0040] Step S101: Hardware Configuration and Multi-Source Data Acquisition The multi-source time-series alignment module includes an edge gateway deployed on the air compressor station side, a factory-level energy consumption acquisition module, a carbon emission data interface module, and a time-series database.

[0041] The air compressor's local controller is connected via an industrial bus or Modbus protocol to collect high-frequency pneumatic operating parameters at a first sampling frequency. This first sampling frequency is in the range of seconds or milliseconds, and is preferably 1 second in this embodiment. The high-frequency pneumatic operating parameters include, but are not limited to: air compressor outlet pressure, main pipe pressure, terminal pressure, instantaneous flow rate, motor current, active power, and exhaust temperature.

[0042] The factory-level energy consumption acquisition module is connected to the factory's main electricity meter via a power gateway, and collects low-frequency energy consumption data in a second sampling period. The second sampling period is the sampling period T of the factory's main electricity meter data. meter In this embodiment, 15 minutes is preferred. Low-frequency energy consumption data includes the factory's total power consumption and cumulative power consumption.

[0043] The carbon emission data interface module is connected to the power grid carbon factor data platform, the factory manufacturing execution system (MES), and the factory energy management system (EMS) via API interfaces. The second sampling period is used to obtain the green electricity ratio data and the daily power structure prediction value (i.e., the predicted output ratio P of each power generation resource at time t). i, t Simultaneously, the carbon emission data interface module reads the preset inherent carbon emission factors (EF) of each power generation resource (coal, gas, hydropower, photovoltaic, wind power, etc.) from the platform's local database. i Based on the above data, the carbon emission data interface module calculates the dynamic power grid carbon factor EF within the current time window. t :

[0044] in, Let be the predicted output percentage of the i-th power generation resource at time t. The inherent carbon emission factors of their corresponding power generation resources, and the corresponding carbon emission factors of photovoltaic and wind power. When the power structure forecast data is not released by the grid, estimations are made based on historical power output statistics for the same period in the regional grid and the daily weather forecast (solar intensity, average wind speed). The green electricity ratio data directly participates in the Energy Emission Standard (EF). t The calculation is not stored separately in the fused feature dataset; the calculated EF t Fields from the fused feature dataset are written into the corresponding time window records.

[0045] The number of gas-consuming devices currently in operation is obtained from the MES during the second sampling period. (This data and the sampling period T of low-frequency energy consumption data) meter If consistent, directly write to the corresponding time window record.

[0046] Establish a communication connection with the factory's energy management system (EMS) through an API interface, and obtain the carbon emission quota limit allocated to the current time window in real time based on the current production schedule. t This parameter characterizes the hard carbon constraint of the macro-zero carbon target on the current production cycle. After being acquired, it will be synchronously merged into the local cache of the edge gateway and encapsulated into the fused feature dataset, serving as the extreme boundary condition for the subsequent aerodynamic carbon contribution evaluation module to calculate the baseline carbon emission equivalent.

[0047] The platform configuration database stores the following static parameters: the rated air output of each air compressor. (Dynamically accumulated based on the current operating unit combination) Total number of gas-using equipment in the factory The inherent carbon emission factor EF of each power generation resourcei Pipeline pressure drop under design conditions p design The target loading rate R corresponding to the optimal running strategy opt And the target start-stop frequency C opt .

[0048] Of the above data, high-frequency aerodynamic parameters are cached locally at the edge gateway, awaiting subsequent time window merging and feature extraction processing; low-frequency energy consumption data, dynamic power grid carbon factor EFt, and the number of operating gas-consuming devices are also included. and carbon emission quota cap t All in T meter For periodic updates, the data is directly written to the fused feature dataset after being associated and matched with time window identifiers; static parameters are stored in the platform configuration database and do not change with time windows, but are read as needed during subsequent module calculations.

[0049] Step S102: Adaptive Time Window Merging Based on Operating Condition Fluctuations The edge gateway performs time window merging on the collected high-frequency data streams using a preset window width W. The preset window width W is related to the sampling period T of the low-frequency energy consumption data. meter The time windows are equal, and in this embodiment, they are all 15 minutes. Time window merging refers to dividing a continuous high-frequency data stream into several contiguous data segments with this fixed time length. The windows do not overlap with each other, so as to ensure that each sampling point belongs to only one window.

[0050] Preferably, under unsteady operating conditions with drastic fluctuations in gas consumption at the terminal, a long window of fixed width will smooth out high-frequency features, leading to the loss of details of aerodynamic anomalies; while under steady-state operating conditions, an extremely short window will calculate the load at the edge. Therefore, the edge gateway is configured with an adaptive window width adjustment mechanism. The edge gateway calculates the real-time information entropy H(p) and the absolute value of the transient pressure change rate V of the pipeline pressure signal within the current front-end observation window (e.g., the previous minute) using a microsecond-level polling method in the background. transient The default base window width is set to W. base (Sampling period T of low-frequency energy consumption data) meter Consistent (15 minutes in this embodiment). The edge gateway dynamically calculates the specific width W of the current merging time window based on the following adaptive logic. dynamic :

[0051] in, The minimum resolution window (e.g., 1 minute) is set to be subject to the Nyquist sampling theorem. The sensitivity gain constant characterizes the dynamic inertia of the system.

[0052] The above mechanism ensures high-fidelity retention of high-frequency aberration features in subsequent feature vector extraction. For multiple sub-windows generated by dynamic compression, during the temporal interpolation mapping in step S104, an adaptive spline function is used to fit and align them to the corresponding low-frequency period T. meter Inside.

[0053] Step S103: Steady-state identification and feature vector extraction for each window For each subset of data within a window, the edge gateway first performs steady-state identification. The purpose of this is to eliminate the interference of transient disturbances such as instantaneous valve opening and closing and equipment start-up and shutdown on the window feature values, so as to ensure that the extracted feature vector can truly reflect the continuous operating state of the compressed air system and avoid transient events being misjudged as a degradation of system performance.

[0054] The specific method for steady-state identification is as follows: calculate the standard deviation σ of pressure fluctuation within the window. p,t and flow rate Q / t. When σ p,t Less than the steady-state pressure threshold and Q / If t is less than the steady-state flow threshold, the window is marked as a steady-state window; otherwise, it is marked as a non-steady-state window. The data subset of the non-steady-state window does not participate in the construction of subsequent feature vectors.

