External wall thermal insulation monitoring and data acquisition system based on Internet of Things sensing

Abstract

The invention relates to the technical field of building structure monitoring, and particularly discloses an external wall thermal insulation monitoring and data acquisition system based on Internet of Things sensing, which comprises a multi-physics field sensing array, an edge computing gateway, a data fusion analysis center and an early warning and visualization platform. Temperature, heat flow and humidity data are collected through the multi-physics field sensing array, after the data are preprocessed through the edge computing gateway, spatio-temporal data fusion, thermal performance degradation evaluation and risk prediction are carried out through the data fusion analysis center, and finally graded alarm and visual display are carried out through the early warning and visual platform. According to the invention, comprehensive, dynamic and quantitative monitoring and risk assessment of the performance of the thermal insulation layer of the external wall of the building are realized, and data support is provided for preventive maintenance.

Description

Technical Field

[0001] This invention belongs to the field of building structure monitoring technology, specifically relating to an external wall insulation monitoring and data acquisition system based on Internet of Things (IoT) sensing. Background Technology

[0002] In the field of building energy conservation and structural health monitoring, IoT sensing technology has become a key means to achieve real-time data acquisition and remote monitoring. This technology aims to improve building energy efficiency, ensure structural safety, and optimize operation and maintenance management by deploying various sensor nodes in the building structure to build a sensing network.

[0003] Performance monitoring and evaluation of external wall insulation systems are crucial components of building energy conservation and operation. The external wall insulation layer plays a decisive role in maintaining a stable indoor thermal environment and reducing energy consumption; its long-term reliability is directly related to the building's overall energy consumption level. Continuous monitoring and scientific evaluation of the insulation layer's condition are fundamental to achieving energy-saving management throughout the building's entire lifecycle.

[0004] Deploying a single type of temperature sensor on or inside the insulation layer to obtain local temperature data indirectly reflects insulation performance. However, this method has significant limitations: a single temperature data point can only reflect instantaneous and localized thermal conditions, failing to comprehensively characterize the overall thermal resistance changes caused by aging, moisture absorption, or damage to the insulation material, and even less so to reveal the spatial distribution characteristics of internal material performance degradation. Due to the lack of comprehensive perception and analysis of multi-physics information, existing systems cannot effectively distinguish between normal temperature changes caused by environmental fluctuations and abnormal thermal signals caused by localized failures in the insulation layer (such as hollowing, detachment, and water seepage), resulting in insufficient early warning capabilities for potential risks. Furthermore, traditional data acquisition methods often process temperature information in isolation, failing to deeply correlate and model it with time, spatial dimensions, and environmental parameters, making the assessment results one-sided and lagging, unable to provide accurate decision-making basis for preventive maintenance. Summary of the Invention

[0005] The purpose of this invention is to provide an external wall insulation monitoring and data acquisition system based on Internet of Things (IoT) sensing, in order to solve the problems of existing technologies that rely on a single temperature data point, cannot comprehensively characterize the overall performance degradation and spatial distribution characteristics of the insulation layer, have difficulty distinguishing between normal environmental fluctuations and local failure anomalies, and have one-sided and lagging evaluation results.

[0006] This invention provides an external wall insulation monitoring and data acquisition system based on Internet of Things (IoT) sensing. The system includes a multi-physics field sensor array, an edge computing gateway, a data fusion analysis center, and an early warning and visualization platform.

[0007] The multiphysics sensing array is fixedly deployed in key areas of the building's exterior wall insulation layer using a preset gridded topology, enabling synchronous, in-situ acquisition of multi-dimensional physical parameters of the target monitoring area. The array consists of multiple sensing node units, each integrating at least three different types of miniaturized sensors. Specifically, the first type of sensor is a high-precision digital temperature sensor with a measurement accuracy of no less than ±0.1 degrees Celsius, used to acquire the absolute temperature value of the insulation layer surface or a specified depth. The second type of sensor is a thin-film heat flux density sensor, with a heat flux measurement density range covering 0 watts per square meter to 200 watts per square meter, used to directly measure the heat flux intensity perpendicular to the insulation layer surface. The third type of sensor is a capacitive humidity sensor, with a relative humidity measurement range covering 0% to 100%, used to monitor changes in moisture content within the insulation layer or at the interface layer. Furthermore, each sensing node unit also integrates a microcontroller and a low-power wireless communication module. The microcontroller is responsible for coordinating the synchronous sampling timing of each sensor and performing local analog-to-digital conversion and preliminary encapsulation of the collected raw temperature, heat flow, and humidity data. The low-power wireless communication module adopts a wireless personal area network protocol operating in a specific unlicensed frequency band and is responsible for periodically or event-triggered uploading the encapsulated sensor data packets to the edge computing gateway in a multi-hop self-organizing network manner.