[0055] Preferably, the steady-state pressure threshold is not a fixed absolute value, but rather depends on the system's rated pressure P. rated The steady-state pressure threshold is set adaptively and its value ranges from 3% to 6% of the rated pressure. In this embodiment, when the system rated pressure is 0.8 MPa, the steady-state pressure threshold is set to 0.03 MPa; when the system rated pressure is 1.2 MPa, the steady-state pressure threshold is adaptively adjusted to 0.05 MPa to ensure that systems of different pressure levels have consistent steady-state discrimination sensitivity. The steady-state flow rate threshold is set based on the total rated flow rate of the currently operating air compressor unit, at a ratio of 3% to 5%.

[0056] For the data subset labeled as a steady-state window, the edge gateway further performs feature extraction and calculates the following feature values ​​to characterize the operating state of the window: Pressure characteristics: average pressure, maximum pressure, minimum pressure, standard deviation of pressure fluctuation σ p,t ; Flow characteristics: mean flow rate, peak-to-valley flow rate, and rate of change of flow rate. Q / t, and the actual gas production calculated based on the cumulative instantaneous flow rate. ; Operating status characteristics: percentage of air compressor loading time and percentage of air compressor unloading time, number of start-stop cycles or loading / unloading switching cycles per unit time window, and the number of air-consuming devices currently in operation obtained from the factory's Manufacturing Execution System (MES) via the data interface module. (This data uses a sampling period T consistent with that of low-frequency energy consumption data) meter Update (directly write to the corresponding time window record); Temperature characteristics: average exhaust temperature and temperature rise slope.

[0057] The aforementioned eigenvalues ​​constitute the aerodynamic eigenvector F of this window. k (k is the window number), used to characterize the overall aerodynamic operating state of the compressed air system within that window. Furthermore, the aerodynamic feature vector F... k It also includes the current window rotation speed n directly collected by the edge gateway. t With exhaust pressure P t The above two items are the original acquisition parameters and are not involved in feature extraction calculations.

[0058] Among the above characteristic values, the standard deviation of pressure fluctuation σ p,t Actual gas production Number of gas-consuming equipment in operation Rotational speed n t and exhaust pressure P t These parameters serve as core inputs to the subsequent dynamic carbon-efficiency coupling modeling and aerodynamic carbon contribution evaluation modules, participating in the calculation of additional carbon emission equivalents and contribution index. Flow rate average and exhaust temperature-related parameters are used as auxiliary monitoring parameters for system condition identification and auxiliary monitoring of equipment operating status, and do not participate in the aforementioned core calculation processes.

[0059] The high-frequency aerodynamic feature vectors of multiple consecutive windows are arranged in chronological order to form a high-frequency aerodynamic feature vector sequence.

[0060] Step S104: Timing Alignment (Delay Compensation and Interpolation Mapping) After completing the high-frequency data preprocessing, the edge gateway performs a time alignment process to map the high-frequency aerodynamic feature sequence to the time axis of the low-frequency energy consumption data.

[0061] The specific method of delay compensation is as follows: The edge gateway first merges the high-frequency aerodynamic feature sequences (with a time granularity of T) after the time windows have been completed. meter (The sampling period is consistent with that of low-frequency energy consumption data) and time-series comparison is performed with the low-frequency energy consumption data sequence.

[0062] Based on the fusion feature dataset, the generated T meter F is a high-frequency feature vector sequence with intervals. k(k is the window number) and the corresponding low-frequency energy consumption sequence E k Using the input as input, calculate the cross-correlation function R between the two. FE (m):

[0063] Where m is the time offset in units of time windows (ranging from -M to M, where M is the preset maximum allowable time window offset; in this embodiment, M = 3, corresponding to a theoretical latency search range of ±45 minutes). The search makes R FE (m) The maximum value of m Calculate the delay compensation amount τ=m ×T meter .

[0064] Selection of cross-correlation window length: The length K of the data sequence involved in the cross-correlation calculation. window It should cover at least three complete theoretical maximum time delays, i.e., K. window ≥ 3M, and at least 30 time windows (corresponding to 7.5 hours) to ensure the cross-correlation function has sufficient statistical significance. In this embodiment, K window The value is set to 48 time windows (corresponding to 12 hours) to cover a complete production shift cycle.

[0065] Triggering conditions and update frequency: The recalculation of the delay compensation amount τ adopts an event-triggered mechanism. Specifically, the system triggers delay recalculation in the following two situations: (1) when the system is put into operation for the first time or when the equipment configuration changes (such as adding an air compressor or modifying the pipeline); (2) when the time deviation between the high-frequency aerodynamic feature sequence and the low-frequency energy consumption sequence in the fused feature data exceeds the preset tolerance range. The preset tolerance range is ±0.5 × T meter (In this embodiment, the time interval is ±7.5 minutes). When the deviation exceeds this range, it indicates that the timelines of the two sets of data have become out of sync due to communication jitter, data cache buildup, or system clock drift. In this case, the cross-correlation calculation is re-executed to update the τ value. Between two trigger events, τ remains constant to maintain consistency in timing alignment. When the cross-correlation function cannot detect a significant peak (i.e., the ratio of the maximum peak to the second largest peak is lower than a preset threshold of 1.5), the system determines that the current data does not meet the cross-correlation calculation conditions, keeps the current τ unchanged, and generates a prompt message to notify maintenance personnel to check for any abnormalities in the data acquisition link.

[0066] When the time deviation between the high-frequency aerodynamic feature sequence and the low-frequency energy consumption data sequence exceeds the preset tolerance range, the cross-correlation time delay estimation method is used to calculate the time delay compensation τ, and the high-frequency aerodynamic feature sequence is shifted by τ as a whole to align the two sets of data on the time axis.

[0067] The specific method of interpolation mapping is as follows: using the timestamp T of low-frequency energy consumption data k As interpolation nodes, the aerodynamic feature vectors of each window are mapped to the time axis of the low-frequency data. The interpolation method can be any of piecewise linear interpolation, spline interpolation, or preserve-the-moment interpolation. In this embodiment, piecewise linear interpolation is used when the data changes gradually, and spline interpolation is used when the data has obvious peaks or troughs.

[0068] After the above steps, the multi-source time-series alignment module generates a time-synchronized fused feature dataset. This dataset uses the sampling time T of low-frequency energy consumption data as the basis. k The primary key is used for each record, which contains the variance σ of the pressure fluctuation corresponding to that time window. p,t Average flow rate, number of loading / unloading switching times, and actual gas production. Number of gas-consuming equipment in operation Rotational speed n t Exhaust pressure P t and load factor L t (Calculated from the proportion of loading time) and other aerodynamic characteristic parameters, energy consumption data such as the factory's total power consumption and cumulative power consumption at the same timestamp, and the dynamic power grid carbon factor EF calculated in step S101. t And the current time window carbon emission quota limit obtained through the data interface. t Rated gas production Total number of gas-consuming devices The static configuration data is stored in the platform configuration database and does not change with the time window. It is read on demand when the aerodynamic carbon contribution evaluation module calculates the production load coefficient.

[0069] The aforementioned time-synchronized fusion feature dataset is persistently stored in the time-series database and serves as the standard input for the subsequent dynamic carbon-efficiency coupling modeling module. This dataset ensures the temporal consistency of subsequent carbon emission equivalent calculations: all parameters correspond to the same time window, eliminating the cumulative errors introduced by temporal misalignment of inter-frequency data.

[0070] The dynamic carbon efficiency coupling modeling module is configured to receive the fused feature dataset, construct a pressure efficiency-carbon emission mapping model, which is used to convert the aerodynamic operating state into dynamic carbon emission equivalent and output the converted dynamic carbon emission equivalent.

[0071] The dynamic carbon efficiency coupling modeling module includes a nonlinear performance mapping sub-model for air compressors, a real-time estimation sub-model for pipeline pressure drop and leakage, a power consumption increment conversion sub-model, and a dynamic carbon emission equivalent conversion sub-model. Specifically, the nonlinear performance mapping sub-model for air compressors and the real-time estimation sub-model for pipeline pressure drop and leakage respectively receive corresponding fields from the time-synchronized fused feature dataset and output the specific power characteristic value and real-time leakage rate and pipeline pressure drop value under the current operating condition in parallel. The power consumption increment conversion sub-model receives the output results of the above two sub-models and calculates the additional power consumption by combining it with the standard deviation of pressure fluctuations. The dynamic carbon emission equivalent conversion model receives the additional power consumption and converts it into a dynamic additional carbon emission equivalent by combining it with the dynamic power grid carbon factor.

[0072] In this embodiment, the specific implementation methods of the four sub-models are as follows: (I) Nonlinear performance mapping sub-model of air compressor The nonlinear performance mapping sub-model of the air compressor is configured to receive the rotational speed within the current time window t. Exhaust pressure and load rate The nonlinear function outputs the specific power characteristic value under the current operating condition. .

[0073] This sub-model is used to characterize the specific power variation of the air compressor under different speeds, different exhaust pressures and different loading rates. It expresses the energy consumption per unit of air production as a nonlinear function of speed, exhaust pressure and flow load to reflect the actual energy consumption difference between variable frequency operation and loading / unloading operation.

[0074] The relationship between power consumption and operating speed of an air compressor is not a simple linear or cubic one, but rather exhibits complex nonlinear characteristics. This nonlinear relationship is simultaneously influenced by the coupling effect of exhaust pressure and load rate: at the same exhaust pressure, the exhaust volume decreases as the speed increases; at the same speed, the exhaust volume decreases sharply as the exhaust pressure increases; and the specific power growth rate increases with both speed and exhaust pressure. Therefore, it is necessary to construct a multidimensional nonlinear mapping surface to accurately describe the above relationship.

[0075] The specific power characteristic value The calculation formula is:

[0076] Where, f( To characterize the specific power as a nonlinear function of speed, exhaust pressure, and load rate, a multidimensional mapped surface is constructed. For operating points not measured in actual measurements, radial basis function interpolation or Kriging interpolation is used for estimation, i.e.:

[0077] in, ( ) are radial basis functions (such as Gaussian functions or multiple quadratic functions). Given the known measured operating points, is the interpolation weighting coefficient, and N is the number of measured operating points.

[0078] Radial basis function interpolation is suitable for isotropic and uniformly distributed data scenarios, while Kriging interpolation is suitable for data distributions with spatial correlation. Both can effectively handle nonlinear interpolation problems in high-dimensional spaces, thereby obtaining high-precision power-to-weight ratio mapping surfaces across the entire operating range. In this embodiment, the measured operating points are obtained through factory performance tests of the air compressor unit, or through variable operating condition tests during the calibration phase at the initial stage of system commissioning. The number of measured operating points is no less than 30 sets to ensure interpolation accuracy.

[0079] (II) Sub-model for real-time estimation of pipeline pressure drop and leakage The real-time estimation sub-model for pipeline pressure drop and leakage is configured to receive the total pipeline pressure, terminal pressure and flow rate change fields in the current time window from the fused feature dataset, estimate the pipeline pressure drop in real time, and estimate the real-time leakage rate based on the pipeline pressure decay data collected during non-production periods. The leakage rate is a dynamic variable that changes over time.

[0080] The real-time leakage rate is estimated by acquiring pipeline pressure decay data during low-load or non-production periods at night, and calculating the equivalent leakage flow rate based on the pressure drop rate.

[0081] in, The total volume of the pipeline system (unit: m³) is obtained by calculating the internal volume of each pipe segment based on the pipe length and diameter data in the pipeline design drawings and then summing them up to obtain the theoretical total volume; or by using a gas tracer method (such as nitrogen tracer) to measure the effective volume on site; when the design drawings are missing or the pipeline network is inconsistent with the drawings after modification, the measured value shall be used first. Standard atmospheric pressure; / This refers to the pressure drop rate during non-production periods. The pipeline pressure decay data is not included in the time-synchronized fusion feature dataset, but is instead collected and cached separately by the edge gateway during non-production periods.

[0082] As a supplement, the leakage rate of a branch can be calculated based on the difference in flow rate before and after isolation by obtaining isolation test data of each branch within a preset time period; or the difference between the total pipeline flow rate and the sum of the metered flow rates of each end gas-using device can be used as an estimate of the total pipeline leakage.

[0083] The method for estimating pipeline pressure drop is as follows: obtain the total pipeline pressure at the air compressor station outlet from the fused feature dataset. and the terminal pressure at the most unfavorable end-user gas equipment The actual pipeline pressure drop is obtained by calculating the difference between the two values. :

[0084] At the same time, based on the current total pipe flow Rated flow rate under design conditions The ratio of these values ​​is used to correct the design pressure drop with flow rate, yielding the theoretical pipeline pressure drop:

[0085] in, The pipeline pressure drop under design conditions is provided in the pipeline design documents and pre-stored in the platform configuration database. The actual measured value and the theoretical estimate are weighted and fused to obtain the pipeline pressure drop. :

[0086] Wherein, γ is a preset weighting coefficient, with a value range of 0.5 to 0.8, and preferably 0.7 in this embodiment. The actual measured value is the main basis, while the theoretical estimated value is used as verification and supplement.