[0008] The edge computing gateway, acting as an intermediate processing node between the sensor array and the cloud data center, is fixedly installed inside the building in a low-voltage electrical room or equipment platform near the exterior wall. The edge computing gateway integrates a first-level data processing module and a local storage unit. The first-level data processing module performs preprocessing and feature extraction on the raw data stream received from the multiphysics sensor array. The preprocessing includes outlier removal and Kalman filtering smoothing of temperature, heat flux, and humidity data based on a sliding time window to suppress measurement noise. The feature extraction process calculates key feature values ​​for each sensor node within a single sampling period based on the preprocessed time-series data. These feature values ​​include, but are not limited to: the spatial difference of the temperature gradient, the instantaneous ratio of heat flux density to the temperature difference, and the temporal difference of the humidity change rate. The local storage unit caches the complete feature value sequence and raw data snapshots for the past 7 days, allowing for data retransmission and offline analysis in case of network interruption.

[0009] The data fusion and analysis center, deployed on a cloud server cluster, is the core analysis engine of the system. It includes a spatiotemporal data fusion module, a thermal performance degradation assessment module, and a risk prediction module. The spatiotemporal data fusion module receives feature value data streams uploaded from all edge computing gateways and appends unified timestamps and 3D spatial coordinate labels based on Building Information Modeling (BIM) to construct a four-dimensional spatiotemporal dataset covering the entire exterior facade and integrating temperature, heat flow, and humidity field information. This module further runs spatial interpolation algorithms to perform kriging interpolation to complete missing grid data caused by node failures or communication blind spots, generating a continuous spatial distribution cloud map.

[0010] The thermal performance degradation assessment module, based on the four-dimensional dataset output by the spatiotemporal data fusion module, performs a quantitative assessment of the thermal performance of the insulation layer. The core of this module is a built-in dynamic baseline thermal resistance calculation model. This model first selects a continuous 30-day period with relatively stable ambient temperature fluctuations during the first full heating or cooling season after system initialization as the baseline period. Using stable heat flux density data collected by all sensor nodes during this period and corresponding indoor-outdoor temperature difference data, it calculates the baseline thermal resistance value and its confidence interval for each grid cell under healthy conditions using least squares linear regression, and stores this data in the baseline database. In subsequent daily monitoring, the module acquires the heat flux density and temperature difference data of each grid cell in real time and calculates its instantaneous apparent thermal resistance value. Furthermore, the module compares the instantaneous apparent thermal resistance value with the corresponding baseline thermal resistance value in the baseline database to calculate its relative degradation percentage. Meanwhile, this module introduces a performance degradation trend index based on time series analysis. This index is used to quantify the rate and acceleration of performance degradation by linearly fitting the percentage of thermal resistance decay of each grid cell over the past 90 days and extracting its slope and intercept.

[0011] The risk prediction module, coupled with the output of the thermal performance degradation assessment module and real-time humidity field information, performs qualitative and quantitative predictions of insulation layer failure risk. This module defines and calculates two key risk indicators. The first indicator is the damp-heat coupling risk coefficient, which is obtained by weighted multiplication of the percentage thermal resistance attenuation of a specific grid cell with the duration for which its humidity sensor reading exceeds a critical threshold of 80%. The weighting coefficients are calibrated based on material damp-heat aging test data. The second indicator is the spatial anomaly diffusion index, which is calculated by analyzing the spatial autocorrelation of thermal resistance attenuation values ​​between adjacent grid cells. This index is used to identify whether the boundary of a local failure area is expanding outwards. The risk prediction module sets multiple risk thresholds. When the damp-heat coupling risk coefficient or spatial anomaly diffusion index of any grid cell exceeds its corresponding first-level warning threshold, the area is determined to have potential risk; when it exceeds the second-level alarm threshold, the area is determined to have a high failure risk and requires immediate intervention.

[0012] The aforementioned early warning and visualization platform, serving as the system's human-computer interface, receives assessment results and early warning information from the data fusion and analysis center, and comprehensively displays and distributes alarms. This platform integrates a 3D building model rendering engine, capable of accurately mapping thermal resistance distribution cloud maps, performance degradation trend maps, humidity distribution maps, and risk hotspot maps onto the corresponding 3D building exterior wall models through layer overlay. The platform sets tiered alarm rules: areas identified as potential risks are marked with a flashing yellow icon in the visualization interface, and an early warning work order is generated and pushed to the mobile application of operation and maintenance management personnel; areas identified as high-risk requiring immediate intervention are marked with a continuously highlighted red icon, triggering both audible and visual alarms, and automatically generating an emergency repair work order containing specific location coordinates, risk type, and recommended measures, which is synchronized to the building operation and maintenance management system via the application programming interface.

[0013] As one embodiment of the present invention, the gridded topology deployment of the multiphysics sensing array follows the following principles: Densified monitoring nodes are set at the external corners, internal corners, around windows, on both sides of expansion joints, and in the central areas of walls facing different directions on the building's exterior walls, with a node spacing of no more than 1 meter; on other large flat wall surfaces, the node spacing is extended to 2 to 3 meters. Each sensing node unit is installed using a non-destructive fixing process. For external wall insulation systems, the sensors are adhered to the outer surface finish of the insulation layer or to the gaps between insulation boards using a special thermally conductive adhesive; for internal wall insulation systems, the sensors are installed on the surface of the interior wall, ensuring close contact with the wall.