[0087] (III) Power Consumption Calculator Model The power consumption increment conversion sub-model is configured as the received specific power characteristic value. Real-time leakage rate Standard deviation of pressure fluctuation Pipeline pressure drop and overall efficiency parameters Calculate the additional power consumption within the current time window. E t This sub-model converts the pressure efficiency deviations between the production and transmission ends into additional power consumption, enabling multi-dimensional pressure efficiency deviations to be summarized under the same dimension.

[0088] Preferably, the comprehensive efficiency parameter The value is obtained from the air compressor unit's factory performance curve: based on the rotational speed n within the current time window. t Exhaust pressure P t and load factor L t Interpolate the values ​​in the efficiency map provided by the manufacturer to obtain the overall efficiency value corresponding to that operating point. When a factory performance curve is unavailable... The efficiency can be obtained through calibration tests during the initial commissioning phase: During the first stable operating cycle after the system is put into operation, the average ratio of the measured input power to the theoretical isentropic compression power is used as the benchmark comprehensive efficiency of the unit. In subsequent operation, a calibration test is performed every fixed period (such as 30 days) to track efficiency decay.

[0089] The physical mechanisms by which the decrease in compressed air system pressure efficiency leads to an increase in additional power consumption are as follows: Increased pressure fluctuations lead to an increase in the frequency of compressor loading and unloading, and an increase in the proportion of idle running time. This part of the idle power consumption is completely ineffective energy consumption; Increased pipeline pressure drop causes the compressor to increase the discharge pressure setpoint to compensate for delivery losses, and the increase in discharge pressure directly leads to an increase in power consumption per unit of air production, i.e., specific power; Increased leakage rate causes the compressor to produce additional air to maintain the terminal pressure, and the power consumption corresponding to this part of air production is completely ineffective energy consumption.

[0090] The additional power consumption E t The calculation formula is:

[0091] Among them, w1, w2, w3, and w4 are weighting coefficients, which correspond to the contribution weights of pressure fluctuation, pipeline pressure drop, leakage and specific power coupling, and pressure fluctuation and pressure drop interaction terms to additional power consumption, respectively; b is the fixed loss term under the baseline operating condition, including the inherent energy loss under the baseline operating condition such as motor iron loss and mechanical friction loss.

[0092] The above formula includes an interaction term between pressure fluctuation and pipeline pressure drop. The physical significance of this is that when the pipeline pressure drop already causes the air compressor to operate in a high-pressure zone, if the gas consumption at the terminal also fluctuates drastically, the air compressor not only needs to operate in the high-pressure zone but also needs to be frequently loaded and unloaded. The no-load loss under high pressure is much greater than the no-load loss under low pressure, and the superposition of the two produces a nonlinear amplified additional power consumption effect.

[0093] As the air compressor rotor wears, the roughness of the pipe inner wall increases, and the filter resistance grows, the physical mapping relationship between aerodynamic parameters and power consumption increments undergoes an irreversible, slow drift, and static weights lead to a continuous amplification of the conversion error. Therefore, the system employs a recursive least squares algorithm with a forgetting factor, using the observation data vector of the current time window... The input is the actual additional power consumption feedback value ΔE, calculated by subtracting the theoretical power consumption under the baseline operating condition from the total power consumption measured by the factory's main power meter in the current time window. act To monitor the signal, the weight matrix W is updated online. t =[w 1,t , w 2,t , w 3,t , w 4,t ]T This causes the model to track the drift in mapping relationships due to the performance degradation of the tracking equipment and the aging of the pipeline network.

[0094] (iv) Dynamic carbon emission equivalent conversion sub-model The dynamic carbon emission equivalent conversion sub-model is configured to obtain the dynamic grid carbon factor EF corresponding to the current time window from the time-synchronized fused feature dataset. t Based on the dynamic grid carbon factor, the additional power consumption is converted into a dynamic additional carbon emission equivalent ΔC. t The model outputs the converted dynamic additional carbon emission equivalent. This sub-model converts the additional power consumption (energy consumption deviation) into the additional carbon emission equivalent (carbon emission deviation), realizing a quantitative correlation between pressure efficiency and carbon efficiency.

[0095] The dynamic carbon emission equivalent ΔC t The calculation formula is:

[0096] Among them, EF t The dynamic power grid carbon factor corresponding to the current time window t is calculated by the carbon emission data interface module in the multi-source time series alignment module and is passed into the sub-model as a field of the fused feature dataset.

[0097] The dynamic grid carbon factor fluctuates in real time with changes in the power generation energy structure (e.g., the carbon factor decreases when photovoltaic / wind power output increases, and increases when thermal power generation increases). Using a fixed emission factor for conversion would introduce significant bias. Therefore, this invention uses a dynamic grid carbon factor for conversion to accurately reflect the differences in grid carbon emission intensity at different times.

[0098] Figure 3 This is a flowchart of the dynamic carbon-efficiency coupling modeling module of the present invention. Figure 3 As shown, combining the above four sub-models, the dynamic carbon-efficiency coupling modeling module performs calculations according to the following process: Step S201: Obtain the rotational speed n within the current time window from the fused feature dataset. t Exhaust pressure P t and load factor L t Input the nonlinear performance mapping sub-model of the air compressor to obtain the specific power characteristic value S under the current operating condition. t .

[0099] Step S202: Obtain the main pipe pressure P within the current time window from the fused feature dataset. supply Terminal pressure P end The flow rate change rate is input into the real-time estimation sub-model of pipeline pressure drop and leakage to obtain the real-time leakage rate λ. t and pipeline pressure drop Δpt .

[0100] Step S203: The specific power characteristic value S t The real-time leakage rate λ t The pipeline pressure drop value Δp t and the standard deviation of pressure fluctuation within the current time window σ p,t The additional power consumption ΔE is calculated by inputting the power consumption increment conversion sub-model together. t . Step S204: Obtain the dynamic power grid carbon factor EF corresponding to the current time window from the fused feature dataset. t The additional power consumption ΔE t With the dynamic grid carbon factor EF t Input the dynamic carbon emission equivalent conversion sub-model to obtain the dynamic additional carbon emission equivalent ΔC corresponding to the current time window. t .

[0101] Step S205: The dynamic additional carbon emission equivalent ΔC t As the output of this module, it is passed to the aerodynamic carbon contribution evaluation module.

[0102] Considering that the performance of air compressors may degrade over time, the dynamic carbon-efficiency coupling modeling module is also equipped with a model calibration mechanism: When the output S of the nonlinear performance mapping sub-model of the air compressor t When the deviation between the measured specific power and the actual specific power exceeds the preset calibration deviation threshold, the calibration process is triggered.