[0014] In one embodiment of the present invention, the feature extraction process in the first-level data processing module of the edge computing gateway specifically includes the following calculation steps: For each sensing node, take the time window data of the 10 minutes before the current sampling time, calculate the standard deviation of the temperature data within the window as the temperature fluctuation intensity feature; calculate the absolute mean of the temperature difference between the node and four preset adjacent nodes as the local temperature gradient feature; calculate the 30-minute moving average of the ratio of the heat flux density reading to the instantaneous ratio of the indoor and outdoor temperature difference as the dynamic thermal resistance feature; calculate the 1-hour change of the humidity sensor reading as the humidity change rate feature. These feature values ​​constitute a feature vector, which is uploaded along with the node identifier and timestamp.

[0015] As one embodiment of the present invention, the dynamic benchmark thermal resistance calculation model of the thermal performance degradation assessment module incorporates a solar radiation heat gain correction factor when calculating the instantaneous apparent thermal resistance value. This correction factor is calculated by combining real-time total solar irradiance data obtained from a small weather station deployed on the building roof with the solar radiation absorption coefficient of the exterior walls facing different directions. When calculating the instantaneous apparent thermal resistance, the contribution of radiation temperature rise calculated by this correction factor is subtracted from the measured indoor-outdoor temperature difference, thereby eliminating the interference caused by the periodic changes in solar radiation on the thermal resistance assessment and improving the assessment accuracy, especially in the transitional seasons of hot summer and cold winter regions.

[0016] As one embodiment of the present invention, the specific method for calculating the spatial anomaly diffusion index by the risk prediction module is as follows: A neighborhood range with a radius of 2 grid units is defined, centered on the target grid cell; the difference between the current thermal resistance attenuation percentage of the target cell and the average thermal resistance attenuation percentage of all cells within its neighborhood is calculated; the ratio of this difference to the corresponding difference 24 hours prior in the previous calculation period is calculated; this ratio is weighted and summed with the rate of change of the coefficient of variation of the thermal resistance attenuation values ​​between cells within the target cell's neighborhood, with weighting coefficients of 0.7 and 0.3 respectively. The summation result is the spatial anomaly diffusion index of the target cell at the current moment. An index greater than 1 indicates that the anomaly is spreading, while an index less than 1 indicates that the anomaly is stabilizing.

[0017] In one embodiment of the present invention, the system operates within a layered collaborative framework, comprising a sensing layer, an edge layer, and a cloud layer. The sensing layer, composed of the multiphysics sensor array, is responsible for sensing and initial digitization of signals from the physical world. The edge layer, composed of the edge computing gateway, is responsible for localized data preprocessing, feature extraction, and short-term caching, achieving data burden reduction and initial real-time response. The cloud layer, composed of the data fusion analysis center and the early warning and visualization platform, is responsible for deep fusion of massive amounts of data, complex model calculations, long-term trend analysis, risk assessment, and global decision support. The three layers interact via an encrypted communication protocol, and a breakpoint resume mechanism ensures data integrity between the edge layer and the cloud layer.

[0018] Compared with the prior art, the advantages and positive effects of the present invention are as follows: 1. This invention achieves multi-dimensional, high spatial resolution, in-situ synchronous monitoring of the insulation layer's condition by deploying a multi-physics field sensing array integrating temperature, heat flux, and humidity sensors, and using a gridded topology to cover key areas of the exterior wall. This system not only collects data at single temperature points but also directly acquires heat flux density to calculate thermal resistance. Combined with humidity information, it comprehensively and directly quantifies the thermal performance and moisture status of the insulation material, fundamentally overcoming the limitations and inaccuracies of relying on a single temperature parameter to infer performance.

[0019] 2. This invention, by constructing a cloud-based data fusion and analysis center and introducing a dynamic benchmark thermal resistance calculation model and time series analysis, achieves quantitative assessment and trend prediction of insulation layer performance degradation. By establishing a health status benchmark and continuously tracking the relative decay and change trend of apparent thermal resistance, the system can clearly distinguish between continuous performance decline caused by material aging and damage and normal changes caused by environmental fluctuations. This represents a technological leap from static state description to dynamic performance degradation process tracking, providing accurate and forward-looking data for preventative maintenance.