[0103] Preferably, the specific value of the calibration trigger threshold is determined jointly based on the factory performance accuracy level of the air compressor unit and the historical statistical distribution of the measured specific power on site: Reference value: The lower limit of the reference value is the specific power accuracy (usually ±2%~±5%) given in the factory performance curve of the air compressor unit; Statistical correction: Calculate the standard deviation σ of the deviation between the measured specific power and the model output specific power within the first stable operating cycle (e.g., 30 days) after the data acquisition system is put into operation. dev , with 3σ dev As a confidence upper limit for random fluctuations; The final calibration trigger threshold is the larger of the two values ​​mentioned above, to ensure that the threshold will not be frequently falsely triggered due to fluctuations in the inherent accuracy of the device, while effectively capturing the true degradation of the device's performance.

[0104] In the typical industrial frequency air compressor unit scenario described in this embodiment, the calibration trigger threshold obtained by combining the above methods is 3%. For variable frequency units or centrifugal air compressors with higher precision levels, this threshold can be narrowed to 1.5%~2.5%. Air compressors of different brands, models, or ages can all have their own exclusive thresholds calculated independently using the above methods. The system will automatically complete the statistical calculation and write it into the configuration database during the initial stage of operation.

[0105] When calibration is triggered, the system acquires actual operating data from the most recent preset time period (e.g., the last 7 days) as an update sample set, and uses recursive least squares or Kalman filtering to adjust the weighting coefficients α of the radial basis function interpolation in the nonlinear performance mapping sub-model of the air compressor. i Adaptive correction is performed. The weight of the updated sample set in the total training samples increases monotonically with the increase of device running time to reflect the impact of device aging and performance degradation on model accuracy.

[0106] Preferably, the recursive least squares method employs an exponentially weighted form with a forgetting factor. This forgetting factor is used to attenuate and weight historical data, ensuring that recent measured data has a greater influence on model parameter updates than older data, thus tracking the slow drift of equipment performance. The specific value of the forgetting factor is calibrated within the conventional range in this field based on the actual sampling period and equipment aging rate. When using the Kalman filter method, its state equation and observation equation are constructed according to standard forms in this field, using model parameters as state variables and measured specific power as observations. Online optimal estimation of model parameters is achieved through a two-step recursive process of prediction and update.

[0107] After calibration, the corrected α i The original parameters are replaced and used for the calculation of the specific power characteristic value in subsequent time windows, so that the nonlinear performance mapping sub-model of the air compressor can track the impact of equipment performance degradation on the model accuracy.

[0108] The pneumatic carbon contribution evaluation module is configured to receive the dynamic carbon emission equivalent, calculate the pneumatic carbon contribution index based on the benchmark carbon emission equivalent, determine the degree of contribution or drag on the zero-carbon target of the current operating state of the compressed air system according to the numerical range of the pneumatic carbon contribution index, and output early warning information.

[0109] In this embodiment, the benchmark carbon emission equivalent is the carbon emission equivalent that the compressed air system should produce when it is in its theoretical optimal operating state under the same production load and air supply constraints. Its physical meaning is the theoretical minimum allowable value of carbon emissions under the current production task.

[0110] The theoretical benchmark carbon emission equivalent The upper limit of carbon emission equivalents allowed to be produced by the compressed air system within the current time window is determined by the following two constraints: first, based on the current gas production rate. Specific power characteristic value under current operating conditions and dynamic grid carbon factor The first is the calculated theoretical minimum carbon emission equivalent; the second is the upper limit of carbon emission allowances allocated to this time window based on the factory's zero-carbon target baseline and the current production plan. The actual setting value is the smaller of the two values ​​mentioned above, that is... .

[0111] Actual carbon emission equivalent benchmark carbon emission equivalent With the aforementioned dynamic additional carbon emission equivalent ΔC t The sum of the two components. Current actual carbon emissions consist of two parts: the theoretical minimum allowable value (benchmark) and the additional increase (additional) caused by inefficient operation. The theoretical minimum allowable value is determined by the production task (gas production) and the theoretical optimal energy efficiency level, while the additional increase is jointly determined by operational deviations such as pressure fluctuations, pipeline pressure drop, and leakage.

[0112] To avoid misjudging the system status due to low absolute energy consumption during low-load phases, the pneumatic carbon contribution evaluation module incorporates a production load coefficient. This is used to reflect the relative level of compressed air demand in response to the current production task intensity. Under low load conditions, even if aerodynamic characteristic parameters (such as pressure fluctuation standard deviation and leakage rate) are relatively high, the absolute deviation of the actual system operation may be small. The relative level of the current load is characterized by the weighted gas production ratio and the equipment operation ratio. The benchmark value of the PCCI index is dynamically adjusted to match the actual load, thereby avoiding false alarms caused by an excessively low benchmark during low load periods.

[0113] The production load factor The calculation formula is:

[0114] in, This represents the actual gas production within the current time window. This refers to the rated gas production under design conditions. The number of gas-consuming equipment in operation. This represents the total number of gas-using devices. and For preset weighting coefficients and .

[0115] Preferably, the values ​​of the weighting coefficients α and β are determined based on the distribution characteristics of the factory's gas-consuming equipment: When a factory has a large number of small gas-consuming devices (many devices, small gas consumption per unit, and frequent start-up and shutdown), the influence weight of β should be increased, and it is recommended that α:β = 5:5 to 4:6. When a factory has a small number of large gas-consuming devices (few devices, large gas consumption per unit, and continuous operation), the proportion of gas production is more indicative, and it is recommended that α:β = 7:3 to 6:4. In this embodiment, α is 0.6 and β is 0.4.

[0116] As an alternative, α and β can also be determined by least squares fitting: collect at least 30 historical data samples under different load conditions, use the proportion of gas production and the proportion of equipment operation as independent variables and the total power consumption of the compressed air system as dependent variable, perform binary linear regression fitting, and normalize the regression coefficients as the values ​​of α and β.

[0117] The aerodynamic carbon contribution evaluation module is equipped with a dual-mode evaluation mechanism based on the load boundary: (1) Relative contribution evaluation model: when ≥L min At that time, the carbon contribution index PPCI t The calculation formula is:

[0118] in, This represents a correction term to prevent the denominator from being zero; θ is a preset load compensation coefficient, whose physical meaning is the minimum operating baseline load rate, used to prevent... The value approaches 0, which causes the evaluation to fail. The range of values ​​is 0.05 ≤ θ ≤ 0.15 (preferably 0.1).