[0020] 3. This invention, by designing two key risk indicators—the humidity-heat coupling risk coefficient and the spatial anomaly diffusion index—and developing a risk prediction module, achieves comprehensive assessment and graded early warning of localized failure risks in the insulation layer. This mechanism combines thermal performance degradation with humidity conditions and analyzes the spatial diffusion trend of anomalies, enabling earlier and more accurate identification of potential defects such as hollow areas and water seepage. It transforms the operation and maintenance model from reactive post-construction repair to proactive early warning, significantly improving the safety and durability of building exterior wall insulation systems and reducing the total life-cycle operation and maintenance costs. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall technical solution architecture of the present invention; Figure 2 This is a schematic diagram of the core principle framework for dynamic reference thermal resistance calculation and performance degradation assessment in this invention; Figure 3 This is a flowchart illustrating the data acquisition and preprocessing logic of the multi-physics sensing array and edge computing gateway working together in this invention. Figure 4 This is a schematic diagram of the multi-level interaction and data flow of the data fusion and analysis center in this invention for spatiotemporal data fusion and risk prediction. Detailed Implementation

[0022] Example 1: The overall technical architecture of the external wall insulation monitoring and data acquisition system based on Internet of Things (IoT) sensing proposed in this invention is shown in the attached figure. Figure 1 As shown, it consists of four main parts: a multiphysics field sensor array, an edge computing gateway, a data fusion and analysis center, and an early warning and visualization platform. These parts collaborate through a hierarchical mechanism to achieve high-precision, high-timeliness, and high spatial resolution perception, analysis, and decision support for the external wall insulation status. The following will be combined with the attached... Figure 1 To be continued Figure 4 The specific implementation methods of each component of the system are described in detail.

[0023] The multiphysics sensing array is deployed in key areas of the building's exterior wall insulation layer, using a pre-defined gridded topology for fixed installation. The array consists of several sensing node units, each integrating at least three different types of miniaturized sensors for collecting three core physical parameters: temperature, heat flux density, and humidity. The first type of sensor is a high-precision digital temperature sensor with a measurement accuracy of at least ±0.1 degrees Celsius. It can be deployed below the outer surface finish layer of the insulation layer, at a specified depth inside the insulation board, or at the interface layer to obtain absolute temperature values. The second type of sensor is a thin-film heat flux density sensor with a heat flux measurement density range covering 0 watts per square meter to 200 watts per square meter. It is vertically attached to the surface of the insulation layer to directly measure the heat flux intensity perpendicular to the wall direction. The third type of sensor is a capacitive humidity sensor with a relative humidity measurement range covering 0% to 100%, used to monitor changes in moisture content inside the insulation material or at the interface between the insulation layer and the structural layer in real time. Each sensing node unit also integrates a microcontroller and a low-power wireless communication module. The microcontroller is responsible for coordinating the synchronous sampling timing of the three types of sensors, ensuring that the acquisition timestamps of the three types of physical quantities are strictly aligned in each sampling period, and performing local analog-to-digital conversion and preliminary data encapsulation. The low-power wireless communication module operates in the unlicensed frequency band and uses a wireless personal area network protocol to build a multi-hop ad hoc network topology, sending the encapsulated raw sensor data packets to the edge computing gateway in a periodic upload or event-triggered upload manner.

[0024] In terms of deployment strategy, the gridded topology of the multiphysics sensor array follows a differentiated encryption principle. Encrypted monitoring nodes are installed at the external corners, internal corners, around windows, on both sides of expansion joints, and in the central areas of walls facing different directions, with a node spacing of no more than 1 meter. On other large, flat wall surfaces, the node spacing is extended to 2 to 3 meters. For external wall insulation systems, the sensor node units are fixed using a non-destructive process, adhering to the outer surface finish of the insulation layer or the gaps between insulation boards using a special thermally conductive adhesive, ensuring good thermal contact between the sensor and the insulation material. For internal wall insulation systems, the sensor node units are installed on the interior wall surface, and an elastic compression structure ensures a tight fit between the sensor and the wall surface, avoiding additional thermal resistance interference introduced by air gaps.

[0025] Please refer to the attached document. Figure 3The edge computing gateway, acting as an intermediate processing node between the sensing layer and the cloud layer, is fixedly installed inside the building in a low-voltage room or equipment platform near the exterior wall. It is internally configured with a first-level data processing module and a local storage unit. After receiving the raw data stream from the multiphysics sensor array, the first-level data processing module first performs preprocessing operations. This preprocessing includes outlier removal based on a sliding time window and Kalman filtering smoothing for temperature, heat flux, and humidity data. Specifically, outlier removal uses the 3σ criterion: within a time window consisting of five consecutive sampling points, if a point deviates from the window mean by more than three times the standard deviation, it is identified as an outlier and removed. Kalman filtering uses a one-dimensional linear state-space model, weighting and fusing the predicted and observed values ​​at the current moment to effectively suppress high-frequency noise interference and improve data stability.

[0026] After preprocessing, the first-level data processing module enters the feature extraction stage. This stage calculates multiple key feature values ​​for each sensor node based on the preprocessed time-series data. Specifically, it takes data from a 10-minute time window prior to the current sampling time and calculates the standard deviation of the temperature data within that window as the temperature fluctuation intensity feature; it selects four pre-defined neighboring nodes (upper, lower, left, and right) in the grid topology and calculates the average absolute value of the temperature difference between these nodes and the target node as the local temperature gradient feature; it calculates the 30-minute moving average of the ratio of the heat flux density reading to the instantaneous indoor-outdoor temperature difference as the dynamic thermal resistance feature; and it calculates the change in humidity sensor readings over one hour (i.e., the current value minus the value from 60 minutes ago) as the humidity change rate feature. These four feature values ​​together constitute a four-dimensional feature vector, which, along with the sensor node's unique identifier, timestamp, and original data snapshot summary information, is encapsulated into a feature data package and uploaded to the data fusion and analysis center via a secure transmission protocol. The local storage unit is used to cache the complete feature value sequence and original data snapshot of the most recent 7 days. When the network connection is interrupted, the missing data can be automatically retransmitted after the connection is restored according to the breakpoint resume mechanism.

[0027] Please refer to the attached document. Figure 2 With appendix Figure 4The data fusion and analysis center, deployed on a cloud server cluster, is the core intelligent analysis engine of the system, comprising three main functional units: a spatiotemporal data fusion module, a thermal performance degradation assessment module, and a risk prediction module. The spatiotemporal data fusion module first receives feature data streams uploaded from all edge computing gateways and attaches a Coordinated Universal Time (UTC) timestamp and a 3D spatial coordinate label based on Building Information Modeling (BIM). This spatial coordinate label is generated through a pre-established mapping relationship between the digital model of the building's exterior walls and the physical installation locations of sensor nodes, ensuring that each data point can be accurately located to a specific grid cell on the exterior facade. Based on this, the spatiotemporal data fusion module constructs a four-dimensional spatiotemporal dataset covering the entire exterior facade and integrating temperature, heat flow, and humidity field information.

[0028] To address data loss caused by node failures, communication blind spots, or sensor malfunctions, the spatiotemporal data fusion module further runs a spatial interpolation algorithm to complete the missing grid data. This algorithm employs ordinary kriging interpolation, using the measured values ​​of neighboring valid nodes as known sample points, and estimates the optimal unbiased value of the missing points based on a semi-variogram model. The semi-variogram uses a spherical model with a range set to 3 meters, and the nugget effect is dynamically adjusted based on the historical data noise level. After interpolation completion, the system can generate continuous and seamless temperature distribution cloud maps, heat flux density distribution cloud maps, and humidity distribution cloud maps, providing a complete data foundation for subsequent analysis.

[0029] The thermal performance degradation assessment module performs a quantitative assessment of the thermal performance of the insulation layer based on the aforementioned four-dimensional spatiotemporal dataset. The core of this module is a dynamic baseline thermal resistance calculation model. This model selects a 30-day period of relatively stable ambient temperature fluctuations as the baseline during the first full heating or cooling season after system initialization. During this period, the system continuously collects stable heat flux density data and corresponding indoor-outdoor temperature difference data from all sensor nodes. For each grid cell, the least squares method is used to perform a linear regression on the heat flux density q and temperature difference ΔT, and the fitted relationship is:

[0030] in, This is the baseline thermal resistance value of the mesh element under healthy conditions. It is obtained through regression calculation. The optimal estimate and its 95% confidence interval are obtained, and the results are stored in the benchmark database. In subsequent daily monitoring, the system acquires the heat flux density of each grid cell in real time. With temperature difference Calculate its instantaneous apparent thermal resistance.

[0031] To eliminate the interference of periodic variations in solar radiation on temperature difference measurements, especially during the transitional seasons in hot-summer and cold-winter regions, a solar radiation heat gain correction factor is introduced into the system. This correction factor is based on real-time total solar irradiance provided by a small weather station deployed on the building roof. Solar radiation absorption coefficients of different orientations from the exterior walls The calculation shows that, ,in The empirical heat transfer coefficient is taken as 0.025 degrees Celsius per square meter per watt. The final corrected temperature difference used to calculate the apparent thermal resistance is... .

[0032] The thermal performance degradation assessment module further evaluates the instantaneous apparent thermal resistance value. The corresponding reference thermal resistance value in the reference database Compare and calculate their relative attenuation percentage. Defined as:

[0033] Simultaneously, this module introduces the Performance Degradation Trend Index (TDI), which is calculated by analyzing the percentage decrease in thermal resistance of each grid cell over the past 90 days. Perform linear fitting Extracting the slope As a rate of degradation, the intercept Reflects the initial offset.

[0034] TDI comprehensively reflects the long-term trend and accelerating characteristics of performance degradation, providing a time-series basis for risk prediction.