[0119] (2) Absolute carbon drag evaluation model: when <L min When the system is determined to be in a dormant or shutdown period, the calculation of the relative index PCCI will be automatically stopped. t Instead, they calculated the absolute carbon drag equivalent. :

[0120] At this point, the aerodynamic carbon contribution evaluation module directly bases its evaluation on the absolute carbon drag equivalent. The system is compared with the preset carbon emission tolerance threshold during the shutdown period to determine whether the compressed air system is hindering the zero-carbon target.

[0121] This formula represents the degree of deviation of the actual carbon emission equivalent from the benchmark carbon emission equivalent, weighted by the production load factor to obtain a normalized index. Specifically, the numerator reflects the absolute carbon emission increment of the compressed air system's current operating state from the benchmark; the denominator reflects the coupling strength between the benchmark level and the current load, normalizing the absolute deviation into a relative deviation rate; multiplied by... This allows the index to scale linearly with the load, avoiding inconsistencies in evaluation criteria under high and low load conditions.

[0122] Based on the above calculations, PPCI t The deviation of actual carbon emissions from the benchmark carbon emissions is normalized into a dimensionless index, providing a quantitative basis for subsequent classification and evaluation.

[0123] To achieve a dynamic early warning balance between carbon quotas and energy efficiency fluctuations, the aerodynamic carbon contribution evaluation module incorporates a threshold elastic scaling mechanism. The technical logic of this mechanism is that when the plant's overall carbon emission progress is significantly ahead of schedule, the system automatically tightens the tolerance threshold of the evaluation module. This allows the monitoring platform to not only objectively reflect the current energy efficiency status but also proactively detect the risk of depletion of the plant's macro-level carbon quota.

[0124] Figure 4 This is a flowchart of the threshold elastic scaling mechanism in the aerodynamic carbon contribution evaluation module of this invention. Figure 4 As shown, the mechanism includes the following process: Step S301: Calculate the urgency factor using the plant-wide carbon quota consumption data as input; Step S302: Generate candidate dynamic thresholds based on the baseline threshold; Step S303: Output the valid threshold after lower limit protection judgment.

[0125] The above process enables the module to dynamically adjust the tolerance boundary of the graded evaluation based on whether the carbon emission progress of the entire plant is ahead or behind, thus achieving adaptive transmission of macro carbon quota risks to the micro evaluation scale.

[0126] Specifically, the local database of the aerodynamic carbon contribution evaluation module pre-stores a first baseline threshold Thresh. 1,static Second baseline threshold Thresh 2,static This serves as the initial decision boundary under the standard carbon quota consumption rate. In this embodiment, the Thresh... 1,static The value range is -5% to 0%, preferably -3%; the Thresh 2,static The value ranges from 5% to 15%, with 8% being preferred.

[0127] The system obtains the total carbon emission allowance quota for the current accounting cycle of the factory in real time through the data interface. total Cumulative carbon emissions C consumedconsumed Real-time calculation of dynamic carbon constraint urgency factor. :

[0128] Among them, T elapsed T represents the elapsed time in the current period. total The total time for the accounting period is κ, which is the risk amplification factor (κ>1). It provides a direct reflection of whether the current rate of carbon consumption is outpacing the rate of time.

[0129] The carbon contribution evaluation module uses the urgency factor to dynamically reconstruct the aforementioned baseline threshold in real time, generating an effective threshold for the current judgment:

[0130]

[0131] Preferably, to prevent An excessively large threshold can lead to over-compression of the threshold, causing the system to become oversensitive. Therefore, the effective threshold is equipped with a lower limit protection mechanism. Specifically, Thresh is set... 2,dynamic Thresh (lower limit) 2,min Its value ranges from 1% to 3%, and is preferably 2% in this embodiment.

[0132] The aerodynamic carbon contribution evaluation module performs a graded evaluation of the compressed air system status based on the numerical range of the aerodynamic carbon contribution index and the dynamically reconstructed effective threshold. The effective threshold includes a first effective threshold. Second effective threshold : when When the compressed air system is deemed to have made a positive contribution to the zero-carbon goal, it is determined that the compressed air system contributes positively to the goal. when At that time, it is determined to be an acceptable operating area; when When the compressed air system is deemed to be hindering the zero-carbon target, a warning signal is triggered.

[0133] when In this process, the aerodynamic carbon contribution evaluation module will proactively trigger the drag-off cause decomposition process to assist managers in intervening before carbon quotas are exhausted. In this embodiment, the drag-off causes are decomposed into three normalized weights.

[0134] The three weights are calculated as follows: Air compressor body efficiency deviation contribution W body Based on the specific power characteristic value S under the current operating condition output by the nonlinear performance mapping sub-model of the air compressor tCompared with the theoretical optimal specific power Deviation calculation:

[0135] This indicator reflects the degree of decline in the energy conversion efficiency of the air compressor unit. When S t Deviation The larger the value, the more severe the efficiency loss caused by equipment aging, wear and tear, or improper selection. When S... t ≤ When the contribution is 0, it indicates that the ontology efficiency does not contribute to the drag.

[0136] Contribution of pipeline transportation efficiency deviation W pipe The current actual pipeline pressure drop Δp is output based on the real-time estimation sub-model of pipeline pressure drop and leakage. t Design pressure drop Δp design The deviation, combined with the current leakage rate λ t calculate:

[0137] in, This is a conversion factor for the leakage rate, used to unify the dimensionless leakage rate to the order of magnitude of the pressure drop deviation. In this embodiment, it is preferably 0.5. This indicator reflects the degree of decline in the pipeline's transport efficiency. When the pipeline pressure drop is higher than the design value or the leakage rate increases, it indicates that the pipeline is blocked, the filter is malfunctioning, or there is a leak in the pipeline.

[0138] Operational strategy deviation contribution W strategy Based on the current actual loading rate R t Start-stop frequency C t The value R corresponding to the optimal operating strategy opt C opt Deviation calculation:

[0139] Among them, R t The percentage of loading time in total runtime, and C t R represents the number of starts / stops or add / unload switches within a unit time window. opt C represents the target loading rate corresponding to the optimal operating strategy under the current working conditions. opt The target start-stop frequency corresponds to the optimal operating strategy under the current working conditions. Both are pre-stored in the configuration database by the platform based on equipment design parameters and optimal operating data. δ is a conversion factor for the start-stop frequency, used to unify the relative deviation dimension of the start-stop frequency with the relative deviation dimension of the load rate to the same numerical range. In this embodiment, it is preferably 0.3. This indicator reflects the degree of deviation between the operating strategy and the optimal strategy. When the load rate is too low or the start-stop is too frequent, it indicates that there is room for optimization in the group control strategy or pressure bandwidth setting.