[0035] The risk prediction module, coupled with the output of the thermal performance degradation assessment module and real-time humidity field information, comprehensively assesses the risk of insulation layer failure. This module defines and calculates two key risk indicators. The first indicator is the humidity-heat coupling risk coefficient. The calculation method is as follows: when the humidity sensor reading of a certain grid cell continuously exceeds the 80% relative humidity threshold... At what time does its thermal resistance decrease by a certain percentage? and Perform a weighted product, that is

[0036] Among them, the weighting coefficient Based on the accelerated damp heat aging test data of thermal insulation materials, a typical value is taken as 0.005 (unit: % / hour). -1The second indicator is the Spatial Anomaly Diffusion Index (SADI), used to identify whether a localized failure region is expanding outwards. Its calculation method is as follows: Define a neighborhood with a radius of 2 grid units centered on the target grid cell; calculate the current percentage of thermal resistance decay for that target cell. The average percentage of thermal resistance attenuation of all units in its neighborhood The difference between Then calculate the difference between this value and the corresponding difference 24 hours prior to the previous calculation period. ratio Simultaneously calculate the coefficient of variation of the thermal resistance attenuation value between neighboring units. And calculate its rate of change over 24 hours. Ultimately, the spatial anomaly diffusion index When SADI > 1, it indicates that the anomalous region is spreading; when SADI < 1, it indicates that the anomalous region is stabilizing.

[0037] The risk prediction module sets two levels of risk thresholds. When the HRC or SADI of any grid cell exceeds the first-level warning threshold (e.g., HRC>15, SADI>1.1), the area is determined to have potential risks; when it exceeds the second-level alarm threshold (e.g., HRC>25, SADI>1.3), the area is determined to have a high risk of failure and requires immediate intervention.

[0038] The aforementioned early warning and visualization platform serves as the system's human-computer interaction terminal, deployed in the operation and maintenance management center or accessible via a web browser and mobile applications. This platform integrates a 3D building model rendering engine, capable of loading building information models and overlaying thermal resistance distribution cloud maps, performance degradation trend maps, humidity distribution maps, and risk hotspot maps as independent layers onto the corresponding 3D geometric surfaces of the exterior walls. Each layer supports transparency adjustment, color mapping, and numerical annotation, facilitating technicians' intuitive understanding of the spatial distribution characteristics of the insulation status.

[0039] The platform incorporates a tiered alarm rule engine. Areas identified as potential risks are marked with a flashing yellow icon on the 3D model, and an early warning work order is automatically generated, containing location coordinates, risk type, current HRC and SADI values, and suggested inspection measures. This order is sent to the mobile terminals of designated maintenance personnel via push notifications. Areas identified as high-risk requiring immediate intervention are marked with a continuously highlighted red icon, triggering both audible and visual alarms and generating an emergency repair work order. This work order includes precise 3D spatial coordinates, failure mode inferences (such as suspected water seepage causing a sudden drop in thermal resistance or the spread of hollow areas), recommended repair solutions, and a materials list. It is synchronized to the building maintenance management system via a standardized application programming interface, enabling automatic work order dispatch and closed-loop tracking.

[0040] The entire system operates within a layered collaborative framework, as shown in the attached diagram. Figure 1 As shown, the system is clearly divided into a sensing layer, an edge layer, and a cloud layer. The sensing layer consists of a multi-physics sensor array, responsible for in-situ sensing and preliminary digitization of signals from the physical world. The edge layer consists of edge computing gateways, undertaking localized data cleaning, feature extraction, and short-term caching, significantly reducing uplink data traffic and improving system real-time response. The cloud layer consists of a data fusion analysis center and an early warning and visualization platform, focusing on the deep fusion of massive heterogeneous data, computation of complex physical models, long-term degradation trend mining, and global risk decision-making. The three layers interact with each other through an encrypted communication channel based on a transport layer security protocol. The edge layer and the cloud layer use a breakpoint resumption protocol with verification and retransmission mechanisms to ensure data integrity and consistency in unstable network environments.

[0041] In summary, this embodiment constructs a complete closed-loop system from data acquisition to intelligent decision-making through multi-physics synchronous sensing, edge-cloud collaborative computing, dynamic benchmark modeling, and dual-dimensional risk prediction of humidity and heat and space. It achieves high-precision, high-timeliness, and high-spatial-resolution monitoring and early warning of the performance degradation process of external wall insulation layers, providing solid technical support for building energy conservation and safe operation and maintenance.

[0042] Example 2: Based on Example 1 above, this example further optimizes the power supply and communication mechanism of the multiphysics sensing array and enhances the local decision-making capability of the edge computing gateway to adapt to building application scenarios in remote areas with no stable mains power supply or weak network coverage.

[0043] Specifically, each sensing node unit, in addition to the existing microcontroller and low-power wireless communication module, integrates a micro energy harvesting module and a supercapacitor energy storage unit. The energy harvesting module is composed of a flexible photovoltaic film and a thermoelectric generator. The flexible photovoltaic film is attached to the outer surface of the sensing node shell, generating electricity using natural sunlight during the day; the thermoelectric generator is sandwiched between the insulation layer and the outer decorative layer, generating electricity using the Seebeck effect based on the temperature difference between the inner and outer surfaces of the insulation layer. The two energy sources are powered by a power management integrated circuit for maximum power point tracking and multi-source fusion, prioritizing power to the sensor and microcontroller, with excess energy stored in the supercapacitor.

[0044] Regarding communication protocols, the low-power wireless communication module employs LoRa modulation technology, operating in the unlicensed frequency band of 470 MHz to 510 MHz, with software-configurable transmit power up to a maximum of 19 dBmW. In mesh deployment, the sensor node unit uses an adaptive routing algorithm to dynamically select the next-hop relay node based on link quality indicators and remaining energy levels, ensuring that data packets are transmitted to the edge computing gateway via the lowest energy-consuming path. When a node detects that its own battery level is below the 20% threshold, it automatically switches to an ultra-low-power monitoring mode, only receiving query commands from the gateway during a preset wake-up window, and shutting down the RF module at other times to extend standby life.

[0045] In this embodiment, the edge computing gateway is upgraded to an intelligent edge node with local early warning capabilities. In addition to performing the original preprocessing and feature extraction, its first-level data processing module adds local risk assessment logic. This logic is based on a simplified version of the risk indicator calculation rules: when the thermal resistance attenuation percentage of a sensor node exceeds 20% for three consecutive sampling cycles and the humidity reading is higher than 75%, a local first-level early warning is triggered; if the local temperature gradient characteristic value suddenly increases by more than 50% simultaneously, a local second-level alarm is directly triggered. This type of local early warning information does not rely on cloud analysis and can be immediately uploaded to the operation and maintenance platform through the gateway's built-in narrowband IoT module, or issued as an alert through local audio-visual devices, making it suitable for emergency scenarios where network interruptions or cloud service unavailability occur.

[0046] Furthermore, the edge computing gateway's local storage capacity has been expanded to 64 gigabytes, and it supports a cyclic overwrite write strategy. When storage utilization reaches 90%, the earliest original data snapshot is automatically deleted, preserving the integrity of the feature value sequence. Simultaneously, the gateway incorporates a lightweight machine learning inference engine that can load distilled and compressed anomaly detection models to perform real-time classification of feature vectors, identifying typical failure modes such as rapid water seepage, progressive aging, and construction defects. The mode labels are uploaded along with the feature data, providing prior knowledge for cloud-based analysis.

[0047] On the data fusion and analysis center side, the spatiotemporal data fusion module has added fusion processing for energy status data. The data packets uploaded by each sensor node contain the remaining power percentage and energy harvesting efficiency indicators. The spatiotemporal data fusion module maps this information to a 3D model to generate a node health status layer, helping maintenance personnel determine whether missing data is due to physical failure or energy depletion, thus avoiding misjudgments.

[0048] The early warning and visualization platform has added an energy management view to display information such as real-time power consumption, average daily power generation, and estimated runtime for each sensor node. It also supports sorting and filtering low-power nodes by power consumption, facilitating maintenance scheduling. For nodes that enter dormancy due to insufficient power, the platform automatically marks their data as unreliable and reduces their weight or temporarily excludes their impact in risk assessment.

[0049] Through the above improvements, this embodiment significantly enhances the robustness and autonomy of the system in complex power supply and communication environments, expands the application scope of the present invention in the field of external wall insulation monitoring, and is particularly suitable for scenarios with limited infrastructure conditions such as rural self-built houses, historical building protection, and plateau border outposts.

Claims

1. An external wall insulation monitoring and data acquisition system based on Internet of Things (IoT) sensors, characterized in that, include: A multiphysics field sensing array is used to deploy in a preset gridded topology in key areas of the building's exterior wall insulation layer to achieve synchronous in-situ acquisition of multi-dimensional physical parameters. An edge computing gateway is fixedly installed inside the building and serves as an intermediate processing node between the multiphysics sensing array and the cloud data center. The data fusion and analysis center, deployed on a cloud server cluster, is the core analysis engine of the system. The early warning and visualization platform, as the human-computer interaction interface of the system, is used to receive the evaluation results and early warning information issued by the data fusion analysis center, and to comprehensively display and distribute alarms.

2. The external wall insulation monitoring and data acquisition system based on Internet of Things sensing according to claim 1, characterized in that, The multiphysics sensing array is composed of multiple sensing node units, each of which integrates a high-precision digital temperature sensor, a thin-film heat flux density sensor, and a capacitive humidity sensor. The high-precision digital temperature sensor is used to collect the absolute temperature value of the surface of the insulation layer or a specified depth. The thin-film heat flux density sensor is used to directly measure the heat flux intensity perpendicular to the surface of the insulation layer; the capacitive humidity sensor is used to monitor the moisture content changes inside the insulation layer or the interface layer. Each sensor node unit also integrates a microcontroller and a low-power wireless communication module. The microcontroller is used to coordinate the synchronous sampling timing of each sensor and to perform local analog-to-digital conversion and preliminary encapsulation of the collected raw data; The low-power wireless communication module is used to upload the encapsulated sensor data packets to the edge computing gateway in a multi-hop self-organizing network manner.

3. The external wall insulation monitoring and data acquisition system based on Internet of Things sensing according to claim 2, characterized in that, The edge computing gateway has a built-in first-level data processing module and a local storage unit. The first-level data processing module is used to perform preprocessing and feature extraction on the raw data stream received from the multi-physics field sensor array. The preprocessing includes outlier removal and Kalman filtering smoothing of temperature, heat flow, and humidity data based on a sliding time window. The feature extraction is based on the preprocessed time-series data to calculate the key feature values ​​of each sensor node in a single sampling period. The local storage unit is used to cache the complete feature value sequence and original data snapshot for a preset number of days.