[0140] Normalizing the above three contribution values ​​yields the normalized causal weight vector:

[0141] The softmax function maps the three non-negative contributions to a probability distribution that sums to 1, making the relative magnitude of each weight intuitively reflect the proportion of each cause in the total drag.

[0142] The final output warning information of the aerodynamic carbon contribution evaluation module specifically includes the following instantiations: The current PCCI value and its graded evaluation result (e.g., PPCI = 12.5%, judged as a dragging state); contribution or dragging state identifier; PCCI trend prediction results for the next 1-2 time windows based on historical PCCI data using first-order exponential smoothing or linear regression extrapolation (e.g., PPCI prediction value for the next 15 minutes is 10.2%~14.8%); normalized causal weight vector (body efficiency accounts for 35%, pipeline transportation accounts for 48%, and operation strategy accounts for 17%); and investigation guidance labels automatically generated based on the component with the largest proportion in the causal weight vector.

[0143] The matching logic for the investigation guidance tags is as follows: When the contribution of the main body efficiency deviation is the largest, output the troubleshooting label for the air compressor main body efficiency, such as prompting to check the main unit wear, lubricating oil quality and cooling system status; When the deviation in pipeline transportation efficiency contributes the most, output troubleshooting labels for pipeline transportation efficiency, such as prompts to check for pipeline leaks, whether filters are clogged, and pipe diameter matching. When the deviation of the operating strategy contributes the most, output the troubleshooting label for optimizing the operating strategy, such as prompting to check the pressure bandwidth setting, loading and unloading threshold and unit combination strategy; When multiple weights are close (for example, the difference between the largest and second largest weights is less than a preset proximity threshold, such as 5%), a comprehensive investigation label is output, indicating that there is room for optimization in the efficiency of the main body, pipeline transportation and operation strategy.

[0144] The aforementioned early warning information is output by the aerodynamic carbon contribution evaluation module and pushed to the monitoring dashboard of the factory energy management system in a structured data format, or presented in the form of visual charts on the human-computer interaction interface of the aerodynamic carbon contribution evaluation module for energy management personnel to view and make decisions in real time.

[0145] In addition, when an early warning signal is triggered, the aerodynamic carbon contribution evaluation module simultaneously generates an early warning event record and writes it into the early warning event log table in the time series database for subsequent statistical analysis, historical tracing, and trend judgment.

[0146] Figure 5 This is a flowchart of the aerodynamic carbon contribution evaluation module of the present invention. Figure 5 As shown, this module performs calculations according to the following process: Step S401: Obtain the actual gas production, number of operating gas-consuming devices, and dynamic grid carbon factor within the current time window from the fused feature dataset; read the rated gas production and total number of gas-consuming devices from the platform configuration database; and calculate the production load factor.

[0147] Step S402: Obtain the specific power characteristic value under the current operating condition from the nonlinear performance mapping sub-model of the air compressor, and calculate the theoretical baseline carbon emission equivalent within the current time window.

[0148] Step S403: Receive the theoretical baseline carbon emission equivalent and the dynamic additional carbon emission equivalent output by the dynamic carbon efficiency coupling modeling module, and calculate the actual carbon emission equivalent within the current time window.

[0149] Step S404: Calculate the aerodynamic carbon contribution index.

[0150] Step S405: Obtain the current urgency factor, dynamically reconstruct the baseline threshold, and calculate the effective threshold.

[0151] Step S406: Perform a graded evaluation based on the numerical range of the aerodynamic carbon contribution index to determine the current system status.

[0152] Step S407: When the aerodynamic carbon contribution index exceeds the dynamic effective threshold Thresh 2,dynamic When the time comes, the causal weight decomposition submodule is triggered to calculate the three normalized causal weights.

[0153] Step S408: Generate and output the warning information.

[0154] The above dynamic additional carbon emission equivalent ΔC t Together with the PCCI index, it forms a complete quantitative chain from aerodynamic operational deviations to carbon emission impacts and then to deviations from the zero-carbon target. Among them, ΔC t The final output of the dynamic carbon efficiency coupling modeling module characterizes how much additional carbon emissions are generated by inefficient operation; the PCCI index, the final output of this module, characterizes the relative deviation of these additional carbon emissions from the zero-carbon target. The combination of these two transforms the impact of compressed air systems on the zero-carbon target from a qualitative description into a quantifiable, traceable, and gradeable numerical indicator, providing precise data support for factory carbon management.

[0155] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A zero-carbon factory energy consumption and air pressure efficiency integrated monitoring platform, characterized in that, It includes a multi-source time-series alignment module that is interconnected, a dynamic carbon-efficiency coupling modeling module, and an aerodynamic carbon contribution evaluation module; The multi-source time-series alignment module is configured to synchronously collect high-frequency aerodynamic operating parameters of the compressed air system and low-frequency energy consumption data and green electricity ratio data of the factory power grid, and obtain the current time window carbon emission quota limit issued by the factory energy management system through a data interface. The high-frequency aerodynamic operating parameters and low-frequency energy consumption data are timestamped and aligned, and the green electricity ratio data and the carbon emission quota limit are also combined. The dataset is associated and matched with the corresponding time window to output a time-synchronized fused feature dataset; The dynamic carbon efficiency coupling modeling module is configured to receive the fused feature dataset and construct a pressure efficiency-carbon emission mapping model, which is used to convert the aerodynamic operating state into a dynamic additional carbon emission equivalent. Output the converted dynamic additional carbon emission equivalent ; The dynamic carbon efficiency coupling modeling module specifically includes: a nonlinear performance mapping sub-model for air compressors, a real-time estimation sub-model for pipeline pressure drop and leakage, a power consumption increment conversion sub-model, and a dynamic carbon emission equivalent conversion model; The nonlinear performance mapping sub-model of the air compressor is configured to receive data on speed, exhaust pressure, and load rate within the current time window, and output the specific power characteristic value under the current operating condition. ; The real-time estimation sub-model for pipeline pressure drop and leakage is configured to receive the fields of total pipeline pressure, terminal pressure and flow rate change rate in the current time window from the fused feature dataset, estimate the pipeline pressure drop value in real time, and estimate the real-time leakage rate based on the pipeline pressure decay data collected during non-production periods. The power consumption increment conversion sub-model is configured to be based on the specific power characteristic value. The real-time leakage rate, the pipeline pressure drop value, and the standard deviation of pressure fluctuation are used to calculate the additional power consumption caused by the inefficient operation of the compressed air system; The dynamic carbon emission equivalent conversion sub-model is configured to obtain the dynamic grid carbon factor corresponding to the current time window from the time-synchronized fused feature dataset. Based on the dynamic power grid carbon factor The additional power consumption is converted into dynamic additional carbon emission equivalents. Output the converted dynamic additional carbon emission equivalent ; The aerodynamic carbon contribution evaluation module is configured to receive the dynamic additional carbon emission equivalent. The aerodynamic carbon contribution index is calculated based on the benchmark carbon emission equivalent. The contribution or drag of the current operating state of the compressed air system to the zero-carbon target is determined according to the numerical range of the aerodynamic carbon contribution index, and early warning information is output. In the aerodynamic carbon contribution evaluation module: Actual carbon emission equivalent benchmark carbon emission equivalent With the aforementioned dynamic additional carbon emission equivalent sum; The benchmark carbon emission equivalent ,in, The specific power characteristic value, The dynamic power grid carbon factor, This refers to the upper limit of carbon emission quotas for the current time window; The aerodynamic carbon contribution evaluation module is equipped with a dual-mode evaluation mechanism based on the load boundary: (1) Relative contribution evaluation model: when At that time, the aerodynamic carbon contribution index The calculation formula is: in, Indicates the production load factor. The preset production load boundary threshold, This represents a correction term to prevent the denominator from being zero. The preset minimum operating baseline load compensation coefficient; (2) Absolute carbon drag evaluation model: when If the system is in a dormant or shutdown period, it will automatically stop calculating the aerodynamic carbon contribution index. Instead, they calculated the absolute carbon drag equivalent. : At this point, the aerodynamic carbon contribution evaluation module directly bases its evaluation on the absolute carbon drag equivalent. The compressed air system is compared with the preset carbon emission tolerance threshold during the shutdown period to determine whether it is hindering the zero-carbon target. The production load factor The calculation formula is: in, This represents the actual gas production within the current time window. This refers to the rated gas production under design conditions. The number of gas-consuming equipment in operation. This represents the total number of gas-using devices. and For preset weighting coefficients and .