4. The external wall insulation monitoring and data acquisition system based on Internet of Things sensing according to claim 3, characterized in that, The data fusion and analysis center includes a spatiotemporal data fusion module, a thermal performance degradation assessment module, and a risk prediction module. The spatiotemporal data fusion module receives feature value data streams uploaded from all edge computing gateways and adds a unified timestamp and a three-dimensional spatial coordinate label based on building information model to construct a four-dimensional spatiotemporal dataset covering the entire exterior facade. The thermal performance degradation assessment module performs a quantitative assessment of the thermal performance of the insulation layer based on the four-dimensional dataset output by the spatiotemporal data fusion module. The risk prediction module couples the output of the thermal performance degradation assessment module with real-time humidity field information to perform qualitative and quantitative prediction of the failure risk of the insulation layer.

5. The external wall insulation monitoring and data acquisition system based on Internet of Things sensing according to claim 4, characterized in that, The early warning and visualization platform integrates a 3D building model rendering engine, which can accurately map thermal resistance distribution cloud maps, performance degradation trend maps, humidity distribution maps, and risk hotspot maps onto the corresponding 3D building exterior wall models in a layer overlay manner.

6. The external wall insulation monitoring and data acquisition system based on Internet of Things sensing according to claim 5, characterized in that, The thermal performance degradation assessment module has a built-in dynamic benchmark thermal resistance calculation model. The dynamic benchmark thermal resistance calculation model is used to select 30 consecutive days with gentle ambient temperature fluctuations as the benchmark period during the first full heating or cooling season after system initialization. Using the stable heat flux density data collected by all sensor nodes during this period and the corresponding indoor and outdoor temperature difference data, the benchmark thermal resistance value and its confidence interval of each grid cell in a healthy state are fitted and calculated by least squares linear regression and stored in the benchmark database. In subsequent daily monitoring, the dynamic reference thermal resistance calculation model acquires the heat flux density and temperature difference data of each grid cell in real time, calculates its instantaneous apparent thermal resistance value, and compares the instantaneous apparent thermal resistance value with the corresponding reference thermal resistance value in the reference database to calculate its relative attenuation percentage.

7. The external wall insulation monitoring and data acquisition system based on Internet of Things sensing according to claim 6, characterized in that, The thermal performance degradation assessment module also introduces a performance degradation trend index based on time series analysis; The calculation process of the performance degradation trend index is as follows: obtain the thermal resistance attenuation percentage sequence of each grid cell in the past 90 days, and perform linear fitting on the sequence to extract its slope and intercept, which are used to quantify the rate and acceleration of performance degradation.

8. The external wall insulation monitoring and data acquisition system based on Internet of Things sensing according to claim 7, characterized in that, The dynamic reference thermal resistance calculation model incorporates a solar radiation heat gain correction factor when calculating the instantaneous apparent thermal resistance value. The solar radiation heat gain correction factor is calculated by combining real-time total solar radiation irradiance data obtained from small weather stations deployed on building roofs with the solar radiation absorption coefficients of different orientations of the exterior walls. When calculating the instantaneous apparent thermal resistance, the contribution of radiation temperature rise calculated by the solar radiation heat gain correction factor is subtracted from the measured indoor-outdoor temperature difference.

9. The external wall insulation monitoring and data acquisition system based on Internet of Things sensing according to claim 8, characterized in that, The risk prediction module defines and calculates two key risk indicators; the first indicator is the damp-heat coupling risk coefficient, which is calculated as follows: The percentage of thermal resistance attenuation of a specific grid cell is weighted and multiplied by the duration for which the humidity sensor reading exceeds 80% of the critical threshold. The weighting coefficients are calibrated based on the material's damp heat aging test data. The second indicator is the spatial anomaly diffusion index, which is calculated as follows: a neighborhood with a radius of 2 grid units is defined with the target grid cell as the center. Calculate the difference between the current thermal resistance decay percentage of the target cell and the average thermal resistance decay percentage of all cells in its neighborhood; Calculate the ratio of this difference to the corresponding difference 24 hours ago in the previous calculation period; The ratio is weighted and summed with the rate of change of the coefficient of variation of the thermal resistance attenuation value between units in the neighborhood of the target unit. The summation result is the spatial anomaly diffusion index.

10. The external wall insulation monitoring and data acquisition system based on Internet of Things sensing according to claim 9, characterized in that, The risk prediction module sets multiple risk thresholds; when the damp-heat coupling risk coefficient or spatial anomaly diffusion index of any grid cell exceeds its corresponding first-level warning threshold, it is determined that there is a potential risk in the area; when it exceeds the second-level alarm threshold, it is determined that the area has a high failure risk and immediate intervention is required.

Images