2. The integrated monitoring platform for zero-carbon factory energy consumption and air pressure efficiency according to claim 1, characterized in that, The multi-source time-series alignment module includes an edge gateway deployed on the air compressor station side, a factory-level energy consumption acquisition module, a carbon emission data interface module, and a time-series database; The edge gateway connects to the local controller of the air compressor via an industrial bus or Modbus protocol to collect high-frequency pneumatic operating parameters, including air compressor outlet pressure, main pipe pressure, terminal pressure, instantaneous flow rate, motor current, active power, and exhaust temperature data. The factory-level energy consumption acquisition module connects to the factory's main electricity meter via a power gateway to collect low-frequency energy consumption data. The carbon emission data interface module connects to the power grid carbon factor data platform via an API interface to obtain green electricity ratio data and power structure prediction data, and connects to the factory energy management system via an API interface to obtain the carbon emission quota ceiling allocated to the current time window in real time. The time-series database is used to store the fused feature dataset after time-series synchronization.

3. The integrated monitoring platform for zero-carbon factory energy consumption and air pressure efficiency according to claim 1, characterized in that, The multi-source time-series alignment module is further configured to: perform time window merging on the acquired high-frequency data stream with a preset window width, perform steady-state identification and feature extraction on the data subset within each window, and calculate the feature vector used to characterize the operating state of the window; and map the feature vector of each window onto the time coordinate axis of the low-frequency energy consumption data using the timestamp of the low-frequency energy consumption data as the interpolation node, thereby forming a time-stamp aligned same-source time window dataset.

4. The integrated monitoring platform for zero-carbon factory energy consumption and air pressure efficiency according to claim 1, characterized in that, The aerodynamic carbon contribution evaluation module is also equipped with a threshold elastic scaling mechanism: The threshold elastic scaling mechanism stores a first baseline threshold and a second baseline threshold, which correspond to the upper and lower decision boundaries of the hierarchical evaluation, respectively. The threshold elastic scaling mechanism obtains the total carbon emission quota and the cumulative carbon emissions consumed in the current accounting cycle of the factory, and calculates the urgency factor. The threshold elastic scaling mechanism dynamically reconstructs the first baseline threshold and the second baseline threshold according to the urgency factor to generate a first effective threshold and a second effective threshold. The aerodynamic carbon contribution evaluation module is based on The comparison results with the first effective threshold and the second effective threshold are used for hierarchical evaluation: when When the value is less than or equal to the first effective threshold, it is determined to be a positive contribution. When the value is greater than the first effective threshold and less than or equal to the second effective threshold, it is determined to be an acceptable operating area. When the value exceeds the second effective threshold, it is determined to be a dragging state and an early warning signal is triggered.

5. The integrated monitoring platform for zero-carbon factory energy consumption and air pressure efficiency according to claim 4, characterized in that, The effective threshold is equipped with a lower limit protection mechanism: when the second effective threshold is lower than the preset lower limit value, the second effective threshold is locked to the lower limit value.

6. The integrated monitoring platform for zero-carbon factory energy consumption and air pressure efficiency according to claim 1, characterized in that, When the pneumatic carbon contribution evaluation module outputs the drag-causing causal weights, it is specifically configured to calculate the contribution of three deviations: the contribution of air compressor body efficiency deviation, the contribution of pipeline transportation efficiency deviation, and the contribution of operation strategy deviation. The softmax function is then used to perform probability distribution mapping on the above three deviation contributions, and a normalized causal weight vector with a sum of 1 is output.

7. The integrated monitoring platform for zero-carbon factory energy consumption and air pressure efficiency according to claim 6, characterized in that, The normalized causal weight vector is used to trigger the generation of the warning information, which includes one or more of the following: The aerodynamic carbon contribution index within the current time window Numerical values ​​and grading evaluation results; A status indicator indicating whether a contribution or a hindrance is made; Normalized causal weight vector; The trend prediction results are calculated based on historical aerodynamic carbon contribution index data using the first-order exponential smoothing method or the linear regression extrapolation method. Suggested investigation directions are automatically generated based on the component with the largest proportion in the causal weight vector.

8. An integrated monitoring device for energy consumption and air pressure efficiency in a zero-carbon factory, comprising: One or more processors, a memory, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the computer program, it implements the functions of an integrated monitoring platform for zero-carbon factory energy consumption and pressure efficiency as described in any one of claims 1 to 7.

Citation Information

Patent Citations

  • Blockchain-based carbon marketplace

    WO2025210465A1

  • Method and apparatus for measuring, auditing and monetizing energy savings for carbon credits

    WO2026035834A1